CM~YET- * ~SchoollRV G M 4", A

Transcription

1 * ~SchoollRV G M 4", A JUN1VERSTV OF STRATHCLYDE Department of Ship and Marine Technology The Ship Stability Research Centre [SSRC September 1998 CM~YET- tjf-r~ffofand Mobdiy of RmeaSdlrsF

2 ABOUT WEGEMT WEGEMT is a European Association of 43 Universities in 18 countries. t was formed in 1978 withi the aim of increasing the knowledge base and updating and extending the skills and competence of engineers and postgraduate students working at an advanced level in marine technology and related sciences. WEGEMT achieves this aim by encouraging universities to be associated with it, to operate as a network and therefore actively collaborate in initiatives relevant to this aim. WEGEMT considers collaborative research, education and training at an advanced level, and the exchange and dissemination of information, as activities which further the aim of the Association. NB For marine technology and related sciences, WEGEMT includes all aspects of offshore oil and gas exploration and production, shipping and shipbuilding, underwater technologies and other interdisciplinary areas concerned with the oceans and seas. ABOUT TLE PUBLCATON This publication represents a series of lecturers' papers, which were presented at a fiveday event entitled Total Stability Assessment of Damaged Passenger Ro/Ro Vessels. The event was hosted by the University of Strathclyde Department of Marine Technology from le~ to i8th September 1998 at The Ship Stability Research Centre, Glasgow, UK. The event was partly sponsored by the European Commission through the Training and Mobility of Researchers Programme under a project submitted by WEGEMT entitled "Supporting European Marine Research and ndustry through Training and Mobility (EMTY'. Published by WEGEMT SBN Number: This volume has been made available so that it contains the original authors' typescripts. The method may from time to time display typographical limitations. t is hoped however, that they do not distract the attentions of the reader. Please note that the expressed views are those of the individual authors and that WEGEMT as publishers cannot accept responsibility for any errors or omissions.

3 The Ship Stability Research Centre University of Strathclyde PROBABLSTC CRTERA Preamble 1. How does the topic - "Probabilistic Criteria" relate to the course subject heading: "Total Stability Assessment of Damaged Passenger/RoRo Vessels"? 2. Make-up of the talks on "Probabilistic Criteria" Part Historical Development Subdivision standards, post-titanic..2 The need to establish minimum standards of'residual Stability'.3 The search for a better way to assess 'survivability'..4 Passenger ship Equivalent Regulations..5 (Dry) Cargo Ship Regulations..6 'Harmonized' Regulations..7 A / Amax calculations..8 Compartmental Standard. Part 2 Theoretical Background. A listing of present (and future) 'probabilistic' regulations..2 The major underlying principle in using 'probabilistic' methods..3 The movement at MO away from 'deterministic' methods towards 'probabilistic' methods..4 "Subdivision of Ships" by Prof K.Wendel..5 Collision data - it's use in 'probabilistic' regulations..6 Main conclusions drawn from an analysis of the collision data..7 The factor 'p' - longitudinal extent of damage.8 The factor 'r' - penetration inboard.9.10 nfluence of ship condition on 'probability of survival'. nfluence of damage location on 'probability of survival'. 11 nfluence of'sea state' on 'probability of survival'..12 The factor's'- estimate of the 'probability of survival'..13 The factor 'v' - maximum vertical extent of damage..14 The Attained Subdivision ndex, A.15 dealised Damage Distribution along a Ship's Length The Factor "a". 16 The Treatment of Simultaneous Flooding Probabilistic Criteria

4 The Ship Stability Research Centre University of Strathclyde Part Historical Development 1. The First Bulkhead Committee - post "Titanic" 2. The need to establish minimum standards of 'Residual Stability' 3. The search (at MO) for a more suitable way to assess the survivability of a damaged passenger ship. 4. The Passenger Ship Equivalent Regulations - Resolution A.265 (V) 5. The (Dry) Cargo Ship Regulations Reg.25 of Chap.H-, Part B-i, of SOLAS A future set of 'Harmonized' regulations, to replace the existing sets of Passenger and (Dry) Cargo Ship regulations. 7. A 'one-off use of a simplified version of Res.A.265 (V), in order to up-grade the stability of 'existing' RoRo passenger ships. (By means of A / Amax calculations 8. Compartmental Standard Probabilistic Criteria 2

5 The Ship Stability Research Centre University of Strathclyde Preamble PROBABLSTC CRTERA 1. How does the topic - "Probabilistic Criteria" -relate to the course subject heading: "Total Stability Assessment of Damaged Passenger/RoRo Vessels"? 1.1 t may not appear immediately obvious why it is necessary to include talks on 'Probabilistic Criteria', when almost all ships have been constructed to SOLAS (or regionally agreed) standards of subdivision/stability such as SOLAS 90 or SOLAS 90 +'water-on-deck. There are two main reasons why it is useful to have a little knowledge of 'probabilistic' regulations, besides the general one that everyone needs to be aware of their existence. 1.2 Since the early 70's, it has been possible to build a passenger ship to a set of regulations which is equivalent to, and a complete alternative to, the corresponding SOLAS regulations. These are referred to as the Passenger Ship Equivalent Regulations, or (more commonly) as Resolution A A much more common instance of the application of probabilistic principles is the use of the A / Amax ratio, to decide the order in which 'existing' RoRo passenger ships are to be upgraded in order to comply with SOLAS 90 / SOLAS 90 + 'water-on-deck. t should be emphasized at this point that A / Amax is not a standard! 1.4 t may not, at first sight, be apparent why the next instance of subdivision standards based on probabilistic principles - the (dry) Cargo Ship Regulations - is relevant to passenger ships, (including RoRo passenger ships). For reasons given in the next sub-paragraph, it is advisable to mention these regulations at this point. Probabilistic Criteria 3

6 The Ship Stability Research Centre University of Strathclyde 1.5 You may be aware that there has been a lot of activity recently in SLF sub-committee meetings at MO, concerning the merging of the two present sets of probabilistically-based regulations into A single set of'harmonized' regulations. The primary intention is to develop a common procedure for both passenger and (dry) cargo ships, in order to estimate the level of survivability achieved by a particular ship design. n parallel with the introduction of these harmonized' regulations, it is intended to remove from SOLAS both the subdivision provisions - which have fare into disrepair - and the deterministically-based residual stability criteria. Probabilistic Criteria 4

7 H The Ship Stability Research Centre University of Strathclyde H Part Historical Development 1.1. The First Bulkhead Committee - post "Titanic" The major outcome of the discussions held at meetings of the First Bulkhead Committee was the introduction of the 'Factorial' method, which lays down a procedure for determining statutory minimum levels of subdivision which are to be achieved by all sea-going passenger vessels - which meant, in essence, Trans-Atlantic liners! Assuming total flooding of a main compartment - (in certain cases, two, or even three adjacent compartments) - it is necessary to show that a ship will retain sufficient residual/reserve buoyancy that it will not sink. An alternative way in which this provision can be interpreted is as follows: "The distance between adjacent bulkheads is to be such that, (assuming total flooding), the bulkhead deck/margin line will not be immersed." The level of subdivision to be attained is based upon a function of ship length, passenger density and LSA provision. All passenger ships are to be provided, as a minimum, with a 'onecompartment' standard of subdivision. For ships having a high passenger density, (or insufficient boatage), the requirement is for a 'two-compartment' standard, (or even three- in rare cases) The need to establish minimum standards of 'Residual Stability' n the early 50's, passenger ships, (particularly RoRo passenger ships), were shown - by damaged stability calculations - to be at significant risk of (rapid) capsize, rather than sinkage, if (side-collision) damage were to occur around the midships region. The problem arose because Probabilistic Criteria - Part 1 5

8 The Ship Stability Research Centre University of Strathclyde of the growing tendency to build more 'beamy' ships, i.e. ships having a relatively low L/B ratio t was generally agreed that it was necessary to introduce statutory minimum levels of residual stability which, in retrospect, were pitched at too low a level. Additionally, the wording of the regulations is quite indefinite in places, leading to significant variations in interpretation by various Authorities At the present time, the so-called SOLAS 90 standard is the agreed international standard for all passenger ships, whilst the recently ratified Stockholm Agreement, (relating to RoRo passenger ships), incorporates the SOLAS 90 +water-on-deck standard. n both sets of regulations, all RoRo passenger ships must achieve at least a one-compartment standard; those with high passenger numbers and deficient boatage must comply with a twocompartment standard of subdivision. 1.3 The search (at MO) for a more suitable way to assess the survivability of a damaged passenger ship For the last 30/40 years, MO members have become increasingly aware of the inherent flaws in the 'Factorial' method as a suitable means of assessing the 'survivability' of a passenger ship. The use of floodable/permissible length curves nowadays is quite rare, since designers appreciate that damaged stability requirements invariably over-ride those arising out of the subdivision part of SOLAS. Maintaining an adequate level of residual stability is much more important than establishing that there is residual buoyancy available The concept of a statutory maximum damage extent is linked with the regulatory minimum residual stability standard to be met. n effect, this implies a minimum length for any compartment before it can be considered as a 'true' compartment in a regulatory sense. Therefore, most present-day designs of passenger ships tend to have compartment lengths only Probabilistic Critcria - Part 1 6

9 H The Ship Stability Research Centre Universit" of Strathclyde marginally greater than this statutory distance L m in most cases - particularly in the midships region. This preponderance of relatively-closely-spaced bulkheads clearly increases the possibility of a side-damage opening up two, or even three, adjacent compartments. When this undesirable situation is linked, in addition, to a freeboard to the bulkhead deck which only achieves marginal compliance with the specified stability criteria, (as it often does), the overall * result is a relatively poor standard of subdivision MO members were favorably impressed by the argument put forward in a paper read in 1968 at SNAME -'Subdivision of Ships' - presented by Prof K.Wendel. n this paper, there was an outline procedure given which was intended to provide efficient subdivision for ships, by employing 'probabilistic' principles Wendel laid out a procedure whereby it was possible to quantify the level of subdivision achieved in a particular ship design, having a fixed arrangement of main watertight bulkheads. He recommended the use of a Subdivision ndex - this being a number between zero and unity - which is an estimate of the proportion of all possible damages that a ship is likely to survive, i.e. will not sink or capsize How is this proportion determined? Assume that there is a damage confined to a compartment - or group of compartments - and assign a probability of 'p' to this case. Then assume that there is a probability of survival, 's', that the ship can survive this damage extent. Therefore, the overall probability appropriate to this damage scenario is the product (p * s). A similar procedure is adopted for all the other damage combinations, over the entire service range. Giving the Attained Subdivision ndex, (A) as: A=Ep*s Probabilistic Criteria - Part 1 7

10 The Ship Stability Research Centre Universin. of Strathclvde Two further assumptions need to be mentioned at this stage: 'p' is the rehltive probability of damage extent, since 'damage' is assumed to be certain; * 's' varies between zero and unity. s = 0, where the ship is assumed to sink/capsize; s = 1, where the ship is assumed to survive; 0 < s < 1, where the ship is assumed to have a probability of's' of survival Since the factors 'p' and s' are given as different formulae in the two sets of probabilistically-based regulations, a separate consideration of how these factors were arrived at will be given later in the talks. Probabilistic Criteria - Part 1 8

11 The Ship Stability Research Centre University of Strathclvde 1.4 The Passenger Ship Equivalent Regulations - Resolution A.265 (V) The subdivision/stability requirements, as set out in the Passenger Ship Equivalent.. Regulations, Res.A.265(V), follow the outline of the proposed theoretical, probabilisticallybased, approach written in Wendel's paper. The core of the calculation procedure in these regulations is the calculation of an index- the so-called Attained Subdivision ndex (A), such that: A = S (a * p * s) See formula H, reg. 6(a)(i) n the above formula, the factors 'a', 'p' and 's' represent the following: 'a' represents the probability of (centre of) damage location of a compartment, or group of compartments, 'p' represents the probability of a compartment, or group of compartments, being damaged, 's' represents the 'probability-of-survival', after damage as described. n addition, there is a further factor - 'r -which represents the probability of penetration in from the ship's side. This factor may be regarded as a reduction factor to 'p' - which assumes all penetrations as far as the ship's centreline The factors 'a', 'p' and 'r' were based upon an analysis of historic (side-collision) casualty data, and may be seen at the following positions in the regulations: 'a' formulae [1, reg.6(b) 'p' formulae V, reg.6(c)(i) 'p' for two, or more, adjacent compartments flooded simultaneously: formulae V, V and V of reg.6(c)(iii) 'r' formulae X, reg.7(b)(ii) A more detailed consideration of how these factors are derived will be given in part 2. Probabilistic Criteria - Part 1 9

12 The Ship Stability Research Centre University of S(rathclvde The 'probability-of-survival' after assumed damage is strongly influenced by two main factors. Firstly, the residual buoyancy and its distribution are clearly very important. n broad terms, this means that the greater the residual stability lever curve, the greater the probability... that a damaged ship will not sink/capsize. Of course, the environmental conditions at the time of collision will also have a major bearing on the probability of capsize, (or otherwise). Two series of 'damaged' model tests - one in the UK, the other in the United States - provided the raw data from which the factor's' of the regulations was derived: 0.5 Ss,= 4.9 [ tan *(GMr- sj Formula V, reg.6(d) Note that this factor's' was based upon an average of the 'windward' and 'leeward' results! When the Attained Subdivision ndex, A, has been calculated, it is compared to an index - the Required Subdivision ndex, R - which is a function of ship length (Ls) and 'persons-atrisk *(N)*, such that 1000 R l- 4L, +/N± 1500 Formula 1, reg.2(c) *(N)* Actually, N = N + 2 N2, where N = Boatage; N2 = Compl. - Boatage The original intention in developing these regulations was to evaluate a ship's survivability purely on the value of it's Attained ndex. However, there is always a risk, (admittedly rather small), that an insignificant damage occurring at, (or near to), a main transverse bulkhead will lead to the loss of a ship. Administrations, understandably, are unwilling to take this risk where large numbers of people are involved, without a regulation specifically added to address this risk. This is the reason why there is a deterministically-based regulation 5. Failure to comply with regulation 5 means total noncompliance, even when it is * demonstrated that A > R. Probabilistic Criteria - Part 10

13 The Ship Stabilit" Research Centre Universitv of Strathclyde The text of Res. A.265(V) has remained unchanged since it was first introduced in the early 70's, despite the fact that certain weaknesses in these regulations have been identified. Rectifying these weaknesses will form part of the task of producing a set of 'harmonized' regulations To those who wonder - in view of the decision at MO to phase out the 'deterministic' approach, in favor of the 'probabilistic' approach - why has Res.A.265(V) been so little used? The short answer is that, whilst it has remained an equivalent/alternative to the corresponding SOLAS regulations, designers have almost invariably chosen the latter. reasons, this might be a topic of discussion later. For Probabilistic Criteria - Part 1

14 The Slip Stability Research Centre University of Strathcl-de 1.5. The (Dry) Cargo Ship Regulations - Reg.25 of Chap. 11-, Part B-, of SOLAS The (Dry) Cargo Ship Regulations were developed more than fifteen years later than. the issue of the Passenger Ship Equivalent Regulations. Originally, they applied to cargo ships of l00m in length, or above. Recently, this lower limit was dropped to 80m. The calculation procedure in these regulations is similar in outline to that given in Res.A.265(V). That is, an Attained Subdivision ndex, A, is calculated which must be at least as great as the Required Subdivision ndex, R. However, all the factors which go to make up 'A' are different! Why this should be so, will be commented on later. n addition, a new factor, 'v', has been introduced in order to give credit for watertight horizontal subdivision (above the waterline) The formula for the Attained Subdivision ndex, A, is as follows: A = F (p*s) Reg The formula for the Required Subdivision ndex, R, is as follows R = 3O TL Reg The formulae for the factors 'p', 'r', 'v' and 's' are to be found at: Regs , and , respectively The formula for the factor 's' is to be found at Reg Probabilistic Criteria - Part 12

15 The Ship Stability Research Centre University of Strathclyde 1.6 A future set of 'Harmonized' regulations, to replace the existing sets of Passenger and (Dry) Cargo Ship regulations Once it was decided to develop a single, consistent calculation procedure for the ndex, A, for both passenger and (dry) cargo ships, it was clear that, as far as possible, the factors making up the ndex should be identical for the two ship types. This was particularly the case in respect of the factors 'a', 'p' and 'r', since they had been derived from the same set of (sidecollision) data! The latest draft text of the 'harmonized' regulations contains such 'harmonized' factors. (The factor 'a' is a new proposal, being a linearisation of the damage location data. The factor 'p' is that currently used in the Cargo Ship Regulations. The factor 'r' originates from a Polish proposal made in conjunction with a proposal for the factor 'p' which was eventually adopted for the Cargo Ship Regulations.) The current thinking at MO is that credit should be given for horizontal watertight subdivision (above the waterline) for both passenger and cargo ships. respect of the factor 'v' are currently being evaluated by the SDS group. Final proposals in The present text of the draft regulations lays down two 's' factors for passenger and cargo ships. n respect of passenger ships, the 's' formula is a 'probabilistic' equivalent of SOLAS90. For cargo ships, there is a relaxation in the maximum equilibrium angle permitted, together with a GZmax value which does not need to be greater than 0.1 m The number of draughts, (and the weightings to be attached to these draughts for the purposes of calculating a 'weighted' 's'), are still to be decided. The likelihood is that three draughts will be specified, but their positioning is still uncertain. Probabilistic Criteria - Part 1 13

16 H The Ship Stability Research Centre University of Strathclyde 1.7 A 'one-off' use of a simplified version of Res.A.265 (V), in order to up-grade the stability of 'existing' RoRo passenger ships. (By means of A/ Amax calculations) At a meeting of the MO Maritime Safety Committee - MSC 59 - it was decided that all RoRo passenger ships would need to be brought up to the residual stability standards applicable to passenger ships constructed on or after April the so-called SOLAS 90 standard, which was adopted through NO resolution Res. 12(56). The time scale for such an up-grade was still to be agreed - provisionally, a maximum of five years was tabled At that same meeting, parallel discussions took place within a specialist working group concerning the development of a suitable ranking procedure of 'existing' RoRo passenger ships, in order that the difficult task of up-grading should be carried out efficiently and in an orderly and logical manner. For obvious reasons, the enhancement process would be on a 'worst-case-first' basis The over whelming majority of 'existing' RoRo passenger ships are constructed in accordance with deterministically-based regulations. Therefore, such ships are not quantified as to safety level, in contrast to the two sets of probabilistically-based regulations. Therefore, one obvious way to assess the safety level of 'existing' RoRo passenger ships would be to calculate the Attained Subdivision ndex (A), according to the provisions of Res.A.265(V). Obtaining the index A in this manner would allow MO members to rank 'existing' RoRo passenger ships in ascending order of A/R ratio MO members were acutely aware of the limited time - less than a year, the time of MSC 60 - available before a timeframe for up-grading had to be agreed and implemented. A full use of the 'A' calculation as in Res.A.265(V) was considered to be impracticable. Some simplification to the full calculation procedure was needed. Probabilistic Criteria - Part 1 14

17 HThe Ship Stability Research Centre.Uniiversity- of Strathclv de The specialist working group discussed the best way forward and decided that, (whilst preserving the essential features of the fuill calculation for 'A), a simplified calculation procedure for the Attained Subdivision ndex, A, should be developed. n brief, these simplifications can be summarised as follows: * Calculations are performed for the frill subdivision draught only. * The ship's actual KG is used for the initial' set of calcuilations. * Freight space permeabilities to be taken as * Only flooding scenarios involving one and two (adjacent) compartments are investigated for their contribution to the* 'A' value. 0 The factor's used in these calculations is taken from a formula for 's' proposed by the Russian Federation. (This factor is the 'probabilistic' equivalent of the SOLAS 90 criteria, except that GZmax need not be greater than 0. 1 in.) This simplified method of calculating the Attained Subdivision ndex, A, was issued as an MSC circular - MSC/Circ * Two separate sets of computer calculations are to be made for each ship investigated. The first get are made using the ship's actual KG and produces an index, A. The second set are made, using the same flooding scenarios, but employing a notional ship KG! This KG is purely theoretical, and is derived from a study of the first set of calculations in the following manner, The worst damage case is determined by locating the case with the lowest 's' value. The (notional) drop in ship KG which would involve marginal compliance only with the residual * criteria is established; this resulting notional KG is then used in a second set of calculations to produce an index, Amnax. *For the second set of calculations, see 7.6. Probabilistic Criteria - Part 1 15

18 The Ship Stability Research Centre University of Strathclyde The A/Amax ratios for the RoRo ships involved in the world-wide computer exercise were then ranked in ascending ratio order. Obviously, those having the lower ratios were to be up-graded first This computer exercise was always intended as a 'one-off' operation and the circular, MSC/Circ.574 was specifically issued for this one purpose. There seems to be a persistent feeling among some that the ratio A/Amax represents a kind of agreed standard. Perhaps these few comments will make the real position more clear Probabilistic Criteria - Part 1 16

19 The Ship Stability Research Centre University of Strathclyde The use of the Probabilistic concept in aqsessiing the stability standards of 'existini, passenzer ships. Minimum acceptable standards of residual stability for nassenger ships, prior to the introduction of the provisions ofmsc. 12(56). 1.1 n respect of minimum residual stabilit, standards for passenger ships- prior to the wording in the SOLAS regulations is vague and easily misinterpreted n fact. the residual stabilitv criteria. (GZmax. range and area), applied by individual Administrations appear to vary wildly. 1.2 n the early 80's, the UK Administration developed a set of criteria, (usually referred to as the STAB 80 standard).. The minimum requirements in respect of GZmax. range and area represent a significant enhancement compared to the standards which have been acceptable by many other Administrations. There appear to be examples where positive residual GM alone was deemed to be sufficient n view of this very unsatisfactory position members considered the matter in some depth and eventually decided to introduce a significant improvement in this section of the SOLAS regulations. The agreed text of this SOLAS amendment relates to all passenger ships constructed on. or after. April This amendment was adopted per resolution MSC. 12(56): it is commonly referred to as the SOLAS 90 standard 2 The concerns of the UK regarding the residual stabilitv standards of existing. ro-ro passenger ships. 2.1 During the extended discussion period at 1410 on this topic. the tragic accident involving the "Herald of Free Enterprise" occurred Opinion in the UK at this time was that all ro-ro passenger ships should be able to comply with the residual stability standard currently being proposed for 'new' passenger ships. (This standard - with only rather minor modifications -is essentially the SOLAS 90- standard). Probabilistic Criteria - Part 1 17

20 i The Ship Stability Research Centre University of Strathclvde i i 2.2 A computer study, involving a sample of existing. ro-ro passenger ships. showed quite clearly that it was unlikely that a ro-roferry built prior to 1990 would survive damage to a vulnerable part of the ship. in normal sea conditions - i.e. other than in calm water. 2.3 The Steering Committee. commissioned to oversee the UK ro-ro passenger ship research programme. recommended that MO members be persuaded to adopt a significantly higher standard of residual stabilitv for 'existing'. ro-ro passenger ships in as short a timeframe as was considered practicable. Parallel efforts were to be made with our EUpartners. in order to develop a regional solution acceptable to the UK. 3.3 A description of the method adonted by MO to up-grade the stability standards of existing' ro-ro passenger ships to those contained in MSC. 12(56) - the so-called SOLAS 90 standard. 3.1 The specialist working group dealing with (damaged) stabilitv matters was given the task of developing an acceptable method of ranking the respective safety levels of existing, ro-ro passenger ships. They favoured the use of probabilistic methods, and considered that Res.A.265(V) should form the essential basis of such a method, by calculating the attained subdivision index, A. (Al is generally regarded as an objective measure of the level of safety achieved). 3.2 Because of the ver restricted time frame within which a worldwide computer exercise was to be completed. a fidl application of the provisions of Res.A. 265(V1) was considered impractical: accordingly, a number of simplifications were made which, (whilst retaining the essential merits of the fidl A. 265 calculation procedure). considerably reduced the number of calculations required 3.3 These simplifications involve * the use of the deepest subdivision draught only: * ignoring the trim range, by using design trim only: * GZmax to be up to O.lm. (The number of persons on board not to be taken into account). Probabilistic Criteria - Part 18

21 T1he Ship Stability Research Centre University of Strathclyde 3-4 The 'probability of survival' is represented in probabilistic regulations by the.factor "s". For the simplified calculation procedure. this factor is given by a finction of GZmax. range and area. 3.5 The agreed simplified calculation procedure is contained in the annex to MSC Circ.574: it was used in a world-wide computer exercise to approximate to the safety levels achieved by current roro passenger ships. n order to achieve an objective ranking.for the purpose of establishing the order of up-grading, the use of a ratio A/Amax was proposed in this MO circular. A brief description of the significance of this ratio follows. 3.6 Using the operational ship KG.for the deepest subdivision drought condition. an attained index. A. is calculated, such that: A = Al + A2. (Contributions from all single and two adjacent compartments. respectively) 3.7 Clearly, all ro-ro s which fidly complied with the SOLAS 90 standard were excluded from the up-grading procedure. However. in all other cases there will be at least one critical damage case where "s', is less than unity - bearing in mind the note marked * in 3. 4 above. For a (deterministic) 6ne-compartment standard ship. this critical case will be for a single compartment damage. For a (deterministic) two-compartment standard ship. this critical case will be for a damage to two adjacent compartments. 3.8 There is a theoretical ship KG position which, when applied to the critical damage case. which will correspond to an 's ". factor of unit: this implies compliance with SOLAS 90. "relaxed" as described at 3.6. This notional KG is then used to re-calculate the same damage scenarios referred to at 3.6. in order to produce a new index, labeled A max. For a fixed arrangement of internal watertight boundaries, this Amax is the maximum possible. Therefore. Amax = Amax] + Amax2 (Compare with the identity at 3.6) Note that GZ max is a fixed value m - and is not related to the persons on board. Probabilistic Criteria - Pan 19

22 The Ship Stability Research Centre University of Strathclvde 3.9 The basic reasoning is then that the ratio A-Amax. being linked directly with the ship KG. provides an objective measure of the relative safety levels of ro-ro *s When MSCCirc was.first released this ratio was stated to be given by the ratio: A Al + A2 A max A max + A max Subsequent studies revealed that the use of this ratio in the cases involving one-compartment standard ships tended to penalise such ships. Therefore.- it was decided to apply the following ratio for one-compartment standard ships A A max A1±+A2 A max l + A2 Probabilistic Criteria - Part 1 20

23 4 The Ship Stability Research Centre Universijv of Strathclyde Regional and nternational Agreements concerning the application of the AAmax ratio to the up-grading of 'existing" ro-ro Passenger ships. 4.1 World-wide solution agreed at MO Despite difficult and protracted discussions with other MO members. the UK were unable to achieve their declared objective -which was the up-grading of the stability of all ro-ro passenger ships. worldwide. to the SOLAS 90 standard in the shortest possible time scale. (The SOLAS 90 standard applies to passenger ships. including ro-ro passenger ships. constructed on or after 29 April and is given in LWO res.msc. 12(56). Nevertheless. the arguments put forward byv the UK in favour of the earlv adoption of higher stabiliy standards for such ships were generally accepted as valid by 1AO members: the particular proposal that this higher standard should be the SOLAS 90 standard was originally accepted, in principle at the Maritime Safey Committee. At a later MSC meeting, however, the majority of members decided that fidl compliance with the "new "passenger ship standard was too stringent. For a fuidl text of the SOLAS amendment. see res.msc. 26(60). Summarising the MO solution, to which the UK entered a formal objection. 'existing' ro-ro passenger ships i.e. those built prior to April were to be up-graded. where necessary, within an agreed time-scale so as to achieve a standard at least as high as follows Final flooding stage GZmax to be not less than 0. 09m: A range of 15 deg. or more. although a lesser range may be acceptable to an Administration: Area of the residual stability curve to be at least m.rad. ntermediate flooding stages GZmax to be not less than 0. 05m: A range of at least 7 deg. Probabilistic Critcria - Part 21

24 The Ship Stability Research Centre University of Strathclvde A schedule of up-grading is specified in the SOLAS amendment, such that those ro-ro passenger ships having A.Amax ratios up to 70% need to be up-graded b, October: Where the A-Amax ratio is at least 95%. this-new regulation states that an zip-grade is not necessary. This latter provision was totally unacceptable to the UK. since it's effect is that a significant proportion of existing. ro-ro passenger ships. (having. in some cases a stability standard substantiallv worse than the SOLAS. standard), will be permitted to continue trading indefinitelv without any upgrade whatsoever The complete table of scheduling dates is Value of AJA max Date of compliance Less than 70% 1 Oct % or more but less.than 75% 1 Oct % or more but less than 85% 1 Oct % or more but less than 90% 1 Oct % or more but less than 95% 1 Oct % or more need not comply Probabilistic Criteria - Pail 1 22

25 ! The Ship Stability Research Centre University of Strathclvde 4.2 A Regional Solution agreed by the European countries Whilst efforts were being made within the MO.forum to convince other Administrations to adopt SOLAS 90 standards.for all ro-ro passenger ships within an agreed time frame, the UK were making parallel efforts to come to a suitable agreement with the European maritime nations' Comparatively recently, an agreement was achieved with 13 other European countries. whereby all existing. ro-ro passenger ships will need to be up-graded to the SOLAS 90 standard. (or it's agreed 'fiture equivalent), in a specified time scale. A significant difference in this scheduling compared to the 1MO solution is that there are two schedules -given - one for one compartment standard ships and the other for two compartment standard ships. These regulations are contained in S no entitled : the Merchant Shipping (Ro-Ro Passenger Ship Survivability) (No.2) Regulations A copy of the Agreement between the 13 countries and the UK. together with the designated area within which it is to apply and details of the agreed stability standard are given in Merchant Shipping Notice no issued in March, Time scale for compliance Value of A/Amax Date for compliance one-compt. std Two-compt. std. less than 70% 1 Oct Oct % or more but less than 75% 1 Dec Dec % or more but less than 85% 1 Dec Dec % or more but less than 90% 1 Dec ]Dec % or more but less than 97% 1 Dec Dec % or more 1 Dec Dec Probabilistic Criteria - Part 1 23

26 The Ship Stabilit, Research Centre University of Strathclyde 1.8 Compartmental Standard The concept of 'compartmental standard' is introduced at regulation 8.1 of Part B,. Chapter 11-1, of SOLAS. This is linked to the notion of a 'statutory' compartment, which has been based upon a statutory maximum extent of damage, as defined at regulation 8.4. t is to be demonstrated that a ship has sufficient residual buoyancy/stability to withstand the effects of flooding to the extent as specified in reg This deterministic approach to evaluating a damaged ship's ability to survive is therefore linked to the notion of'a 'statutory-compartment'. On the other hand, the probabilistic approach is intended to give credit for all transverse subdivision, irrespective of the distance between adjacent bulkheads All Authorities are aware of the dangers involved in damage at, or near to, a main transverse bulkhead. The ensuing wider extent of flooding would greatly increase the probability of serious flooding, probably leading to the loss of the ship, even in quite calm weather conditions. Where there is the potential for heavy loss of life, the regulations specify that a two-compartment standard is to be complied with The outcome of these understandable worries, as far as the Passenger Ship Equivalent Regulations are concerned, was the addition of (deterministic) regulation5. This corresponds to regulation 8. 1, Pan B, chapter - of SOLAS Note that there is no equivalent to reg.5 of Res.A.265 in the (Dry) Cargo Ship Regulations. Since the problem of"nminor" damage still remains, how was this to be resolved? The present regulations make no provision for this. However, the matter will be dealt with when the two present sets of probabilistically-based regulations are hannonized. Probabilistic Criteria - Part 1 24

27 The Ship Stabilit" Research Centre Universit, of Strathclyde The present text of the proposed 'harmonized' regulations makes provision for 'minor' damage. A defined extent of damage -including penetration inboard- is applied anywhere to the ship's side (and bottom). t is to be demonstrated that the ship will survive. Probabilistic Criteria - Part 1 25

28

29 The Ship Stability Research Centre University of Strathclvde * Part 2 Theoretical Background 1. A List of those present, (and future), regulations which are based upon, (or at least contain an element of), probabilistic principles. U.1 Passenger Ship Equivalent Regulations - 4MO Resolution A.265 (V).2 (Dry) Cargo Ship Regulations - SOLAS Chap.1-1, Part B-1, Reg.25. Still-developing regulations which will replace and.2.4 A / Amax ratio, used in a ranking procedure for up grading 'existing' RoRo passenger ships. See MSC circular, MSC/Circ What is the essential principal involved in making up any set of probabilistically- * based regulations? 3. The growing disquiet at MO at the continued use of deterministically-based regulations in assessing minimum levels of subdivision/stability. 4. The paper - "Subdivision of Ships", presented by Prof. K.Wendel to SNAME in A brief review. 5. The use of collision casualty data in developing a suitable method of assessing appropriate levels of subdivision/stability, employing probabilistic principles. 6. What principal conclusions were drawn from an analysis of the collision data? 7. Determination of a factor 'p', which is intended to estimate the probability of longitudinal extent of damage. 8. Determination of a factor 'r, which is intended to estimate the probability of penetration inboard. 9. How the 'probability of survival' is influenced by the ship condition - i.e. mean draught, trim, ship KG - at the time of collision. 10. How the 'probability of survival' is influenced by the position and extent of the space(s) damaged, and the contents within the space(s). Sby 11. How the 'probability of survival' is influenced, very significantly, the sea state/weather conditions at the time of collision. Probabilistic Criteria - Part 2 26

30 The Ship Stability Research Centre University of Strathclyde 12. By taking due account of.7 to.11, how a factor's' may be determined, which is intended to be an estimate of the 'probability of survival'. 13. The factor 'v' - statutory maximum extent of damage. V 14. How to calculate the Attained Subdivision ndex, A. 15. dealised Damage Distribution along a Ship's Length The Factor "a" 16. The Treatment of Simultaneous Flooding

31 The Ship Stability Res earch Ccntre University of Strathcivde 1. A List of those present, (and future), regulations which are based upon, (or at least contain an element of), probabilistic principles. 1.1 Passenger Ship Equivalent Regulations - NMO Resolution A.265 (V) 1.2 (Dry) Cargo Ship Regulations - SOLAS Chap.l-1, Part B-1, Reg Still-developing regulations which will replace. 1 and A / Amax ratio, used in a ranking procedure for up-grading 'existing' RoRo passenger ships. See MSC circular, MSC/Circ.574. Probabilistic Criteria - Part 2 28

32

33 UThe Ship Stability Research Centre University of Strathclyde 1.1 The Passenger Ship (Equivalent) Regulations. Res.A.265(V). Abrief resume of the First intern atio nally-agreed set of safety regulations setting out minimum acceptable levels of subdivision / residual stability.-.x H The main purpose of this set of regulations is to indicate the calculation procedure for determining the Attained Subdivision ndex, A, and to demonstrate that this is not less than the Required Subdivision ndex, R. A short summary of the layout of these regulations:. Reg. idefinitions 2. Reg. 2. Required Subdivision ndex, R 4. 3 Reg. 3 Special Rules regarding Subdivision Reg. 4 Permeability Reg. 5 Subdivision and Damage Stability (Deterministic Part) Reg. 6 Attained Subdivision ndex, A 7. (Formulae for "a" "p" and "S') Reg. 7 Combined Longitudinal and Transverse Subdivision (Formulae for "r") 8. Reg. 8 Stability nformation 9. Reg. 10 Peak and Machinery Bulkheads, Shaft Tunnels, etc. 10. Reg. Double Bottoms 11. Regs. 13 to 20 All these regulations concern the maintenance of watertight integrity 12. Regs. 9,12, 21 Sundry regulations Probabilistic Criteria - Par 2 29

34 The Ship Stability Research Centre University of Stratliclvde 2.2 What are the princip al factors to be taken into considerations when making up any set or pro bab ilistically- based regulations?p f a damaged ship - that is, a ship which is breached, allowing flood water to enter - is to remain afloat in an upright, or nearly upright, attitude, then two essential conditions have to be met. Firstly, unless the loss of buoyancy of the space flooded is at least in balance with the remaining (i.e. residual) buoyancy, then total sinkage is inevitable, even in calm water. The second condition is fulfilled if a ship retains sufficient righting ability after deducting the effect of the flooded spaces. A complete solution to this complex problem - irregular seas, wave direction and length, trapped floodwater, details of the flooded space, etc. - is clearly unattainable. However, it is generally accepted that there is a direct correlation between the characteristics of a still-water statical residual stability lever curve for a damaged ship and it's ability to resist capsize. Therefore, if it can be demonstrated that the residual stability lever curve is adequate - which may be interpreted as the ability to survive in a reasonable sea state - then the second condition is assumed to be met The first internationally-agreed statutory minimum standard of subdivision for passenger ships is to be found in the 1929 Safety Convention. n essence, the procedure for determining this minimum standard has remained the same since that time. Terms such as 'floodable length', 'permissible length', 'factor of subdivision', 'criterion of service numeral', have been in common usage for the last 70 years in the application of the Factorial Method to decide the positioning of bulkheads on passenger ships. Note that this method was only intended to deal with the first of the two conditions mentioned previously. Why should that be? The manner in which the 'Titanic' was lost is crucial in answering this question. The great trans-atlantic liners of that age were all very similar in size and proportions. For an assumed, but realistic, maximum damaged length, the relatively slender proportions of these liners - together with significant asymmetric flooding be rather unlikely - meant that there was quite a low probability of such a damaged ship being lost due to capsize. Sinkage was the much more likely cause of loss. At the time of the 1929 Convention, maximum statutory limits for assumed damage extent were Probabilistic Criteria - Part 2 30

35 The Ship Stability Research Centre.University of Strathclvde O0.02L (in) and B/5, for longitudinal extent and penetration inboard, respectively. The corresponding present day values are O.03L (in) and B/ The obvious dangers associated with damage occurring at, or near to, a main transverse bulkhead have been recognized as far back as the 1914 Convention. t was stated that all passenger ships were to be capable of floating - that is, to retain some residual buoyancy - after the flooding of any single compartment. Where such a ship has a factor of subdivision of 0.5 or less, then it has to be shown that she remains afloat with two adjacent compartments flooded; in the case of a factor of subdivision of 0.33 or less, the ability to remain afloat with three adjacent compartments flooded is to be demonstrated Passenger ships have tended to become more 'beamy' since the days of the great transatlantic liners. Therefore, the second of the two conditions has gradually become the over-riding one when deciding on appropriate minimum levels of subdivision. Quite low standards of residual stability (for passenger sh~ips) were introduced for the first time at the 1948 Safety Convention. The present Convention (SOLAS 74 + amendments) contains the socalled SOLAS 90 standards for passenger ships constructed after April Nowadays, although the minimum spacing between adjacent bulkheads is stiff governed by 0.03L (in), the acceptance, or otherwise, of a passenger ship design is determined by compliance, or non-compliance, with the residual stability criteria. Because of the statutory limitation of the penetration inboard, please note that a designer need not investigate those cases involving penetration inside B/ Another concept, which is still applicable to the present day, is that of a 'compartmental standard'. The logic behind this concept may be summarised as follows- Consider a high-density passenger ship which is damaged on, or near to, a main transverse bulkhead. t is unacceptable, in the event of such an accident occurring, that there should be a high probability that the ship will capsize. Accordingly, where large passenger numbers are Probabilistic Criteria - Part 2 31

36 The Ship Stabilitn Research Centre University of Strathclvde involved, it needs to be shown that there is sufficient residual stability for all possible damage cases involving two adjacent compartments Serious flooding may arise when a breach in the hull of a ship occurs as a result of * extreme fore end damage, due to 'ramming' another ship; a bottom / stranding damage; * "scraping" damage to a ship's bottom or shell; * side-collision damage clear of a main transverse bulkhead;. in way of a main transverse bulkhead As we all know only too well, there is one further mode of flooding which is not included. - n respect of RoRo types, such as the "Herald of Free Enterprise" and "Estonia", there are the potentially-disastrous effects of bow-door(s) failure brought about by serious mal-function or mal-operation. The failure-modes fisted above have one feature in common - the damages are all caused by strildng - or being struck by - another ship or fixed object. t is assumed that the damaged ship will be brought to rest as soon as possible after collision. n both disasters, the ships were travelling at significant forward speed when they capsized. n addition, the extremely rapid build-up of floodwater on the vehicle decks of both RoRo's was different in character to that which could be expected after a side-corision damage. n the latter case, a varying "water-in / water-out" situation develops, which appears to govern both the probability and actual onset of 'capsize. t is quite impractical to attempt to set a residual stability standard for flooding cases similar to the "Herald of Free Enterprise" and "Estonia". A comprehensive set of procedures, (considered to be 100% effective), including 'adequate local strength', 'sufficient door redundancy' & 'separation, 'fail-safe' means of closure and positive means of reporting, are all necessary to avoid future disasters of this nature. Raising the residual stability standards is pointless, since the water inflow in a future incident is completely unspecified. All that can be said is that, undoubtedly, considerable amounts of water will be taken aboard a RoRo under way with it's bow doors open and - unlike the side collision case - NO outflow! Probabilistic Criteria - Part 2 32

37 The Ship Stability Research Centre University of Strathciyde 2.3 The growing dis-quiet at MO regarding the continued use of l deterministically-based regulations in assessing statutory minimum levels of subdivision / stability By the 60's, MO (MCO) members were increasingly concerned that the existing SOLAS regulations relating to minimum subdivision/stability standards - for passenger ships - had become obsolescent and were in need of a complete overhaul After much discussion and debate, it was decided to introduce an alternative set of regulations: " Regulations on Subdivision and Stability of Passenger Ships ". These are contained in a (blue) booklet, together with a complete set of explanatory notes concerning their derivation and purpose t is recommended that those wishing to learn more about probabilistically-based regulations should study these notes, in order to be aware of the basic thinking behind them. An extract from these notes - pp 42 & 43 - follows on next pages 33, 34. Probabilistic Criteria - Part 2 33

38 The Ship Stability Research Centre University of Slrathclyde n after many preliminary studies, the Sub-Committee established the Ad Hoc Group.for the preparation of new Regulations. Since establishment of the Sub-Committee. and including the Ad Hoc Group. more than 25 -\- international meetings have been held and more than 150 relevant papers have been considered. The background theoretical and experimental work has been extensive and has involved a high degree of co-operation among countries both in undertaking this work and in the discussions within the Sub- Committee. t ma also be said that the new Regulations reflect a -considerable degree of contribution from all members of the Sub-Committee. n preparing the new Subdivision Regulations the following main premises were taken as a basis: - The level of safeiy provided by the new Regulations should, in general be similar to that given by the 1960 Safety Convention. - The degree of safety should increase with the ship length and the total number of persons which the ship is certified to carr. - As a criterion of the degree ofsafe t an index of subdivision which is a measure of the ship's abilitv to survive after damage should be used. This index should reflect the effect of bulkhead spacing, stability, and other features relevant to this abiliy'. - Since a purely probabilistic approach to evaluation of safety would lead. in some cases, to design of ships with unacceptable vulnerability in some part of their length. the Regulations should include specific minimum standards for compartmentation and damage stability leading to approximately the same degree of safety in these respects, as is provided by the Regulations of the 1960 Safety Convention. The most important and principal distinction of the new Regulations is the use of the probabilistic approach. As it is well known, many factors actually affecting the final consequences of ship hull damage are random and their influence is different for ships with different characteristics. For example. it is obvious that in ships of similar size carrying different amounts of cargo, quite similar damages may lead to different results because of differences in the range of permeability and draught during service. Due to this fact, the effect of damage to a ship with given watertight subdivision depends on the following circumstances: - which particular compartment or group of adjacent compartments is flooded: - the draught and intact stabilitv at the time of damage: - the permeability ofaffected spaces at the time of damage: Probabilistic Criteria - Part 2 34

39 The Ship Stability Research Centre University of Strathclyde - the sea state at the moment of'damage: -- other.factors such as possible heeling moments due to unsymmetrical weights. Some of these circumstances are interdependent and the relationship between them and their effects may vary in different cases. For these reasons and because of mathematical complexity as well as insjfficient data. it is not practicable to make an exact or direct assessment of their effect on the probability that a particular ship will survive a damage if it occurs. However. accepting some approximations or qualitative judgments. a logical treatment may be achieved by using the probability approach as the basis of a comparative method for the assessment and regulation of ship safety. The probability of ship's survival includes the.following probabilities: - the probability offlooding each single compartment and each possible group of two or more adjacent compartments: - the probability that the residual buoyancy after flooding of a compartment (or a group of two or more adjacent compartments) under consideration will be sufficient to provide for survival: - the probability that the stability after flooding a compartment (or a group of two or more adjacent compartments) will be sufficient to prevent capsizing or dangerous heeling due to loss of stability or to heeling moment t may be demonstrated by means of the probability theory that the probability of ship survival should be calculated as a sum of probabilities of her survival after flooding each single compartment. each group of two, three etc. adjacent compartments multiplied, respectively, by the probabilities of such damages as lead to the flooding of the corresponding compartment or group of compartments. Probabilistic Criteria - Part 2 35

40 The Ship Stability Research Centre University of Strathclvde Why decide Levels of Subdivision on a Probabilistically-based Method? 1. Earlier, there are allusions to a growing need for the minimum allowable level of subdivision (passenger ships) to be founded upon a probabilistically-based method. The principal reason for this is that the current SOLAS deterministically-based regulations incorporate a (fixed) maximum statutory damage length Ls (m), together with a (fixed) maximum statutory penetration inboard - B / 5. Ls is the subdivision length B is the maximum beam 2. To illustrate the reason why the use of a fixed damage extent is fundamentally flawed, consider the design of a one-compartment-standard passenger ship, 1 00m in length. Assume that the basis (first) design is subdivided into 16 compartments of equal length - therefore, each 6.25m long. Further assume that the subdivision draught chosen is such that the critical damage - i.e. critical compliance with the residual stability criteria (no margin) - applies to at least one of the ship's compartments. Now consider an alternative (second) design where an extra compartment is introduced, such that there are 15 equal-length compartments, plus two end compartments as in the first design. The non-end compartments are therefore all 5.83m long. 3. The basis design complies (critically) with the SOLAS subdivision requirements. The second design does not comply, since the distance between adjacent bulkheads is (marginally) below the statutory damage length of 6m for a design of this length. This statutory length has always been no more than an inspired "guesstimate"; it could, just as easily, have been made greater or smaller - kl * Ls + k2, in general. 4. A more flexible way of estimating subdivision levels is needed urgently, in order to place watertight divisions in their optimum positions. An objective calculation procedure, Probabilistic Criteria - Part 2 36

41 The Ship Stability Research Centre University of Strathclvde- (which estimates the proportion of all possible damages a ship is likely to survive), is a clear necessity. quantitative comparison of two similar designs is not possible. However, it is possible to 5. Because of the pass non-pass nature of deterministically-based regulations, a assess the two designs in a qualitative manner, thus 5.1 A damage of up to 5.83m in length, not involving damage to a bulkhead. Clearly, the 2 nd design is slightly better. 5.2 A damage of up to 5.83m in leng-th, involving damage to a bulkhead. The 2 J'd design is at least as good as the ' 5.3 A damage of more than 6.25m, involving flooding of two (or even three) adjacent compartments., The 2 nd design is at least as good as the 1 'Z, and maybe slightly better where some two-compartment damages in the forward part of the ship indicate a higher probability of survival. 5.4 Damages between 5.83mn and 6.25m in length. When damage does not include a bulkhead, the ' design is better; however, when two-compartment flooding is considered, the 2nd design is at least as good as the ". This, (admittedly qualitative), analysis highlights the need for an objective calculation procedure that quantifies the level of subdivision attained by a particular ship design. Probabilistic Criteria - Pant 2 37

42 The Ship Stability Research Centre University of Strathclvde 2.4 The paper - "Subdivision of Ships", presented by Prof. K.Wendel to SNAME in A brief review For more than 30 years, increasing numbers of members of the nternational Maritime Organisation (MG) have expressed disquiet concerning those SOLAS regulations which specify the minimum statutory subdivision standards for passenger ships- Post-"Titanic", these regulations have been based upon the provision of adequate reserve buoyancy after assumed flooding. Comparatively recently, additional requirements have been included to ensure that a reasonable level of residual stability is also provided. For passenger ships (monohulls) of normal proportions, the specified standards of residual stability generally govern the placement of the main transverse bulkheads, particularly in the midship region n the late 60's, Prof K.Wendel (Hannover/Hamburg Universities) outlined his ideas on the "Subdivision of Ships" in a paper to SNANE. Members at MO, (then the ntergovernmental Maritime Consultative Organisation, MCO), who had been assigned the task of improving the then-existing subdivision regulations, were favorably influenced by the arguments contained in this paper- Future work by a specialist MP group used the basic ideas expressed therein to develop a new set of subdivision regulations. The outcome of this work was the issue of the "Equivalent Passenger Ship Regulations", commonly referred to as Resolution A The basic principles embodied in the paper are quite simply expressed. Two main assumptions are vital to the calculation procedure outlined in the paper. n cases of side collision damage, which involves internal flooding of a ship: - There is a probability, say "p", that the flooding is confined to a particular compartment, (or group of compartments) (0 < p <=1) Probabilistic Criteria - Part 2 38

43 The Ship Stability Research Centre University of Strathclyde - There is a conditional probability, say "s", for flooding as described above, that the 'probability-of-survival' can be estimated. (0 < s <=) Then, by aggregating all the products (p * s), for all possible damage cases over the full draught service range, this sum of products may be termed: the Attained Subdivision ndex, (A) Expressed as an identity A = Z (p*s) ; for all compartments (& compartment groups) over the entire draught service range t needs to be emphasized that the ndex A is intended to be a measure of side collision damage! Historically, protection against bottom or grounding damage has always been achieved by providing a double bottom of a specified minimum height. Additionally, protection against damage to the extreme fore end, (if the ship is the ramming" one in a collision incident), has always been deterministically-based. The empirical approach to bottom and extreme end protection is still generally regarded as the optimum way to provide local subdivision in these areas of a ship For a given ship design, the factor "p" may be regarded as non-draught-dependent; a specified arrangement of internal watertight boundaries uniquely defines the "p" factor for each compartment (or compartment group). On the other hand, the factor "s" is highly dependent on draught, as a study of residual stability lever curves over the draught service range clearly demonstrates Damage to the side shell may be adequately defined by the following parameters: Centre of damage position, relative to the aft end, (say) related to a factor "a" Probabilistic Criteria - Part 2 39

44 The Ship Stability Research Centre University of Strathclvde Longitudinal extent of damage Damage penetration inboard Upper limit of damage above baseline, (say) related to a factor "p" related to a factor "r" related to a factor "v" The first three factors above - "a", "p" and "Y'- are non-draught-dependent The factor "v", is intended to represent the proportion of an damages that will be confined to a particular compartment (or compartment group), when a upper watertight boundary is considered to be fully effective. This factor, therefore, is best considered in conjunction with the factor "s". Probabilistic Criteria - Parn 2 40

45 The Ship Stability Research Centrc Univ ersity of Strathiclyde 2.5 The use of collision c asualty data in developing a suitable method of assessing appropriate levels of subdivision/stability, employing probabilistic principles. The Collection of Collision Data The Definition of 'Statutory Damage' U2.5.1 f a statistically valid sample of historic side-collision data can be assembled, then it is reasonable to derive an idealized cumulative density function for the probability of damage occurring between aft and forward boundaries. The principal factor, ("p' say), relates to damage length only in the case of passenger ships. The reduction factor, ('r' say), relates to the penetration inboard. ship-side ; r =1 centerline) MCO issued a Damage Card to obtain the raw data, which comprises: The damage dimensions and it's location relative to a datum. The sea state at the time when the collision incident occurred The data sample to derive the "p' and 'r' factors consisted of 296 cases of a rammed ship. The damage particulars were represented by the following parameters: length Ls Ship breadth B *Damage location (from AT to centre of damage) x Damage length y *Damage penetration inboard z Probabilistic Criteria - Pant 2 41

46 The Ship Stability Research Centre University of Strathclyde Only the side-structural arrangements of the rammed ship were considered to be significant in deciding on the "p" and "r" factors; -all other elements connected with each case were considered to be random. U Since 'vertical extent of damage' is not recorded on the Damage Cards, side collision * damage extent is essentially defined by the three parameters: x y z centre of the damage location: longitudinal extent: penetration inboard. However, any set of regulations concerning minimum levels of subdivision residual stability requires, in addition, that the assumed damage shape is rigidly and unambiguously defined. Otherwise, uniform interpretation of the regulations is unlikely to be achieved Statutory Damage - for the deterministically-based SOLAS regulations - assumes a shape such that the aft and forward boundaries are plane, vertical and normal to the ship centerline, whilst the 'damage' front is also plane and vertical, with the surface parallel to the subdivision waterline. n reality, because of the deterministic nature of the regulations, only a penetration inboard of precisely B/5 is significant. Longitudinal watertight boundaries outboard of the B/5 line are automatically assumed to be damaged, whilst those inboard of the B/5 line - for statutory purposes - are to be taken as intact To penalise the fitting of a main watertight transverse boundary which is 'stepped' - i.e. is 'recessed' outboard of B/5 - the normal SOLAS requirements entail that 'damage' is to be assumed such that the compartments either side of that 'step' are to be considered as flooded. The SOLAS regulations are also based upon a concept of a maximum statutory assumed damage length Ls + 3.0, in metres. t follows that where adjacent transverse boundaries Probabilistic Criteria - Part 2 42

47 The Ship Stability Research Centre University of Strathclyde * boundaries. are closer together than this, all damage scenarios must include flooding on both sides of such When probabilistically-based regulations are considered, the 'damage shape' is decided upon similar grounds to those of paragraphs 5 and 6. Clearly, however, the concept of statutory maximum longitudinal extent and penetration inboard needs a rethink. n respect of longitudinal extent, this maximum extent becomes X,. - i.e. the (non-dimensional) damage length, which is very unlikely to be exceeded, after a study was made of the collision data. n respect of penetration inboard, analysis of the same data suggests that, (with no significant error), the penetration may be taken as limited to the ship centerline How are those cases dealt with where a longitudinal watertight boundary is not parallel to the subdivision waterline? The (Dry) Cargo Ship Regulations contain the parameter 'b', which is the mean (average) penetration inboard in the case of damage to a single compartment - or compartment group. 'For the correct interpretation ofb' the values to be used with various boundary configurations, it is recommended that the illustrations given in figure A-7, Appendix 2 of Annex H of the Merchant Shipping Notice M are consulted - the guidance contained in the Notice was issued in respect of the regulations issued as MSC Resolution, MSC. 19(58). Note that, once the decision is made concerning the appropriate 'b' to be used for an assumed single compartment damage, henceforth the corresponding longitudinal boundaries are - for the purpose of calculating the reduction factors - assumed to be parallel to the subdivision waterline. When estimating the mean 'b' for two (or more) adjacent compartments, the minimum 'b' for the group is to be used - ignoring those compartments having no longitudinal division to the damage-side of the ship centerline. Probabilistic Critcria - Part 2 43

48 The Ship Stability Research Centre University of Strathclyde 2.6 What principal conclusions were drawn from an analysis of the collision data? The Original Analysis of Collision Data The parameters specifying the damage extent are non-dimensionalised such that Non-dim. Ship length Ls Ship breadth B Damage location x x/ls Damage length y y/ls Damage penetration z z/b Analysis of the n/d collision data leads to the following conclusions: c There is no significant change in damage length, (except at the ends), for a change in damage location. End damages will be discussed later..2 Provided that ship length does not exceed 200m, damage length is approximately independent of ship length. (Lack of data for greater ship lengths precludes any conclusions being drawn for such ships.).3 Analysis of the location data reveals a significant bias towards damage occurring in the forward half of the ship. Further consideration later.4 n respect of penetration inboard, there is no significant dependence of nondimensional penetration, ( z / B ), on ship breadth, (B).5 The ratios z/b and y/ls are highly correlated. (i.e. A relatively long damage is usually associated with a relatively deep penetration.).6 t can be assumed, with very little error, that all penetrations lie between the ship-side and the centerline..7 There is a rather low probability that damage length will exceed 0.24 Ls. (n absolute terms, 48m.) Probabilistic Criteria - Part 2 44

49 The Ship Stability Research Centre University of Strathclyde n the analysis carried out prior to the adoption of Res. A. 265, 'damage location' - see item.3 of the previous paragraph - was plotted in the form of a histogram. A conversion was then made to an idealised distribution - see Reg.6(b), formula having the following characteristics : '. - Starting from a density of 0.4 at the aft end, then a linear rise to mid-length (density 1.2); - Constant density over the forward part i.e. density = 1.2 For a more thorough treatment of the derivation of the factor 'a', see -later.

50 The Ship Stability Research Centre Univecrsity of Strathciyde n the original analysis of'damage extent', note that 'extent' was separated from 'damage location', (which was already discussed in the previous paragraph.) Bearing in mind the items.1,.2.6 and.7 of the previous paragraph, the data relating to damage extent' were idealised into two formulae - see Reg. 6(c), formulae TV..x. p =W (1 ) ( * 1/?) where /X <= 0.24 p = W (1.072 *1/? ) 1/%X> 0.24 n the above: W = X = Ls ; where Ls <= 200 W = 184/(Ls - 16) X = 200 Ls > A reduction factor 'r' needs to be introduced in those cases where penetration inboard falls short of the ship centerline - see items..4 and.5 of the previous paragraph. The idealised distribution is given at Reg.7, para.(b)(ii) and is represented as a function of n/d compartment length and penetration. Probabilistic Criteria - Part 2 46

51 The Ship Stability Research Centre University of Strathclyde 2.7 Determination of a factor p, which is intended to estimate the probability of longitudinal extent of damage. * Transverse Subdivision Only Any method of assessing a ship's level of subdivision needs to be practical, and therefore should allow for local subdivision by transverse, longitudinal and horizontal watertight boundaries. (The latter, only when significantly above the draught position!) Nevertheless, the basic reasoning behind the probabilistically-based calculation procedure is the same when applied to a ship having plane, side-to-side, transverse bulkheads only, extending from baseline to upper deck level (figure A). Figure 1. _=_ L = _-. "rtei P 2-4 Probabilistic Criteria - Pant 2 47

52 The Ship Stability Research Centre University of Strathclyde For the example illustrated, there are a total of 15 possible damage scenarios, provided that all damages are considered to be continuous and unbroken. There are 5 cases of single compartment damages, 4 comprising two adjacent compartments, 3 for three adjacent compartments, 2 for four adjacent compartments and for all five compartments. These 15 damages are, in figure 1, represented diagrammatically by the same number of tniangles and parallelograms ('diamonds'), which are all included in the circumscribing triangle. (The latter is of unit height and unit base length, and contains all possible damages between the terminals of the ship's length.) Triangles (1,1) (2,1) (3,1) (4,1) (5,1) -are single compartment flooding cases 'Diamonds' (1,2) (2,2) (@),2) (4,2) -are simultaneous flooding cases for two adjacent compartments 'Diamonds' (1,3) (2,3) (3,3) -are simultaneous flooding cases 'Diamonds' (1,4) (2,4) 'Diamond' (1,5) for three ad accent compartments -are simultaneous flooding cases for four adjacent compartments -Simultaneous flooding of all 5 cpts For statutory purposes, it is reasonable to assume that very long damages are extremely rare, and may be assigned zero probability. The historic collision data suggests that damages above 0.24 Ls may be disregarded. (X 1.~= 0.24 ) The longest recorded damage was 48m. Since the greatest ship length in the data sample was 200m, X 1.~ is taken as 48, Ls for slips more than 200m in length Using the data set, (296 side-collision cases), 'damage extent' was analyzed to arrive at an associated probability 'p' of the cumulative frequency of occurrence up to a specified (nondimensional) length. Probabilistic Criteria - Part 2 48

53 The Ship Stability Research Centre University, of Strathclyde The assumption that damage will occur somewhere along a ship's length means that the trapezoid OAF contains all damages where 'p' is non-zero. ntegrating the area OAF gives unity. That is, Z p = This property is useful when calculating the ndex A for a particular ship design. Summing up the 'p' contributions for all single compartment damages, then for 2 adjacent compartments, etc., until there are no further non-zero contributions to be made to the summation, should eventually lead to a total of unity Assume that p(l,1) represents the probability that all damages are confined within compartment 1. Similarly, p(2,l) p(3,l) p(4,) p(5,) for compartments 2, 3, 4, Further assume that p(1,2) represents the probability that all damages confined within the aft limit of compartment and the forward limit of compartment 2. Similarly, for p(2,2) p( 3,2) p(4,2) By analogy, p(1,3) p(2,3) p(3,3) for damages within aft of cpt. 1 and fwd end of cpt.3, etc By analogy, p(1,4) p( 2,4) for damages within aft of cpt. 1(2) and fwd end of cpt.4(5) Finally, p(1,5) for all damages within the ship length. p = Consulting the illustration A, it can be seen that the 'diamonds' (2,3) (3,3) (1,4) (2,4) (1,5) all have p = 0 Probabilistic Criteria - Part 2 49

54 The Ship Stability Research Centre University of Strathclyde Assign the probability of the simultaneous flooding of'n' adjacent compartments to be p(i-n), where: ST n' is the number of the after most compartment, numbered from aft. is the number of adjacent compartments considered flooded Then a study of the illustration reveals that Singles p(l-1) = p(o,1) etc 2 adj.cpts. p(l-2) = p(l,2) -p(l,1) -p(2,1) etc 3 adj. cpts. p(1-3 ) = p(1,3) - p(l,2) - p( 2, 2 ) + p(2, 1) etc 4 adj.cpts. p(1-4) = p(1,4) - p(1,3) - p( 2,3) + p(2,2) etc 5 adj.cpts p(l- 5 ) = p(1,5) - p(l,4) - p(2,4) + p(2,3) Until now, comments in this part of the notes have concerned only plane side-to side watertight divisions. Such divisions - i.e. ones without 'steps' or 'recesses'- are clearly the best from a safety viewpoint, since the probability is at a minimum where a relatively short damage may 'open-up' two adjacent compartments. Conversely, where a step in such a division runs all the way from side-to-side - see illustration B the probability of a relatively short damage 'opening-up' two adjacent compartments is maximized. n this case, the proportion of all damages confined within adjacent watertight boundaries is reduced and the proportion of two adjacent compartments flooded is increased The deterministically-based regulations employ similar reasoning in the case of a 'step' in a main transverse bulkhead. (A 'step' is a 'recess' outboard of the B/5 line). When the probabilistically-based Passenger Ship Regulations were introduced, the presence of the deterministic Regulation 5 meant that the aft and forward boundaries are defined in a similar manner. However, the (Dry) Cargo Ship Regulations do not have the equivalent of this Regulation 5, and therefore no distinction is needed between a 'step' and a 'recess', when defining transverse and longitudinal watertight boundaries. Probabilistic Criteria - Part 2 50

55 The Ship Stability Research Centre University of Strathclyde 2.8 Determination of a factor 'r, which is intended to estimate the probability of penetration inboard. Combined Transverse and Longitudinal Subdivision A normal ship design is very likely to have at least some longitudinal subdivision at a point within the ship's length. A practical calculation procedure for the Attained ndex, A, needs to cater for this. t should also be noted that, where transverse bulkheads are recessed, the method needs to be suitable for estimating the probability of penetrating inboard to that 'recess', irrespective of whether there is a watertight longitudinal division at that position The true and correct way to calculate the factors 'p' and 'ry is to treat the calculation as comprising of two parts: for a particular damage zone - Firstly, all penetrations from the ship's side to the first longitudinal boundary, - Secondly, all penetrations from the first longitudinal to the next inner boundary. Of course, when there is no longitudinal watertight division at all on the damage side of the ship's centerline, the second set of'p' and 'Y' values do not apply Consider the following compartment, or compartment group, configurations, Figure 2 Figure 3 x2 oiner T X1l inbb inner oter For Res.A,265. the 'recess' is ignored when calculating the factor 'p': this is theoretically incorrect Probabilistic Criteria - Part 2 51

56 The Ship Stability Research Centre University of Stralhclvde n studying the figures 2 and 3, it is self-evident that, for the general case of 'damage' to a compartment, or compartment group, the contribution to the index 'A' is in two parts - namely, penetration to a specified position, say b(outer) inboard, and then the remaining cumulative probability of penetrations beyond this position to a deeper penetration, say b(inner). Therefore, two sets of calculations are involved over different boundary limits and positions. Each separate calculation for 'outer' or 'inner' portion itself consists of lower and upper limits in order to integrate the shaded areas. For the 'outer' case, the lower integration limit is at the ship's side - and is therefore zero. Where the 'inner' case is concerned - and also those cases where there is no longitudinal boundary at all - the upper integration limit is at the ship's centerline and is therefore unity. rbblsi rtei at25

57 The Ship Stability Research Centre University of Strathclvde 9. How the 'probability of survival' is influenced by the ship condition - i.e. mean draught, trim, ship KG - at the time of collision. \ 10. How the 'probability of survival' is influenced by the position and extent of the space(s) damaged, and the contents within the space(s). 11. How the 'probability of survival' is influenced, very significantly, by the sea state/weather conditions at the time of collision. Comments on items 9, 10 and 11.1 tems 9, 1 0 and all relate to draught-dependent elements which will have a significant influence on the 'probability-of-survival' in the event of a side-collision incident. That is, they influence the choice of an appropriate factor 's' - see item Of even greater significance, however, is the limiting sea state assumed to apply to a future collision incident. n practical terms, therefore, a realistic 's' formula always needs to be linked with an appropriate limiting sea state..3 To achieve an efficient placement of main transverse bulkheads, it is necessary to optimise the Attained ndex, A, for a particular ship design. Examining a whole range of ways of internal subdivision, it is possible to draw a number of qualitative conclusions as what constitutes efficient subdivision. Probabilistic Criteria - Part 2 53

58 The Ship Stability Research Centre Universit, of Stratliclyde.4 Consider the following cases of flooding:. 'Side-collision' damage versus 'Ramming' damage..2 Variation in height of main transverse bulkheads..3 Longitudinal subdivision..4 Honizontal subdivision. *.5 Size of ship..6 Ship's condition - mean draught, trim and KG..7 Transverse bulkheads. Equal versus Unequal spacing..8 Assumed 'damage' position. Contents of the 'damaged' compartment..9 Transverse bulkheads having Steps and Recesses'..10 Beneficial effect of'reserve buoyancy' & 'residual freeboard'..11 Comparative benefits of : 'number of transverse bulkheads' versus 'freeboard'. rbblsi rtei a125

59 The Ship Stability Research Centre Universit" of Strathclvde 12. By taking due account of.9 to.11, how a factor s' may be determined, which is intended to be an estimate of the 'probability of survival'. Estimating the "Probability of Survival" The Factor "s" When the Equivalent Passenger Ship Regulations were being developed in the late 60's early 70's, an essential part of the work was the use of results obtained from 'damaged' model tests - carried out in the UK and USA - to estimate the 'probability of survival' after assumed side-collision damage'. After the test results were analysed, the specialist SDS group at that time proposed that 's' should be a function of residual (mean) freeboard and residual GM, such that: 'Probability of Survival' Factor 's'= k-- F GMe. where : Fý GMNe B is the effective flooded freeboard, corrected for heel due to flooding; is the effective flooded metacentric height; is the ship's breadth (maximum). At reg.6(d)(i), res.a.265(v), formula (V), the factor's' is given by: Si =4.9. -tan(!)).(gm, -MM)j where: Fi is the effective mean damaged freeboard, (corrected for deck side, over 2/3Ls amidships); B 2 a (GM,-MMS) is the extreme moulded ship breadth at mid-length (bulkhead deck level); is the heel angle due to asymmetrical flooding, (in the final flooding stage); is the residual metacentric height i.e. after flooding. Probabilistic Criteria - Part 2 55

60 The Ship Stability Research Centre University of Strathclyde At the time when Res.A.265(V) was issued, the corresponding SOLAS regulations were entirely empirical in nature - that is, the minimum standards specified were completely unrelated to past casualties. When present-day SOLAS standards of subdivision / residual stability are considered, their empirical nature is still evident - although, of course, the standards themselves have been enhanced considerably. n essence, standards are set by assigning minimum values to some of the principal parameters of the residual stability lever curve. For instance, equilibrium heel angle, area, range, GZmax Ever since 1990, the SOLAS residual stability standard for 'new' passenger ships the so-called SOLAS 90 standard - has included the concept of linking the minimum GZmax requirement with the 'Persons at Risk', (but in no case to be less than 0. in). After the loss of the 'Herald of Free Enterprise', it was agreed that 'existing' RoRo passenger ships should be enhanced such that they would have a standard equivalent to that of 'new' ships, within a realistic timeframe. The order in which the 'existing' RoRo's were to be enhanced was based upon a specific, one-off, calculation of the Attained ndex, A, as laid down in MSC/Circ.574. This calculation procedure is a simplified set of instructions taken from regs. 6 and 7 of Res.A.265(V). A ratio A / Amax is obtained by two separate sets of calculations - the first, (using the actual ship's KG), leads to A; the second, (using a theoretical KG whereby the critical one (or two-) compartment damage case just complies with the criteria), leads to Amax. For one- (two-) compartment-standard RoRo ships, single (two) compartment damage applies. Then the enhancement procedure is determined in ascending order of A/Amax t is evident that any enhancement procedure which is probabilistically-based requires a factor 's' which is also probabilistically-based. Russia had previously proposed such a 's' factor in their paper SLF35/4/9; therefore, this factor was used in the exercise to calculate the A/Amax ratios for the world 's existing RoRo passenger ship fleet. The formula is: s = 2.58.(GZ,,.-Range-Area) 0 ' 2 5 See pp.56-59of the corresponding explanator' notes Probabilistic Criteria - Part 2 56

61 The Ship Stability Research Centre University of Strathclvde where: GZma <= 0. 1 m. Range => 15deg. Area => m.rads When the (Dry) Cargo Ship Regulations were issued, the factor 's' was in a similar format to the Russian proposal, except that the 'Area' criterion was omitted" s = c.(0.5-gz,,.range) 0, 25 Where: c=0 for S, > 30 deg. else, c = [ (30-ae) / 5] 0.5 c = 1 for a. => 25 deg After the 'Estonia' tragedy, MO members re-examined the residual stability standards which applied to RoRo passenger ships at that time. As a result, they introduced the concept of adding fixed amounts of 'water-on-deck into the damaged stability calculations, in order to allow (empirically!) for flooding of the vehicle space. Of course, the basic idea of basing residual standards upon minimum acceptable values of GZ,,, etc., remains For both sets of probabilistically-based regulations, it should be noted that the factor 's', being highly draught-dependent, needs to be 'averaged' over the entire draught service range, otherwise the probabilistic aspect of these type of regulations would be greatly undermined. For the Passenger Ship Regulations - Res.A-265(V) - three draughts, dl, d2 and d3, are specified, with weightings of 0.45, 0.33 and 0.22, respectively. For the (Dry) Cargo Ship Regulations, there are two draughts, dl and dp, with corresponding weightings of 0.5 and 0.5. See reg Probabilistic Criteria - Part 2 57

62 The Ship Stability Research Centre University of Strathclvde Concerning a future set of 'harmonised' regulations, no decisions have yet been made regarding either the number of draughts or the corresponding weightings, although three draughts for both ship types seems to be the preferred option. n respect of the factor 's' itself, decisions post-estonia will only be made after a thorough study and evaluation of the work emanating from the North West Europe Research Program, as yet incomplete. However, it is reasonable to suppose that the residual stability curve will be used in establishing minimum residual stability standards for the foreseeable future. n addition, supplementary standards will apply to RoRo passenger ships, both 'new' and 'existing'. Specifically, the water-on-deck' concept will probably be developed further for such ship types. t is very likely that specified minimum criteria, such as 'equilibrium angle', 'GZmax' will vary according to ship type, but the general aim will be to have a common set of criteria, applying to all ship types. Probabilistic Critcria - Part 2 58

63 The Ship Stability Research Centre University of Strathclyde 2.13 The factor 'V - statutory maximum vertical extent of damage. Consideration of Horizontal Subdivision The assumption concerning minimum vertical extent of damage has always been without limit, in both sets of passenger ship subdivision regulations. When the (Dry) Ship Cargo Regulations were issued, however, the concept of a statutory maximum vertical extent of damage was introduced, in order to give credit for watertight horizontal divisions situated well above the draught position. The height of a watertight horizontal division - say H - may be regarded as 100% effective in avoiding the spread of flooding when H > Hmax, where Hmax is defined as the maximum vertical damage extent liable to be encountered. Conversely, where H < H max, but above the draught position, the spread of flooding is assumed to be avoided for a specified proportion of all damages within the damage zone. Probabilistic Criteria - Part 2 59

64 H The Ship Stability Research Centre University of Strathcyde 2.14 The Attained Subdivision ndex, A, for Passenger Ships. A Summary of the Factors used in the Calculation Procedure To obtain the Attained Subdivision ndex, 'A, for a passenger ship, it is necessary to follow the calculation procedure outlined in regulations 6(a)(i), (ii), (iii) and regulations 7(b),(c) ofres.a.265(v) n order to provide realistic, (but still generous), limits for the parameters used in the calculation itself, it is suggested that: the ship should have not less than 5, nor more than 25 main compartments - damage zones - within the ship length; - the number of adjacent compartments assumed to be flooded simultaneously is to be no more than 5; Then, if NAG is the number of adjacent compartments assumed to be flooded simultaneously, and if NMC is the assigned number of the aftermost compartment of a group, (numbering such that aft cpt = 1, etc.); NAG NMC ndex, A = Y da where da is the contribution to A iag=1 ia= of a particular cpt.gp. flooding. Probabilistic Criteria - Part 2 60

65 The Ship Stability Research Centre University of Strathclvde Each 'da value is to be obtained from the product of 6 variables, such that da= a pr v.wt s Factor Regulation Formulae a 6(b) p 6(c) V V V V r 7(b) & 7(c) X wt 6(d)(ii) X v Does not apply None s 6(d)(i) & (ii) V X n general, 'da' may be composed of 3, 6 or 12 elements, for the following reasons: - For the Passenger Ship Regulations, (unlike the Cargo Ship Regulations), three draughts - dl, d2, d3 - are used to determine the weighted' factor 's'. To represent the (idealised) draught distribution density assumed, 'weightings' of wtl(0.45),wt2(0.33),wt3(0.22) are specified. - Where there is at least one longitudinal watertight boundary to the 'damage' side of the ship's centerline, (within the damage zone being considered), there are two 'p' factors say, p(outer) and p(inner). - Where there is at least one horizontal watertight boundary, (within the damage zone being considered), - which is higher than the draught position, but lower than Hmax) there may be two 'v" factors - say, v(lower) and v(upper). Vertical extent of damage in the present set of regulations is without limit. However, it is highly likely that the Hmax concept will be extended to passenger ships when the 'harmonised' regulations are finalised. Probabilistic Criteria - Part 2 61

66 The Ship Stability Research Centre University of Strathclyde Summarising, therefore, the general expression for 'da" is: Laprwfle,) * V' J, S( / /U) SJLWt2 ]s 3 (o/ii/u)j Probabilistic Criteria - Part 2 62

67 The Ship Stability Research Centre University of Strathclyde 2.14 How to calculate the Attained Subd' ' ion ndex, (A)*, *accordin g to the provisions of: The Passenger Ship 'Equivalent Regulations and the (Dry) Cargo Ship Regulations. The Passenger and (dry) cargo ships regulations similarities and-differences Common Objectiý,e 1.1 For both sets of regulations, the basic objective is to ensure that, the subdivision arrangements for a ship attain an appropriate minimum standard, and are associated with an adequate residual fireeboard. The latter is needed to provide an acceptable degree of reserve buoyancy. 1.2 To achieve this objective, each set of regulations comprises distinct elements which may be summarised as probabilistic and deterministic, plus those portions of the regulations. which are related to watertight integrity and safe/efficient ship operation Probabilistic elements *2.1 Both sets of regulations include a probabilistic element, intended to assess the subdivision characteristics of a ship in relation to assumed side damage. This is achieved by calculating an attained subdivision index 'A; it is then necessary to show that this index is at least as great as a required index 'R' 2.2 The 'R' formulae for passenger and cargo ships are of different format, but share certain similarities in the parameters used. For both ship types, 'R' is proportional to ship length - (actually subdivision length Ls) - and the slope of the 'R' curve decreases as ship length increases. At zero ship length, KR has a positive value in both cases. When ship length Probabilistic Criteria - Part 2 63

68 The Ship Stability Research Centre University of Strathclyde becomes very large, 'R tends to unity without reaching it, in the case of passenger ships. (Note that for cargo ships, when Ls exceeds 10009m, 'R' becomes more than unity). 2.3 However, whilst cargo ships have 'R' values which depend solely on ship length, passenger ships, on the other hand, have 'R' values dependent also on the number of persons permitted to be on board, 'N'. 2.4 The permeabilities used in damage stability calculations are usually fixed values, and are decided according to a space's intended function i.e. whether appropriate for cargo, fuel, fresh water, machinery, stores, etc. For the cargo ship regulations, the permeability to be used for cargo spaces is For passenger ships, it should be noted that cargo space permeabilities vary between 0.60 and 0.95, dependent on draught - see ref 4(b). 2.5 n the procedure used to obtain the attained index 'A' the general form of the summation equation is as follows: The attained subdivision index, NAG NC A = Z Z da j~l i= A = NAG NC j=1 i=1 E a.p(r).si(v) Where: NAC the greatest number of adjacent main compartments which are assumed to be flooded simultaneously. NC the number of main compartments in the ship (equivalent to the number. of main transverse bulkheads in the ship, plus one). da the positive contribution to the 'A' value, relating to a particular damage tone. Probabilistic Criteria - Part 2 64

69 HThe Ship Stability Research Centre Univcrsitv of Strathclvde a the positi on (measured from the after terminal) of the centre of assumed damage. p the probability factor related to the assumed longitudinal extent of damage. (r) a reduction factor to be applied to 'p',, where longitudinal subdivision exists to the damaged side of the ship centreline. (An additional contribution to da may be made for the case of simultaneous flooding both inboard an' outboard of such a longitudinal division). Si a probability factor' related to the estimated probability of survival, taking into account the assumed flooding extent and ship condition, (i.e. draught, trim and KG). (v) a reduction factor to be applied to 's', where horizontal subdivision - above the waterline - is assumed to restrict the flooding extent. (An additional contribution to da may be made for the case of simultaneous flooding both above and below such a horizontal division). 2.6 As can be seen from the formula above, the summation process is carried out initially on the assumption that flooding is confined by the aft and forward boundaries of -a single damage zone - Normally, (but not always these boundaries are the positions of main transverse bulkheads. 2.7 The summation of the contributions (da) is then made for assumed simultaneous flooding of two adjacent damage zones, then three, etc. 2.8 This process is continued until it is demonstrated that A > R and/or there are clearly no more non-zero contributions to be made to the 'A' value. Probabilistic Criteria - Part 2 65

70 The Ship Stability Research Centre U Lniversity of Strathclvdle 2.9 Considering the formula for 'A', it should be noted that probabilistic principles are applied somewhat differently for passenger and cargo ships For passenger ships, the centre of assumed damage and its longitudinal extent are represented by two distinct factors, la, and 'p' respectively; on the other hand, the cargo ship regulations combine the centre of damage and extent into a single factor' p.ý 2.11 The reduction factor 'r', which accounts for the effect on restricting the flooding by the fitting of a longitudinal division, has very similar formulae in both the passenger and cargo ship regulations. However, in the case of cargo ships, it was considered necessary to increase the r factor for comparatively shallow penetrations and therefore the formula was adjusted slightly to cater for t~his, (the practical effect is to increase Yi by 0. 1) The factor 's' is an estimate of the probability of survival and is dependent on the residual stability characteristics of a particular assumed damage case The correction factor 'V', accounts for horizontal subdivision, (above the waterline ). t appears in the cargo-ship regulations only. When the passenger ship, regulations were developed, it was considered that the vertical extent of damage should not be dealt with probabilistically and that the 'worst' vertical extent of damage was to be used. U2.14 Probability of survival after damage depends on the ship condition at time of damage. The survival probability factor s, is highly draught dependent For a passenger ship having an operational draught range 'd,'., to 'd,', see reg. 1 (g) (i) and (ii) - three representative draughts 'dl', 'd2' and 'd3' see reg.l (g)(iii) - are, defined, with * associated weighting factors as follows: si = 0.45 sl s s3 Formula (MX of reg.6(d)(ii). Probabilistic Criteria - Part 2 66

71 The Ship Stability Research Centre University of Strathclvde 2.16 A similar approach is used in the cargo ship regulations, except that only two representative draughts 'd 1 ' and 'd4' are used - see reg and 1.3 -so that: si = 0.50 sj sp Reg * (Note that equal weighting is assigned to the two draughts) 2.17 The 's', formulae for both sets of regulations are square root functions. They are: For a passenger ship, si = tan(fl. -(GMr - MM). H For a cargo ship, si = c.(0.5. GZma. Range) The 's' formula for passenger ships was derived from an analysis of damage model tests; the cargo ship 's' damage data. formula involved Rahola-type criteria, unsupported by historical Probabilistic Criteria - Part 2 67

72 The Ship Stability Research Centre Deterministic Elements Umv-ersity of Strathclyde 3.1 There are more deterministic elements in the passenger ship regulations, -primarily ~ because it is considered that there is a greater possibility of heavy loss of life if a serious sidecollision incident occurs. 3.2 Befof~e discussing those deterministic elements which are unique to the passenger ship regulations (at section 6), it is proposed to summarise some of the similarities and the differences between the two sets of regulations. 3.3 t was decided that, in respect of vertical extent of damage, the 'worst' extent, (i.e. that which produces the least favorable result), should apply. Because of the greater draught variation, on average, of cargo ships vis-a-vis passenger ships, the vertical extent of damage is ~limited by a distance 'Ha.' in the cargo ship regulations. 3.4 There is a requirement in the passenger ship regulations that the fact or should be unity, when combined flooding of the foremost and the adjacent compartment is assumed. This s, value is to be achieved at the deepest draught condition and with the most adverse bow trim anticipated in service. There is a similar requirement for cargo ships, except that the flooding in this case is for the foremost compartment only and trim is not considered. U3.5 The most important deterministic features of the passenger ship regulations are contained in regulation 5. Before describing the features of this regulation in more detail, it should be noted that: - - all assumed damages are to be associated with the appropriate service trim, the collision bulkhead has to lie within specified limits see reg.lo(a)(i). Probabilistic Criteria - Part 2 68

73 The Ship Stability Research Centre University of Strathclyde 3.6 Regulation 5 of the passenger ship regulations specifies compartmental standards which are linked ; to a minimum damage length, as well as a maximum damage penetration. A deterministic one-compartment standard is assumed where 'N' is 600 or less. A two-compartment standard is required throughout where 'N' is 1200 or more. For values of 'N' between 600 and 1200, there is to be a proportion given as (N/600-1) - of the ship's length, measured from forward, of two-compartment standard - see reg.5(b)(ii). n applying reg.5(b)(ii), only compartments having a minimum length > 0.03Ls + 3.0, (limited to 11.0 m.), are to be considered. n addition, only longitudinal subdivision at least B/5 inboard from the ship side is to be considered effective when determining the extent of flooding - see reg.5(b)(i). 3.7 With the above in mind, it must be established that the resultant damage GM's corresponding to the compartment flooding cases which are to meet the compartmental subdivision standard are to be at least as great as those given by the formula values of reg.5(c)(i)(1). Probabilistic Critcria - Part 2 69

74 The Ship Stability Research Centre Universitv of Strathclvde 2.15 dealised Damage Distribution along a Ship's Length The Factor "a" A histogiam was prepared - (for a more detailed description, see the explanatory notes.x which are appended to the Equivalent Passenger Ship Regulations, p. 9 1) - from the 296 cases of the collision data set, using the non-dimensionalised 'damage location' values. On studying this histogram, it becomes evident that there is a definite bias for damage to occur in the forward half of a ship The details of the histogram were idealised into a frequency density which rises linearly up to mid-length, and then remains constant over the forward part. Since the assumption is that damage is certain, the aft density is 0.4, whilst the constant density in the forward half of a ship is For passenger ships, formula H, reg.6(b) of Res.A.265(V) indicates how to obtain the factor 'a' i.e. the idealised 'damage location' factor. a=0.4 [ a2+ 12] where = x /Ls, but not more than 0.5 ý2 = x.2/ Ls, but not more than = (xl+x2)/ls but not more than 1.0 t should be noted that - for the Equivalent Passenger Ship Regulations -'damage location' is defined by a separate factor 'a', whilst the factor 'p' represents 'longitudinal extent' only. Probabilistic Criteria - Part 2 70

75 The Ship Stability Research Centre Universitv of Strathclvde When the (Dry) Cargo Ship Regulations were being developed, it was agreed that., damage location would be merged with 'damage extent', to form a composite 'p' factor. However, the idealised distribution function for 'damage location' remained unchanged, even.. \. though the formula given - at reg is in a different format. a = E, but not more than 1.2 where E=xl/Ls+x2/Ls During the 40"' session of the SLF sub-committee, the SDS group agreed that a 'harmonised' factor 'a' - based upon a Netherlands proposal originally put forward some years previously - should be recommended, first to plenary and then to the Maritime Safety Committee (MSC). f adopted, any future set of'harmonised' regulations would then contain a factor 'a' which would remove the 'kink' at mid-length and substitute a density distribution which rises from 0.6 at the aft terminal to 1.4 at the forward terminal. The group noted that correlation with collision data was essentially about the same, and that incorporating this simplified distribution into a composite factor 'p' made the calculation of 'A' more straightforward. The modified 'a' formula would then be a = E where E xl/ls + x2fls Probabilistic Criteria - Part 2 71

76 The Ship Stability Research Centre University of Strathclyde 2.16 The Treatment of Simultaneous Flooding Firstly, consider the formulae V, V and V of reg.6(c)(iii) of res.a.265(v). These..x. are identical to the corresponding 'pi" formulae of the (Dry) Cargo ship Regulations. ndeed, not only are they applicable to both sets of regulations, but also to any future 'harmonised' set of regulations which are yet to be developed For the record, these formulae are: V For 2 adjacent compartments: p = p12 - p - -p2 etc. V For 3 adjacent compartments: p = p p 12 - p23 + p2 etc. V For 4 adjacent compartments: p = p p123 - p234 + p23 etc. Actually, V and V are two variants of the same formula. They also apply to 5 (or more) adjacent compartments Although paragraph is true, it should be noted that, since tentative agreement has already been reached at MO that a composite 'p' will be used in future 'harmonised' regulations, the formulae are to be interpreted as including the factor 'a'. (n other words, both 'location' and 'damage extent' are to be catered for, when these formulae are used.) Everything so far assumes damage to the ship centreline, or beyond. A reduction factor 'r' is to be applied in those 'damage' cases where penetration inboard is a distance greater than zero, but less than B /2. Probabilistic Criteria - Part 2 72

77 The Ship Stability Research Centre University of Strathcl.vde 2.17 Treatment of'existing' non/ro-ro passenger ships Work at MO on the subiect. Some MO delegations, including those of Sweden and the Russian Republic, expressed the opinion in past sessions that higher stability standards should be applied to all existing, passenger ships, not just ro-ro passenger ships. A computer exercise, similar to that already completed for ro-ro passenger ships, was conducted - Sweden acting as collator of the submitted calculations. MSC/Circ.574 again formed the basis of the calculation procedure used to establish the ratio A/Amax for the sample ships The current position This subject was discussed at the 38th session of the sub-committee on Subdivision and Loadlines and on Fishing Vessels Safety. 1MG members had before them the Swedish collation paper (SLF 38/4/4). After an evaluation of the sample cases submitted in this paper, it was decided that there was no compelling need to proceed further with up-grading at this stage. Any future decisions on this matter would be dealt with under the agenda item "Harmonisation * of damage stability requirements". Probabilistic Critcria - Part 2 73

78 H The Ship Stability Research Centre University of Strathciyde The Passenger Ship (Equivalent) Regulations. Res.A.265(VH) A brief resume of the first internationally-agreed set of safety regulations setting out minimum acceptable levels of subdivision / residual stability... x. The main purpose of this set of regulations is to indicate the calculation procedure for determining the Attained Subdivision ndex, A, and to demonstrate that this is not less than the Required Subdivision ndex, R. A short summary of the layout of these regulations. 1. Reg. Definitions 2. Reg. 2 Required Subdivision ndex, R 3. Reg. 3 Special Rules regarding Subdivision 4. Reg. 4 Permeability 5. Reg. 5 Subdivision and Damage Stability Deterministic Part 6. Reg. 6 Attained Subdivision ndex, A (Formulae for "a" 'P " and "S") 7. Reg. 7 Combined Longitudinal and Transverse Subdivision (Formulae for "r") 8. Reg. 8 Stability nformation 9. Reg. 1 0 Peak and Machinery Bulkheads, Shaft Tunnels, etc. Probabilistic Criteria - Part 2 74

79 The Ship Stability Research Centre University of Sirathclyde 10. Reg. Double Bottoms 1. Regs. 13 to 20 All these regulations concern the maintenance of watertight integrity 12. Regs. 9,12, 21 Sundry regulations Probabilistic Criteria - Part 2 75

80

81 CONTENTS 1. NTRODUCTON 1 2. BREF REVEW OF RESDUAL STABLTY STANDARDS 2 3. SOLAS'90 DAMAGE STABLTY STANDARDS 7 SOLAS 92 Amendments 7 SOLAS 95 Amendments 8 4. THE STOCKHOLM AGREEMENT 10 Main Observations 11 Danish Proposal 12 Constant Height of Water on Deck Standards STRONG AND WEAK PONTS OF SOLAS90 AND STOCKHOLMAGREEMENT 17 APPENDX A APPENDX B SOLAS 90 STANDARDS STOCKHOLM AGREEMENT

82

83 1. NTRODUCTON n the traditional Archimedean Naval Architecture, the Ship designer attempts to satisfy primarily two important qualities: Floatabilty: the ability of the vessel to support a given weight W, by means of the hydrostatic pressure acting on the underwater surfaces, giving rise to the buoyancy force, B. Furthermore, to achieve a conditi on of upright equilibrium, the weight and force vectors have to act along the same vertical line on the centreplane of the vessel. ~Hydrostatic Stability: the ability of the vessel to return to a state of equilibrium (preferably the upright) in still water when disturbed from it. As Naval Architects learn very early in undergraduate courses, a necessary condition for hydrostatic stability is that the metacentric height, GM, should be positive. Naval architects also know that whilst positive G3M is the necessary condition for stability at small angles of heel, it may not be so for larger angles. For this, they use a different measure of stability, the righting lever GZ, the variation of which with increasing heel provides a better indication of the hydrostatic stability characteristics of a vessel. n relation to the above, regulators request that in addition to a positive GM, the GZ curve must satisfy certain criteria. These include maximum GZ and the angle at which it occurs, angle of vanishing stability and area under the GZ curve. f a vessel is damaged so that part of her internal volume is flooded, she will sink and heel until she reaches a condition in which reserve buoyancy (i.e. buoyancy above the initial waterline) has been brought into effect to offset the lost buoyancy with equilibrium again restored, but with the vessel possibly taking a steady angle of heel. For most onerous cases GM is likely to decrease and the GZ curve characteristics are likely to diminish in all regards. Capsize of the vessel occurs if GM became negative with the vessel upright and there were no positive righting levers in the residual GZ curve (GZ curve after damage). On the basis of this reasoning, damage survivability can be enhanced by ensuring that the residual GZ curve satisfied certain critenia. The survivability of ships is, in general, related to both intact and damage stability requirements. For passenger vessels this is normally assessed by their stability in a damaged condition, i.e. by their residual stability. However, in spite of the unavoidable reference of Naval Architecture to Archimedes, quantitatively specific standards of residual stability after damage have been introduced for the first time as recently as the

84 Safety Of Life At Sea (SOLAS) Convention addressing, in spite of better knowledge, only the metacentric height (0.05m GM). the deterministic approach assumes that ships of common standard of subdivision are equally safe The tragic accidents of the Herald of Free Enterprise and the Estonia were the strongest indicators of yet another gap in assessing damage survivability concerning subdivision above the bulkhead deck of large undivided deck spaces. t has also brought about a realisation that "ship survival" might have to be addressed separately from 'passenger survival" in that the deterioration in the stability of such vessels when damaged could be 'catastrophic' rather than one of gracefful degradation, thus allowing virtually no time for evacuating passengers and crew. 2. BREF REVEW OF RESDUAL STABLTY STANDARDS. Tables and 2 show the progress of regulations on damage stability cniteria and the assumed extent of damage, pertaining to the UK, (2]. The question of survivability standards of a ship following damage was considered towards the end of the nineteenth century by several Select Committees of the House *. of Commons. n 1890, the first Committee on watertight subdivision recommended that passenger vessels over 425 ft (129.5m) in length should be capable of floating with any two adjacent compartments open to the sea. t is interesting to mention that the 1890 Committee also recommended one compartment standard for cargo ships but this * was not implemented due to lack of support. Following the loss of Titanic in 1912, the second Bulkhead Committee was set up to consider in depth the question of damage survivability. A year later, the first nternational Convention for the Safety of Life at Sea met in London. This Convention laid down an empirical system for deciding on the appropriate bulkhead spacing, consistent with the number of compartments which could be flooded without submerging the so-called "margin line", located at 3" (76mm) below the bulkhead deck. Again, it is interesting to note that the 1890 Committee used a percentage of depth rather than 3" as the margin line. The number of adjacent compartments considered to be flooded (between one and three) was based on an empirical relationship involving ship length and type. The work, interrmpted by the Fist World * 2

85 War, was finally completed atthe 1929 SOLAS Convention which also laid down the first details on the extent of damage. For the first time, there was a quantifiable standard for ships aimed to ensure that realistic damage could be sustained without the prospect of immediate progressive -flooding and eventual loss of the ship. However, there were no specific requirements for damage stability criteria. U The 1932 Supplement to the 1928 "nstructions as to the Survey of Passenger Steamships' went one stage fturther by imposing restrictions on the amount of heel permissible after flooding and following counter flooding ( ý 70). t is interesting to note again that the effect on stability of flooding in way of the bulkhead deck was also to be calculated and considered on its merits. The extent of damage and damage stability criteria on the final condition after damage and equalisation was detailed at the 1948 SOLAS Convention introducing also the requirement for a positive residual GM. The UK Department of Transport formulated for the first time construction rules for passenger ships in 1952 in order to implement the 1948 Convention. The first specific criterion on residual stability standards was introduced at the 1960 SOLAS Convention with the requirement for a minimum residual GM (O.05m). This represented an attempt to introduce a margin to compensate for the upsetting environmental forces.,t "Additionally, in cases where the Administration considered the range of stability in the damaged condition to be doubtful, it could request further investigation to their satisfaction". Although this was a very vague statement, it was the first attempt to legislate on the range of stability in the damaged condition. is interesting to mention that a new regulation on "Watertight ntegrity above the Margin Line" was also introduced reflecting the general desire to do all that was reasonably practical to ensure survival after severe collision damage by taking all necessary measures to limit the entry and spread of water above the bulkhead deck. At about the same time as the 1974 SOLAS Convention was introduced, the nternational Maritime Organisation (NMO), published Resolution A265 (V), [3]. These regulations used a probabilistic approach to assessing damage location and extent drawing upon statistical data to derive estimates for the likelihood of particular damage cases. The method consists of the calculation of an Attained ndex of 3

86 Subdivision, A, for the ship which must be greater than or equal to a Required Subdivision ndex, R, which is a function of ship length, passenger/crew numbers and hifeboat capacity. ndex "A" is in turn a function of three different probabilities, "a", P " and "S". The Equivalent Regulations raised new damage stability criteria addressing equilibrium as well as recommending a minimum GZ of 0.05m to ensure sufficient residual stability during intermediate stages of flooding. The 1980 Passenger Ship Constmuction Regulations introduced requirements on the range of the residual stability curve as well as on the stability of the vessel at intermediate stages of flooding. The loss of the Herald of Free Enterprise in 1987 drew particular attention to Ro-Ro ferries in which the absence of watertight subdivision above the bulkhead deck is a particular feature. The implications of this feature were highlighted by the Court of nquiry which observed that the SOLAS Conventions and UK Passenger Ship Construction rules had been aimed primarily at conventional passenger ships in which there is normally a degree of subdivision above the bulkhead deck, albeit of unspecified ability to impede the spread of flood water. n response to this, the Department of Transport issued Consultative Document No 3 in 1987 which outlined a level of residual stability that required all existing Ro-Ro ferries to demonstrate compliance with the 1984 Passenger Ship Construction Regulations. This standard had previously formed the basis of a submission by the UJK and other Governments to MO which considered the question of passenger ship stability in some detail. This was the fore-runner to SOLAS '90. The much deliberated SOLAS '90 introduced new higher standards of residual stability, to be applied to all passenger ships including Ro-Ro's and, for the first time, cargo ships 4

87 .19 E-E 1 Q _ o -o.. O -.. E~ ii~* sp o,.- J3..h -~n f E -- a ic.5 t-, 00 l h a g o... -.^o,, >0 U A2 A 0 0

88 Table 2: ASSUMPTONS ON THE EXTENT OF DAMAGE (UK CRTERA) LONGTUDNAL TRANSVERS VERTCAL OTHER NOTES E 1914 Convention Spacing of BHD BHD Recess - No mention NL 1. Longitudinal extent of Art X (6) recesses 3.05m + sufficient damage is assumed: 0.02L distance from.1 clear of main WT ship's side bulkheads if FOS>0.5.2 to include one WT BHD if FOS between 0.33 and nstructions 3.05m L B/5 measured No mention NL.3 to include two main WT paras'78 and 79 inboard from BHDS if FOS< Convention Rees V() & ship's side at deepest sub Convention not V(7) division implemented 1942 nstructions loadline Paras 69(3) & Convention - extent 69(7) of damage assumed was based on minimum spacing of bulkheads and recesses Convention 3.05m L or Do From top of Any lesser Regn 7(d), 10.67m whichever double extent Convention increased 1952 MS is less bottom up to which longitudinal extent of (Construction) margin line would damage and introduced Sch 3 para l(c), result in a vertical extent nstructions more severe Convention increased condition vertical extent of damage and acknowledged other lesser damages which could result in worse damage conditions Convention Do Do From base Do Regn 7(d) line upwards Convention, 1965 MS without limit although accepting that (Passenger Ship side damage on small Construction) ships would be more Rules severe than on big ships, and that a higher transverse requirement would lead to congestion of fittings and to larger angles of heel under some conditions - decided on reflection to make no change in the B/5 extent of damage Convention Do Do Do Do 1980 MS (PSC) Rules 1984 MS (PSC) Regs 6

89 3 SOLAS 90 DAMAGE STABLTY STANDARDS SOLAS 90 requires the following in the final condition after damage: * A minimum range of 15 degrees beyond the angle of equilibrium which should not exceed 12 degrees for two-compartment flooding and 7 degrees for one compartment flooding.. A minimum area of m.rad under the residual GZ curve. 0 A minimum residual GM of 0.05m with a maximum GZ of at least 0.10m, increased as necessary to meet certain stipulated heeling moments due to wind heeling, passenger crowding and lifeboat launching. Passenger ships, including Ro-Ro Passenger vessels constructed on or after 29 April 1990 should comply with SOLAS 90 standards. SOLAS 1992 Amendments SOLAS 90 rules were to be applied for the new passenger vessels. n 1992 SOLAS amendments were introduced to improve the survivability characteristics of the existing passenger ships with Ro-Ro cargo space. For passenger ships with Ro-Ro cargo spaces defined in regulation 1H-2/3, constructed before 29 April 1990 shall comply with the provisons of this regulations within the time table, which is based on the A/Amax value. A/Amax value was calculated by using the simplified method based on A.265(V) developed by the Maritime Safety Committee at its 59 th session in June 1992 Table3 Time table for existing ferries to comply with the MSC. 12(56) A/Amax Date of Compliance less than 70% 1 October 1994 Less than 75 % 1 October 1996 Less than 85% 1 October 1998 Less than 90% 1 October 2000 Less than 95% 1 October 2005 * 7

90 c CD * *CD *(OD S S U *" - BS

91 The provision of this regulation need not be applied to ships having the value of A/Amax of 95% or more. SOLAS 1992 Amendments also stated that Administration may allow *Reduction of the minimum range of the residual fighting lever Curve defined in Paragraph Range may be reduced to minimum of 10 degrees, in case where the area under the righting lever curve is increased by the ratio: 15 Range *Calculation of the residual righting lever (GZ) referred to in paragraph by the following formula GZ HeelingMoment Displacement provided that in no case GZ will be less than 0.09 m SOLAS 1995 The Tragic accident to Estonia on 28 September 1994 led to the new stability requirements. Following the accident, considering the urgency of the situation and the public outcry for an immediate solution of the problem with the existing ships, the MO set up a panel of experts to consider the issues carefully and make suitable recommendations. Following the work carried out by the M40 panel of experts recommended SOLAS '90 as the new global standard for all existing ferries with dates of compliance ranging from 1 October 1998 to 1 October 2010 depending on a combination of the vessel's A/Amax value, the number of persons carried and age. n November 1995 SOLAS 90 standards were accepted as new global standards for existing vessels. 2 Compartment RoRo Passenger Vessels t was agreed that two compartment standard vessels should comply with SOLAS 90 standards according to the time table, which is based on the A/Amax value 8

92 Table4 Time table for existing ferries to comply with SOLAS 90 standards as accepted at SOLAS 95( 2 compartment vessels) A/Amax Date of Compliance Less than 85% 1 October 1998 Less than 90% 1 October 2000 Less than 95% 1 October 2002 Less than 97.5% 1 October 2004 More than 97.5% 1 October Compartment RoRo Passenger Vessels t was agrred that the number of persons carried on carried by one compartment standard Ro-Ro passenger vessels to be restricted to 400. t was also agreed that no new one compartment passenger ships would be permitted after 1 July 1997, to carry more than 400 persons. All existing one compartment ships carrying more than this number would now have to be modified to a two compartment standard, reduce the number of passengers they carry to the agreed amount or phase out operation, which is based on the A/Amax value, number of persons carried and the ageof the vessel. 1- Table 5 Time table for existing ferries to comply with SOLAS 90 standards as accepted at SOLAS 95( 1 compartment vessels) A/Amax. Date of Compliance Less than 85% 1 October 1998 Less than 90% 1 October 2000 Less than 95% 1 October 2002 Less than 97*5% 1 October 2004 More than 97.5% 1 October Table 6: Time table for existing ferries to comply with SOLAS 90 standards as accepted at SOLAS 95( 1 compartment vessels) No of Persons Date of Compliance 1500 or more <No of Pass> October <No of Pass> October <No of Pass>600 1 October Age of the ship equal or greater than 20 years. 9

93 4 THE STOCKHOLM AGREEMENT 4.1 Joint West European Project With the Estonia tragedy shaking once more the foundations of shipping, it forced the profession to provide answers "immediately" and, in attempting to do so, to use the right expertise and experience to provide the right answers. The focus and effort in RoRo and passenger ships capsize safety have suddenly reached the deserved and long overdue intensity. As mentioned in the foregoing, the Nordic countries reacted quickly in undertaking this responsibility leading to a wider-based project within a very short period, referred to here under as the Joint R&D Project. Taking onboard the fact that, in addressing the probability of a ship surviving a given damage, the problem of ~damage survivability does not end with quantifyuing the probability of damage and the consequences of damage. The Estonia disaster was the strongest indicator yet of the urgent need to define acceptable risks and maximum tolerable consequences as well as to identifying procedures for managing such consequences and dealing with the residual risks. During the project effects of different structural arrangements, loading conditions, freeboard, sea-state have been considered Folowings are the brief outcome of the Task 5 of Joint West European Project related to the development of.stockholm Agreement CRTCAL AMOUNT OF WA TER ON DECK To validate the main observation that the point of no return occurs at a heel angle corresponding to 4,ar, comparative studies between the static and dynamic predictions of critical amount of water on deck were carried out for SHP 2 for four subdivision arrangements of the vehicle deck and three lengths of a flooded compartment located amidships. Additional calculations were also carried out for SHP 1 with open vehicle deck (0) only, for three lengths of a midship compartment. Comparisons are presented in Figures 6.1 to 6.5, based on data provided in Appendix C. The static prediction is understood as the amount of water on deck needed to heel the ship, damaged only below the vehicle deck, to the angle Omar~. The dynamic prediction on the other hand is obtained from the numerical simulations and corresponds to the amount of water on deck accumulated at the point of no return. As can be observed, the comparison is cxcellent, particularly at the operational range of KG-values. The explanation of this impresive agreement is simple if one considers the quasi-static behaviour of the ship at the point of no return. The heeling moment acting then on the ship derives from the water accumulated on deck. The nature of this * moment is almost static though the moment itself is the result of dynamic effects, as the elevation of water on deck is produced entirely by the dynamic action of waves. As a result of the static nature, the maximum sustainable angle of heel cannot be greater than approximately the angle 4,,x, depending on the variation of this moment with the heel angle. n accordance with the nature of a critical sea state, the critical amount of water on deck is also a random quantity - a fact which should be kept in mind when examining Figures 6.1 to 6.5. Because of its random nature, points corresponding to the dynamic prediction do not always lie on a smooth curve. to

94 MAN OBSER VA TONS One of the most valuable observations which was made during the model tests results is that the damaged ship in waves behaves quasi-statically when it reaches the 'point of no return'-but reaching this point and the time to reach this point are determined by the dynamics of the ship. Quasi-static behaviour of the ship at the point of no return denotes lack of roll motion, which is marginal at this position of the ship, while the other ship motions such as sway, heave and possibly pitch (in case of flooded compartments other than those midships) remain in practice unaffected by the quasistatic heel of the ship. Heave is then particularly important. Appreciating the existance of the point of no return, so characteristic of the behaviour of the ship in sea states close to the critical one, was a turning point in developing the static equivalent method. Having analysed the theoretical model for the prediction of the dynamic behaviour of a damaged ship in natural waves, and examined a large number of results for different loading conditions, sea states and deck arrangements, the following additional observations were made: " The point of no return is reached when the ship has attained approximately the angle of heel Oma&, at which the maximum of the GZ-curve occurs. This angle, at a large majority of cases, is less then 10 degrees. Reference is made here to the GZcurve calculated traditionally, using the constant displacement method and allowing for free-flooding of the vehicle deck when the deck edge is submerged. * The amount of water on deck at this point can be predicted from statical stability calculations considering a flooding scenario, in which the ship is damaged only below the vehicle deck but with some amount of water on the (undamaged) deck inside the upper (intact) part of the ship. The critical amount of water on deck is such that the ship assumes an angle of loll (angle of equilibrium) 6 e that equals the angle Omax, determined previously. " n the process of seeking a generalised damaged stability criterion, the following quantities describing the above scenario at the point of no return were considered: => the elevation of water on deck above the sea level, h, and =' the depth of the deck edge below sea level, f, measured at the centre of damage at the inner shell of wing spaces, if any. 11

95 / 8T-J 41., WL o DANSH PROPOSAL The DM experiments were designed to investigate the water ingress phenomena, comprehensively. To this end, the water level inside the deck as well as the water elevation outside the damage opening are measured using an array of wave gauges, together with roll and pitch motions including static heeling and trim. The analysis of these results was also to be used to calibrate the developed numerical water ingress model in a range of sea states, conditions and compartmentation as indicated below: * Open RoRo deck * Centre Casing * Side Casings * Size of damage opening (25%, 100% and 200% SOLAS) * Location of damage (midship and forward) e Freeboardfood (0.5m, 1.0m and 1.5m) o Loading conditions (KG ranging from 9.5m to 12.0m) e Transverse Bulkheads (Partial and full height) * Sea States (Hs=l.3m, 3.Om and 5.0m) * RoRo deck damage only n addition to information pertaining to water ingress, valuable information was also obtained concerning the survivability of SHP 2 in these conditions. Video records of all the above were also obtained. 12

96 Arrangement Number of Number of Survival Survival Tests Causizes Heel to ntact Heel to Damaqe Open Deck Open Deck w th Bulkheads Cente Casing Centre Casing with Bulikhead i Side Casig Side Casing with Bulkhead *Second line shows tests with reduced height bulkheads. Summary of Outcome of Model Tests.(Avýtw E Water on Ro-Ro Deck. CW, _. a.)0 >, Actual freeboard (m) Test Data Points for Water Elevation as a Function of Actual Freeboard. tad ý

97 Based on measurements undertaken during these experiments a computer code was written by the SUSS Group to analyse the flooding process with view to producing time histories of the flood water on the vehicle deck. This is explained in Appendix B and sample results are presented leading to predictions of the critical amount of water on deck causing a vessel to capsize. A comparison is also presented between various methods concerning the above. The Estonia tragedy once more shook the foundations of shipping, forcing the profession to provide answers pertaining to the Ro-Ro safety problem "immediately". n this pace of developments and following considerable deliberations and debate, a new requirement for damage stability has been recently agreed among north-western European Nations to account for the risk of accumulation of water on the Ro-Ro deck. This new requirement, known as the Stockholm Agreement, demands that a vessel satisfies SOLAS '90 with, in addition, water on deck. The Stockholm Agreement should be complied by the existing ferries according to the time table, which is based on the calculation of A/Amax. (Table 1). Table 7: Time table for existing ferries to comply with the Stockholm Agreement A/Amax Date of Compliance less than 85% 1 April 1997 Less than 90% 31 December 1998 Less than 95% 31 December 1999 Less than 97.5% 31 December % or higher 31 December 2001 but in any case not later than 1 October 2002 The Stockholm Agreement requires deterministic stability calculations to be carried out by assuming a constant height of water on deck as explained in the following. 2.1 Constant Height of Water-on-Deck Standards Based on a 4 m significant wave height, the height of water on deck should be: 0.5 m if the residual freeboard at the damage opening is 0.3 m or less, 0.0 m if the residual freeboard at the damage opening is 2.0 m or more, intermediate values can be determined by linear interpolation. For ships operating in restricted areas where the significant wave height is less than 4 m, the height of water on deck will be: 13

98 0.0 m if the significant wave height is 1.5 m or less, intermediate values of height of water on deck for significant wave heights between 1.5 and 4 m can be determined by linear interpolation. \ Height of 5 Water on DeckH=4.m M Residual Freeboard (m) Figure 2: Stockholm Agreement (Height of Water On Deck) Details and the application of the Stockholm Agreement are given in Appendix B 14

99 i STRONG-AND WEAK-PONTS OF SOLAS-90 STRONG PONTS * Overall Good.Standard * Takes into Account the Limited Wave Effect e Simple to Understand and Apply WEAK PONTS * Can Not provide Consistent Standard Against the Environmental Conditions for Each Vessel * Deterministic Method STRONG AND WEAK PONTS. OF STOCKHOLM AGREEMENT STRONG PONTS t takes into account the water on deck effect WEAK PONTS * Based on Wrong Sample of Experimental Data o nfluenced by the Politics * Extremely Conservative Compared to the Reality (Model Experiments) * t damages the development of Ro-Ro Concept * The Standard for a Ferry With Freeing Port is Unjustifiable The Ship Stability Research centre 17

100

101 APPENDX A: SOLAS 90 STANDARDS.

102

103 co O. S% <U6 m 0Z tu2 Z Q4 U) 0~r lion *. t o UJ0 L60

104

105 D L~ 4) E t.2 0 S= u. - U" U "E 2. 7 '4 =, m~ 0., UO 0 U E.2. 5 > > 0 ý L. 0, - r¾ 0 Ol0 od = 0, 0. &.C40 ) - "U E 1;>< o

106 2 -E E 0 0 o E -.- o o o ýo 7,2.2 0 z LO t3 -rz ý.o r 0 u ý.0 um 80 t -0 C3 o t.2 -a E 0, M-0 - a- -0 ý : E o= 77 -,7,' Q > > z t 5; 0:2 = ý= tz Zý :2 -a r, 0 r -Z 0 0 L) rl u 0 0 M 0 > 0 o 0 0 o u 0 > 1 > s! ý > mu 2 0 > 0 0 C) > o :rz W 0 -Z 0- j -u CL o -u CL CL oz 0 CL, > ý.o 0 E

107 - 4 CC) - S

108

109 W 0) r.. 0ý' 0)~ " ~ o r.o.oo Jaw - - Do E Mo ~o E~00 0' E0 00. ~' 0. -= ; 0-0 D -c -w; 0) C U mo s cz =:'r- -Do=~A - ~ ~ ~ C- =Z0Z C CO C 0 C > - >~C.m CDs~ w",c 0)x0 0) 0 C) H' x.00 CL= m 0-0) 00-0N ~C U r -- Cc Or aoo > )5 CoSL ~ -au, CO~~r C~ 0 tl a>u' C--~ 1 -C> Sn> =~ ~=z9-..n J~ 0~ o U co 0ZCL C ~ ~.SE 0) '<c CZ CCO c5 to w = 0 0 Qi a -coo : >w -= C) *.C,; 0.

110 MSC Add. 1 ' ~ANNEX 2 Regulation 8 - Stability of passenger ships in damaged condition. 3 'MTe following is added at the end of paragraph : "This range may be reduced to a minimum of 10', in the case where the area under the righting lever curve is that specified in paragraph 2.3.2, increased by the ratio: 15 Range where the range is expressed in degrees." 4 The words "range specified in 2.3.1" in paragraph are replaced by the words "range of positive stability". Regulation Application 5 The following sentence is added at the end of existing paragraph 1: "T'he requirements in this part shall also apply to cargo ships of 80 mn in L, and upwards but not exceeding 100 m in L, constructed on or after July 1998." Regulation Required subdivision index R 6 Existing paragraph 2 is replaced by the following: "2 The degree of subdivision to be provided shall be determined by the required subdivision index R, as follows: for ships over 100 m in L,: R = ( L ), where L, is in metres; and.2 for ships of 80 m in L, and upwards but not exceeding 100 to in length L,: L RO R.-1 - [ 1( _.u *.] R. where R. is the value R as calculated in accordance with the formula in subparagraph. 1." Regulation 45 - Precautions against shock, Fire and other hazards of electrical origin 7 'Me words "55 VW in paragraph are replaced by "50 V".

111 APPENDMX B: STOCKHOLM AGREEMENT

112

113 i Session 6 The new survivability requirements for RoRo passenger vessels from the 1995 SOLAS conference and the 1996 Stockholm regional conferences Tom Allan Chief Surveyor (Ship Construction), Marine Safety Agency UK u (Member of MO Panel of Experts) TOM ALLAN is the Chief Surveyor (Ship Construction) at the United Kingdom Maritime Safety Agency, formally the Marine Directorate of the Department of Transport. He leads the departments responsible for construction, stability, load line matters and safety equipment for all types of merchant ships. He co-ordinates the Department of Transports efforts at the MO subcommittee on Subdivision and Load Line (SLF); Design and Equipment (D&E) and Fire Protection (FP). He leads the United Kingdom delegation to both the D&E and FP sub-committees and takes over as Chairman of the SLF sub-committee this year. Mr Allan has a honours degree in naval architecture and has been with the Marine Safety Agency and its predecessors since 1971 initially as an outport surveyor but latterly in the headquarters policy branches. His experience includes the survey during construction of roro passenger ferries through to the current position of setting new standards for their design and construction. He was also a member of the Panel of Experts set up by MO to develop recommendations and proposals to improve the safety of roro passenger ferries.

114

115 THE NEW SURVVABLTY REQUREMENTS FOR RO-RO PASSENGER VESSELS FROM THE 1995 SOLAS CONFERENCE AND TE1996 STOCKOLM4 REGONAL. CONFERENCES Following the ESTONA tragedy, in which over 900 persons lost their lives, the Secretary-General of the nternational Maritime Organisation (MO0) recommended to the Maritime Safety Committee (MSC) the urgent need to investigate all means to enhance the safety of ro-ro passenger ferries. The MSC accepted the recommendation and agreed to establish a Panel of Experts to develop recomendations and proposals. The Panel proposed a wide range of measures affecting many of the aspects of construction and operation of this type of ship. The most significant change would affect their ability to withstand an amount of accumulated water on the ro-ro deck. This paper addresses the proposals presented by the Panel of Experts to increase the survivability aspects of ro-ro passenger ferries, the procedure by which the Panel addressed them and decisions taken at both the SOLAS and Stockholm Conferences to bring those recommendations into force. ntroduction On 28 September 1994 the Estonian ro-ro passenger vessel ESTONA Capsized and sank in the Baltic Sea. How and why it happened is the subject of the Joint Accident nvestigation Commission (JAC) of Estonia, Finland and Sweden. Could it happen again? To answer this it was clear that the maritime world was facing a major international problem, and that 1240 needed to act - but what could be done? The Secretary General of MO invited all member Governments, with an interest up ofthe recommended Panel of Experts and relevant organisational matters. Temeeting was held at 12M0 on Saturday 10 December The Steering Committee were tasked to determine the terms of reference for the Panel; to supervise their activities; consider the outcome of the Panel's deliberations and to recommend to the next MSC meeting actions to be taken to improve the safety of ro-ro passenger ships. The panel members were then set "terms of reference": 1. to carry out a thorough review of the safety of ro-ro passenger ships; 2. to de~'elop safety measures, even if the risk of incidents was low, for the purpose of lowering the total risk to ro-ro passenger ships taking into account that even incidents with a low probability could have unacceptable consequences for such ships; 3. n carrying out that review the panel were requested to: 1 review the basic design and operation principles of ro-ro ships,.2 identify hazardous situations the ship may encounter during its active life and the associated potential consequences thereof on safety; 3 identify potential consequences deemed unacceptable from a safety standpoint or which need to be mitigated; U 2 1

116 .4 assess present requirements and identify corrective measures in areas not properly covered; and.5 finalise a set of recommendations. 4. Take into consideration the following available information:.1 background information on design and operational experience;.2 results available from research studies; and.3 the outcome of additional studies or model tests performed upon request, if necessary. S. Concentrate on the most important issues of existing ro-ro passenger ships as a matter of high priority in order to submit a set of recommendations to the next Maritime Safety Committee meeting in May Submit proposals to the Steering Committee for consideration, proposing.1 which of the Panel's recommendations should be taken immediately;.2 which recommendations might necessitate amendments to mandatory instruments; and *.3 which items should be tasked to relevant Sub-Committees for further consideration.. The main questions raised were assembled into groups and allocated to specific "expert" sub-groups within the Panel. They covered the whole range of safety concerns and fitted broadly into the following categories: stability of the ship in intact and damaged conditions; the implications of accidents resulting in water on the bulkhead (ro-ro) deck;. measures to prevent such accidents and their consequences; construction of the ship, especially of hull doors and closing devices; basic ship design, including the design of the ship to facilitate evacuation; the human factor and potential areas of human error; * operational factors, including the closure of watertight doors and the lashing of vehicles; - the interface between passengers and the shipboard safety system; * crisis management, including access to information needed to manage * crises effectively; * training required for personnel to enable them to deal with the special characteristics of, and special requirements on board ro-ro ships; U communications, both within the ship and between the ship and the outside world;.. the adequacy of lifesaving appliances; * the adequacy of search and rescue arrangements; overall safety assessment and risk analysis; 3

117 * the ship/port interface. ship/operator relations including the international safety management code; and survey and inspection. This paper concentrates on the first two of these subjects which collectively come under the heading of survivability. Panel of Experts Proposals on Survivability New regulations proposed by the Panel of Experts which would have had a major impact on ferry survivability were the new proposed regulations 8-1, 8-2 and 8-3. The regulations applied new standards to ferries in both the intact and damaged conditions as follows: Regulation 8- - stability in the damaged condition Regulation stability in the intact condition Regulation one compartment ships Regulations 8-1 and 8-2 both addressed the problem of water on the car decks. REGULATON 8-1 The main difficulty for the Panel members was that they had identifying been tasked solutions with in a very short period of time. t possible was not considered for them to propose anything which would require further t research. was however accepted that from a naval architectural point of revised view cz that criteria would be the best that way it forward would however not be possible it was also to accepted devise and validate new criteria in the time available. t was at this stage the sub-group on survivability considered items, proposals within the from Panel, the Nordic countries and the Society Architects of and Naval Marine Engineers in North America (SNAME) to address of accumulated a quantity water-on-deck. the Panel to accept t was one agreed of the among main conclusions the stability of the sub-group UK post HERALD of ENTERPRSE research. That conclusion OF was FREE that:- Ro-ro vessels constructed capable of to avoiding meet the rapid standards capsize after of SOLAS an assumed '90 were moderate sea states with a significant extent of wave damage height in up to 1.5m.* The sub-group therefore agreed that, as a minimum, all ferries should have to comply with the SOLAS '90 standard. The dilemma was how to increase that standard to a level which would be expected to address the problem of wateron-deck in a reasonable sea state. An example of "SOLAS '90" as against the previous residual stability standards can readily be seen in fig 1: L FigT j. 4 1

118 Having accepted wave heights of 1.5m as a minimum, in association with a residual freeboard of 0.3m, in what sea state should we expect ferries to survive? After considerable debate this was taken as 4.Om and came from available statistics that 99% of all collisions occur in sea states up to 4.Om significant wave height. t was also felt that this was a sea state that could be recognised as a significant standard for survival. While views were expressed that further study and information was needed to decide on the amount of water-on-deck the group agreed to require 0.5m'/m 2 as a maximum based on: 1 the proposal by the Nordic countries; 2 a study by SAHME suggesting 0.5m was a reasonable level for 4.Om significant wave height (N); 3 model tests in Finland which indicated 0.75m water on deck at 4.Om h,, and 4 recent work within the North West European research programme which had also confirmed that the panel's proposal was of the right order. t wa loaccepted that residual freeboard was a primary factor in 2.m survival. was agreed as the residual freeboard to provide sufficient reserve to prevent SNAME. water accumulating on the freeboard deck, again based on the study by From earlier work carried out by the United Kingdom Marine Safety Agency (MSA) it was considered essential that the vessel's survivability standard be unmproved first before considering very low water-on-deck. survivability This standards work large showed numbers that for necessary before any of significant transverse barriers increase may in survivability be could be gained. The panel therefore proposed a distinct four step process: 1. First the vessel should comply with SOLAS '90; 2. calculate the minimum residual freeboard (fig 2.); 3. Add the appropriate amount of water on deck depending on residual freeboard and sea state in the area of operation (figs 3&4); and 4. Comply with SOLAS '90. From this, benefit was given to those vessels with enhanced residual freeboards and took account of those ships which may operate in restricted sea states. A series of guidance notes was developed to explain the step process and other areas which required clarification.... " ~..L Fig 2 Fig fr, 2.0 metres. water on deck (x) = 0.0 m 2 f,5 0.3 metres, water on deck (x) = 0.5 m

119 Fig 4 1. h, a 4.0 metres. water on deck caicutated as per Figure 3 2. hs, 1.5 metres. water on dec (x) = 0.0 -m For e;arnjle, -. Where: fr = 1.t5M, and h, = 2.75m Hypothetical water on deck (x) = 0.125m Regulation 8-2 Regulation 8-2 was an attempt to address the "HERALD" and "ESTONA" type casualty. That is water on intact the car condition deck but below with this the deck. ship remaining The volume in of an water assumed to accumulate at the fore end was proposed to be l.om 3 /m 2 (fig 5). The greater amount was to take account of the scooping effect from forward speed. Likewise an amount of water was assumed to enter through the stern, in the event of stern door damage, to account for following seas but as scooping would not be a feature then the amount assumed to enter the aft end was reduced to 05m3/S/m 2 (fig 6) This regulation as written virtually forced the fitting of a transverse barrier aft of the inner door and proved to be a very contentious point before and during the SOLAS Conference. mv' Barner\ Ramp or inner door (emension to colion bhd) aft barer P -Y t s - ' T n Wl.f. mno 1, intac i fn ýr h a signjr"g; wm h eiht hen..e W0'en :0.U T..x :0 rn9mt Fig 5 men, L " oorn'14 Fig 6 Regulation 8-3 This new regulation addressed the one compartment standard passenger ship. t was the aim of the Panel to prevent the construction ships of new intended one compartment to carry large out numbers such existing of passengers ships which and to carry eventually large numbers phase of passengers. While there was unanimous agreement within the Panel on how to deal with new constructions there was much debate on the number of passengers which or should persons be permitted to be carried on existing one compartment ships and on the possible phase out of ships carrying Panel proposed large numbers restricting of passengers. the numbers The other proposals of passengers for limiting to 250. the numbers There were to 400 persons or 500 persons. 6 -.rqj l 'l ~ ille ln l m n m~ a u m

120 NTERSESSONAL MECTNG As no agreement on a survivability was not sufficient standard could time be achieved to consider at MSC and many there of the other issues raised Panel by it the was agreed to hold an -intersessional" meeting prior Conference. to the SOLAS However while significant progress was made intersessional on many items at meeting the as far as survivability was concerned agreement was hampered by the need to consider additional alternative proposals. Alternative Survivability Proposals to the SOLAS Conference Other proposals were put forward to the SOLAS Conference, one by one France by the and Russian Federation. The states proposal associated by France with was 95% based of on collisions the sea instead of the panel's 99%. reduced This the relevant sea state to 2.2m with an associated reduction of the amount of assumed accumulated water-on-deck ie: Panel France Min stability std SOLAS '90 SOLAS '92 Min freeboard 0.30m 0.30m Max freeboard 2.00m 1.10m Min. 1.50m 1.50m Max l 4.00m 2.20m Water on deck O.Smi/m2 0.14m3/m2 The Russian Federation proposed an increase in the GZ curve criteria the assumption based on that SOLAS '90 was sufficient for survival in sea 3.0m states significant up to wave height. This was a simplification of research just completed in Canada. The Canadian conclusion related to freeboards ranging from 0.5m Panel' to 1. 5m s proposals. in sea states from 1. 5m to 3.Om which was in line with the The Canadian research also showed that to achieve the range criteria of the GZ curve the area under the GZ curve was about 3 to 4 times that required to comply with SOLAS '90. The Russian Federation option however only proposed an area marginally in excess of that required for SOLAS '90 ie an increase in area from 0.015m.rads. to m.rads for sea states up to 5.0m h, with an interpolation between wave heights of 3.Om and 5.Qm but no interpolation for residual freeboard SOLAS CONFERENCE At the SOLAS Diplomatic recommendations Conference the were vast consolidated majority of the into Panel's modification of existing a series regulations; of new SOLAS regulations; Conference Resolutions, Assembly Resolutions and MSC Circulars. However those areas relating specifically "survivabilityto still proved to be the most contentious. Regulation 8-1 t was always considered unlikely that any of the above proposals on survivability standards would achieve a majority acceptance at the Conference and that regional solutions, which would allow groups of countries to apply the higher or lower standards in their area, was a more realistic possibility. 7

121 H This in fact proved to be the case. At the SOLAS Conference a majority countries of rejected the Panel's proposal for a worldwide survivability of SOLAS standard '90 plus S0cms of water on deck. After extensive debate a majority of countries proposed that the SOLAS standard '90 be accepted as the worldwide standard. that Many the countries application considered of the SOLAS '90 standard to all existing major ferries step forward was a and should be regarded however as a a significant sufficient number standard of countries worldwide; this in view North and West wished Europe to disagreed apply a higher with standard in their areas. The final conclusion of the SOLAS Conference on survivability matters was: 1. A.11 existing ro-ro passenger ships shall comply with the SOLAS survivability '9a standard based an the vessel' s A/Annx H value not than later the date of the first periodical survey after the following dates ofcompliance: HValue of A/Aman Data of compliance less than 85% 1 October % or more but less than 90% 1 October ormore bt less than 95% 1 October %or more but less than 97.5% 1. October r m r 1 O t b r A SOLAS Conference Resolution (Resolution 14) to permit two or more Contracting Governments to apply specific stability requirements to all ro-ra passenger ships undertaking regular scheduled voyages between designated ports of those Contracting Governments. The Resolution was strange in that it only permitted specific requirements stability that did not exceed those proposed by the Panel of Experts were detailed which in an annex to the Resolution. An Agreement based on the Resolution was not to come into force until 12 had months been after notifiled it's to acceptance the Secretary General of M(0. all The Contracting Resolution Governments also urged to apply the provisions of any such Agreement on ro-ro passenger ships entitled to fly their flags when engaged scheduled on regular voyages between designated ports covered by such Agreements. Regulation 8-2 At the intersessional meeting, held prior to the SOLAS Conference, countries many spoke against this proposed regulation mainly they felt proposals that other addressed the same issue arid that in some cases as many doors as would four be needed to address the ingress of water from the fore Bearing end. in mind the application of the revised nternational Classification Association of Societies (ACS) Unified Requirements to bow door arrangements and the amendment to regulation 10 of SOLAS it was eventually sufficient agreed additional that protection at the fore end bad now been The established. Conference therefore agreed to the deletion of this proposed regulation. Regulation 8-3 A compromise to this proposal was quickly reached. t agreed that the number of persons carried on one compartment standard ro-ro passenger vessels restricted be to 400 persons. t was also readily agreed that no new one compartment passenger ships would be permitted, after 1 July 1997, to carry more than 400 persons. Likewise all existing one compartment ships carrying more than this number would now have to be modified to a two compartment: 8

122 standard, reduce the number of passengers they carry to the agreed amount or phase out of operation. The phase-out programme eventually agreed was based vessel's on a A/Ama~x combination value, of the the number of persons carried age and and not a final later criteria than the of date of the first periodical of compliance survey after prescribed the date in 1, 2 or 3 below which occurs the latest: 1. Value of A/lemax Date of compliance less than 85% 85% or 1 October more but 1998 less than 90% 1 90% October or more 2000 but less than 95% 95% 1 or October more but 2002 less than 97.5% 1 October 97.S% 2004 or more 1 October Number of persons permitt~ed to be carried 1500 or more 1 October o 1000 or more oebtls but less than h Otbr or more but less than 1000 October October Age of the ship eanal to or greater than 20 years THE FRST STOCKHOLM CONFERENCE Tefirst Stockholm Conference was instigated by seven countries which became Norway, known as Sweden the "Gothenburg and the Group" United viz Denmark, Kingdom. Finland, t was Germany, primarily reland, of this group through of the countries efforts that the SOLAS Conference confirmed Resolution and accepted 14 was at the SOLAS Conference, t was agreed that Sweden, who lead had and suffered host most a Conference as the result to of take the ESTONA forward tragedy, the should demand take for the higher survivability standards for ro-ro passenger ships operating in North West Europe. Little progress was made on the acceptance of a higher standard to address the effect of water on the car deck. On the first day of the Conference two alternatives to the Panel's proposal were tabled by Belgium and Denmark. The Belgian proposal was a compromise proposition assumed which accumulated simply reduced amount the of water-on-deck (as proposed deck edge by the was Panel) submerged. as the At each angle of heel the would amount naturally of water have accumulated which on the car deck was calculated to be deducted amount. from This the could in some cases have the negative effect of amount having of water-on-deck. a The Panel'*s calculated proposal amount, maintained as additional the water, on top of the angles water accumulated of. heel where at the deck edge was submerged. Conference By the end the of Belgian the proposal had little technical support. The proposal put forward by Denmark, as a counter to that from Belgium, was based on the initial results of research being carried out within the North West European research project for application to new vessels. This research, sponsored by Denmark, Finland, Norway, Sweden and the United Kingdom, identified a relationship between the height of water entrained on the car deck similar to that proposed by the Panel. The main difference however was that the amount of water on deck related to a hydrostatic pressure head rather than a fixed volume of water. The Conference debated all three proposals, however would not it was be possible accepted in that the it time available to reach an agreed solution. t 9

123 wasmtherefore agreed that each country shoul.d examine the proposal emanat ing frmthe North West European research, compare it with the Panel's proposal adreport back to the second Stockholm Conference to be held On 27 and 28 February t was additionally agreed, as all three survivability proposals depended on the significant wave height in the area of operation, that all participating Administrations should meet before the second Conference to agree the si.gnificant wave heights for year round operation on all routes operating from their ports. THE SECOND STOCXOLM CONFERENlCE The Regional Agreement was finally accepted at the second Conference held Stockholm in on 27 and 28 February A more consensual view was prevalent at this meeting. There was an eagerness to come to a satisfactory conclusion for the benefit of all, particularly for the operators who had been left in a state bf limbo since the SOLAS Conference in November The final Agreement, which will come into force in 1 April 1997, applies formally a new survivability standard to ro-ro passenger ships operating regular scheduled services between, or to or from, parts within an area its southern with boundary at Cape Finisterre (Spain) and its northern *Northern boundary at Norway. The area includes celand, the Farces and the Baltic The Sea. Agreement does not extend to the Mediterranean though France is pressing that it should. state it to will decide apply to third how flag this ships is though to it be will done. be for each ToAnnexes are appended to the Agreement, the first relates tosinfct wave heights and the second to the standard of survivability to applyica. Significant wave Heights The significant wave heights to be used to calculate the height of water-ondeck a ship will have to withstand are defined by area and were based on allyear round statistics. Maps detailing the agreed wave heights for North West Europe including the Baltic are at Annex 1 to this paper. Provision has been made for established operators to operate ships on a seasonal basis with a significanit wave height relevant to that period. particular This provision seeks to exclude predatory seasonal operators who wish may to operate summer only service at a reduced standard. Survivability ii The results of the new calculations using the proposal from the North European West research were presented and considered, a sample of the examined ships is at table 1. The general view was that this new method calculation of was technically more sound than that proposed by the Panel and appeared to more accurately reflect the phenomenon of flooding onto the car deck after a collision. Those members of the Panel's stability sub-group in attendance at the Conference agreed with that conclusion. f the information from the research had been available during the Panel's considerations there was little doubt that this new method would have been preferred by the Panel.- t was unanimously agreed that the standard was equivalent to that proposed by the Panel since the overall effect on existing ships would not be different to the effect from the original Panel proposal. 10

124 WATER ON DECK COMPARSON OF PROPOSALS SHP RESOLUTON 14 STOCKHOLM AGREE4ENT (PANEL OF EXPERTS) SOLAS '90 3 Barriers 2 Barriers MODFED B STAB '80 1 Barrier + No Barriers m sponson 1.25 m saonson C (A265) 1 Barrier No Barriers D SOLAS '90 1 Barrier 1 Barrier MODFED E SOLAS '90 AS BULT 2 Barriers No Barriers F SOLAS '90 4 Barriers + 4 Barriers + MODFED C.L. Bulkhead C.L. Bulkhead G STAB '80 4 Barriers + 5 Barriers + C.L. Bulkhead C.L. Bulkhead H SOLAS '90 3 Barriers + 2 Barriers + MODFED C.L. Bulkhead C.L. Bulkhead STAB '80 3 L Barriers + 2 Barriers C.L. Bulkhead Table The presentation of the calculations showed that the vessels built to SOLAS '90, with a higher inbuilt survivability, performed better in the calculations in that fewer modifications were necessary to meet the new standard as against that proposed by the Panel. While those vessels which only just comply with SOLAS '74 would probably require more modifications. The effect vessels on those in the mid ranges between SOLAS '90 and SOLAS '74 was generally neutral. The main advantage in the new method was that benefit was now being given to those craft with a higher inbuilt survivability standard (GM) in addition to the benefit of larger freeboards. The New Standard The new standard is essentially the same as proposed by the Panel in that the initial amount of water assumed to accumulate on the car deck is the same taking account of residual freeboard and relevant sea state in the area of operation. The main difference is in how the water is treated within the calculation.

125 n the former proposal when a barrier an the car the deck amount was of assumed water constrained damaged then within over the both largest the compartment damaged compartments was distributed (fig 7). n the new requirement of the the accumulated height water is maintained as a constant plane on both sides of the damaged barrier (fig 8). d = 12 A' A, + A, W fh d t nws bilnd be,9 ý tmmý e, = dpth = d, L F W4 Fig 8 Fig 7 At each angle of heel the 'Panel maintained a constant volume of water within the space (fig 9); in the new proposal the height of water is maintained as a constant plane above the deck edge or above the still water level if the deck edge is immersed (fig 10). Resolution 14 Fig 9 12 P

126 b. Fig 10 n simplistic terms the amount of water assumed to accumulate on the car deck is initially greater, in those cases where a barrier is assumed damaged, under the new requirements however this amount reduces with each angle of heel the until deck edge is immersed when the amount of water then starts to increase again as water naturally enters the car deck space. As stated earlier the ships with a high survivability standard have the advantage of the larger freeboard and GM to withstand the reduced amount of water on deck while the other ships have to withstand larger quantities and will therefore require more modifications to survive. Transverse Bulkhead (Barrier) Heights. A further change of note within the Stockholm Agreement is the height of any transverse or longitudinal barriers fitted to enable the vessel to achieve the new standard. Resolution 14 proposed a standard height of not less than 4.0m. The new proposal relaxes that requirement in that 4.Om the in height cases shall where be the at water least depth is 0.5m; for the lesser bulkhead depths may the be height calculated of in accordance with the following: B,= h where Bs = bulkhead height; and h.. = height of water on the car deck. n any event the minimum height of the bulkhead shall be not less than 2.2m. This height was considered to be the average height of hanging car decks when in their lowered position. 13

127 To address the problem of "splash over" in ships with hanging car decks the mini~mum height of the bulkhead is to be not less than the height to the underside of the hanging car deck when in its lowered position. On the carried out within the North West European research project and a recent model test on an existing ferry for compliance as an equivalent to the new standard, would strongly recommend that all transverse and/or longitudinal barriers be as near full height as possible within the car deck space. The problem of splash over of considerable amounts of water into adjacent spaces should not The agreed implementation procedure for the new standard is again based on the vessels A/Amax value but on a shorter proposals: timescale than the SOLAS Conference less than 85% 1 April % or more but less than 90% 31 December % or more but less than 95% 31 December % or more but less than 97.5% 31 December % or more 31 December 2001 but in any case not later than 1 October Finally the Stockholm Agreement has a further Resolution attached which acknowledges implementation of the specific stability requirements shall apply not later than the above dates. The Resolution further that urges Administrations to bring one compartment re-ro passenger ships up to compliance as soon as possible and agrees that Contracting Governments can, by agreement between them, apply earlier implementation dates than those specified above for ships trading between their ports. Value of A/Amax Date of compliance CONCLUSON The acceptance of SOLAS '90 as the new global st~andard for all existing ferries is a justification of the stand taken by the United Kingdom back in 1992 and is a significant step forward in the survivability of re-re passenger ships worldwide. Those of us within the "Gothenburg Group" did not achieve the enhanced worldwide standard we would have wished but at least the MO0 Resolution (14) permitted the opportunity to achieve our goal through a Regional Agreement. Such Agreements which come under the auspices of 1M0, allowed us to take advantage of current research and take forward the overall aim of the Panel of Experts to apply higher survivability standards to address the fundamental problem of water-on-deck for ferries operating to and from ports in any area. The new Regional Agreement urges all Contracting Governments to enforce the provisions of the Agreement on ro-ro passenger ships entitled to fly their flags when engaged on regular scheduled voyages between designated ports. While there was some disappointment in the UK and Scandinavian countries in not achieving our primary aim that the new higher survivability standard would apply the water-on-deck scenario worldwide, a great deal of satisfaction can be felt for the overall increase in the safety standards standard of existing re-re passenger ferries and the setting of standards for new ferries. The views and opinions expressed in this paper are those of the author and may not necessarily be those of the United Kingdom Marine Safety Agency. 14

128 sai Aberdee O NORTH SEA PELAND Dubin Kolyliead epc Sounhamptm, J.j 3.-~ssL n o -*W r fils 1 a ---, -

129 rv- The, wave heights state, on this map are ti significant wave heights (Hr) which is not exceeded tva probability * of mbre than 10% and siauld be used * f~ordi termining the height of water. t AV =

130

131 511 40/NE.14 NTERNATONAL MARTME ORGANZATON 5 July 1996 SUB-COMMTTEE ON STABLTY AND LOAD LNES AND FSHNG VESSELS SAFETY' MO Ct~jellENGLSH ONLY 40th session Agenda item 19 ANY OTHER BUSNESS Agreement concerning specific stability requirements for ro-ro passenger ships undertaking regular scheduled international voyages between or from designated ports in North West Europe and the Baltic Sea Guidance notes on the annexes of the Agreement Submitted by Denmark, Finland, Norway, Sweden and the United Kingdom Reference is made to MO circular letter 1891 by which Member States have been informed of the above-mentioned Agreement made in accordance with SOLAS 1995 Conference resolution 14. t is considered that annexes and 2* of this paper are the appropriate guidance notes for the uniform application of annexes 1, 2 and 3 of the Agreement. The Sub-Committee is invited to note the contents. Due to the limited number of copies available, annexes 1 and 2 are distributed on the basis of one copy per delegation. 1.,SLF\40\NF. 14 o ; easons wofeconomy, this document is printed in a limited numb~er. ojeiea.tes arm kindly, sked obrinq their cu-oies to meetings and not to reauesi additional coies,

132

133 Page 2 CONTENTS Page Number 1. AMS OF THE MODEL EXPERMENTS MODULE 3 2. NTRODUCTON 4 3. MODEL TEST PREPARATON General Particulars of SHPA Selection of Damaged Cases Construction of the Model 3.4 Loading Condition and Ballasting 7 4 MODEL TEST EXECUTON nclining Experiment Determination of Roll Radius of Gyration Determination of Pitch Radius of Gyration Determination of Natural Roll Period in Damaged Condition Environmental Conditions Survivability Tests MODEL TEST RESULTS POTENTAL PROBLEMS AND RECOMMENDED SOLUTONS Paragraph 2 - Ship Model Paragraph 3 - Procedure for Experiments CONCLUSONS 29 ANNEX MO Circular letter No. 1891and SLF 40/NF APPENDCES APPENDX A Hydrostatics and Static Stability Calculations 38 APPENDX B Drawing for Test Damage Conditions 48 APPENDX C Photographs of the Model 55 APPENDX D Wave and Roll Motion statistics 59 APPENDX E Time Histories of Model Tests 66 APPENDX F Physical Modelling and Similitude of Marine Structures 73 APPENDX G Weather Statistics 86 Total Stabilitv Assessment of Damaged Passenger/Ro-Ro Vessels [Model Experiments Route]

134

135 Page 3 1. AMS OF THE MODEL EXPERMENTS MODULE The principal aim of this module is to introduce to the industry the concept of performance-based criteria for assessing damage survivability of passenger/ro-ro vessels by performing experiments of damaged models of ships which comply with the SOLAS regulation -1/8. Specific objectives of this module include the following:. To present a typical test case and associated relevant information concerning model test preparation, execution and reporting procedures. To highlight potential problems and make recommendations concerning necessary amendments to safeguard the uniformity of the model test method as well as improvements aiming to promote the adoption of a more scientific basis in the model experiments route, leading ultimately to the standardisation of such tests. 7t9al Stahilitv Assessment of Damaged PassengerlRo-Ro Vessels 11Model Experiments Routej

136 2. NTRODUCTON Pg The limited understanding of the complex dynamic behaviour of a damaged vessel and the progression of flood water through the ship in a random sea state has, to date, resulted in approaches for assessing the damage survivability of ships that rely mainly on hydrostatic properties with potentially serious consequences concerning the lass of life and property and danger to the environment. One, only too well known example and the focus of this course, concerns the serious flooding of ships with large undivided deck spaces, such as Ro-Ro vessels, where the loss could be catastrophic as a result of rapid capsize, rendering evacuation of passengers and crew impractical, with disastrous (unacceptable) consequences. Concerted action to address the wateron-deck problem in the wake of the tragic accidents of the Herald of Free Enterprise and more recently of Estonia led to the proposal of new stability requirements, known as the Stockholm Agreement, or more commonly as SOLAS '90+50, pertaining to compliance of existing Ro-Ro vessels with SOLAS '90 requirements whilst accounting for the presence of a maximum 0.5 mn height of water on the vehicle deck, [ANNEX]. The dates of compliance with the provisions of the agreement range from April 1, 1997 to October 1, To "cushion" the impact produced by what is perceived to be too high a price to pay to comply with the proposed new measures and in view of the uncertainties in the current state of knowledge concerning the ability of a vessel to survive damage in a given sea state, an alternative route has been allowed which provides a non-prescriptive way of ensuring compliance and one hopes enhanced survivability, namely the "Equivalence" route, by performing experiments of damaged models of ships which comply with the SOLAS regulation 11-1/8, [ANNEX]. n response to these developments, the shipping industry, slowly but steadily, appears to be favouring the model experiments route, implicitly demonstrating mistrust towards deterministic regulations, which admittedly lack solid foundations. n relation to these developments, it is worth noting that a clear tendency of moving from prescniptive to performance-based safety regulations is emerging internationally. ntroduction of performance standards is seen as beneficial from industry as these allow considering alternative designs as well as a rapid implementation of technological innovation. n this respect, the analysis of alternative design solutions, particularly in the case of Ro-Ro ferries, requires the development of a standardised approach to demonstrate compliance with the requirements of intact and damage stability regulations by a combination of physical model experiments and numerical model tests whilst applying probabilistic techniques in the process of assessing damage survivability. This module concentrates on the development and implementation of model experiments as an altrnaiveroute to achieving compliance with the Stockholm Agreement requirements. This entails performing a series of physical model experiments in the relevant operational sea state and in the two worst damage conditions as defined in ANNEX. To facilitate an illustration on the model test preparation as well as the execution and reporting of model experiments, SH-PA, a SOLAS '90 ship will be used as an example in the following. The operational sea state for the vessel in question, characterised by the significant wave is 4m. Some relevant information on weather statistics is provided in Appendix G. Total &tabilzev Assessment of Damaged Passenger/Ro-Ro Peýssels [M1odel Experiments Route]

137 3. MODEL TEST PREPARATONPage 5 A check-list pertaining to model building, model preparation and experimental set-up and trial tests is given in Tables 4 to 6 to provide guidance and information on what is necessary for this phase of the procedure. 3.1 General Particulars of SHPA. Full Scale SHPA is a Car/Passenger ferry working in the West Coast of Scotland and Western sles. She was built in 1993 according to SOLAS '90 damage stability requirements. Table 1: Principal Particulars of SHPA (Figure 1) Length Overall, LOA m Length Between Perpendiculars, LBp m Breadth 15.8 m D dk m D=_,'dk 4.50 m draught m Displacement tonnes Operational KG 7.18 m Maximum Allowable KG m ntact Design GM 2.21 m She has B/ side tanks below the car deck and these side tanks are cross-connected. There are also side casings at the car deck and mezzanine deck level. There is in addition a centre casing, which is slightly off the centre towards the starboard side. Model Scale A model of SHPA was designed and built in the scale of 1:23.5, resulting in a model of 4 m in length. According to the guidance notes of the model test method, the model length ought "to be at least that corresponding to 1:40 scale but not less than 3m. Therefore, a 3m model would have been acceptable in this case. Opting for a larger model, however, will contribute to reducing scale effects. Relevant information on physical modelling and similitude is provided in Appendix F. The model particulars are given in Table 2 below. Length Overall, LOA Length Between Perpendiculars, L 8 p, Breadth mm Table 2:_ Model Particulars of SHPA (SCALE 1:23.5) mm mm mm D dc~k mm Total Depth mm draught 134 mm Displacement (Target) kg Maximum Allowable KG mm ntact Design GM 94.0 mm Total Stabilitv Assessment of Damaged Passenger/Ro-Ro Vessels [Model Experiments Route]

138 Page 6 The model was constructed on the basis of drawings provided by the owner as indicated in Table 3. Table 3: List of Drawings Title Drawing No Date Lines Plan D -01 (1 of 4) General Arrangement (1 of 2) General Arrangement (2 of 2) Capacity Plan Docking Plan Skeg 589 -H Profile and Decks H - 01 (1 of 2) Pumping Plan 589 -E 03 (1 of 2) Supplementary nformation: Damage Stability Analyses (Official submission of Damage Stability Analysis for Approval by the relevant shipyard) Level Keel Hydrostatics of ntact Vessel For the selection of the damage test cases, KG curves will also be necessary. 3.2 Selection of Damage Cases The damage conditions to be tested were selected according to the information provided in ANNEX 1, as defined in the following: * The worst damage is to be taken as that which gives the least area under the residual stability curve up to the angle of maximum GZ. * f the worst damage location according to SOLAS 90 is outside the range ±10% LBP from amidships, a second even keel damage condition is to be selected in the midship area with regard to residual freeboard. Worst SOLAS Damage (Figure 2) According to the documents provided by the relevant shipyard the critical maximum allowable KG is m at even keel condition. This KG value is determined from the worst damage condition, which is damage case 3 (Appendix A). Using this critical KG value, case 3 was also confirmed as the worst damage case according to calculations carried out at the SSRC. Midship Damage (Figure 3) From calculations carried out at the SSRC as well as the damage stability results from the relevant shipyard, it was found that case 6 is the worst midship damage condition with regards to the minimum residual freeboard in the midship area. Results are provided in Appendix A. Total Stabifitv Assessment of Damaged Passenger/Ro-Ro Vessels [flodel Experiments Route]

139 Pa Page 7 Permeabilitie The permeabilities taken into consideration are as follows, unless stated otherwise. Machinery = 0.85 Provision Compartment = 0.95 Tanks = 0.95 Car Deck = Construction of the Model The model was constructed by using GRP, plexiglass and foam. The shell of the hull is made of 3 mi thick GRP. The skeg, fenders and bilge keels were modelled, but appendages such as rudders, thrusters, fin stabilisers were not. The transverse watertight bulkheads below the car deck were constructed by using plywood covered with fibreglass. The intact tanks below the car deck were constructed using foam. The car deck was made of 5 mm clear plexiglass to observe any leakage and thus take action immediately. The centre casing and the side casings were again constructed by using foam. Those side casings, which had to be damaged, were also made of plexiglass. Two vertical bars were fixed at the aft and forward ends of the model with a horizontal bar, which can be moved vertically, attached to these two bars and carrying most of the ballast weights. The vertical centre of gravity can be adjusted by moving the horizontal bar vertically as shown in the photographs of Appendix C. Aluminium bars are attached along the top edges of the model both longitudinally and transversally to increase the stiffness of the model. 3.4 Loading Condition and Ballasting From the damage stability calculations carried out by the relevant shipyard the following intact loading condition was used. Shipyard SSRC Draught (moulded) 3.15 m 3.15 m Displacement (moulded) 3,112 tonnes 3,090 tonnes GM 1.65 m 1.65 m KG mn m t should be noted that the full displacement of the vessel including the shell thickness is approximately 3,125 tonnes at 3.15 m draught. The intact draught and the loading of the vessel was checked and approved by the authorities. Because of the method of construction of the model, it is difficult to achieve the displacement exactly. n this case, the mass of the model was measured to be 237 kg (fresh water), which corresponds to 3152 (seawater) tonnes of displacement in full scale. The difference of 27 tonnes can, in part, be accounted for by the fact that thrusters and fin stabilisers were not modelled. Total Stabilitv Assessment of Damaged PassengeriRo-Ro Vessels [Alodel Experiments Route]

140 L z i * 0 QL LLL L LL) cc z ws P:<F-Z Cý L z L 0U<cý <

141 L L * -- z < 0 - *0 z QL / L ] L L 0.1 <-zcw C4C Cc > r Z0C/ -n z r, <6 <L L L L- C)

142 nl w - * Cý nljl U ý 2 c C,,L w nz F -J c ) L nv w jc LjL Cj u L L z:z w L C /) Li/) 0 a 6n ~-a '

143 -,- g1 om / L ', L.1 \ =..,, 81=1\X R ;' " ' "a 7 C Ca

144 -,at -' / - U \~>~,W \ atat U ~ *.~.w CD o = * XX Ci) - S C _ U N Cl -~ * r t L2 -= -

145 LiLU H - H * oj L C LU -i - C LU '-ix L /JN o c A,~rALU 1 L L Nj LU w * * 'T~7 Cft *,' U, * U z -- 1 S L'G ~ -~ -. en * -- 8 *1 1~ * ft

146 4. MODEL TEST EXECUTON Page 14 This phase of the procedure is summarised in Tables 14 and 15, referring to the programme of the model tests and the format of the protocol for these tests, which is submitted to the authorities beforehand, for approval purposes. 4.1 nclining Experiment n order to achieve the required KG (GM) value, an inclining test was performed. Target GM=66.8 mm Table 7: Model Particulars for the nclining Test - - tem Full Scale Model Scale Displacement 3152 (sea water) 237 kg (fresh water) KM 9.10 m mm Target GM 1.57 m 66.8 mm Horizontal Distance between weights (d)652.5 mm Vertical position of weights mm from base line mm Procedure 2 kg extra weight was used for the inclining test, 1 kg at the port side and kg at the starboard side of the model. Following standard practice, the inclining test was performed by transferring the 1 kg weight first to the port side and then to the starboard side, measuring each time the ensuing angle of inclination by using an inclinometer and subsequently repeating this process in reverse order. The average inclination was used to find the uncorrected GM and by deducting the effect of the added weights the corrected GM was determined. This is shown next. wd GM - d A tan(q) KG = KM-GM KG cor = KG A- 2 wh A-2w GMor =" KM-KGcor The results of the inclining test are given in Table 8. Total StabilitvAssessment of Damnaged Passenger/Ro-Ro Vessels [Alodel Experiments Route]

147 Page 15 Table 8: nclining Test Results tem Average inclination Uncorrected GM Uncorrected KG Corrected KG (KG,,) Corrected GM (GM 0,) Model Scale 2.45 degrees 63.8 mm mm mm 65.7 mm As the table shows, the GM value of the model was determined to be 65.7 mm which is slightly lees than the target value of 66.8 mm. 4.2 Determination of Roll Radius of Gyration According to the model test procedure, the roll radius of gyration of the model should be less than 0.4B, where B refers to the beam of the vessel. t is also important to ensure an accurate modelling of the vessel's dynamic characteristics so that an essential departure from 0.4B must be avoided. n order to determine the roll radius of gyration (i), free rolling tests in calm water were carried out to allow for an estimate of the natural roll period. By using the following expression the roll radius of gyration can then be determined:. ýggm.. 27r in 2%t mm or m The following results were derived from the experiments: Natural roll period (T.) 2.04 sec Roll radius of gyration (i) : mm i/b (< 0.4B) Recordings for free rolling tests are given in Appendix E. 4.3 Determination of Pitch Radius of Gyration M According to the model test procedure, the pitch radius of gyration of the model should be less than 0.25L, where L refers to the length of the vessel. n order to determine the pitch radius of gyration the following two methods can be used: Bi-filar Suspension System Bi-filar suspension can be described as a platform suspended from both ends to a high ceiling by using long wires as shown in Figure 4- The model is then put on the platform on its side and the platform is offset by an angle relative to the fore-and-aft axis of the model left to oscillate freely. The natural period of the platform with the model is measured from which total mass moment of inertia is calculated as shown in the following. The same procedure is repeated for the platform only to find the mass moment of inertia of the platform and by deducting this from the total inertia to determine the pitch mass moment inertia for the model. The pitch radius of gyration can then be determined. Total StabilitvAssessment of Damaged PassengeriRo-Ro Vessels [fa'odel Experiments Route]

148 * Page 16 Mass moment of inertia of the plat(form U p = M Pgr 2 TP = M rt. (27Et2h h Mass moment of inertia and radius of gyration of the model in pitch PC H = (M s p) ' P h K = 1 PTH Ms (check if ic, <0.25Lo4) where, Mp mass of the platform (kg) H height (wire length) between the suspension point of the wire on the ceiling and the platform (m) TP natural pitch period of the platform (sec) M, mass of the model (kg) lphid Pitch mass moment of inertia of the model (kg m 2 ) T,..v natural pitch period of the platform + model (sec) kv pitch radius of gyration (m) Table 9: Bi-filar Test Results M kg h 4.3 m Tv sec MK k _147 kg m2 T 1.74 sec kyy 0.78 m (0.21L )m As the results indicate the pitch radius of gyration is 0.21L, which is less than 0.25 L. Such difference in pitch radius of gyration is considered acceptable, on the basis that the model maintains in the main a beam on direction to the waves. Total Stability Assessment of Damaged Passenger/Ro-Ro Vessels [A/odel Experiments Route]

149 Page 17 FDr W ire./7: Si S "/... Test -Model Plan View Side View --_ Top view Figure 4: Bi-Filar Suspension System Single Suspension System n order to determine the pitch radius of gyration the model could alternatively be suspended from the horizontal ballast bar, which is above the vertical centre of gravity (KG). The longitudinal position of the suspension point on the bar is then adjusted, until a horizontal position of the model in air is achieved and the model pushed down and let free to pitch around the rotation point (Figure 5). By measuring the total time for a number of pitching periods, the average natural pitch period of the model in the air can be determined and the pitch mass moment of inertia calculated by using the expressions given below. Pitch mass moment of inertia of the model P ý M gh U2 (2rT)2 Total Stability Assessment of Damaged Passenger/Ro-Ro Vessels [Model Experinents Route]

150 Page 18 Pitch radius of gyration of the model in pitch mode of motion _ = PZH (check if ic, < 0.25L) Sc V M s, where, h the distance between the centre of rotation and the centre of gravity (in) Tp natural pitch period of model (sec) M, mass of the model (kg) pitch pitch mass moment of inertia of the model (kg M 2 ) kyy pitch radius of gyration (in) Centre of Rotation *Centre o gravity Figure 5: Single Suspension System for Measurement of Pitch Radius of Gyration 4.4 Determination of Natural Roll Period in Damaged Condition n order to decide on the peak period of the long waves tests (i.e., JONSWAP spectrum with y =1), the natural roll period of the damaged model must first be determined. According to the model test procedure, the peak period in this case must be the least between the natural roll period for the damaged model and 6Hs" ' 2. For this purpose, free rolling tests were carried out for damaged condition 3 and the results are given in the following. Records for the free roiling test are given in Appendix E. Table 10: Free Rolling Tests for Damage Case 3 Damage Case Model Roll Period Full scale Roll Period (sec) (see) Case 3 Cae (Forward Damage) Considering the intact rolling period (10.0 sec) and the forward damage rolling period (16.06 sec), the authorities did not request measurement of the rolling for the midship damage condition in foreseeing that it would have been greater than 12 seconds, which is the maximum allowable wave period for the tests. Total Stability Assessment of Damaged Passenger'Ro-Ro Vessels[Model EpenmentsRoute] ses[ oe EprmnsRue

151 Page Environmental Conditions n accordance with the model test procedure, the irregular wave environment is modelled by using JONSWAP spectra as specified below. Waves are assumed to be coming from the beam into the damage opening. Waves with Y=3.3 The peak period is calculated as Tp = 4iF (sec), and the zero crossing period as To= T= (sec), Table 11: Wave Characteristics (JONSWAP with,=3.3),.5*... metres seconds seconds where, Tp: Peak period (sec) To: Zero crossing period (see) H,: Significant wave height (m) y: Peakness parameter Waves with Y- The peak period is calculated as the smaller of T = 6.f]T (see) and Tp = Natural roll period in damaged condition Furthermore, the zero crossing period should be less than To i = 1.4 -P (sec) Table 12: Wave Characteristic (JONSWAP with y-1.0) H, TP TO metres seconds secornds Total Stability Assessment of Damaged PassengerlRo-Ro Vessels [Model Experiments Route!

152 4.6 Survivability Tests Page 20 Experimental Setup M The test section of the seakeeping tank is 90 m long, 7.0 m wide and 2.7 m deep with the wave maker having 5 independent flaps capable of generating regular and irregular waves using in-house software. Wave realisations are generated in the presence of authorities and for each record the spectral characteristics are automatically checked to ensure adherence to the pre-specified sea states. Waves are measured by using two wave probes as shown in Figure 6. The first probe was fixed to a position, 17 m away from the wave maker. The second probe is fixed to a point on the carriage and is moving with it following the freely drifting model. The position of the second wave probe on the carriage is 1.0 m in front and 0.5 m to the side of the front end of the model. A yaw control mechanism has been devised by using bars and loose strings attached to the aft and forward ends of the model along the centreline and at the vertical centre of gravity level. The loose strings can be pulled easily to control yawing of the model. Drift was measured by recording the carriage speed. The latter is adjusted according to the drift of the model aiming to keep the position of the model relative to the carriage constant. Motions were measured by using an infrared camera system, a sophisticated non-contact motion measuring system available at the Centre. This system can measure with extreme accuracy the motions of a free model in six-degrees-of-freedom. Roll and pitch motions are also measured by using an electronic inclinometer attached to the centre of gravity of the vessel to readily provide motion records on site, during the test. Experimental Procedure The model is positioned 20 m away from the wave maker. f the model in the damaged condition was inclined to less than degree of static heeling towards the damage opening, a static heeling of degree was obtained by moving the ballast weight off the centreline. When the set-up is ready, random wave realisations were produced in the computer, which were witnessed by the authorities. Following the completion of the realisations all equipment were switched on and wave signals were sent to the wave maker. Measurements of wave realisation, roll and pitch motions of the model can be observed in the available monitors while experiments are being carried out and recorded by the video camera. The total test time was around seconds, which corresponds to approximately 33 minutes in flill scale. After the completion of each test the measured wave and roll statistics are examined to ensure compliance with the requirements of the model test procedure. Test Propgramme Table 13: Summary of the Test Programme DAMAGE CONDTON WAVES: JONSWAP SPECTRUM - Hs = 4.0 m y=3.3 Y=1.0 CASE 3 (Forward Damage) X (5 times) X (5 times) CASE 6 (Midship Damage) X (5 times) X (5 times) Total Stabilitv Assessment of Damaged Passenger/Ro-Ro Vessels [Alodel Experiments Route]

153 Page 21 \ \ Shallow End Carriage Mode m Travelling Wave Probe Fixed Wave Probe "Maker Carriage Wave Figure 6: Experimental Set-up (Top View) Total Stabilitv Assessment of Damaged Passenger/Ro-Ro Pessels fa'odel Experiments Route]

154 J Page 22 Table 14: SHPA MODEL TESTS PROGRAMME [DENNY TANK - DUMBARTON] Time Task DAY 09:00-12:00 NSPECTON OF MODEL Dimensions Damage locations Displacement, intact draught, even keel nclining test in even keel intact condition to determine GMT Free rolling test to determine roll radius of gyration, k., Forward damage case 3 Draught, trim in damaged condition 12:00-13:00 LUNCH 13:00-17:00 EXPERMENTS FOR FORWARD DAMAGE Experiments with long crested irregular waves (JONSWAP SPECTRUM) y-=3.3, Hs=4.0 m (Short Waves) Experiments with long crested irregular waves (JONSWAP SPECTRUM) y-=1.0, Hs=4.0 m (Long Waves) DAY 2 09:00-10:30 NSPECTON OF MODEL Displacement, intact draught, even keel nclining test in even keel intact condition to determine GMT Free rolling test to determine roll radius of gyration, k,, Midship damage case 6 Draught, trim in damaged condition 10:30-12:00 EXPERMENTS FOR DAMAGE AMDSHPS Experiments with long crested irregular waves (JONSWAP SPECTRUM) y--3.3, Hs=4.0 m (Short Waves) Experiments with long crested irregular waves (JONSWAP SPECTRUM) y1l.0, Hs=4.0 m (Long Waves) 12:00-13:00 LUNCH 13:00-17:00 EXPERMENTS FOR DAMAGE AMDSHPS (continued) Total Stabilitv Assessment of Damaged Passenger"Ro-Ro Vessels [Model Experiments Route]

155 Page 23 Table 15: PROTOCOL SHP MODEL "Explanatory Checked/Comments Main dimensions (ntact) Notes LOA (i) LBp (i) BLD (m) DMLD (m) to upper deck DmD (m) to car deck TMLD (m) from BL Damage opening Main dimensions (Damaged) 3%Ls + 3 (m) longitudinal B/5 (m) transverse keel-upper deck Longitudinal positions of damage cases Aft Damage Case 3 (distance measured from AP) Fr... (m) Fr... (m ) Fr... (m) Midship Damage Case 6 (distance measured from AP) Fr... (m) Fr... (m ) Fr... (m) ntact stability characteristics vertical aft bulkhead centre fwd bulkhead aft bulkhead centre fwd bulkhead Worst SOLAS damage (Case 3) TAFF (m) (TFwo (m) Displacement Fresh water at design draft, (tonnes) even keel GMT (m) kxx (m) < 0.4 BMLD Total Stabilitv Assessment of Damaged PassengerRo-Ro Vessels [fmodel Erperiments Route]

156 Page 24 kyy (m) _< 0.25 LOA Damage stability characteristics TM)r (im) TFWD (M)... static heel (0) starboard/port TROLL (s) Selected Sea States (5 nrns in each case) Significant wave height, Hs = 4 m JONSWAP spectrum with y = 3.3 and peak period Tp = 8.0 s Zero crossing period (Tz) Range 6.25 s to 6.67 s Significant wave height, Hs = 4 m JONSWAP spectrum withy = 1.0 and peak period Tp = 12.0 s or equal to TROLL if the latter is smaller Zero crossing period (Tz) Range 8.57 s to 9.23 s short waves long waves Midship damage (Case 6) ntact stability characteristics TMEAN (M) (M) TFWD (M) Displacement Fresh water at design draft, (tonnes) even keel GMT (M) k (M) 0.4 B\hLD ky (M) 0.25 LOA Damage stability characteristics (ME.A..N TAFT (M) TF)D (M) i static heel (0) starboard/port TotalStabiltv Assessment of Damaged Passenger/Ro-Ro Pessels [Model Experiments Rotute oa

157 ' TROLL (s) Page 25 Selected Sea States (5 runs in each case) Significant wave height, Hs = 4 m JONSWAP spectrum with y = 3.3 and peak period Tp = 8.0 s short waves Zero crossing period (Tz) Range 6.25 s to 6.67 s Significant wave height, Hs = 4 m long waves JONSWAP spectrum with y = 1.0 and peak period Tp = 12.0 s or equal to TROLL if the latter is smaller Zero crossing period (Tz) Range 8.57 s to 9.23 s Total Stabili~v.Assessment of Damaged PassengeriRo-Ro Vessels [Alodel Experiments Route]

158 5. MODEL TEST RESULTS Page 26 K Two damaged conditions were tested in two different wave environments. Therefore, deriving from the foregoing, a total of 20 tests were carried out. n all the conditions considered, SHPA survived all 20 tests. The summary of the tests is given in Table 1. Waves and recorded roll motion statistics are given in Appendix D while samples of time histories of waves and roll motion are provided in Appendix E. Table 16: Summary Results of the Experiments Run Wave y Measured-wave Damage Number Realisation Height, Hs Condition Outcome R3C(m) RUN 2 MAC CASE 3 SURVVED RUN 2 MAC CASE 3 SURVVED RUN 3 MAC CASE 3 SURVVED RUN 4 MAC CASE 3 SURVVED RUN 5 MAC CASE 3 SURVVED RUN 6 MAC CASE 6 SURVVED RUN 7 MAC CASE 6 SURVVED RUN 8 MAC CASE 6 SURVVED RUN 9 MAC CASE 6 SURVVED RUN 10 MAC CASE 6 SURVVED RUN MAC CASE 3 SURVVED RUN 12 MAC CASE 3 SURVVED RUN 13 MAC CASE RUN 14 SURVVED MAC CASE 3 SURVVED RUN 15 MAC CASE 3 SURVVED RUN 16 MAC CASE 6 SURVVED RUN 17 MAC CASE 6 SURVVED RUN 18 MAC CASE 6 SURVVED RUN 19 MAC CASE 6 SURVVED RUN 20 MAC CASE 6 SURVVED Total Stability Assessment of Darmaged PassengerlRo-Ro Vessels [Mlodel Experiments Route]

159 6. POTENTAL PROBLEMS AND RECOMMENDED AMMENDMENTS Page 27 Through implementation of the physical model testing "Equivalence" route to compliance with Stockholm Agreement requirements to a number of ships and discussion with both the authorities as well as colleagues in the scientific community, several point have emerged since the introduction of the model test method.. These include: 6.1 Paragraph 2 - Ship Model Model Building Recent experiments at Strathclyde with model used in earlier research have revealed that considerable flexural response (sagging/hogging) of the model hull could take place. Related to this, it must be specified that the model ought to be constructed so as to ensure that such response is eglgil. Extra caution must also be exercised to ascertain that water leakage into the intact part of the hull is prevented. This is rather difficult to cater for completely and experimenters must be warned of this and advised accordingly, e.g. filling the double bottom with foam. Model Scale Even though a minimum scale of 1:40 is defined, the specification of a minimum model length for obvious reasons makes it necessary to recommend a maximum scale as a compromise concerning the physical impossibility in wave tanks to create high waves at very low frequencies. Alternatively, this could act as incentive to authorities who are not willing to compromise the model test procedure by proposing a minimum length of Ro-Ro ships below which this testing is not mandatory. Roll and Pitch nertia n both cases a maximum radius of gyration is specified related to the beam and length of the model. Notwithstanding the lack of justification for proposing these limits for Ro-Ro vessels which in itself is in need of attention, there is clearly a need to provide the following additional specifications: * A lower limit of' these radii of gyration- Problems in attaining the recommended values show a dangerous drift of the values to unacceptably low levels if claims to attempting in reproducing the dynamic behaviour of the real ship were to remain valid. a Measurement of the pitch radius of gyration is normally an afterthought. This must be rectified, particularly in cases where the model appears to be pitching substantially.. The maximum of the pitch radius of gyration is specified as 0.25L. A more specific definition of L is needed to avoid ambiguity. t is recommended that LOA be used. 6.2 Paragraph 3- Procedure for Experiments Wave Spectra Experience so far has clearly shown the following to be generally true: * Models in damaged condition have roll periods considerably higher than peak periods of naturally occurring waves. Total StahilitvAssessment of Damaged PassengervRo-Ro Vessels /X'fodel Experiments Route]

160 Page 28 Models in damaged condition whilst being subjected to progressive flooding are hardly rolling and will certainly not resonate with the impinging waves. Considering the above, it is recommended to abandon testing in the longer waves spectrum. Furthermore, according to the testing procedure, the generated wave realisation needs to be measured in two positions, one towards the wavemaker end and the other drifting with the model down the tank. However, the position where to check whether the characteristics of the generated waves are acceptable is again dubious. n this respect, whilst it makes sense to check the significant wave height in the vicinity of the model, the tolerance for the zero crossing period of ±5% appears to be impractical. t is, therefore, recommended that the latter be checked at a fixed point near the wavemakers. Position of the Wave Probe near the Model The guidance currently in place must be given some qualification concerning an optimum position to avoid wave reflection. n the diagram shown the reflection from the tank walls of tanks less than 10m wide of waves generated by the pitching and heaving model may be considerable to suggest a compromised (improved) position outside the 900 sector shown. 10 mposed Heeling This requirement was initially intended to help experimenters to waste time in conducting experiments that produced no useful information, namely when the model heels away from damage. However, the imposition of a mandatory heel makes nonsense of the physical system and the intention to prove equivalence. More specifically, how can one try to prove that a SOLAS '90 ship can survive 4m waves if by imposing a 10 heel the ship fails to meet SOLAS '90 standards? Damage Test Cases _. fk The specified test cases must be subjected to continuous evaluation and assessment as knowledge in * the subject advances. t is found, for example, time and again that the vessel resistance to capsize is not a minimum at the specified cases. Criteria such as areas under the GZ curve and freeboard need to be critically re-examined. Duration of Test Run * The time of a test run, corresponding to 30 minutes full scale derives from stationary processes. As the motion of vessel subjected to flooding represents a non-stationary process, this time is need of re-consideration and further discussion. Minimum Number of Runs Without resorting to a theoretical treatise on this issue, it can be shown that 5 runs, as is currently specified, offer assurance on the probability of vessel survival of 50% with a confidence that such an outcome is right of 95%. Moreover, to achieve a probability of survival of 95% with a * confidence of 95%, 59 runs will be required. This simple fact, highlights, how important it is to appreciate what this method of testing is trying to achieve and information it actually provides. Totl Sability Asesetof Damaged Passenger'Ro-Ro Vessels [Afodel Experiments Routej To sssmn

161 i 7. CONCLUSONS Page 29 The following conclusions address only the survivability tests of SHPA. Survivability tests were carried out in sea states characterised by a significant wave height of 4.0 m (Hs = 4.0 m) according to the model test procedure specified by MO and the relevant guidance notes provided by the authorities. Based on the results of these tests, the following conclusions may be drawn: " For the worst SOLAS damage condition, which is case 3, and the worst midship damage condition, which is case 6, SHPA survived all the required tests. * The midship damage condition is more onerous than the forward damage. _\.. Random waves with the smaller peak period (y-3.3) appear to be considerably more dangerous than those with the larger peak period (y,=). Therefore, testing the model in only the shorter waves should have been adequate from the survivability point of view. Total Stability Assessmen, of Damaged Passenger/R o-ro Vessels [A.[odel Experiments Route]

162

163 q

164

165 r"ternatonal MARTME _ ORGANZATON GSLF 40NF. 14 SUB-COMMTTEE ON STABLTY AND LOAD LNES AND FSHNG VESSELS SAFETY 40th session Agenda item 19 MO ow 5 July 1996 ENGLSH ONLY ANY OTHER BUSNESS Agreement concerning specific stability requirements for ro-ro passenger ships undertaking regular scheduled international voyages between or from designated ports in North West Europe and the Baltic Sea Guidance notes on the annexes of the Agreement Submitted by Denmark, Finland, Norway, Sweden and the United Kingdom Reference is made to MO circular letter 1891 by which Member States have been informed of the above-mentioned Agreement made in accordance with SOLAS 1995 Conference resolution 14. t is considered that annexes and 2* of this paper are the appropriate guidance notes for the uniform application of annexes 1, 2 and 3 of the Agreement. The Sub-Committee is invited to note the contents. Due to the limited number of copies available, annexes and 2 are distributed on the basis of one copy per delegation. :\SLF\40UNF14 For reasons of economy. is docurment is oritted in 0 irtited numtber. Delgatcs aes kindly asked to bring, thei' conses tn me-tinps aned n l,,,iiona :5.,unsl cr.es. /

166

167 Circular letter No ANNEX2 STABLTY REQUREMENTS PERTANNG TO THE AGREEMENT Preamble Application n accordance with this Agreement, passenger ships with ro-ro cargo spaces or special category spaces as defined in regulation 11-2/3 of the nternational Convention for the Safety of Life at Sea, as amended, shall comply with the provisions of this Agreement not later than at the first yearly inspection following the date of compliance prescribed below, according to the value of A/Amax as defined in the annex to the Calculation Procedure to Assess the Survivability Characteristics of Existing Ro-Ro Passenger Ships When Using a Simplified Method Based Upon resolution A.265(V), developed by the Maritime Safety Committee at its fifty-ninth session in June 1991 (MSC/Circ.574): Value of A/Amax Date of Compliance Stability Standard Less than 85% 1 April 1997 Less than 90% 31 December 1998 Less than 95% 31 December 1999 Less than 97.5% 31 December % or higher 31 December 2001 but in any case not later than 1 October 2002 n addition to the requirements of SOLAS regulation 1-1/8, ro-ro passenger ships shall comply, subject to the provisions of paragraph 2, if applicable, with the following:.1 the provisions of paragraphs 2.3 regulation 8 shall be complied with when taking into account the effect of a hypothetical amount of sea water which is assumed to have accumulated, on the first deck above the designed waterline of the ro-ro cargo space or special category space as defined in regulation fl-2/3 assumed to be damaged (referred to as "the damaged ro-ro deck" hereinafter). The other requirements of regulation 8 need not be complied with in the application of the stability standard' contained in this Agreement. The amount of assumed accumulated sea water shall be calculated on the basis of a water surface having a fixed height above: (a) (b) the lowest point of the deck edge of the damaged compartment of the ro-ro deck, or when the deck edge in way of the damaged compartment is submerged then the calculation is based on a fixed height above the still water surface at all heel and trim angles; :\CL\ 891. Guidance notes on the standard to be developed.

168 ' Circuiar lerter No,. i 9 i ANNEX 2 Page 2 as follows: 0.5 m if the residual freeboard (Q is 0.3 in or less; 0.0 m if the residual freeboard (f) is 2,0 m or more; and intermediate values to be determined by linear interpolation, if the residual freeboard (Q) is 0.3 m or more but less than 2.0 m; where the residual freeboard (f) is the minimum distance between the damaged ro-ro deck and the final waterline at the location of the damage in the damage case being considered without taking into account the effect of the volume of assumed accumulated waler on the damaged ro-ro deck;.2 when a high-efficiency drainage system is installed, the Administration may allow a reduction in the height of the water surface in accordance with the guidelines to be developed by the Organization'; m.3 for ships in geographically defined restricted areas of operation, the Administration may reduce the height of the water surface determined in accordance with subparagraph 1 substituting such height of the water surface by the following: mn if the significant wave height (hj defining the area concerned is 1.5 in or less;.3.2 the value determined in accordance with subparagraph.1 if the significant wave height (hi defining the area concerned is 4.0 mn or above;.3.3 intermediate values to be determined by linear interpolation if the significant wave height (h) defining the area concerned is 1.5 m or more but less than 4.0 in; provided that the following conditions are fulfilled:.3.4 the Administration is satisfied that the defined area is represented by the significant wave height (h) which is not exceeded with a probability of more than 10%; and.3.5 the area of operation and, if applicable, the part of the year for which a certain value of the significant wave height (h) has been established are entered into the certificates; and m :\C_L\1891. Refer to the "guidelines...

169 kcircujar ileter zž0.1 i 69 ANNEX 2 Page 3.4 as an alternative to the requirements of subparagraph.1 or subparagraph.3, the Administration may exempt application of the requirements of subparagraph.1 or subparagraph.3 and accept proof, established by model tests carried out for an individual ship in accordance with the model test method developed by the Organization', annexed to this document justifying that the ship will not capsize with the assumed extent of damage as provided in paragraph 4 of regulation 8 in the worst location being considered under paragraph 1. 1 in an irregular seaway, and -'.5 reference to acceptance of the results of the model test as an equivalence to compliance with subparagraph.1 or subparagraph.3; the value of the significant wave height (h,) used in the model tests shall be entered into the ship's certificates..6 the information supplied to the master in accordance with paragraphs 7.1 and 7.2 of regulation 8, as developed for compliance with paragraphs 2.3 to 2.3.4, shall apply unchanged for ro-ro passenger ships approved according to these requirements. 2 For assessing the effect of the volume of the assumed accumulated sea water on the damaged ro-ro deck in paragraph 1, the following provisions shall prevail:.1 a transverse or longitudinal bulkhead shall be considered intact if all parts of it lie inboard of vertical surfaces on both sides of the ship, which are situated at a distance from the shell plating equal to one-fifth of the breadth of the ship, as defined in regulation 2, and measured at right angles to the centreline at the level of the deepest subdivision load line;.2 in cases where the ship's hull is structurally partly widened for compliance with the provisions of this regulation, the resulting increase of the value of one-fifth of the breadth of it is to be used throughout, but shall not govern the location of existing bulkhead penetrations, piping systems, etc., which were acceptable prior to the widening;.3 the tightness of transverse or longitudinal bulkheads which are taken into account as effective to confine the assumed accumulated sea water in the compartment concerned in the damaged ro-ro deck shall be commensurate with the drainage system, and shall withstand hydrostatic pressure in accordance with the results of the damage calculation. Such bulkheads shall be at least 4 in in height unless the height of water is less than 0.5 mn. n such cases the height of the bulkhead may be calculated in accordance with the following: where Bh, l = bulkhead height, and = height of water n any event, the minimum height of the bulkhead shall be not less than 2.2 mn. However, in the case of a ship with hanging car decks, the minimum height of the bulkhead shall be not less than the height to the underside of the hanging car deck when in its lowered position..\c_l\ Refer to the "model test method attached to this document"

170 -Circuiar letter No.1891 ANNEX 2 Page 4.4 For special arrangements such as, e.g., full width hanging decks and wide side casings, other bulkhead heights may be accepted based on detailed model tests;.5 The effect of the volume of the assumed accumulated sea water need not be taken into account for any compartment of the damaged ro-ro deck, provided that such a compartment has on each side of the deck freeing ports evenly distributed along the sides of the compartment complying with the following:.5.1 A Ž is the length of the compartment in m; where A is the total area of freeing ports on each side of the deck in m 2 ; and.5.2 the ship shall maintain a residual freeboard of at least 1.0 m in the worst damage condition without taking into account the effect of the assumed volume of water on the damaged ro-ro deck; and.5.3 such freeing ports shall be located within the height of 0.6 m above the damaged ro-ro deck, and the lower edge of the ports shall be within 2 cm above the - damaged ro-ro deck; and :1\CL\ such freeing ports shall be fitted with closing devices or flaps to prevent water entering the ro-ro deck whilst allowing water which may accumulate on the ro-ro deck, whilst allowing water which may accumulate on the ro-ro deck to drain; and.6 when a bulkhead above the ro-ro deck is assumed damaged, both compartments bordering the bulkhead shall be assumed flooded to the same height of water surface as calculated in paragraphs 1.1 and 1.3 above.

171 Circular letter No.] 1891 ANNEX 2. Page 5 Appendix MODEL TEST METHOD Objectives n the tests provided for in paragraph 1.4 of the stability requirements pertaining to the agreement, the ship should prove capability to withstand a seaway defined in paragraph 3 hereunder in the worst damage case scenario. 2 Ship model 2.1 The mode! should copy the actual ship for both outer configuration and internal arrangement - in particular of all damaged spaces, having an effect on the process of flooding and shipping of water. The damage should represent the worst damage case defined for compliance with paragraph of SOLAS regulation 11-1/8 (SOLAS 90). An additional test is required at a level keel midship damage, if the worst damage location according to SOLAS 90 is outside the range %0 Lpp from the midship. This additional test is only required when the ro-ro spaces are assumed to be damaged. 2.2 The model should comply with the following:.1 length between perpendiculars (Lpp) is to be at least 3 m;.2 hull is to be thin enough in areas where this feature has influence on the results;.3 characteristics of motion should be modelled properly to the actual ship, paying particular attention to scaling of radii of gyration in roll and pitch motions. Draught, trim, heel and centre of gravity should represent the worst damage case;.4 main design features such as watertight bulkheads, air escapes, etc., above and below the bulkhead deck that can result in asymmetric flooding should be modelled properly as far as practicable, to represent the real situation;.5 the shape of the damage opening shall be as follows:.5.1 rectangular side profile with a width according to SOLAS regulation 11-1/8.4.1 and unlimited vertical extent;.5.2 isosceles triangular profile in the horizontal plane with a height equal to B/5 according to SOLAS regulation 11-1/ Procedure for experiments 3.1 The model should be subjected to a long-crested irregular seaway defined by the JONSWAP spectrum with a significant wave height Hs, defined in paragraph 1.3 of the stability requirements and having peak enhancement factor y and peak period Tp as follows: :\CL\1891,

172 Circular letter No ANNEX 2 Page 6 1 Tp =4V with y=3.3; and.2 Tp equal to the roll resonant period for the damaage ship without water on deck at the specified loading condition but not higher than 6v/H and with y = The model should be free to drift and placed in beam seas (900 heading) with the damage hole facing the oncoming waves. The model should not be restained in a maimer to resist capsize. f the ship is upright in flooded condition, 1* of heel towards the damage should be given. 3.3 At least 5 (five) experiments for each peak period should be carried out. The test period for each rnshall be of a duration such that a stationary state has been reached but should be run for not less than 30min in full-scale time. A different wave realization train should be used for each test. * 3.4 f none of the experiments result in final inclination towards the damage, * be the repeated experiments with 5 should runs at each of the two specified wave conditions or, alternatively, oe given an the additional model should V~ angle of heel towards the damage and the experiment * each repeated of the with two 2 specified runs at wave conditions. The purpose of these additional * demonstrate, experiments in the is to best possible way, survival capability against capsize in both directions. 3.5 The tests are to be carried out for the following damage cases: - the worst damage case with regard to the area under the GZ curve according to SOLAS; * and.27 the worst midship damage case with regard to residual freeboard in the midship area if required by Survival criteria 4.1 The ship should be considered as surviving if a stationary state is reached for the successive test runs as required mn 3.3 but subject to Angles of roll of more than 3Q0 against the vertical axis, occurring more frequently than in 20% of the rolling cycles or steady heel greater than 200 should be taken as capsizing events even if a stationary state is reached. 5 Test approval 5.1 t is the responsibility of the Administration to approve the model test programme in advance. t should also be borne in mind that lesser damages may provide a worst case scenario. 5.2 Test should be documented by means of a report and a video or other visual record containing all relevant information of the ship and test results. A copy of the video and report should be submitted to the Organization, together with the Administration's acceptance of the test. \CL\1 891.

173 SLF 40/Qnf.** ANNEX 1 RO-RO PASSENGER SHP SAFETY GUDANCE NOTES ON ANNEXES 1 AND 2 OF THE AGREEMENT CONCERNNG SPECFC STABLTY REQUREMENTS FOR RO-RO PASSENGER SHPS UNDERTAKNG REGULAR SCHEDULED NTERNATONAL VOYAGES BETWEEN OR TO OR FROM DESGNATED PORTS N NORTH WEST EUROPE AND THE BALTC SEA. GENERAL 1NThe most dangerous problem for a ro-ro ship with an enclosed ro-ro deck is undoubtedly that posed by the effect of a build-up of significant amount of water on that deck. The principle of additional water-on-deck has been adopted to account for the risk of accumulation of water-on-deck as a result of the dynamic behaviourin a seaway, of the vessel after sustaining side collision damage. 2 t is considered that the problem of water accumulating on deck when entering through bow, stern and side doors has been addressed by the increased standards now required with respect to strength, closing and locking systems, as well as by the new requirements relating to the position of the extension to the collision bulkhead. 3 The damage stability requirements applicable to ro-ro passenger ships in 1990 (SOLAS '90) implicitly include the effect of water entering the ro-ro deck in a sea state in the order of 1.5 m significant wave height. n order to enable the ship to survive in more severe sea states those requirements have been upgraded to take into account the effect of water which could accumulate on the 3 ro-ro deck. 4 n developing the new requirements the following basic elements were taken into account: 1. MSC/Circ. 153 confirms that 99% of all recorded collisions occur in sea states up to 4m significant wave height (h,). This was therefore taken as the most severe sea * state to be considered;.2 compliance with SOLAS '90 standard is assumed to be equivalent to survival of the damaged ship in sea states of up to 1.5m significant wave height (h,) which according to the distribution collisions; function in MSC/Circ.153, covers 89% of all.3 sea states between 1.5m to 4.Om significant wave height (h,) would be covered by the additional damage stability requirements to take into account the effect of "water-ondeck"; and 1

174 in which according to the statistics available be expected collisions to occur, can a reduction has been permitted in the requirement for "water-on-deck" geographically for defined ships restricted operating areas. in height (h,) is the The qualifying significant parameter, wave in association with a 90% probability that h, is not exceeded in that area or route. 5. When considering the amount of water to be assumed as accumulating on the ro-ro deck the figure of up to 0.5m, depending on the significant wave height and residual freeboard, was agreed based on consideration of the following information:.1 an initial Nordic proposal which suggested 0.Sm for the amount of "water-on-deck";.2 a study by the Society of Naval Architects and Marine Engineers (SNAME) suggested that 0.5m 3 /m 2 was a reasonable level for 4.0m significant wave height on a vessel with low damaged freeboard; Ferry with Water on the Car Deck (M-304)) which indicated the volume at the significant wave height of 4.Om was approximately equal to 0.75m 3 /m 2 ;.3 model tests carried out in Finland (Model Tests of a Car -.4 investigations carried out in the United Kingdom, which indicated that the- corresponding amount of water would be about 10% of the ship's displacement; and.5 research carried out during the Joint North West European Project (Safety of Passenger Ro-Ro vessels) which related to a static pressure head relevant to a head of water above'the deck or above the still water level. 6. However it was considered more appropriate to assume a variable quantity of water on deck depending not only on the residual freeboard and significant wave height, but also on a variable angle of heel. With this in mind the basic assumption of up to 0.5 metres height of accumulated water corresponding to residual freeboard and significant wave height was retained. 7. Research has clearly shown that the residual freeboard had a significant effect on the amount of water assumed to be accumulated on deck. The maximum residual freeboard (f,) to be taken into account was agreed as 2.Om based on both the nstitute for Marine Dynamics (Canada) (MD) model tests and the SNAME Analytical predictions which indicated that the height of water on deck goes to zero as the residual freeboard/significant wave height rises ratio above 0.5. Therefore in order to assume zero accumulation, in a significant wave height of 4.0m, a residual freeboard of 2.Cm would be required. The residual freeboard (f') in this case is defined as "the minimum distance between the damaged ro-ro deck and the waterline at the location of the damage without taking into account the additional effect of sea water accumulated on the damaged ro-ro deck". 2

175 S. A new requirement on damage stability for ro-ro passenger ships taking into account additional flooding above the ro-ro deck has been developed which should be clearly understood to apply to existing as well as to new ships. Existing Ships built complying with MO RESOLUTON A265(V) 9. MC has accepted the principle that the probabilistic residual standard within A265 is equivalent to the SOLAS '90 deterministic standard of residual stability. Such existing ships therefore are not required to be upgraded to SOLAS '90. However such shipsm.. apprvedforcompliance with A265 before the date of entry into force of these requirements must, in addition, be capable of complying with the new damage stability requirement for all cases required for compliance with regulation 5(b) of A265(V), the worst of which may be a one or two compartment case. Scope of application of the new requirement 10. The new damage stability requirement should, in principle, be ~) applied to all such passenger ships with ro-ro decks covered by the - definition "special category spaces and ro-ro cargo spaces as - defined in regulation 11-2/3"1 with the proviso that spaces which have sufficient permanent openings for water freeing purposes may be exempted from the application of the requirements of "water-ondeck". Details of the requirements for freeing ports are given in the attached notes. Bulkhead height including a standard for testing 11. The general requirement for the minimum height of bulkheads which may need to be additionally installed on the ro-ro deck shall apply to all ro-ro passenger ships. However the new requirements provide for the possibility for an Administration to accept lower heights for innovative designs of bulkheads, based on the results of model experiments. *12. Any transverse and longitudinal bulkheads which are fitted to enabl the ship to meet these stability regulations must be in 3 place and secured at all times when the ship is at sea. Accesses Wi*thin such bulkheads may be opened during the voyage but only for sufficient time to permit through passage for the, essential working of the ship and only at the express authority of the master. Modifications which may be consequential to compliance with the new standard 13. Passenger accesses; escapes; fire extinguishing, detection and monitoring systems; car deck drainage; ventilation; cargo securing etc must comply with the same safety standards as are applicable to the vessel after the fitting of any ro-ro car deck modifications. Provision must also be provided such that any accesses in transverse or longitudinal bulkheads /barriers cannot be obstructed. 3

176 -GOZAjNCZ 14QTZa ON X-NNiZ l TO THE AGREZMZNT- The route, routes or areas concerned have been determined by the Administrations at each end of the route or all Administrations within a defined area. The defined route or area is one in which the determined significant wave height would not be exceeded with a probability of more than 10% over a one year period for all year round operation... (insert). GUDANCE NOTES ON THE SURVVABLTY REOVREMENTS CONTANED N ANNEX TO THE AGREEMENT PREAMBLE greement - Application t should be noted that vessels which may operate solely in areas where the significant wave is less than 1.5m and which do not have to comply with the additional water-on-deck requirements (ie comply only with SOLAS '90) are to comply with the dates of compliance set out with the Agreement. Agreement Para 1 As a first step all ro-ro passenger ships must comply with the "SOLAS '90" standard of residual stability as it applies to all passenger ships constructed on or after 29 April t is the application of this requirement that defines the residual freeboard f, necessary for the calculations required in paragraph 1.1. Agreement Para 1.1 i. This paragraph addresses the application of a hypothetical amount of water accumulated on the bulkhead (ro-ro) deck. The water is assumed to have entered the deck via a damage opening. This paragraph requires that the vessel in addition to complying with the full requirements of SOLAS '90 further complies only with that part of the SOLAS '90 criteria contained in paragraphs 2.3 to of regulation Chapter 8 of 11-1 Part B of SOLAS with the defined amount of water on deck. Fcr this calculation no other requirements of Chapter -1 regulation 8 need be taken into account. example For the vessel does not, for this calculation, need to comply with the requirements for the angles of equilibrium or -9 non-submergence of the margin line. 2. The accumulated water is added as a liquid load with one common surface inside all compartments which are assumed flooded on the car deck. The height (h,) of water on deck is dependent on the residual freeboard (f,) after damage, and is measured in way of the damage (see fig 1). The residual freeboard f, is the minimum distance between the damaged deck ro-ro and the final waterline (after equalisation measures if any have been taken) in way of the assumed damage after examining all possible damage scenarios in determining the compliance with SOLAS '90 as required in para 1 of Annex 2 to the Agreement. No account should be taken of the effect of the hypothetical volume of water assumed to have accumulated on the damaged ro-ro deck when calculating f,. 4

177 2. if f, is 2. Om or more, no water is assumed to accumulate on the ro-ro deck. f fr is 0.3m or less, then height h, is assumed to be 0.5 metres. ntermediate heights of water are obtained by linear interpolation (see fig 2) Agreement Para 1.2 Means for drainage of water can only be considered as effective if these means are of a capacity to prevent large amounts of water from accumulating on the deck ie many thousands of tonnes per hour which is far capacities beyond the fitted at the time of the adoption, of these regulations. Such high efficiency drainage systems may be developed and approved in the future (based on guidelines to be developed by the nternational Maritime Organisation) Agreement Pana The amount of assumed accumulated water-on-deck may, in addition to any reduction in accordance with paragraph 1. 1, be reduced for operations in geographically defined restricted areas. These areas are designated in accordance with significant the wave height (h,) defining the area and are detailed in Annex 1. to the Agreement f the significant wave height (h,), in the area concerned, is 1.5m or less then no additional water is assumed to accumulate on the damaged ro-ro deck. f the significant wave height in the area concerned is 4.Om or more then the height of the assumed accumulated water shall be the value calculated in accordance with paragraph 1.1. ntermediate values to be determined by linear interpolation (see fig 3). 2. The height h., is kept constant therefore the amount of added water is variable as it is dependent upon the heeling angle and whether at any particular heeling angle the deck edge is immersed or not. (see fig 4). it should be noted that the assumed permeability of the car deck- spaces is to be taken as 90% (MSC/Circ. 649 refers), whereas other assumed flooded spaces permeabilities are to be those prescribed in SOLAS. f the calculations to show compliance with the Agreement relate to a significant wave height less than 4.Om that restricting significant wave height must be recorded on the H vessel's passenger ship safety certificate.. Agreement Para 1.4 / 1.5 As an alternative to complying with the new stability requirements of paragraphs 1.1 or 1.2 an Administration may accet poofof compliance via model tests. The model test requirements are detailed in Annex 2 to the Agreement. Guidance notes on the model tests are contained in Appendix 2 to this document. Agreement Para 1.6. Conventionally derived SOLAS'SO limiting operational curve(s) (KG or GM) may not remain applicable in cases where "water on deck" is assumed under the terms of the Agreement and may be necessary to determine revised limiting curve(s) which take into account the effects of this added water. To this effect sufficient calculations corresponding to an adequate number of operational draughts and trims must be carried out.

178 Reisoed limiting operational KG/GM Curves may be derived by iteration, whereby the minimum excess GM resulting from damage stability calculations with water on deck is added to the input KG (or deducted from the GM) used to determine the damaged freeboards (f,), upon which the quantities of water on deck are based, this process being repeated until the excess GM becomes negligible. t is anticipated that operators would begin such an iteration with the m~aximumn KG/minimum GM which could reasonably be sustained in service and would seek to manipulate the resulting deck bulkhead arrangement to minimise the excess GM derived from damage stability calculations with water on deck. Agreement Para 2.1 As for conventional SOLAS damage requirements bulkheads inboard of the B/5 line are considered intact in the event of side collision damage. Agreement Pana 2.2 f side structural sponsons are f itted to enable compliance with this regulation, and as a consequence there is an increase in the breadth (B) of the ship and hence the vessel's - B/5 distance from the ship's side, such modification shall not cause the relocation of any existing structural parts or any existing penetrations of the main transverse watertight bulkheads below the bulkhead deck. (see fig 5) ~. Agreement Part Transverse or longitudinal bulkheads /barriers which are fitted and taken into account to confine the movement of assumed accumulated water on the damaged ro-ro deck need not be strictly "watertight". Small amounts of leakage may be permitted subject to the drainage provisions being capable of preventing an accunulation of water on the "other side" of the bulkhead/barrier. n such cases where scuppers become inoperative as a result of a loss of positive difference of water levels other means of passive drainage must be provided. 2. The height (B.) of transverse and longitudinal bulkheads/ barriers shall be not less than (8 x b,.) metres, where h. isq the height of the accumulated water as calculated by application of the residual freeboard and significant wave height (paras 1.1 and 1.3 refers). However fhnho case is the height of the bulkhead/barrier to be less than the greatest of: (a) 2.2 metres; or (b) the height between the bulkhead deck and the lower point of the underside structure of the intermediate or hanging car decks, when these are in their lowered position. t should be noted that any gaps between the top edge of the bulhead deck and the underside of the plating must be "plated-in" in the transverse or longitudinal direction as appropriate. (see fig 6). Bulkheads /barriers with a height less than that specified above, may be accepted if model tests are carried in accordance with Annex 3 to confirm that the alternative design ensures appropriate standard of survivability. * 6

179 Care needs to be taken when fixing the height of the bui~khead/ barrier such that the height shall also be sufficient to prevent progressive flooding within the required stability range. This range is not to be prejudiced my model tests. Note :The range may be reduced t-3 10 degrees provided the corresponding area under the curve is increased (MSC 64/22 refers) Agreement-Para The area "All relates to permanent openings; t should be noted that the "freeing ports" option is not suitable for ships which require the buoyancy of the whole or part of the superstructure in order to meet the criteria. The requirement is that the freeing ports shall be fitted with closing flaps to prevent water entering, but allowing water to drain. These flaps must not rely on active means. They must be selfoperating and it must be shown that they do not restrict outflow to a significant degree. Any significant efficiency reduction must be compensated by the fitting of additional openings so that the required area is maintained. Agreement Part For the freeing ports to be considered effective the minimum distance from the lower edge of the freeing port to the damaged waterline shall be at least l.0m. The calculation of the minimum distance shall not take into account the effect of any additional water on deck. (see fig 7) Agreement Para Freeing ports must be sited as low as possible in the side bulwark or shell plating. The lower edge of the freeing port opening must be no higher than 2cm above the bulkhead deck and the upper edge of the opening no higher than 0.G6m. (see fig B) fl Spaces to which paragraph 2.5 applies, ie those spaces fitted * with freeing ports or similar openings, shall not be included as intact spaces in the derivation of the intact and damage stability jcurves. Agreement Part The statutory extent of damage is to be applied along the length of the ship. Depending on the subdivision standard the damage may not affect any bulkhead or may only affect a bulkhead below the bulkhead deck or only a bulkhead above the bulkhead deck or various combinations. 2. All transverse and longitudinal bulkhead s/barriers which constrain the assumed accumulated amount of water must be in place and secured at all times when the ship is at sea. 3. n those cases where the transverse bulkhead/ barrier is damaged the accumulated water-on-deck shall have a common surface level on both sides of the damaged bulkhead/ barrier at the height h.. (see fig 9).. -

180 O* f aoelnt f rf iur 8

181 1. f f,_ 2.0 metres, height of water on deck 0) = 0.0 metres 2. f f, 40.3 metres, height of water on deck (h) = 0.5 metres Figure 2 < 1. f h, 4.0 metres, height of water on deck is calculated as per fig 3 2. f h, i 1.5 metres, height of water on deck (hj = 0.0 metres For example f f, = 1.15 metres and h, =2.75 metres, height h, = metres 9 Figure 3

182 ~bri =si rr a dkat ~ ~ ~ be anileorvibiu o mmd h hwl xagle (d~k =Sitlpoint of einmeon) hmel =sice (deck edge imannred) Figure 4 10

183 -~ r xisthg -ip Figure 5d rgn D A 3r xsn P85rwJ -r ore H / w Fiur

184 Ship without hinging car decks Example Height of water on deck = 0.25 metres Minimum required height of barnier = 2.2 meutre Gaps to be plated-in Top of bulkhead - x a 2.2 metres. Ship with hanging car deck (in way of the barrier). Example 2 Height of water on deck Ch.) = 0.25 me es Mimimurn required height of barrier = x Figure 6 12

185 fnjstage damage WL minimum req'd freeboard to freeing port =1.0 mn Figure 7, 0 ( ~ ~Length of comportment (t Figure 8

186 i. Deck edge not immersed Deck edge immersed Figure 9

187 SLY 40/nf.14. ANNEX 2 Rb-Rb PASSENGER SHP SAFETY GUDANCE NOTES ON ANNEX 3 OF THE AGREEMENT CONCERNNG SPECFC ~STABLTY REQUREMENTS FOR Rb-NO PASSENGER SHPS UNhDERTAKNG REGULAR SCHEDULED NTERNATONAL VOYAGES BETWEEN OR TO OR FROM DESGNATED PORTS N NORTH WEST EUROPE AND THE BALTC SEA. The purpose of these notes to ensure uniformity in the methods employed in the construction and verification of the model as well * as in the undertaking and analyses of the model tests, while * Th appreciating that available facilities and costs will affect in some way this uniformity. * The content of paragraph 1 of Annex 3 to the Regional Agreement is self explanatory. Paragraph 2 - Ship Model 2. 1 The material of which the model is made is not important in itself, provided that the model both in the intact and damaged condition is sufficiently rigid to ensure that its hydrostatic properties are the same as those of the actual ship and also that the flexural response of the hull in waves is negligible. t is also important to ensure that the damaged compartments are modelled as accurately as practicably possible to ensure that the correct volume of flood water is represented. -Since ingress of water (even small amounts) into the intact parts of the model will affect its behaviour, measures must be taken that this ingress does not occur Model particulars.1 n recognising that scale effects play an important role in the behaviour of the model during tests it is important to ensure that these effects are minimised as much as practically possible. The model should be as large as possible since details of damaged compartments are easier constructed in larger models and the scale effects are reduced. t is therefore recommnended that the model length is not less than that corresponding to 1:40 scale. However it is required that the model is not less than 3 metres long at the subdivision water line. 1

188 .2 (a) The model in way of the assumed damages must be as thin as practically possible to ensure that the amount of flood water and its centre of gravity is adequately represented. t is recognised that it may not be possible for the model hull and the elements of primary and secondary subdivision in way of the damage to be * constructed with sufficient detail and due to these constructional limitations it may not be possible to calculate accurately the assumed permeability of the space. (b) t has been found during tests that the vertical extent of the model can affect the results when tested dynamically. t is therefore required that the ship is modelled to at least three superststructure standard heights above the bulkhead (freefoard) deck so that the large waves of the wave train do not break over the model. (c) t is important that not only the draughts in the. intact condition are verified but also that the. draughts of the damaged model are accurately measured for correlation with those derived from the, damaged stability calculation. After measuring the damaged draughts it may be found necessary to make adjustments to the permeability of the damaged compartment by either introducing intact volumes or by adding weights. However it is also important to ensure that the centre of gravity of the flood water is accurately represented. n this case any adjustments made must err on the side of safety. (d) f the model is required to be fitted with barriers on deck and the barriers are less than the height required as per paragraph 2.3 of Annex 2 of this Agreement the model is to be fitted with CCT`V so that any "splashing over" and any accumulation of water on the undamaged area of) the deck can be monitored. n this case a video recording of the event is to form part of the tests records..3 n order to ensure that the model motion characteristics represent those of the actual ship it is important that the model is both inclined and rolled in the intact condition so that the intact GM and the mass distribution are verified. The transverse radius of gyration of the actual ship is not to be taken as being greater than O.4B and the longitudinal radius of gyration is not to be taken as being more than O.25L. 2

189 H The transverse rolling period of the model is to be obtained by: 7x.4 weegm: metacentric height of the actual (intact) ship * g :acceleration due to gravity X : scale of model B: breadth of actual ship Note While inclining and rolling the model in the damage condition may be accepted as a check for the purpose * of verifying the residual stability curve such tests are not to be accepted in lieu of the intact tests. Nevertheless the damaged model must be rolled in order to obtain the rolling period required to perform the tests as per paragraph The contents of this paragraph are self explanatory. t is assumed that the ventilators of the damage compartment of the actual ship are adequate for unhindered flooding and movement of the flood water. However in trying to scale down the ventilating arrangements of the actual ship undesirable scale effects may be introduced. n order to ensure that these do not occur it is recommended scale than that of the model, ensuring that this does not affect the flow of water on the car deck..5.2 The isosceles triangular profile of the prismatic damage shape is that corresponding to waterline. the load Additionally in cases where side casings of width less than B/5 are fitted and in order to avoid any possible scale effects, the damage length in way of the side casings must not be less than [2) metres. 3

190 Paragraph 3 - Procedure for experiments Wave Spectra The JONSWAP spectrum is to be used as this describes fetch and duration limited seas which correspond to the majority of the cconditions worldwide. n this respect it is important only that not the peak period of the wave train is verified but also that the zero crossing period is correct..1 'Corresponding to a peak period of 4VH, and given that the enhancement factor y is 3.3, the zero crossing period is not to be greater than: [Tp/(1.20 to 1.28)] ± 5%.2 The zero crossing period corresponding to a peak period equal to the rolling period of the damaged model and given that the factor 7 is to be 1, is not to be greater than: [TP/(l.3 to 1.4)) ± 5% ; noting that if the rolling period of the damaged model is greater than EVH,, the peak period is to be limited to G/M,. Note t has been found that it is not practical to set limits for zero crossing periods of the model wave spectra according to the nominal values of the mathematical formulae. Therefore an error margin of 5% is allowed. t is required that for every test run the wave spectrum is recorded and documented. Measurements for this recording are taken to be in the immediate vicinity of the model (but not on the leeside)- see figure - and also near the wavemaking machine. t is also required that the model is instrumented so that its motions (roll, heave and pitch) as well as its attitude (heel sinkage an trim) are monitored and recorded thoughout the test. *m l The "near the model" wave measuring probe to be positioned either on arc A or arc B Fig. 1 4

191 3.2., 3.3., and 3.4 The contents of these paragraphs are considered self explanatory Simulated damages Extensive research carried out for the purpose of developing appropriate criteria for new vessels has clearly shown that in addition to the GM and freeboard being important parameters in the survivability of passenger ships, the area under the residual stability curve up to the angle of maximum GZ is also an other major factor. Consequently in chosing the worst SOLAS damage for compliance with the requirement of paragraph the worst damage is to be taken as that which gives the least area under the residual stability curve up to the angle of trhe maximum GZ. Paragraph 4- survival Criteria The contents of this paragraph are considered self explanatory. Paragraph 5 - Test Approval The following documents are to be part of the report to the Administration: (a) damage stability calculations for worst SOLAS and midship damage (if different); (b) general arrangement drawing of the the model together with details of constuction and instrumentation; (c) inclining experiment and rolling test reports; (d) calculations of actual ship and modal rolling periods; and Ce) nominal and measured wave spectra (near the wavemaking * machine and near the mode respectively) ) (if) representative records of model motions, attitude and drift (g) relevant video recordings. Note All tests must be witnessed by the Administration. 5

192

193 * APPENDX A HYDROSTATCS AND STATC STABLTY CALCULATONS * Critical KG and Trim (Shipyard) o Summary of Stability Calculations (Shipyard) * Summary of Stability Calculations (SSRC) * Damage Stability Calculations for Cases 3 and 6 (SSRC) i

194

195 FW Put n co -r.~ n c rn rr. N =r f nj r t- m- O- N N N N N n% n* m m Q) mo0 70 n n 1 *~ n - r ) a L co1 0'C '- cn a 0 V N ~ 'C\) c NN co No.- o.> Cal fl cf Er 0) 0 C) m o~ n co c o a'- -' oh~ ~~.0~~ ton n oe la N m n n ' LP +l. U-' n -n

196 Z. --. C.. e O N. ) n0 t C - - o C- -' '. - e - 0 C- ' ) N - -.

197 CD D C - (N r- oý o o 0 C _ ON CC c 3C: U- U 0.U , a'r6 41N U NCGehcr 'C4 <;A~oo6w U

198 < 40 ( O, , F " o :"-- ". "0 z z C,, 2: z3, : Xl ~W> lo C'J >* d, \, (Jeau) N )H)8Z

199 ... CASE 6 (FR 72-92) TRM ANGLE... = DEG. FOR ARD HEEL ANGLE... = DEG. PORT '\ LCG... = METRES KG... = SHP HYDROSTATC CALCULATONS COMPARTMENT 9-10 DRAUGHT (M) DSPLACEMENT (T).. = WPA (M--2)... = CB LCA (M)... = CW LCB (M)... = CM TP1CM (T)... = CPL MCHDEG (TM).. = CPV MCT1CM (T.M)... = WETTED S. AREA(M**2)= KB (M) BMT (M)... = GMT (M) KM4T (M)... = BML (M)... = GML (M)... = KML (M)... =

200 CASE 6 (FR H DSPLACEMENT TONNES A LENGTH METRES BREADTH DEPTH LCG = KG CALM WATER RGHTNG LEVER CURVE COMPARTMENT 9-1C AGLE OF HEEL RGHTNG LEVER POSTVE GZ AREA r E t i i , []

201 CASE 3 (FR )... TRM A1GLE... = DEG. FOR ARD HEEL ANGLE... = DEG. SB LCG... = METRES KG... = SHP HYDROSTATC CALCULATONS COMPARTMENT DRAUGHT... = 3.42 (M) DSPLACEMENT (T).. = WPA (M**2)...= CB LCA (M)... = CW LCB (M)... = CM TP1CM (T)...= CPL MCHDEG (T.M).. = CPV MCT1CM (T.M)... = WETTED S. AREA(M**2)= KB () BMT (M)... = GMT (M) fat (NM)... = BML (M) GML (M)... = KML (M)

202 CASE 3 (FR ) DSPLACEMENT...= TONNES LENGTH METRES BREADTH DEPTH LCG...= KG CALM WATER RGHTNG LEVER CURVE COMPARTMENT ANGLE OF HEEL RGHTNG LEVER POSTVE GZ ARE? !1!i i

203 APPENDX B DRWNGS OF TEST DAMAGE CONDTONS

204

205 ~ -N. - V. *,, ', -" / V ¼' E,// /, -- ~ ~ / <1< xx - -/ ~.,,r- -< 'B /~/~ 7-' 1,. "''A '/4 / K",.1 "{/ J t-/ ~ ~'i~' p)x / N 'K ti 77/ * L] Lii K-il '-¾! - )r.. A ~ ' 2 / \j ~1~ '~ ~ih ~jj ~ i %''j'\ /~/K'H//'¾',t'~ K) j'-$<$'.n, - L '~" L12 / -K7>Yxt4t 7'.,'. lp, 9-' K _

206 * ~ C: / :r /\ / vw \! " -. '-' -i L /, A 1 =1./-.. Si -.

207 n <_ <( N N

208 i - -.'.. ~r V iw~i ii / // /1 //~/ ~' ~/, i h/y/( '/A /4 7 j/<q: ;~:. ' -.1 1<~/'-"J'-4 '~ - i 4, it v/'~$ K~ %/jki;;~ ftp- / -tk Li H-ti zf-w ////~i;~>// i; v;> < ~//J~vv4/'ji v~ ~~ hi(2l~j 'ii jk. '//,-~.-.- *,j,%% 3~< i a1 ' P K T:&2'i~sy:; 2 1 r U ' ~' p ''' 5/ - - // - -- '..,.',. -0 '' '-<'-,',.-/ '-'.4 i. VA. / X//? V..s,%i --

209 C-, g F-].- it'.- / -- - ~-.--- / mm- W i w.=an~~~. vi. it / 1' -- / -' - / i -- / C ~1 A" 1/ L * '' C-, C C: C = C 'N = vz~i, - ~: -C C 2 -. "' -, ',/f> -<. A C- ffi K---

210 * U * 0 CC * C to 0 z -J U- to 4., C U U- C ) z z C U

211 i H APPENDX C PHOTOGRAPHS OF THE MODEL

212

213

214 A.

215 V * 'S rx, f ~:1* > >14 0 0

216

217 PED WAE W~APPE NDRL OXN DTTSTC N OLMONSATC

218

219 let Lo c ') CA LOU) C Ck Cl * ~ci c O o LO C) LN O r LoC CD coi cl 6 (oesp' Z W) (M)S

220 D 0o D C A' ~rcdn0 ci Cnen :)e o - or- o E- M~ C' 22 -: 0000 C4, 0.flC:).O.C

221 Table D2: Statistics of the experiments for case 3 damage and short waves Fwd Damage Short Waves Wave 331 Min Max Mean Sd Wave Travellin Wave Fixed Pitch An le Roll Angle Wave 332 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle \ Wave 333 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle H Roll Angle Wave 334 Min Max Mean Sd Wave Travelling Wave Fixed _0_1_1.0 K. Pitch RlAnl-12.5 e Roll Angle Wave 335 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle

222 \ Table HMid D3: Statistics of the experiments for case 6 damage and short waves Damage Short Waves Wave 331 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle Wave 332 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle Wave 333 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle Wave 334 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle Wave 335 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle [ 4.83

223 Table D4: Statistics of the experiments for case 3 damage and long waves Fwd Damage Long Waves Wave Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle Wave 12 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle Wave 13 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle Wave 14 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle Wave 15 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle

224 Table D5: Statistics of the Experiments for case 6 damage and long waves Mid Damage Long Waves Wave!] Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle Wave 12 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle Wave 13 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle Roll Angle Wave 14 Min Max Mean Sd Wave Travelling Wave Fixed Pitch Angle ,2 Roll Ang e Wave 15 Min Max Mean sd Wave Travelling Wave Fixed Pitch Angle Roll Angle

225 APPENDX E SAMPLE TME HSTORES OF MODEL TESTS 9 Free-Rolling tests * Wave and Roll Time Histories =>Forward Damage Short Waves M: Nidship damage Short Waves =>Forward Damage Long Waves M>Nidship Damage Long Waves

226

227 U) -c * C - 'C z U E '4- Zn cc z U.. - zt L 0 CL. 2 U, - U, - C -C - C * S (fl CV '4) - L) 0 14) - 14) (N 6 ~ a r 6 0* 0 6 r ~ * Cr C ~ r... C (flap) iio~ Cr Cr

228 * * i Ct) i t S C) * C & C) tn V XE w 0 U) - C) tno - 0), U C S Ci) a- :1 ~o 0)0 r w 2 C) S C) -~ 001 U~cu E 'U i - C) = * in C)- Ca * C) t * (?) L) C'J ~fl - ~ C - U C5r~ (NJ u~c) - C) (flea) iia~ -t Ci) C- >1 C)

229 C - cn cn *cl* C) 0 ~ E rj~ '~.1<11 Ca ~ * - C ~7 0 1.) 0 CA 0 tj2 U 0~ z U * -r z3m.~ 2: C,, - CaC,, N Ca Ca - Ca - - Ca = Ca 'i zt~ =o~ooc V a 0 0 C , [wj tmhpaqp QAQM (caifiql a6uv U) U)

230 11rr * a: r - cn '0 ~- Cd, fl Ca C" ci ce ~ a- * A - CA C: '0 Cao(3 z CA Ca C) * S R U, UCa = -4~ * C) Q ~ C.J C) ~cn -3- fwj UOWA~P aa% saasfiap~ a6uv ir'u.2- U:

231 - - ~ 2 Cr 'C C 0~ i til en U'C) - C.) * C,, ~ z C,, - C,, = 0 =0 - U c~1 - * - 0~ V C,, Z 000 ~ 4 o C,, (W UOLWA3~ 8~ M U,

232 C U 0, (~ c C.) b - 4) * - * ct 0~ z * ~0 C * ci C Co C: C., = - C) E - -~ '4 - C) * -4) 3 - C) C., - * - C: U - z Ca * Pt *~ - C H 3 * - 'A = = = t -*~=~'=~3P0-,3~' - 1w 1 uor$aaya ~AR~ * N N N N * - (saatapj ~i 6 'w FU.S ~ C,,

233 ii APPENDX F PHYSCAL MODELLNG AND SMLTUDE OF MARNE STRUCTURES

234

235 PHYSCAL MODELLNG AND SMMLTUDE OF MARNE STRUCTURES by Professor Dracos Vassalos The Ship Stability research Centre, Department of Ship and Marine Technology, University of Strathclyde, Scotland, U`K SUMMARY This review paper is aimed at providing information on and at explaining the appropriate use of models in the design of marine structures. Emphasis is placed on the derivation of scaling laws and on limitations of accuracy. Simrilitude, dimensional analysis and the use of governing equations in model experiment scaling are explained before addressing the modelling of some common ocean engineering tests. 1. NTRODUCTON 1.1 GENERAL Systems involving the natural phenomena of the oceans, the atmosphere, geographical features, and so on, are notoriously difficult to investigate in detail. Thus for the purpose of analysis and from the point of view of economics, we are compelled to make use of models of the hiill-scale (often called the "prototype') system. t is possible with laboratory models to ensure some control over the many variables of influence so that those of particular interest can be isolated for special study. Other * reasons for using models include: * To study phenomena defying analytical solution, or those for which no experimental results or empirical methods are available * To improve conceptual and physical understanding of a problem *To verify (or disprove) theoretical results :Data acquisition Calibration of a theoretical (numerical) model *To :To select alternative designs forecast hill-scale behaviour To identify governing parameters. n problems where analytical and/or empirical methods will yield a reliable solution the use of models is jot justified, as being more expensive. n the absence of these, however, it must be understood that the actual cost of model testing will usually be only a small fraction of 1% of the overall capital sums involved and it may well be a great deal less than the cost of corrective action which may be necessary as a result of failure to carry out tests. n the majority of cases that naval architects have to deal with, and especially those involving fluid phenomena, it is very seldom that a problem can be tackled by theoretical means alone. Here the art of engineering must be practised with experience, judgement, ingenuity and patience if useful results are to be obtained and correctly interpreted, and prototype performance predicted from these.

236 Page TYPES OF EXPERMENTS Experiments can be broadly classified into the following two categories, Couch, 1984: (a) Steady State: Typically, resistance and propulsion experiments where the model is tested in a steady condition at forward speed. The data collected is usually timeaveraged to render the results time independent. (b) Unsteady: Typically, seakeeping/ocean engineering experiments where, as a result of the transient loads produced by the ocean environment, such experiments are characterised by their dynamic non-steady nature. Data collected is now time dependent with results shown either as time domain statistics, e.g., rms, significant values, averages, etc., or as frequency domain functions, e.g. RAO's (Response Amplitude Operators) and response spectra. A further distinction can be made between seakeeping and ocean engineering * experiments: a Seakeeping usually implies a dynamic process involving ship like vessels generally with forward speed. * Ocean Engineering usually refers to the dynamic behaviour of stationary or nearly stationary structures (floating or fixed). Finally, in the context of either seakeeping or ocean engineering, the following types of experiment can be considered: * Rigid Body Response - motions, velocities, accelerations, relative motions, pressures. The results will be RAO's and/or motion statistics. * Hydrodynamic Forces - Added mass and damping components (forces excitation at various frequencies), wave excitation component (restrained model). * Flexural Body Response - Experiments investigating stresses or structural deflections of vessels and offshore structures. The emphasis in the paper will be on ocean engineering type experiments. 2. BREF HSTORCAL BACKGROUND Historical records show that experiments with models of ships have been performed since the time of Leonardo da Vinci. However, it was not until W.E. Froude, 1874, that model test predictions were established as a valuable engineering tool - in his case for the resistance of a full-scale ship from model resistance experiments in a towing tank. nterest in the seaworthiness of ships is as old as estimates of resistance and propulsion, but the accurate extrapolation of model test results to full-scale was not possible until the principles of spectral and transfer functions were established, as recorded in the seminal paper of St Dennis and Pierson in Over the years, model tests have demonstrated their usefulness in solving problems relating to both preliminary design and retrofitting. The obvious benefits of a successful model test programme can be an optimum hydrodynamic and/or structural design which will enhance the economic performance of the marine structure. Physical Modelling and Similitude of Marine Structures

237 Page 3 3. MODEL EXPERMENTS SCALNG Even though suitable measurements are generally easier to achieve in testing models that at full-scale, scaling problems can never by completely overcome, and it must be admitted that model tests in the controlled artificial environment of the laboratory always lack something of the uncertain harsh reality of the real world experienced by marine structures. n view of the difficulties in properly scaling an experiment, if a novel one is being planned, it is often prudent first to construct a small crude model for observational purposes only, in order to detect possible sources of problem, and to assist in planning the more detailed model tests. The basic principles of scaling laws and the main approaches to planning model experiments will be considered, under three headings. 3.1 PHYSCAL SMLARTY The concept of physical similarity implies the need to ensure that, in planning the experiment, a certain similarity should be maintained between the model and the prototype. n general, two systems are said to be physically similar (in respect of certain specified physical quantities) when the ratio of corresponding magnitudes of these quantities is the same between the systems. This proposition must be satisfied not only by the contents of the system but also by the system boundary, inputs and output. n other words, it involves not only the physical model itself, but also the test facility, the environment and the model response as shown in Figure 1. Sy~Res'P 'e System boundory, Figure 1 The Physical Similarity Domain The most important problems associated with marine structures involve geometric, kinematic and dynamic similarities. (a) Geometric Similarity (Similarity of Shape) f the specified physical quantities are lengths, the similarity between the systems is known as "geometric similarity". Where boundaries of solid bodies or patterns of motion are considered, the characteristic property of geometric similarity is that the ratio of any length in one system to the corresponding length in the other system is the Physical Modelling and Similitude of arine Structures

238 Page 4 same throughout. This ratio is usually called the scale factor; i.e., adopting the suffix "p" to denote prototype and "in" to denote model. X,(scale factor)= ( L--- Corollaries of geometric similarity are: Area, =? 2 Area. Volume, =X 3 Volume, --- massp = X 3 massm, (if p= pm) P (mk) 2 M.=, kp = k A serious departure from geometric similarity may arise from the failure to scale the roughness of the solid boundaries. Accurate representation of surface roughness entails scaling not only the heights of individual protuberances but also their distribution over the surface, which is usually determined by the manufacturing process. To overcome this deficiency several turbulence stimulators have been devised. (b) Kinematic Similarity (Similarity of Motion) When the flow pattern in a model system is geometrically similar to that in the prototype system then these two systems are said to possess kinematic similarity. Since, however, solid boundaries themselves consist of streamlines, geometric similarity of models is a prerequisite. Similarity of motion also implies geometric similarity and similarity of time intervals. Thus, as the corresponding lengths and corresponding time intervals in two systems are in a fixed ratio, so the corresponding velocities (and accelerations) must also be in a fixed ratio of magnitude at corresponding times. An example of kinematic similarity is found in a planetarium where parts of the Universe are reproduced to a given length scale factor, and planetary motions are copied in a fixed ratio 0f time intervals. (c) Dynamic Similarity (Similarity of Forces) Dynamic similarity exists when the ratio of corresponding forces is constant. The forces that may be relevant include: inertial = (mass x acceleration) * gravitational => (mass x g) * viscous = (shear stress x area) * elastic => (modulus of elasticity x area, or bulk modulus x area, or pc 2 x area - isentropie flow) pressure =:> (AP x area) capillary The fluid inertia force is common to all fluid dynamics problems and consequently any other relevant force is conveniently introduced as a ratio to this force. To form these ratios, the individual forces are first expressed in terms of a number of relevant physical parameters, i.e. Physical Mlkodelling and Similitude of iarine Sructures

239 Page 5 * length - 1 * velocity - v * gravitational acceleration - g * bulk modulus - k * modulus of elasticity - E * surface tension - y Using these parameters, certain important ratios can be formed as shown in Table 1. f one is seeking to preserve dynamic similarity between the model and prototype systems, and hence also geometric and kinematic similarity, it is necessary to preserve the relevant magnitudes of all the relevant forces as the physical parameters change magnitude. TABLE 1: Common Dimensionless Numbers Description Expression Name nertia Force Gravity Force nertia Force Viscous Force V pve pv- pv 2 (v2 nertia Force K or--e wp-) Ca(Ma) 2 Elastic Force Fn Rn nertia Force v Wb Surface Tension Force nertia Force V( P ) Eu Pressure Force f, for example the two forces considered to be relevant are the inertial and viscous fluid forces, then dynamic similarity entails the constancy of Rn between the model and prototype. f, however, inertial, viscous and gravitational forces are all important, this would require a simultaneous scaling of both Fn and Rn, i.e., keeping their ratio R. g'' 2 e constant. However, - and keeping this ratio constant for a substantial F. U change in " " would be impossible. n cases such as this, partial dynamic similarity would normally be adopted by scaling properly what is assumed to be the important quantities, whilst taking steps to ensure that the improperly scaled quantities are of little importance, otherwise correction factors may be used. 3.2 DMENSONAL ANALYSS A common approach to the planning of model experiments lies in carrying out a dimensional analysis of the situation at hand. This enables the relevant factors Physical Mudelling and Similitude of1'farine Structures

240 Page 6 influencing a particular problem to be assembled in a manner which is suggestive of its underlying structure and indicative of requirements that the model should satisfy. Consider, for example, the heave motion of a ship and its model, Lloyd, Without any detailed knowledge of the physical processes involved it might be surmised that heave amplitude will be a function of wave amplitude and frequency, the speed and heading, and the size, shape and inertias of the hull. n addition, fluid properties such as density, gravitational acceleration and viscosity should be relevant. On the basis of the above, and following the standard dimensional analysis methodology, it follows Z.= 4 1 {co,nv, a,l,[ ], ],p0,g g} (1) Where, f = some function of H = hull shape matrix [] = inertias matrix Expression (1) could further be written in the form - fl. 1L ll g'vý- V ' L t L [] ý pvl: ') (2) (i H What Equation (2) is implying is that the non-dimensional heave amplitude will be the same at both model and full scale provided that all the non-dimensional parameters on the RHS of the equation have the same numerical values at model and full scale. This requirement dictates the conditions required for the model experiment and it will be educational to explore this a little further. Non-Dimensional Wave Amplitude * (Lj (Co'L) _ Non-Dimensional Wave Frequency g g From the relationships 27rg 2 (t 2 =--'; -0 2 T=--; m k=- g --> X. ==X 1 /X; T 1 =Tp /-;k.m=kp Non-Dimensional Speed dentical Fn's yield Vm = VP / A. On the other hand, identical Rn's require V. = VpX. Clearly, unless X= the above requirements cannot both be simultaneously satisfied, a conclusion already arrived at. t may be observed, however, that Physical AMOdelling and Similitude of larine Structures

241 Page 7 maintaining Rn constant is not a practical proposition (a prototype speed of 30 knots and a X of 30 would require a model test speed of 900 knots!). Fortunately, viscous forces do not play a prominent part in rigid body motions. The requirement of keeping Rn constant may therefore be waived provided that the right precautions are taken, i.e. adopting as large a model size as possible, and using turbulence stimulators if this is judged appropriate. Finally, assuming that geometric similarity is satisfied and linearity prevails, Equation (2) may be written as Z. =f {o 3, Foa offt F F.,t (3) Extending the above argument to irregular waves, this becomes: RMSZ = f,{4-alto E~~a Summarising the above, the experiment scaling laws are shown in Table 2. TABLE 2: Experiment Scaling Laws (Froude Scaling) Prototype Quantity.. ModeQuantity Mass,mp mp x X-3 Length, Lp Ln Y A- Time, TD Tp x z 1 /2 Velocity, VP Vp X -12 Acceleration, axp op x Angle, 0p Angularl Velocity, 6P Angular Acceleration, 0, 0Pxl 0PXV 0,pxX Pressure/Stress, Pp Pp x X- Frequency, fp fp x X 1/2 Force, Fp Fp x A.3 Moment, Mp Mp x GOVERNNG EQUATONS Model experiment scaling may also be derived by an examination of the governing equations of a problem when these are known. These equations must hold for both the model and the prototype and can be used to develop appropriate scaling laws. Consider, for example, the classic single DOF. spring-mass system, where the equation of motion is given by: mk(t) + cx(t) + kx(t) = F(t) (4) Pkvsical Modelling and Similitude of Marine Structures

242 Page 8 Related parameters of interest include the natural frequency (o n and the damping ratio y defined as: k c W2 (5) Equations (5) must be also held for both the model and the prototype. Applying equation (4) to the model and introducing appropriate scaling factors, e.g., X. = mp /m., etc., yields V KX,X,QmP K +( - -,)ck, + (? k X, )kpx, = ;-'Fp (6) Multiplying (6) by KkXx and comparing the corresponding terms between (4) and (6), * yields - X,- -1 (7) Applying also (5) to model and prototype (?L is constant), Ko(g 6=Kkfl" 8 >.o..,. = z,(8) Solving, for example, (7) and (8) in terms of?.x, -m and Xt kx = X.X F ' ra?lt t Adopting Froude scaling, K x = K, Km = X 3, Kt = K'1/ 2 -Ak = 2 k; Kc = K 25 ; KF=K 3 The above is known as "group theory" approach. 4. MODEL SCALNG LAWS FOR OCEAN ENGNEERNG TESTS 4.1 NON-D[MENSONAL PARAMETERS Nearly all ocean engineering structures, whether they be fixed platforms or floating vessels employ circular tubes for part of their structure. Although steady current situations are of interest for all such structures, the prime design loading arises from wave action as a result of the varying velocity and acceleration due to water particles motion. t is well known that the drag coefficient of long smooth cylinders subjected to steady flow exhibit a large change at Rný5 x n waves, the drag coefficient and added mass coefficient for a circular cylinder appear to vary with both Rn and the Keulegan-Carpenter Number (NKC = UmT/D ; period parameter). The NKC describes the viscous scale effects acting on circular cylinders in sinusoidal flow fields. For small values of the parameter, the values of added mass and drag coefficients vary considerably. At higher values (increasing period of the flow oscillation), the coefficients approach constant values, dependent only on the instantaneous Rn. The majority of offshore structures will experience their severest loading when Rn is in the super-critical range and only a few items, e.g., drill pipes, mooring ropes, etc., will Physical Modelling and Similitude of Marine Structures

243 Page 9 experience maximum loading while in the sub-critical or critical ranges. As the diameter of the member increases, the relative significance of the velocity and acceleration induced forces changes, with the inertia forces beginning to predominate until the stage is reached where there is substantial reflection of the waves with diffraction also occurring as shown in Figure 2. Another viscous phenomenon associated with relatively low flow velocities is the rate at which long, bluff bodies -. such as cables or risers shed vortices. An important parameter for these types of flow is the Strouhal number (S=fD/U; vortex shedding frequency parameter). f the frequency at which the vortices are shed matches one of the natural frequencies of the structure, serious dynamic loading can occur. The subject of time dependent flows acting on slender members, such as tubular structural members, is of high interest to the offshore industry. f viscous forces dominate and Reynolds scaling law, is applied: m =vx 'Vp mm =v',px3; p F = Fp The viscous forces, if scaled properly, are the same at model and full scale while inertia forces are smaller by a factor X z + \ 2.0 \o L') F MORSON >- NERTA OR / DFFRACTON/ DFFRACTON Figure 2: Regions of Validity -Force Prediction Methods for a Fixed Pile Plysical Modelling and Similitude of Marine Structures

244 Page MODELLNG OF WAVE FORCES The general problem concerning the forces acting on a fixed body of characteristic dimension D, when subjected to a regular wave train of length L, height H in water depth d is of fundamental importance. Any time-invariant force F, such as the maximum in-line force, may be expressed in functional form as: -.. F= f (p, Ri,D,H,L,d,g) (9) from which it may be derived that -ghd' L,H/L,D/L,Rn) (10) n model tests the constancy of the dimensionless parameters d/l (wave-depth parameter) and HL (wave steepness) between the model and the prototype ensures that the wave train is suitably represented. A selection of alternative parameters in the literature include: H/d, kh, H/gT 2, HL2/d 3 (Ursell parameter), kd, d/l, d/gt 2 and Td/. Any four independent parameters will suffice. Depending on the size of D/L the following cases arise: D/L>0.2 - wave forces in the diffraction regime When D/L is sufficiently large, wave diffraction or scattering is important. Furthermore, the amplitude of water particle oscillation relative to D is necessarily small and the flow separation effects may be neglected. Therefore Rn may be omitted and the flow may be treated on the basis of potential theory. n addition if H/L is small, F may be considered to be proportional to H and thus H/L may also be omitted. Thus, in the usual linear diffraction problem it follows: pghd 2 F F = f(d /L, D /L) or f(d /d, D /L) ( n deep water, f(d L (12) DL <0.2 - wave forces in the flow separation regime f D/L is small, the body will no longer scatter the incident wave field and D/L has no direct physical significance. Under such conditions, however, NKC assumes importance. Equation (10) now becomes F ighd = f(d /,H/L, N,Rn) (13) This case can further be subdivided into two problems depending on the degree of importance of Rn. (a) Rn Effects Unimportant Provided changes in Rn are taken to be admissible, Equation (13) represents a situation that can be reproduced in the laboratory, i.e., the lack of ability of maintaining high Rn in model tests is assumed to be of no major importance. n this case, under certain Physical Modelling and Similitude of Marine Structures

245 1 Page conditions, a further simplification may be possible. For example, the 2-D sinusoidal flow normal to the axis of a vertical cylinder can be characterised by U m (maximum horizontal wave particle velocity) and T. Then the time-invariant sectional force F' may be written as 1/2PDUm F' =f(nkcrn); NKC =UmT/D; Rnf=UmD/v (14) Correspondingly, the instantaneous time-dependent force F' would be written as F,' = f(nkc,rn, t / T) (b) Rn Effects mportant This case presents the most significant source of difficulties in the modelling of offshore structures. Since it is not possible to model a free-surface flow so as to maintain both Fn and Rn constant, several approaches attempting to by-pass this difficulty have been devised. t should be stated that when no flow separation occurs or if it is located so as not to influence the overall loads on the structure, the Rn effect may be neglected. Some of the approaches worth mentioning are: * restricting the simulation to 2-D -flow by using a piston or a U-shaped tube, Sarpkaya, 1981 * body oscillating in otherwise still water, e.g., Strathclyde ULOC, Wolfram, 1991 * using turbulence stimulators 4.3 MODELLNG OF ELASTC STRUCTURES When taking account of the dynamic response of an elastic member, a number of additional parameters are needed to characterise its behaviour. These include its density Ps, modulus of elasticity E, and damping properties, conveniently characterised by the damping ratio which is partially structural and partially hydrodynamic Ysh. The Cauchy number also becomes yet another dimensionless parameter to consider. n wave force problems (Fn constant), Equation (13) would be extended to F FgLD 2 =f(d/l,h/l, NKC,Rn,p,/D,pUŽ/E, y) (16) deally all three additional parameters should be held constant between the model and the prototype, but in practice this may not be possible. On the basis of Froude scaling, structural deflections will scale by X and the same applies to both moduli of elasticity E and rigidity G: Angular rotations remain unchanged. n many problems the most important effect of E lies in describing the structure's fundamental natural frequency, fn, and it is then convenient to adopt the alternative U/fnL (reduced velocity) in place of pu 2 /E. The required frequency may be obtained by a model which is itself rigid but which is elastically mounted, rather than any flexibility in the model itself When a fully elastic model of the entire structure is required some flexibility is achieved by attempting to scale E correctly rather than E Physical Modelling and Similitude of Marine Structures

246 Page 12 or in isolation. Alternatively, segmented models connected by variable springs or special materials may be used. 5. CONCLUDNG REMARKS Based on the information presented in the foregoing two key remarks are noteworthy: * There is no doubting the usefulness and necessity of physical models in the design of marine structures. * Before any testing is undertaken and more importantly before any results from model tests are to be used, the limitations of the particular test must be clearly understood and appreciated. 6. REFERENCES Couch, R.B. et al, "The Use of Model Basins in the Design of Ships and Marine Structures", Trans. SNAME, Froude, W., "On Experiments with H.M.S. GREYHOUND", Trans. NA, Lloyd, A.R.J.M., "Seakeeping: Ship Behaviour in Rough Weather", Ellis Horwood, * Sarpkaya, T. and saacson, M., "Mechanics of Wave Forces on Offshore Structures", van Nostrand Reinhold, St Dennis, M. and Pierson, W.J., "On the Motions of Ships in Confused Seas", Trans. SNAME, 1953 Wolfram, J., "A Novel Underwater Hydrodynamic Experiment Facility and First Results with Smooth and Marine Growth Covered Circular Cylinders", Trans. RNA, Physica Mfodelling and Simititude of :larine Structures

247 '. U APPENDX G WEATHER STATSTCS

248 p'

249 a W 5 H3 titt amr4 :W-. zz; M ýý, l0 RaN N, *4~fl~~ &NAM# ~a - 4,0 11. A c2 4