LNG / LPG ARMS PROTECTION AGAINST EXCESSIVE SHIP MOVEMENT: A NEW APPROACH FOR ALARMS SETTING AND ESD / ERS ACTIVATION

Transcription

1 LNG / LPG ARMS PROTECTION AGAINST EXCESSIVE SHIP MOVEMENT: A NEW APPROACH FOR ALARMS SETTING AND ESD / ERS ACTIVATION Bertrand LANQUETIN Liquefied Gas Shipping Department Technical Manager TOTALFINAELF Gas Electricity Division ABSTRACT The new OCIMF Design and Construction Specification for MARINE LOADING ARMS (Third Edition 1999) departs significantly from previous editions with regard to ESD and ERS alarms setting. In particular it does not specify any longer figures for ESD and ERS sequences duration and for drifting distances / speeds to be used for first stage and second stage alarms setting on loading arms. On the contrary, it stipulates clearly that: The Owner shall determine the drift velocities for each berth. This will be based on the type and size of vessels using the jetty, wind, current, etc. (Chapter ) It stipulates also that: The first stage alarm will initiate the ERS ball valves closure (Chapter ) This issue of alarms setting was raised by our Company to the OCIMF loading arms Task Force as early as 1997 after considerable work and experience have been gained on the subject in recent years at the opportunity of various LNG / LPG terminal studies characterized by very different weather conditions and ships dimensions (like Indonesia, Yemen, India, etc.). As a summary we found that the traditional way of alarms setting based on the consideration of three parameters fixed arbitrarily and taken independently from each others was not reflecting the reality and could lead either to unnecessary arms envelope wastage or to underestimated drifting speeds and distances. Furthermore no account has been taken until now for a catastrophic scenario like ship breakout. These three parameters are namely: the model of the flanging area and the drifting area (allowance for ship movement at berth), the duration of the ESD and ERS sequences and the estimation of the drifting distances and speeds. A global approach incorporating the accurate estimation of these parameters and examining how they interact is therefore necessary and has been successfully developed. The present paper summarizes the work done on the subject by our Company and gives, through a practical example resulting from a recent project of LNG jetty, our latest methodology for: Establishing the flanging and drifting areas and the setting-up of the alarm 1 st step delimiting a working envelope called safe area. Within this area the ship-mooring pattern remains in sound condition and the cargo operations can take place safely. Both cases of ship reaching this alarm with unnoticeable or with noticeable speeds are considered. (The pre-alarm zone is located within this safe area). Setting-up of alarms 2 nd step and arm mechanical limit beyond the limit of the safe area in a scenario of excessive movement of the ship resulting from the mooring pattern failure. Alarms detection and the re-visited ESD / ERS activation philosophy accordingly. This new approach not only supports the changes done in the Third Edition of the OCIMF Specification, but also goes beyond the OCIMF requirements for the particular case of arms handling liquefied gases. At the same time it eliminates any wasting of unused arms envelope through a good adequacy between ESD and ERS sequences duration, the ship drifting distances and speeds, and the arms design (mechanical envelop limits).

2 PART 1: TRADITIONAL METHOD FOR ARMS DESIGN AND ALARMS SETTING, ALARMS DETECTION AND ESD/ERS ACTIVATION This rather long introduction is necessary in the first part of the paper in order to well understand the traditional method in use for years for arms design and alarms setting in order to identify its insufficiencies and the reasons why improvements are necessary. This method has been used back to 1980 (Design and Construction Specification for Marine loading Arms, 1 st Edition - see Ref [1]) until now and is still continued to be used in many cases as the common practice. The method is summarized on Figure 1 and described hereunder. 1.1 DETERMINATION OF ARM MECHANICAL ENVELOPE (ARM DESIGN) This is done by arm vendor and is based on two sets of data: The flanging area: this area is approximated by a rectangular parallelepiped, the sides of which being based on the extreme positions of the ships flanges (actual flange locations including tide and ship condition: ballast or loaded). The drifting area: this area is also approximated by a rectangular parallelepiped obtained from the flanging area by adding a longitudinal drift and a lateral drift of the ship. There is quite a significant discrepancy on the recommended values to be used for longitudinal and lateral drifts. Furthermore we could not find how these values have been determined in the past: The OCIMF (1 st Edition 1980 and 2 nd Edition 1987 see Ref [2]) recommends +/- 3.1 m to be found as a reasonable approximation of the allowable vessel surge for sheltered areas and +/- 4.6 m for more exposed sites. For the lateral drift a value of 3.1 m is recommended for the vessel sway-off (at any site). It is further assumed that the maximum sway and the maximum surge do not occur simultaneously. The CEN standard EN 1474 (1 st Edition 1997 see Ref [3]) suggests a maximum drift area of 3.0 m for both surge and sway movements of the ship. Note: nor OCIMF or CEN specifies if the above is a quasi static drift or if it includes in addition the movements of the ship caused by the waves. The PIANC Criteria for Movements of Moored Ships in Harbors 1995 (see Ref [4]) has established criteria for safe mooring of gas tankers as being 2.0 m for surge and 2.0 m for sway. These are specified as peak to peak values for surge and zero to peak values for sway with a velocity of 0.3 m/s in both surge and sway to be used for berth design. Whatever the final figure for surge and sway is all here above specifications assume that the mooring pattern is in intact condition. Although integrity of the arms is protected by excessive drift detection and corresponding ESD (Emergency Shut-Down) and ERS (Emergency Release System) sequences activation, the fact to base the arm mechanical envelope (or in other words the arms design) on the assumption of an intact mooring pattern and on drifting values somehow arbitrary is a too simplistic approach. This approach also ignores the fact that a combined longitudinal and lateral drift is a reality. The SIGTTO Information Paper Accident Prevention, the Use of Hoses and Arms at Marine Terminals Handling Liquefied Gas 1996 (see Ref [5]) states very clearly, with statistic calculations to support that statement, that the risk of break-out is high (1.57 for 100 port calls) it identifies a very real risk that should be addressed. 1.2 SETTING UP OF ENVELOPES FOR ALARMS 1 ST STEP AND 2 ND STEP It has to be first clearly mentioned that this responsibility has usually been left in the past to the arm vendor and this is still often the case today, although the OCIMF Design and Construction Specification for Marine loading Arms, 3 rd Edition 1999 (see Ref [6]) departs from the previous Editions and stipulates clearly that: The Owner shall determine the drift velocities for each berth. This will be based on the type and size of vessels using the jetty, wind, current, etc. (Chapter ) The alarm 1 st step and 2 nd step envelopes have usually a spherical shape (although a rectangular shape has been often used in the past to model the alarm 1st step envelope, the PMS arms Position Monitoring System - allowing both rectangular and spherical coordinates). The traditional method for alarms setting is based on the following recommendations: The OCIMF (1 st Edition 1980 and 2 nd Edition 1987, see Ref [1] and [2]) states (page 47) that the signals from these sensors are relayed to a control console where alarms are actuated when

3 FLANGING AREA based on ships data DRIFTING AREA 4.5m or 3.0m ESD 30S ERS 2s TO 30s DRIFTING SPEEDS 15cm/s for ERS, 5cm/s for ESD ARM MECHANICAL ENVELOPE (ARM LENGTH DESIGN) SETTING UP OF ENVELOPES FOR ALARMS 1st STEP AND 2nd STEP DRAWBACKS: SUGGESTED DISTANCES (IE. DRIFTING AREA), SPEEDS (IE. DRIFTING SPEEDS) AND DURATIONS (IE. ESD, ERS) ARE TAKEN INDEPENDENTLY, WHILE THEY ARE IN FACT CLOSELY INTERRELATED DO NOT REFLECT ACCURATLY OPERATING CONDITIONS, IE. WEATHER CONDITIONS AND SHIPS SIZES FOR A GIVEN SITE DRIFTING SPEED PROFILE (IE. SHIP ACCELERATION) NOT PROPERLY MODELLED, IN ADDITION SAME DRIFTING SPEED IS USED FOR LONGITUDINAL AND LATERAL SPEEDS SEQUENCE OF PERC ACTIVATION, LOGICS OF CONTROL (PROXIMITY SWITCHES AND/OR PMS) NOT INCORPORATED IN THE METHOD BECAUSE THEY ARE RELEVANT TO COMPANIES POLICIES CONSEQUENCES: ARM LENGTH MIGHT NOT BE OPTIMIZED WORKING ENVELOPE MIGHT NOT BE OPTIMIZED DRIFTING SPEEDS MIGHT BE GREATER Figure 1: Traditional method for arms design and alarms setting, ESD/ERS detection and activation

4 predetermined slew and drift limits are reached. As said before, the owner and vendor s responsibilities described in Part III of these specifications did not cover the alarms setting and this was left in most of the cases to the arms vendor. Traditionally the figures deemed to be safe and used by the vendors as a standard are 0.5 m between the alarm 2 nd step and the arm mechanical limit (which supposes a drifting speed of 15 cm/s for a typical ERS sequence duration of 5 s) and a distance of 1.5 m between the alarm 1 st step and the alarm 2nd step (which supposes a drifting speed of 5 cm/s for a typical ESD sequence of 30 s). The CEN standard EN 1474 (1 st Edition see Ref [3]) suggests also the same figures to be used, i.e. 0.5 and 1.5 m. Similarly to the values given for the drifting area we could not find how these values of drifting speeds have been determined and what were the hypothesis used (weather conditions, size of the ship, intact mooring or ship break-out, etc.). The fact that the drifting speed increases with time after break-out is somehow taken into account (speed increasing from 5 to 15 cm/s between ESD and ERS sequences) but nothing is said on the hypothesis retained for ship acceleration, although one can easily recognize that the use of constant drifting speeds can be very misleading. Finally it is assumed that the longitudinal and lateral drifting speeds are the same, which is obviously not the case and further, no combined drifting is considered. The deficiencies of the traditional method for arms design (mechanical envelope) and alarms settings are not limited to the above. The major approximations made are in our opinion: The assumption that the drifting area and the drifting speeds are universal: it is clear that they vary with the size of the ship, the mooring pattern composition and its integrity, the environmental conditions and the drifting direction of the ship. The assumption that the longitudinal and lateral drifts are the same while it is not the case and while the limitations with regard of arms design are different: a big lateral drift can be accommodated at design stage by longer arms for example, while a big longitudinal drift cannot be always accommodated, the limit being the maximum arm slewing angle which is limited to approximately 45 (an illustration is given in PART 4 Figure 12 of the paper). The assumption that the ESD sequence has to be taken universally at 30 s. The assumption that the arms design (mechanical envelope based on the flanging area and the drifting area), and the setting of the alarms 1 st step and 2 nd step (based on drifting speeds and ESD and ERS sequences duration) are two independent things, while in practice they are closely interrelated: for example high drifting speeds can be expected on a weather exposed loading jetty berthing big LNG ships. This can be accommodated in several ways: longer arm design (bigger mechanical envelope), shorter ESD and ERS sequences, or ideally by an optimum combination of the two. The above approximations may result in an underestimation of the drifting speeds with, in such case, a serious risk of damaging the loading arms in case of ship break-out (in case the ESD and ERS sequences cannot be timely completed), or at the opposite in a wastage of the arm mechanical envelope: operating envelopes showing a drift of several meters before reaching the envelope alarm 1 st step are not uncommon. It is most likely that such drift could never be reached without a partial or total failure of the moorings. In such scenario the activation of the ESD sequence will be probably too late. On these examples there is an immediate benefit of triggering the ESD sequence earlier by moving the alarm envelop closer to the working area, and hence allowing a better use of the arm mechanical envelope space. The above concerns were conveyed as early as in 1997 to the Task Force in charge of preparing the 3 rd Edition 1999 of the OCIMF Design and Construction Specification for Marine loading Arms (see Ref [6]) and as a result the following new wording was used: The Owner shall determine the drift velocities for each berth. This will be based on the type and size of vessels using the jetty, wind, current, etc. (Chapter ) One should realize that this apparently innocent wording invites in practice to a major change in the approach adopted for years in the field of arms design and alarms setting and indicates clearly were the responsibilities are laying.

5 The OCIMF does not give however any recipe on how to achieve this goal. We trust that our experience described in this paper will be helpful and valuable for terminal owners and designers and for arms vendors. A last point has to be mentioned on the traditional method described above because it has been neglected in the past or fully left to the arms vendors, who have developed the PMS (arms Position Monitoring System). This monitoring tool can be very powerful providing the philosophy of alarm detection is well understood and used in the appropriate way. 1.3 ALARM 1 ST STEP AND ALARM 2 ND STEP DETECTION AND ACTIVATION The way we propose to achieve this is detailed later in PART 2 of the paper. We want to highlight hereunder some of the uses, which have been developed following the appearance of the ERS (1980) and the PMS (1986). As it can be seen, many discrepancies have been found in the various approaches denoting that a serious rethinking of the whole ESD/ERS alarm detection and activation was necessary. Hereunder is a rapid summary on the various uses: Alarm detection: the detection is done traditionally by proximity switches (inductive type), number of which and position being clearly defined in the OCIMF specification. PMS sensors come additionally (potentiometric type or absolute encoder type sensors). These sensors can detect both arms flange position and speed (drifting distances and drifting speed). The position of the PMS sensors is usually left to the arm vendor. Back-up proximity switches by PMS sensors: this is left either to the owner or more generally to the arm vendor. Shape of alarm envelopes: it is clear that the lateral envelope (in case the ship moves away from the jetty) has a spherical shape because proximity switches measure angular co-ordinates. However PMS allows either to measure angular or rectangular co-ordinates. On many installations the alarm 1st step detected by PMS has a rectangular shape (this alarm level being considered as an extension of the arm drifting area, which has also a rectangular shape) while the same alarm 1 st step detected by proximity switches has a spherical shape (this alarm level being considered as a protection before to reach the arm mechanical envelop, which has also a spherical shape. To our knowledge however the alarm 2 nd step detected either by proximity switches or PMS sensors has always a spherical shape. Integration of ship speed to anticipate triggering of alarms: the PMS being able to measure the drifting speed, allows to anticipate the triggering of an alarm. This facility has been used sometimes for the early activation of the alarm 1 st step, however sometimes also for the activation of the alarm 2 nd step and sometimes for both, without taking into account the laws of hydrodynamic for a drifting ship, i.e. drifting acceleration after break-out. In other words, if the alarm 2 nd step is anticipated by the PMS, there is a real risk that an ESD sequence (alarm 1 st step) triggered for example by proximity switches, will not have the time to be completed before activation of the ERS sequence (in case interlock is provided, there is then a real risk to reach the arm mechanical envelope before the ERS sequence is completed. ESD and ERS sequences duration, time delays between ESD and ERS and by-pass flow restriction on the hydraulic activating the ERS sequence: ESD sequence is usually 30 s, sometimes more, rarely less (to take into account the duration of the ship manifold ESD valve closure set at 30 s). The ERS sequence duration is from 2 to 30 s. Some mitigating measures are often taken in order to reduce the risk of surge pressure, depending on sequences duration, time delay between ESD and ERS sequences (see Ref [7] and [8]). Sometimes, and in order to accelerate the ERS sequence in case a high drifting speed is detected by the PMS, a by pass flow restriction on the PERC hydraulic (Powered Emergency Release Coupling) is automatically activated (this device has been installed on some recent installations). Sequence of closing PERC valves: it seems that the majority of owners prefer to have the PERC valves closing during the ESD sequence, with the ERS sequence therefore limited to PERC opening. Some owners however prefer to have the ERS sequence to be entirely done when alarm 2 nd step is triggered, in that case the sequence includes the PERC closing valves following by PERC opening activation.

6 Other considerations: some other considerations must be cited here like hydraulic interlock between ESD and ERS sequences. More recently these traditional interlock systems recommended by OCIMF had to be modified on the so called no spill PERC design, where the interlock system between ESD and ERS sequences had to become both mechanical and hydraulic due to the fact that the upper ball valve of the PERC closes during ESD sequence while the lower ball valve sequence closes during the ERS sequence. It can be seen from the here above that a lot of various approaches if not discrepancies - exist on the particular subject of ESD/ERS alarm detection and activation. This was brought forward also in 1997 to the OCIMF Task Force in charge of the revision of arms design and construction specification and as a result the following changes in the new specification must be highlighted: - No mention is made any more on the ESD sequence duration - The specification stipulates that: The first stage alarm will initiate the ERS ball valves closure (Chapter ) It is clear here again that the aim is to put more responsibility on the owner to establish its own safe and reliable design based on the proper characteristics and operating conditions of his jetty. Here again as well the OCIMF does not give any recipe. We have in the past years, at the opportunity of various projects, revisited the ESD/ERS philosophy with regard to detection, activation, etc. and we trust that our experience described in this paper will be helpful and valuable for terminal owners and designers and for arms vendors. PART 2: A NEW APPROACH FOR ALARMS SETTING, ALARMS DETECTION AND ESD/ERS ACTIVATION This revisited methodology is the result of our experience gained from various recent projects and is now enforced in our Company. These projects were for example the Yemen LNG project characterized by a difficult SouthWest monsoon and the Trombay LNG project characterized by significant tidal currents. For projects like this it appeared very quickly that the traditional method for arms design, alarm setting and detection and ESD/ERS activation was not appropriate and this has been at the origin of our reflection on the subject and on the perfecting of the method described hereunder. The methodology is based on: The laws of hydrodynamic (behavior of the ship at berth and after ship break-out) with the use of sophisticated computer programs to model the behavior of the ship. The logical and the fault tree analysis of the failure events (intact or defective mooring after lines failure). Considerations of the risk analysis (integrity of the arms). Increased level of safety and protection of the installation (anticipation of the events by initiating proper and adequate actions). The concern of maximum use of the arm mechanical envelop without any wasting (in both normal and emergency operations). 2.1 ALARMS ZONES AND ALARMS DETECTION The revisited concept is summarized on Figure 2 to which the hereunder description refers to. To be complete, a description of the type of computer programs we recommend to use for the evaluation of the ships movements at berth (intact mooring conditions) and for the drifting distances and speeds after break-out (total mooring failure) is given in PART 3 of this paper. This will be followed by a PART 4, in which we will give a concrete example of loading arms design and alarms setting based on a recent project where this methodology has been applied. A. Flanging area: This space allows for a variation of ship manifold flanges positions for the range of ships handled. The hip is supposed to be in ideal position on the spotting line and on berthing line, i.e. with no fender compression.

7 SAFE AREA, UNNOTICEABLE DRIFTING SPEED ELEVATION MECHANICAL ENVELOPE (FIXED) ALARM 2ND STEP (FIXED) A FLANGING AREA ALLOWANCE DRIFTING FOR AREA SHIP MOVEMENTS AT BERTH D1 D2 A A ALARM 1ST STEP (FIXED) ALARM 1ST STEP (VARIABLE) ALARM 1st STEP IS IN SPHERICAL COORDINATES (D1 & D2 ARE BASED ON SIMULATED DRIFTING DISTANCES; DRIFTING AREA IS CALCULATED BY MODEL) D'2 D'1 D'1 D'2 FLANGING AREA DETECTION: PMS SENSORS, FUNCTION OF SHIP DRIFTING SPEED SAFE AREA, UNNOTICEABLE DRIFTING SPEED ALLOW. FOR SHIP MOVT DETECTION: PROXIMITY SWITCHES & PMS SENSORS DETECTION: PROXIMITY SWITCHES & PMS SENSORS SECTION A-A MECHANICAL ENVELOPE (FIXED) ALARM 2ND STEP (FIXED) ALARM 1ST STEP (FIXED) ALARM 1ST STEP (VARIABLE) SPOTTING LINE (D'1 & D'2 ARE BASED ON SIMULATED DRIFTING DISTANCES; DRIFTING AREA IS CALCULATED BY MODEL) TO BE ON CONSERVATIVE SIDE, DRIFTING AREA HAS A RECTANGULAR SHAPE IMPORTANT NOTES: A) IN THE SAFE AREA IT IS SUPPOSED THAT THE MOORING PATTERN IS INTACT; BEYOND IT IS SUPPOSED THAT THE MOORING PATTERN HAS A COMPLETE FAILURE B) ALLOWANCE FOR SHIP MOVEMENTS AT BERTH IS USUALLY CALLED "DRIFTING AREA" Figure 2: The revisited methodology - alarms zones and alarms detection

8 It is determined by tabulation of X1, X2, Y, Z1, and Z2 values for the range of ships handled (see illustrations on Figures 3 and 4 enclosed). These values take into account the tide range and the ship draft variations between the loaded and operational ballast conditions. It has a polygonal shape, which we recommend not to be approximated by a rectangular parallelepiped. The reason being that, due to the recommendations for ship manifold standardization (see Ref [9]) and Ref [10]) and the relatively comparable geometrical characteristics of liquefied gas ships in the medium to the big size range, very few ship flanges are in the lower part of the flanging area, and those who are in that area have generally their flanges close to the berth because they correspond to relatively small ships. Approximating the polygonal shape by a rectangular parallelepiped would considerably inflate the volume of the lower part of the flanging area, knowing that the arm length is most of the time determined by the difficulty to reach the lowest flanges due to the high elevations of the loading platform position. To approximate the flanging area by a rectangular parallelepiped would result in an unnecessary over-sizing of the arm length (arm mechanical envelope) for purely geometrical reasons and without bringing more safety in a system (it is a much better and safer approach to accept an increase the arm mechanical envelope because of high expected drifting speeds compared to an increase of this mechanical envelope for unjustified geometrical reasons, i.e. unduly approximation of the flanging area). The Figure 5 illustrates well the here above point, valid in a majority (some exceptions exist) of cases. B. Drifting area, setting up of alarm 1st step, pre-alarm: It is the maximum excursion of the ship under wind, current, and wave, with an intact mooring pattern, rounded up to a value to be given by the owner (the rounding takes into account the fact that some ships may have smaller pre-tensions in their lines or use mooring lines with greater elasticity for example). The drifting area is the result of the ship movement around its ideal position described above, resulting from forces and movements under wind, current and waves action including quasistatic and dynamic movements. As such the traditional wording drifting area is unfortunately very misleading and a confusion can be made with the ship drifting after partial or total rupture of the moorings (ship break-out). A simple wording like allowance area for ship movement at berth would be much more appropriate and will be used in this paper. To be on the conservative side, this drifting area for a given ship with intact mooring is assumed to have a rectangular shape and not a circular shape (arm vendors utilize both; a circular shape implies that the ship movement at berth is the same in all directions, which is obviously not the case. A rectangular shape is a more accurate representation and will lead to longer arms to encompass this rectangular shape). The values of the lateral ship excursion (affecting the arm luffing movement) and of longitudinal ship excursion (affecting the arm slewing movement) are based on the maximum sway and surge motions given by the "Berthing Model Studies" (this will be analyzed in PART 3 of this paper) and are rounded-up to a value to be given by the arm owner. These values (lateral and longitudinal excursions) are added to the lateral Y and to the longitudinal X1, X2 components of the flanging area to determine a polygonal volume, which as a result, is used to set-up the alarm 1st step envelope of the loading arms. The alarm 1 st step envelope has a spherical shape (and not a rectangular shape as it has been done sometimes in the past) and is determined so as to encompass the above polygonal volume (see section "A - A" on Figure 2). It is assumed that this alarm can be reached for a ship drifting ship excursion movement - with an unnoticeable speed (for example the ship can reach an equilibrium position under the action of the elements). In such case the alarm detection is done both by proximity switches and the PMS sensors backing up each other as indicated on Figure 2. (Enough sensors allow detecting luffing and slewing arms movements respectively). Alternately, this alarm envelope can be reached with a noticeable drifting speed of the ship denoting either an abnormal dynamical behavior of the ship or an improper line tensioning (too slack pattern) or a partial failure of the mooring lines. In such case, and depending on the drifting speed, the alarm 1st step is anticipated by the PMS so that the ESD sequence can be triggered earlier, see Figure 2.

9 Y Fender compression Figure 3: Flanging area - tabulation of X1, X2 & Y Figure 4: Flanging area - tabulation of Z1, Z2

10 Alarm 1st Step Alarm 2nd Step MAX.REACH ENVELOPE for 12" Arms MAX.REACH ENVELOPE for 8" Arms MAX.REACH ENVELOPE moves by # 3 m Alarm 1st Step moves by # 3 m 3 m 3 m Figure 5: Consequence on arm length of approximation of polygonal flanging area by a rectangular parallelepiped

11 The area within the alarm 1st step envelope is considered as a safe working area (mooring pattern in intact condition). On the drawings arms operating envelop provided by the arm vendors, we ask specifically not to indicate the so called drifting area anymore because it does not add any useful information and again this wording is extremely misleading. Only the flanging area (polygonal shape) and the alarm 1 st step envelop (spherical shape) are shown in this safe working area. In any case of detection, the alarm 1st step triggers the ESD sequence. Finally, and within the safe area, a pre-alarm is provided, should the ship flange moves beyond 0.75 m (typical figure, can be adjusted) in any direction. This alarm is only audible and visible on the PMS screen. It is provided solely for the operator information and might indicate that during mooring operation the ship has deviated significantly from her spotting line or that a re-tensioning of the mooring lines might be necessary. We further ask in our specification for arm vendors that a re-initialization of the PMS (memory of initial flanges coordinates) is done once the ship is all fast (and not each time one arm is finished to be connected). This is to face a situation where a ship would have slightly moved from her spotting line during the mooring operation. C. Beyond the safe area: a) setting up of alarm 2nd step: Beyond the alarm 1st step envelope, it is assumed to be on a conservative side that the ship faces a complete failure of her mooring pattern (catastrophic scenario called ship break-out). It is further assumed that the ESD sequence (including PERC valves closure time) has been optimized for weather exposed loading terminals where high drifting speeds are expected, as a compromise between arm length (mechanical envelop) and the pressure surge calculations. Generally speaking it is actually not recommended in weather exposed locations to use the standard ESD sequence duration of 30 s because this will allow a break-out event to develop in too great drifting speeds and distances: it can be seen through Drifting Model Studies (see PART 3 of the paper) that, during the first minute of drifting, the ship acceleration is constant, which means that the drifting speed increases linearly with the time elapsed from the break-out event and that the drifting distance versus time follows a quadratic curve! The drifting distances shown on Figure 2 (lateral drift D1 and longitudinal drift D'1) at completion of ESD sequence are given by the Drifting Model Studies. They have further to include a safety margin due to PMS accuracy. These drifting distances D1 and D'1 are used to set up the alarm 2nd step envelope. The alarm 2 nd step detection is done both by proximity switches and the PMS sensors backing up each other. It has to be noted that, due to the quadratic curve of the drifting distance versus time after break-out, it would be unsafe to anticipate also the alarm 2nd step using the PMS sensors; therefore the alarm 2nd step, contrarily to the alarm 1st step, is fixed. The fact that a safety margin is taken on the drifting area (see above chapter 2.1.B.: predicted ship lateral and longitudinal excursions are rounded-up) and that the detection of alarm 1st step is anticipated in case of noticeable drifting speed of the ship in the safe working area (see above chapter 2.1.B.), and finally that the drifting scenarios used for the setting up of the alarm 2nd step are conservative (ship break-out), leads to the conclusion that the ESD sequence should be always completed prior to the detection of the alarm 2nd step. However, should the ESD sequence not be completed for any reason at the time the alarm 2nd step is detected, then an additional safety margin is taken between the alarm 2nd step envelope and the arm mechanical envelope as explained in hereunder chapter 3.2. allowing completing the ESD and ERS sequences sequentially prior to reach the arm mechanical limit. The detection of the alarm 2nd step triggers the activation of the ERS sequence (PERC opening).

12 b) arm mechanical envelope: The definition of the arm mechanical envelope (or arm mechanical limit) and therefore of the arm length required for the project is based on two conservative assumptions in order to build further safety margin in the concept: it is assumed that the ERS sequence (i.e. PERC opening) duration is 5 s (in reality it is 2 s or less). rather than using the predicted drifting distances between alarm 2 nd step and arm mechanical envelope, we use the maximum drifting speeds (lateral and longitudinal) predicted at the time of ERS sequence completion after ship break-out, and this in order to calculate the lateral distance D2 and the longitudinal distance D' 2 between the alarm 2nd step and the arm mechanical envelope. The drifting speeds (lateral speed and longitudinal speed) at the end of the ERS sequence are given in the Drifting Speed Studies. They include on the top a safety margin due to PMS accuracy. Finally it has also to be checked that the part of the arm mechanical envelope close to the loading platform is behind the berthing line with compressed fenders. The value of the fender compression to be used is given by the Berthing Model Studies and is rounded up (i.e. we use predicted fender compression figures rather than the standard conventional figure of 200 to 300 mm used by arms vendors). NOTES: It is important to note that the longitudinal and lateral drifting speeds and distances are not the same. The constraints for loading arms design (mechanical limits) are not the same as well for lateral and longitudinal drifts: - The lateral drift (affecting the arm luffing movement) requirements can usually easily be taken into account by adjusting the arm length - For the longitudinal drift at contrary (affecting the arm slewing movement), very little can be done due to the limit 45 slewing angle (arm length has no effect on this limit). This should be a cause of concern in areas with strong currents for example or for ship where starting the main engine by mistake could be a credible scenario. Again, and for loading terminals, a solution to build more safety in the design is to consider a reduction of the ESD sequence duration compared to the conventional value of 30 s (as said before in PART 1, this ESD sequence duration is no more indicated in the OCIMF Marine Loading Arms specification 3 rd Edition, see Ref. [6]). The most important parameters for arms design and alarms setting are not independent (namely the model of the flanging area and the area for ship movements at berth, the ESD and ERS sequences durations, the drifting distances and speeds). Several tries have generally to be done in order to find a good compromise in particular with regard to the arm length required. The Figure 6 shows how such iterative process can be done in order to reduce an unacceptable arm length. When comparing with Figure 1, it can be seen how the above method departs significantly from the traditional one.

13 CHANGE ARM LENGTH FLANGING AREA + ALLOWANCE AREA FOR SHIP MOVEMENT AT BERTH (UNACCURATELY CALLED "DRIFTING AREA") ESD,ERS SEQUENCE DURATION resp. 30s? 2 to 30s? ESTABLISH MINIMUM POSSIBLE ESD, ERS SEQUENCES DURATIONS taking into account loading or discharging port, piping design (surge), ship valve closing time WORKING AREA WITH INTACT MOORING CONDITIONS (INCLUDES SHIP MOVEMENTS IN OPEN SEAS, MAINLY HEAVE, ROLL, SWAY, SURGE) this area has a polygonal contour. At this stage atypical ships which might unnecessary increase the flanging area can be identified ESTABLISH DRIFTING SPEEDS AND DRIFTING DISTANCES BASED ON ACTUALSHIPS CHARACTERISTICS, ACTUAL WEATHER DATA AND VARIOUS DRIFTING SCENARIOS versus time starting from t = 0s (date of break-out event) SETTING-UP ALARM 1ST STEP ENVELOPE (DELIMITS THE SAFE AREA) SETTING UP OF ENVELOPE 2ND STEP AND ARM MECHANICAL LIMIT (ARM LENGTH) FIXED DATA CHECK IF ARM LENGTH REMAINS WITHIN REASONABLE LIMITS if YES, OK if NO: 1st step: check ships with restrictions in order to possibly reduce the flanging area 2nd step: check if the ESD (and ERS) sequences duration can be reduced this will reduce the distance required between alarm 1st step and arm mechanical envelopes 3rd step (after 1st and/or 2nd step): change arm length Figure 6: Recommended approach for combining arms design and alarms setting in case iterations are necessary

14 2.2 ESD AND ERS SEQUENCES AND ACTIVATION FROM PMS SENSORS AND ALARM PROXIMITY SWITCHES a) ESD and ERS sequences The sequences are summarized on Figure 7. The alarm 1 st step detection triggers the ESD sequence where, among other things, pumps stop and ESD valves close; The sequence duration has to be optimized particularly for loading terminals where it has no reason to be set arbitrarily to 30 s. According to the OCIMF Marine Loading Arms specification 3 rd Edition the ESD sequence must include the closing of the ball valves of the PERC (Powered Emergency Release Coupling also called dry break coupling). The exception to this is for the so-called no spill PERC where the upper ball valve closes first due to the mechanical interlock between upper and lower ball valves. Once the ESD sequence is completed and the alarm 2 nd step is detected, then the ERS sequence is trigger and consists simply in the PERC opening. More and more this ERS sequence is followed (and we recommend to have this facility on all liquefied gas terminals) by a hydraulically controlled raise and retract clear sequence of the arms. It has to be noted here that modern arms for LNG are of DCMA type (Double Counterweight Marine Arms, usually DCMA-S, i.e. supported) while RCMA type arms (Rotating Counterweight Marine Arms) are generally used for other liquefied gases and LPG in particular because they are smaller in diameter and do not require sophisticated DCMA technology. This difference of technologies has to be highlighted when we go through arms balancing and emergency raise and retract sequence, knowing that just before PERC opens arms are in free wheel condition; therefore understanding and analysis of the natural movement of the arm at time of disconnection is important: DCMA arms are easier to balance and due to the fact that the outboard arm is balanced by an independent counterweight, the natural movement of the outer arm after disconnection (at flange or at PERC) is to retract simultaneously upwards and backwards, hence avoiding the risk for the outer arm to hit ship manifold at the time of the disconnection. RCMA arms are different due to the single counterweight. The apex angle (vertical angle between inner arm and outer arm) remains constant after arm disconnection (at outer flange or at PERC) with the natural tendency for the outer arm to hit the ship manifold at the time of the disconnection. Therefore the raise and retract clear mechanism has to be design more carefully for these arms and the absence of such mechanism is not recommended. b) ESD and ERS sequences activation from PMS sensors and proximity switches on arms The logic resulting from the methodology described in above chapter 2.1 is given in Figure 8 showing the PMS CPU (Central Processing Unit), the PLC (Programmable Logic Controller) and the electronic control panel. It illustrates also how the PMS sensors and the proximity switches back-up each other. There has been some discussion on the reliability of PMS sensor in comparison with the reliability of proximity switches, going sometimes up to the decision not to install the PMS and rely only on proximity switches. We are in the opinion that this is a wrong debate because it goes nowhere to compare a piece of a system with another piece of a system. What is important is the way the whole system is conceived from sensors to a multiplexer through signal cables, then connection to PMS board through ILS circuit, then connection to the PLC and finally display of the information on the electric control panel, and what is provided for PMS failure detection. The system of failure detection is called the PMS auto-check and consists of continuous interrogation of values given by the sensors (three sensors per arm), which are compared each other. If the values are not coherent, the wrong sensor information is refused by the micro-processor and an error message is given to the operator (additional information in order to localize the defect is also provided such as default of a PMS card, default on multiplexer, default of sensor on arm, etc.). There is therefore no reason to think that a failure will not be detected. Should a failure being detected on a particular sensor, the alarm 1 st step and the alarm 2 nd step will still be detected by the proximity switch and the ESD or ERS triggered because the alarm detection by PMS sensors or by potentiometric sensors back-up each other in our concept.

15 ESD ERS 30 SEC? < 5 SEC STOP SHIP PUMPS CLOSE SHIP ESD VALVE CLOSE ARM MOV VALVE OPEN ARM VENT VALVE ETC. T.D. CLOSE UPPER PERC VALVE IN 5S (a) & CLOSE LOWER PERC VALVE (b) FOLLOWED BY PERC OPENING, TOTAL SEQUENCE IN < 5 SEC T.D. = TIME DELAY TO BE ADJUSTED DEPENDING ON SURGE PRESSURE CALCULATION THEN AUTOMATIC ARMS RETRACTION CLEAR FROM DRIFTING SHIP ARMS LOCKING NOTE (a): THIS IS FOR THE ENHANCED "NO SPILL" PERC. SHOULD CONVENTIONAL PERC TO BE SELECTED, THEN BOTH PERC VALVES CLOSE SIMULTANEOUSLY IN THAT SEQUENCE. NOTE (b): THIS IS FOR THE ENHANCED "NO SPILL" PERC ONLY. SHOULD CONVENTIONAL PERC TO BE SELECTED, THEN PERC OPENING SEQUENCE CAN BE REDUCED < 2S. Figure 7: Sequences of ESD and ERS activation

16 * * * INBOARD OUTBOARD SLEWING PERC UNLOCKED * POTENTIOMETRIC SENSORS ON ARM * * ALARM 1 ALARM 2 PROXIMITY SWITCHES ON ARM INTRINSICALLY SAFE TRANSMITTER PMS PMS IN OPERATION PMS SENSORS OK * 2ND ALARM * 1ST ALARM * PMS IN OPERATION PMS SENSORS FAILURE * ELECTRIC CONTROL PANEL (OUTPUT SIGNALS) AND AND AND PROXIMITY SWITCHES FAILURE * OR OR PERC UNLOCKED * AND AND PLC PRE-ALARM PRE-ALARM ESD ERS * PER ARM Figure 8: ESD and ERS sequences activation from PMS and alarms proximity switches

17 PART 3: COMPUTER PROGRAMS USED TO MODELIZE SHIP BEHAVIOR AT BERTH AND SHIP DRIFTING AFTER BREAK-OUT The use of computer programs (often completed by physical model tests) is necessary in order to setup properly and accurately the alarms 1 st step and 2 nd step on loading arms. More generally these models allow to understand as well the arm swivels solicitations (with regard to angular amplitudes, period and angular speeds) for the ship at berth in a weather exposed area and are a helpful support for swivel design. 3.1 BERTHING MODELS An interesting paper on the subject is given in Ref. [4]. A lot of work has been done in the past by several working groups on good mooring practices, on design of fenders, on wind and current drag coefficients, establishing very useful guidelines. The use of these documents is satisfactory in most of the circumstances in order to mitigate the risk of accident on a ship at berth. However these documents are not well suited for the design of jetties where more sophisticated tools are required, which include the non linearity of the mooring lines (elasticity curve), the non linearity of the fenders (fender characteristic curve) and factors controlling the moored response to wave action like the added inertia and the dumping. Such models should be also able to predict the dynamic response of the ship at berth under wind gusty phenomena and analyze the effect of long swells. It is also essential to identify at a early stage possible high pick tension values, which may result from coupling of main oscillation modes and apply corrective actions like use of tail ropes for example in order to modify one of the natural frequencies of the system. The programs we usually use are ARIANE developed by Bureau Veritas and ELF of France and widely accepted worldwide or DIODORE developed by PRINCIPIA of France or AQUAPLUS developed by SIREHNA of France. Several similar programs exist, developed by various maritime institutes in different countries such as (MARIN, DHI, CETENA, etc.). It is not possible to mention all of them here. They are well known by the engineering companies. Usually these programs make a two-stage analysis by assuming a summation of the cinematic effects resulting from the two components of the real ship movement: One is dedicated to the low frequency response to weather-induced stress (swell, wind, current), traditionally known as drift effects, including the permanent deformations and thrusts of timeaverage slow sway oscillations. The other is dedicated to the response to the frequency of oscillations imposed by the swell, which constitutes a relatively high frequency exciter of the six degrees of freedom around their mean or slowly varying state. The results of the above programs in various sets of environment conditions of wind (including gusty phenomena), irregular waves and current allow to determine the maximum lateral and longitudinal ship excursions at ship manifold (sum of static plus dynamic effects, combining linear and angular components of ship movements) to be used for setting the alarm 1 st step as explained in above chapter 2.1 B. An example of berthing model results is given in Table 1. Simulations were done using the DIODORE model from PRINCIPIA (France) on a recent project of a LNG jetty located at the edge of a natural channel affected by tidal currents, winds and attenuated waves arriving at location after diffraction and refraction in the channel. The jetty orientation is 200 (ship-moored bow out), more or less parallel to the prevailing tidal flow and ebb currents. The same project will be taken as an example for illustrating the results of the drifting model hereunder and the example of arms design and alarms setting given in PART 4. We have chosen this recent example in relatively sheltered conditions to give a general illustration of the method on a representative project with regard to ship size, weather data, etc. Obviously a project in open sea would give much greater arms envelopes and more spectacular results with regard to drifting speeds. The max lateral and longitudinal motions are considered, i.e. the sum of the quasi-static and dynamic components. The small values obtained by the model reflect the particular characteristics of the site: almost no transversal components due to the berth approximately parallel to the tidal current directions and very attenuated waves. On more weather exposed sites these figures would be significantly higher, remembering that the forces on the ship are proportional to the squared values of the wind and current.

18 DESCRIPTION OF MAIN DIMENSIONNING TESTS IN DYNAMIC CONDITIONS CASE MODE LOADING CONDITION CURRENT WIND WAVES CONDITION SPEED (m/s) DIRECTION TO N SPEED (m/s) DIRECTION FROM N Hs (m) Tp (s) DIRECTION FROM N 7 Dynamic Full Ebb Dynamic Full Ebb Dynamic Full Ebb Dynamic Full Ebb Dynamic Ballast Ebb TABLE 1: RESULTS WITH REGARD TO MANIFOLD MOVEMENTS X MOTIONS (m) Y MOTIONS (m) Z MOTIONS (m) DIMENSIO- NING CASE CASE MEAN MAX Rms MEAN MAX Rms MEAN MAX Rms YES (X) YES (Y) NOTE: THE PROGRAM GIVES DIRECTLY THE X, Y, Z MOVEMENTS AT MANIFOLD FLANGE (AND SPEEDS/ACCELERATIONS IF REQUIRED) TOGETHER WITH THE THREE LINEAR AND THREE ANGULAR MOVEMENTS AT CENTER OF GRAVITY OF THE SHIP

19 3.2 DRIFTING MODELS Until now we have used real time navigational programs in order to model the drifting of the ship in the first minute after ship break-out under various environment conditions of wind, irregular waves and current allow to determine the drifting distances, speeds and ship rate of turn after break-out. Lateral, longitudinal and combined drifts can be modeled. We use usually TRAJNAV model developed by SOGREAH/STNMTE of France or other models developed by various maritime institutes in various countries. A very interesting feature of these programs is that a scenario like engine start can be added to the various drifting scenarios. Also interesting is the fact that drifting distances and speeds can be expressed both in rectangular or angular coordinates. We ask for both as illustrated in Figure 9, knowing that the PMS who will detect the ship flange position can use both types of coordinates. The same Figure 9 gives an example of ship drifting after break-out during five minutes (one plot every 30 s). In practice we are looking only at what happens during the first minute for arms alarms set-up. In order to illustrate the drifting laws, we have chosen an example of pure lateral drift simulated with the real time navigational model SIMON from CETENA (Italy). The LNG jetty studied is part of the studies carried-out on a recent project. The jetty is located at the edge of a natural channel affected by tidal currents and has an orientation 200 of (ship moored bow out), more or less parallel to the prevailing tidal flow and ebb currents. The test conditions were: Wind 20 m/s from 290 Current 0.25 m/s to 110 (reversing tide condition) No waves, no engine, no rudder. The results of the simulation are given on Figure 10. These conditions where the dimensioning case for the lateral drift in this particular project. In the first minute of drifting after break-out, the drifting speed is approximately proportional to the wind speed squared and current speed squared. The curve drifting speed versus time after break-out is linear, which means that the drifting acceleration B is constant and that the drifting distance is a quadratic function of the time elapsed after ship break-out t (after two or three minutes after break-out the drifting speed stabilizes and do not increase anymore). In the example given on Figure 10: Drifting speed: V = A + B.t, with A = cm/s (should be theoretically 0) and B = cm/s 2 (A and B are obtained by linear regression) Drifting distance: D = 1/2 B.t 2 Similar simulations are done for longitudinal drift and combined lateral/longitudinal drifts. Results are summarized in Table 2 attached (negative signs for drifting distances and speeds are due to convention direction of axes).. The following important observations can be made: Average drifting speeds: a) Lateral drift (TEST 3): 8.50 cm/s for the ESD sequence and cm/s for the ESD sequence. b) Longitudinal drift without starting engine (TEST 2): 6.50 cm/s for the ESD sequence and cm/s for the ERS sequence. c) Combined lateral and longitudinal drifts (TEST 1): 8.00 cm/s for the ESD sequence and cm/s for the ERS sequence. Notes: longitudinal drift with starting the engine and ebb current (TEST 4): cm/s for the ESD sequence and cm/s for the ERS sequence. average drifting speeds are given for indicative purpose only because, if used as such in the design, they would be too misleading.