PRESENTED BY K. G. M. PRATLEY EX CABLE GLAND SEALS HUMAN LIVES DEPEND ON THEM AND RECENT RESEARCH IS ASSISTING

Transcription

1 Page 1 of 6 PRESENTED BY K. G. M. PRATLEY EX CABLE GLAND SEALS HUMAN LIVES DEPEND ON THEM AND RECENT RESEARCH IS ASSISTING Abstract: In potentially explosive environments human lives depend upon the integrity of Cable Gland seals. This is particularly true in the case of Flameproof Cable Glands. We have attempted to quantify explosive gas transport mechanisms and we describe some interesting new research involving cutting edge barrier materials. The primary function of a cable gland is to mechanically anchor a cable to an apparatus. This prevents tensile loads from being transmitted to the conductor terminations. An Adjustable Cable Gland and its mechanism for gripping armouring. Modern cable glands generally integrate additional functions. These might, for example, include the provision of earthing continuity or the maintenance of the apparatus IP (Ingress Protection) integrity. In particular, in the case of cable glands for use with Explosion Protected (EP) apparatus, additional functions are always required. The most obvious is that the cable gland and indeed the cable itself, should not compromise the specific Ex- integrity of the apparatus. Any failure to integrate such additional EP functionality is a matter for attention as it potentially endangers both property and life. In the case of Flameproof Equipment (Ex d), entry devices must at least be capable of sealing against the gas pressures induced by conflagrations (or detonations) inside the apparatus. Any hot gas which is allowed to escape must do so only via a flamepath. To achieve this, the majority of Ex d cable glands employ relatively easy-to-fit elastomeric seals. The performance of elastomeric seals and in particular the longevity of the cable-seal interface integrity can vary from excellent to marginal. This depends primarily upon gland and seal design and elastomer material selection. Pratley favours (in the case of armoured cable) a design whereby seals are stretched over the cable bedding. Besides isolating the sealing function from the armour gripping function, this arrangement ensures a long term seal which is independent of unavoidable cable bedding creep. Coupled with careful elastomer selection, this arrangement also leads to reduced internal stress and hence significantly reduced compression set in the rubber. There is the added benefit of relatively easy inspection.

2 Page 2 of 6 Stretch-on Elastomeric Ex d gland seals compensate for creep and are also inspectible. Well designed glands with well designed elastomeric seals are both useful and relatively easy to fit. Notwithstanding this, the drafters of IEC ( Electrical installations in hazardous areas other than mines ) have deemed that in certain special and rather uncommon circumstances compound filled glands (barrier glands) are preferred Specifically, if a cable enters directly into an Ex d enclosure which has a volume of more than 2 l or if gasses requiring Group IIC apparatus are present then a compound filled gland is required. This applies to apparatus with an internal source of ignition and installed somewhere other than in a mine. Such conditions may occur on oil platforms and the like. A compound filled barrier gland. Logic would suggest that the concerns of the drafters of IEC presumably relate to: 1. The possible propagation of hot blast gasses down the interstices of a cable and thereby causing ignition at a remotely connected site. 2. The possibility of explosive gas mixtures migrating down the interstices of a cable connecting an EP (Explosion Protected) apparatus in a classified area to a non EP apparatus in an unclassified area where sources of ignition are present. We examine now each of these concerns in turn. To facilitate analysis, consider a 4-core x 6 mm 2 cable with no filling between the cores. The core insulation diameter is approximately 6 mm.

3 Page 3 of 6 Void Area 4-Core 6 mm 2 cable. The void area is given by 2 (1 - /4) and computes to 7.6 mm 2. This area is approximately the same area as a 3 mm bore, round capillary. For simplicity of calculation, consider a short, say 1 m length of 3 mm capillary tube to be a crude analogue of the interstitial channel in a 1 m length of unfilled cable. We can then use Poiseuille s law of capillary flow to obtain a rough estimate of the steady state flow of combustion products exiting the tube. It computes to 0.03m 3 /sec. This would likely represent the absolute flow limit as steady state conditions cannot be achieved during the short duration conflagration pressure cycle. 4 p R ( V ( ) ) where 8 L V = the volumetric gas flow rate. p = the pressure differential, R = the capillary radius; = the coefficient of viscosity; L = the capillary length; The viscosity has been corrected for an assumed temperature of 700 C Such analysis is patently crude and includes a plethora of simplifying assumptions. It does however indicate the order of magnitude and probably (due to assumptions of laminar flow, combustion gas composition, viscosity and others) provides a rather overstated result. Importantly, note that the analysis assumes that in this instance, the capillary is initially filled with air and not with an explosive mixture. A pressure vs time curve for a conflagration in a closed 2 l vessel exhibits a pressure rise and fall within a relatively short period of time (typically in the order of 0,03 to 0,06 sec). i This is roughly the time it would take for a sonic pulse to traverse the tube. According to Bartknecht (1981) ii most gasses when ignited under atmospheric conditions peak at around 7-8 Bar and this is relatively independent of the vessel size. 1 Considering our 3 mm capillary, it is reasonable to postulate that for most of the duration of the pressure pulse, any flow will also be choked as the required pressure (upstream: downstream) ratio for choked flow in a constriction is in the region of 1.7 to 1.9 for almost all of the commonly encountered ignitable gasses. Additionally, notwithstanding the thermal insulating properties of PVC, the surface area (available to conduct heat) to volume ratio in a capillary is inversely proportional to its diameter. The above, together with the more significant reported experience from test houses (who have witnessed many actual tests) leads us to conclude that: Ignition of a remote gaseous mixture due to conflagration transmission down an air filled interstitial channel in an unfilled cable is unlikely unless the equivalent diameter of the interstitial void is relatively large. To accurately quantify this statement would require a more rigorous analysis. Hopefully someone will take up the challenge. 1 The rate of pressure rise however varies considerably between different gasses and vessel sizes.

4 Page 4 of 6 We turn now to the question of gaseous diffusion. Considering again our 1 m long 3 mm capillary tube analogue, the diffusion rate down the bore of the tube can be calculated. If, for example, a constant supply of methane arbitrarily at its UEL (upper explosion limit) mixture of around 15% (V/V) is assumed at one end of our capillary and pure air is at the other end, then the initial mass flow rate due to diffusion can be calculated using Ficks first law. DA( Ce Co) m Where m = Mass flow rate g/sec D = Diffusion coefficient cm 2 /sec C = Gas concentration g/cm 3 A = Area cm 2 = length cm. For CH 4, D = x 10-2 cm 2 /s. For H 2, D = 76 x 10-2 cm 2 /s. At atmospheric pressure and room temperature a diffusion of about 140 mg of CH 4 per day appears possible. The mass flow rate is inversely proportional to cable length and it will also reduce proportionally if the remote end concentration (partial pressure) starts to build up. The LEL of CH 4 is however just 5% (V/V) which for say a 2 l vessel requires just 71 mg CH 4 in air at atmospheric pressure. Again, these calculations only serve to indicate the significance or otherwise of diffusion as a gas transport mechanism in a cable. Accurate diffusion rate estimation, especially into a closed receiving vessel with rising concentration requires a more rigorous analysis. Nevertheless, we can conclude that gas diffusion along the interstitial capillary paths of a cable is a possibility and the risk may need to be considered especially in the case of H 2 gas which, due to its low molecular weight, diffuses very much faster than many other gasses. At 8pm on Monday 12 January 2004, the South East Australian Sea Gas System Control Centre received an alarm. A water heater installed at Torrens Island had failed. Investigation revealed that a gas explosion had taken place inside a heater control panel located outside of any classified area. Further investigation revealed that an I/P transducer housing was found to be pressurised from the instrument gas system. This gas had migrated along the cable and significantly, had also passed through an epoxy putty filled barrier gland and into the heater control panel. Poor epoxy and poor workmanship were cited as causal factors. The poor workmanship aspect may quite possibly have been related to the difficulty and complexity of the required fitting procedure. Returning to Fick s Law, we consider the case of Hydrogen. We arbitrarily assume H 2 gas at its (UEL) concentration. A 25 mm long epoxy barrier with just a 1 μ gap around half of the four 6 mm diameter cores of our aforementioned cable leads to a calculated gas diffusion rate through this barrier of 0.6 mg/day. 0.6 mg H 2 mixed with air in a volume of about 150 ml would reach the (LEL) for H 2 downstream of such a barrier gland.

5 Page 5 of 6 During the early 1960 s the Pratley Company developed the world s first epoxy putty (now well known as Pratley Putty). A short time later it developed what was presumably the first epoxy putty based barrier gland. This design was superseded by elastomeric seal type glands. Being involved in both adhesives research as well as EP equipment research places the company in a unique position to investigate barrier gland sealing materials. Simple formulations of epoxy putty of the type generally used on barrier glands will pack around the cores of a cable and it will cure. However, it will not stick to the PVC insulation of cable cores. This leads to the very uncomfortable conclusion that invariably, there exists a small yet finite capillary path between the barrier material and the cable cores. Adhesion science explains why this situation exists. Surface wetting is a necessary pre-requisite for chemical adhesion. For wetting to occur, the surface energy of the substrate must be higher than the surface energy (surface tension) of the adhesive. Epoxy resin has a surface energy of approximately 46mJ/m 2 whilst pure PVC has a surface energy of 39 mj/m 2. Plasticised PVC as used for insulation has an even lower energy so surface wetting is thermodynamically not possible. Adhesion to the conductor insulation is clearly difficult to achieve. It nevertheless, in the light of our diffusion calculations, remains a highly desirable goal. Other desirable barrier material attributes are tabulated in table 1 below. Staff at the Pratley Research Facility have been investigating these issues and recent developments are extremely promising. Research Facility where new barrier compound has been developed. They have developed a compound which complies with all of the identified desirable properties. Most importantly, it exhibits exceptional adhesion to plasticized PVC insulation. See Table 1 below.

6 Page 6 of 6 Table 1 PROPERTY EXISTING COMPOUNDS NEW COMPOUND Fast cure time (reduces fitting time/cost) Easy to mix and apply? Fire resistant (during conflagration) Acceptable shelf life? Adhesion to insulation Stability Usable over a wide temperature range Applyable over a wide temperature range Safe to use in industry Its exceptional speed of cure also allows a gland to be put into service just 1½ minutes after installing the barrier. Ultra fast cure compound system showing exceptional adhesion to PVC insulation. Current tests show that for an experimental size number 2 gland, the measured fitting time (to an already prepared cable end) from opening the gland box to having tightened the gland to the apparatus ready for use, is a mere 4½ minutes. Conclusion Most Ex d glanding situations are well served by the use of elastomeric seal type flameproof cable glands. These should be sourced from reputable suppliers and stretch-on seals are preferred. This is especially important when considering long term safety. Barrier glands are seldom necessary but when they are required caution and care is recommended. Current technology does not guarantee an impermeable diffusion barrier especially in the case of low molecular weight gasses such as hydrogen. This is explained by adhesion science surface energies and the consequent difficulty of achieving an adequate bond to cable core insulating materials. Recent research has resulted in a new barrier material which adheres extremely well to cable core insulating materials and is therefore able to guarantee a diffusion barrier. An added advantage is the vastly reduced installation time. This work possibly charts the probable future of barrier gland technology. i Gexcon, Gas Explosion Handbook. ii Bartknecht W. (1981), Explosions Course Prevention Protection, Springer Verlag, Berlin.