Fire-endurance tests of dual-wall fiberglass-resin composite pipe

Fire-endurance tests of dual-wall fiberglass-resin composite pipe

PII: S1359-8368(96)00050-9 ELSEVIER Composites Part B 28B (1997) 295-299 © 1997 Elsevier Science Limited Printed in Great Britain. All rights reserv...

526KB Sizes 2 Downloads 21 Views

PII: S1359-8368(96)00050-9


Composites Part B 28B (1997) 295-299 © 1997 Elsevier Science Limited Printed in Great Britain. All rights reserved 1359-8368/97/$17.00

Fire-endurance tests of dual-wall fiberglass-resin composite pipe Alfred N. Montestruc a, Michael A. Stubblefield a, Su-Seng Pang a'*, Vic A. Cundy b and Richard H. Lea c aDepartment of Mechanical Engineering, Louisiana State University, Baton Rouge, LA 70803, USA bDepartrnent of Mechanical Engineering, Montana State University, Bozeman, MT 59715, USA CSpecialty Plastics, Inc., Baton Rouge, LA 70810, USA (Received 31 July 1995; accepted 9 March 1996) A new method for characterizing fire resistance of composite dual-wall pipes is reported. The technique involves testing a bench-scale pipe sample by exposure to a flame. Detailed temperature profiles and history are obtained. By using this methodology, we have studied the fire resistance of several dual-wall pipe samples. In particular, we have examined the effects of insulation thickness and the use of an intumescent coating. Results show that both insulation thickness and intumescent coating are effective in improving fire resistance. © 1997 Elsevier Science Limited (Keywords: A. polymer-matrix composites (PMCs); D. thermal analysis; fire resistance)

INTRODUCTION There has been tremendous interest in the use of composite pipe systems in the marine and petrochemical industries, especially on open-type offshore oil platforms. Composite pipes are one-fifth the weight of steel pipe, half the cost of copper-nickel pipe, have excellent corrosion resistance and, when properly designed and installed, can offer maintenance-free performance for the life of the platform. Because of these characteristics combined with the demand for strong, lightweight materials on deep-water tension leg platforms, composite piping systems are becoming the 'system of choice' for fire-water systems in the offshore oil and gas industry. Considerable efforts are underway to develop fire standards and/or codes t o assist designers in their efforts to meet the objectives of safety engineers and certifying authorities, while establishing realistic guidelines for the performance of all materials, not merely composite pipe, in essential fire-prevention and exposed systems. Shell Offshore, USA recently installed, with the approval of the US Coast Guard, over 1000 ft of composite pipe for the firewater ringmain system on their MARS Tension Leg Platform in the Gulf of Mexico. With the assistance of the US Navy and Louisiana State University, fire-endurance tests were completed and smoke toxicity levels were performed to satisfy the requirements of the US Coast Guard. Of primary concern regarding the use of composite pipe in these systems is the ability of the pipe to withstand a

* To whom correspondence should be addressed

sustained fire in the case of a hydrocarbon fire and/or an explosion in the case of a 'jet fire.' A major concern regarding the use of composite pipes is to characterize fire resistance. In the past, there have been some efforts in this area; however, previous work has focused on overall system failure modes 1-3. Guidelines which address such concerns are given by the United Kingdom Offshore Operators Association (UKOOA), the International Maritime Organization (IMO) and the American Society for Testing and Materials (ASTM; F-117395) 4"5. Typically, composite pipe lengths are exposed to a flame environment and the time-to-failure (leakage) is measured 6. In general, these tests do not follow a rigorous scientific methodology and they provided no details of the temperature distribution within the system. Generalization of results is, therefore, not easily accomplished. The research reported here is intended to provide: • a more rigorous systematic methodology to characterize fire resistance of selected composite pipe samples; and • a characterization of the effectiveness of various dualwall configurations.


Currently accepted methods~standards Testing methods have been established which address the performance and requirements of plastic pipe and pipe systems 7-I°. Of particular interest are the guidelines given


Fire endurance tests: A. N. Montestruc et al. by the UKOOA, the IMO and the ASTM. The ASTM F1173-95 and International Maritime Organization guidelines are equivalent in requirements. The ASTM F-1173-95 guidelines, developed for offshore platforms in United States waters, were approved in 1995 and developed from the IMO. ASTM F-1173-95, entitled 'Standard Specification for Thermosetting Resin Fiberglass Pipe and Fittings to be Used for Marine Applications', is the only standard to test the fire-endurance requirements established by IMO and UKOOA for composite pipe. It was adopted from the recommendations of the International Maritime Organization Resolution A.753(18), dated 4 November 1993, for shipboard applications. The US Coast Guard is the United States' representative to the IMO and is familiar with this fire test method for commercial shipboard piping. The furnace dry test for Level 1 composite pipe used in named essential services in the engine room and in some other areas of a ship dictates that a pipe withstand, for 60 min, a temperature ramp of 2012°C. This test does not require that a pipe joint be tested. The United Kingdom Offshore Operators Association has developed performance parameters for fire-endurance qualification of composite pipe. UKOOA has designated a fire classification code for a typical fire-water deluge system as 3.2.2/120-(4.9). This classification addresses the performance of this system in an open ventilated area which may be exposed to a hydrocarbon fire and which may be initially empty for a minimum of 1 rain, but becomes water-filled very soon after the fire is detected. The piping system must also be able to maintain a minimum of fluid loss and endure exposure to fire for 2 h. The system tested should be in direct contact with the flames. The air temperatures are to be measured directly adjacent to the test specimen. Test may be performed with either open (e.g. hydrocarbon pool fires) or closed (e.g. furnace) testing facilities. In the current testing methods, a composite pipe section is placed in a flame. The inside of the pipe may be dry or it may be filled with water. Failure is determined by monitoring the internal pressure for a given period of time. The test rigs used in this method are not laboratory scale; rather, they are more pilot scale with pipe lengths varying up to several feet in length. Fire resistance can be characterized by several methods 11, one of which is used by the US Coast Guard. The Coast Guard rates fire resistance according to three levels defined as:

for intelligent designs. Therefore, we propose the following methodology which addresses the deficiencies in the present techniques.

Proposed methodology The dual-wall pipe design was chosen for testing since, in a companion smoke/toxicity study 12, we have shown that this may be an optimal configuration when considering both combustion and strength. Other methods and studies have been done which discuss the importance of smoke and toxicity studies in evaluating fire performance 13-15. Additionally, the dual-wall pipe has greater design flexibility to meet fire-resistance requirements. We have developed a bench-scale experimental technique that supplements the current test methods; however, this new technique is of a more scientifically rigorous nature. This technique not only allows direct comparison of empirical data (high degree of reproducibility), but it also provides a base of data from which a more fundamental analysis can be made. A typical cross-section of the dual-wall pipe system is shown in Figure 1. Our work was limited to glass-fiberreinforced composite pipe (both inner and outer pipes) comprising E-glass and vinyl ester resins. A fire-resistant polyphosphazene foam of varying thickness was used in the annular gap between the pipes. Discussions with the manufacturers led us to believe that this foam would provide good insulating characteristics even if exposed to direct flames. In addition, an intumescent fire-resistant coating was used during some tests. An intumescent coating

Crossection A-A


Pipe s

stant foam

Outer pipe

Burner and flame

• Level 1: at least 1 h without loss of integrity in the dry condition; • Level 2: at least 30 min without loss of integrity in the dry condition; and • Level 3: at least 30 min without loss of integrity in the wet condition. While these tests provide specific information for different pipe systems, they do not provide enough detail needed for rigorous design calculations. Specifically, they lack the detailed temperature profiles and histories needed


TC# =Thermocouple number • =Thermocouple location *** Drawing is not to scale Figure 1

Lay-out of thermocouples for the pipe flame tests

Fire endurance tests: A. N. Montestruc et al.



Pipe test sectioL ]

TC2 .... x--- TC4

1 !


.... +--- TC6 .~ 300 ...... i ~ 2

~ ~

"~ - - - -

Threaded bolt

| Thermocot p l ~ e~qeevl ~ ' ' ~ Figure 2

1/8" Stee!plate


Fire resistance results--dual-wall pipe Effect of intumescent coating.

In this experiment, two dual-wall samples with 1 in foam were tested. One sample was covered with the intumescent coating and the other was not. The thermocouple arrangement for these tests is shown in Figure 1. These tests were conducted with foil inserts as discussed above. The initial test was performed with the intumescent coating. Thermocouples 3 and 7 failed shortly after the beginning of the experiment. As a consequence these data are neglected. We shall first discuss the data from the outer

i i " ~----:-


~ 20o 100 ,,,I


i I,

O b.,

End-plate configurations

is a paint-like coating that foams when it is heated, thus forming an additional layer of insulation. The desired design criterion used in these tests was to maintain the outside temperature of the inner pipe below the glass distortion temperature (approximately 400°F). Each dual-wall pipe sample (1 ft long, 4 in inner pipe diameter, 0.242 in inner wall thickness and 0.095 in outer wall thickness) was subjected to a flame as shown in Figure 1 for a period of 1 h. This is consistent with the US Coast Guard standards mentioned previously. In order to obtain the desired temperature profiles, several thermocouples (type K) were embedded in the pipes as shown in Figure 1. A standard laboratory Meeker burner was used to support the flames. Propane was fed at a constant rate of 12 standard cubic feet per hour (SCFH) to ensure a constant heat-release rate. The sample was supported at a height of 2 in over the burner top. In initial tests, the ends of the pipes were covered with aluminum foil to decrease convective effects. However, as the dual-wall pipe was heated, the fire-resistant foam in the annular region expanded in an uncontrollable manner out of the ends of the pipes. Therefore, these foil inserts were replaced by end-plates (shown in Figure 2) which not only reduced convective effects, but also eliminated the problem of uncontrollable foam expansion. This device was constructed by using two metal plates and an 18 in long threaded bolt. The steel plates are placed on both ends of the pipe sample to prevent the expansion of the insulation and to reduce convection heat transfer in the center of the pipe. A thermocouple sleeve was constructed from a 6 in section of conduit and press fitted into a hole cut in one of the plates.

i 1

Figure 3




1500 2000 2500 Time (sec)


3000 3500 4000

Coated pipe surface temperatures, 1 in insulation






400" ~300"

~ 200



i 500

Figure 4

Tc5 i ~c6 _i • To7 i 1000

1500 2 0 0 0 2500 3000 3500 4000 Time (sec)

Uncoated pipe surface temperatures, 1 in insulation

surface of the inner pipe (thermocouples 2, 4, 5, and 6); the inner surface of the outer shell (thermocouple 1) will be discussed separately. The temperature history on the outside of the inner pipe is shown in Figure 3. As shown in the figure, none of the locations on the outer wall of the inner pipe reached 400°F. This implies that the inner pipe would not fail during the 1 h of heating. This was also verified by a post-experiment examination of the pipes. As expected, location 2 exhibited the highest temperature, location 6 was next, and positions 4 and 5 exhibited the lowest temperature. Thermocouples 4 and 5 exhibited a similar temperature history, indicating the symmetry of the flame with respect to the pipe. Flame temperatures between 1500 and 1800°F were measured. The uncoated pipe was then tested in a similar manner. The resin in the uncoated outer pipe ignited within a minute after the application of flame. This pipe burned with heavy smoking for approximately 12 min after ignition. Intermittent flames continued to arise for an additional 3 min. Figure 4 shows the temperatures measured on the outside of the inner pipe (thermocouples 2, 4, 5, 6, and 7). As expected, the temperature of the inner pipe increased much faster than that of the coated pipe. Also the temperatures in the uncoated sample were consistently above those of the coated sample. In fact, there were times when the experimental design criterion (maximum inner pipe temperature of 400°F) was not met. This clearly shows the value of the intumescent coating.


Fire endurance tests: A. N. Montestruc 1000 . . . . 800

i ....




--~i/11~ t/ /


i !



"'-'~"'" X : - - +'"



,, l,,,,

1t--t \


Crossection A-A

TCl-Coated Mean pipe temp coated . _ . L _ _ TC 1-Uncoated i Mean pipe temp uncoated i

Pipe ~nmnl~

nt foam

600' ~i i



.r" --

. ....


" y ' . .....

)uter wall Burner and flame

. . , . . _ . . , _ _ , _ , _ _ . , , . - -, , - _ . - ,l










200 -





~ I ....

~ I ....

~ I ....

I ....

! J ....






i I

.... 3500 4000

Time (sec)

Figure 5

The effects of an intumescent coating

We note further that while the uncoated pipe burned vigorously for approximately 12 min, the coated pipe was never observed to ignite. Upon exposure to the flame, the intumescent coating bubbles and forms not only an additional heat barrier, but also an effective flame retardant. Figure 5 shows the temperature on the inner surface of the outer pipe (temperature 1 of Figure 1) and the average temperature of the outer surface of the inner pipe (an average of temperatures 2, 4, 5, 6 and/or 7) for both coated and uncoated pipes. Consider the coated case first. The figure shows that the outer pipe temperature exceeded 600°F during the initial exposure to the flame for a brief period and then dropped into the 400 to 500°F range. This behavior may result from rapid initial heat transfer from the flame prior to the bubbling of the intumescent coating. Now consider the uncoated case. The outer pipe temperature exceeds 600°F for a much greater time, as expected, owing to ignition and burning of this pipe. Similarly, the average of the inner pipe temperatures is substantially greater than that of the coated pipe and, therefore, the difference between the outer and inner pipes is much less without the coating. It is noteworthy that at no time did any part of the inner pipe exceed the 400°F limit imposed by the experimental design for the coated pipe. These data show the utility of the dualwall pipe system for fire resistance and also the utility of the intumescent coating.

TC# =Thermocouple number • =Thermocouple location *** Drawing is not to scale

Figure 6

Thermocouple lay-out for insulation thickness tests

800 700



.o 500-





~300 :/






i ......


I .......




















: :





100 : ....

i .....



i ....


J . . . . . . . .



i ....

I ....



n ....

3500 4000

Time (sec) TC1


TC2 .--x---- TC3 TC6

Figure 7






TC7 "'" O--.

Uncoated pipe with 2 in foam insulation

Effect of insulation thickness.

The next step was to change the thickness of insulation on the pipe to develop a better understanding of the effects of adding insulation. In addition it was felt that too little was known about the temperature of the shell, so it was decided to move two of the thermocouples from the pipe to the shell. The new positions of the thermocouples are shown in Figure 6. The new end-plate arrangement of Figure 2 was also used in these tests. An uncoated pipe section with 2 in of insulation was tested. The data of this test are shown in Figure 7. Similar to the previous uncoated test (with only 1 in of insulation), this uncoated pipe ignited and burned vigorously for 12 rain. Thermocouple 1 reflects this in that the temperature rises sharply to 750°F and remains at this level for the duration of the experiment. Thermocouples 5 and 7 show a similar trend and they also exhibit the symmetry expected. Thermo-


couples 2, 4, and 6 provide the data for the outer wall of the inner pipe. These data are somewhat inconsistent in that we would expect thermocouple 2 to exhibit the highest temperatures rather than thermocouple 6. It is, however, not too surprising that some asymmetry may result from the uncontrolled burning of the pipe at the beginning of the experiment. In fact, we observed a flame jet in near proximity to thermocouple 6, which may have influenced this temperature. In any event, the important aspect is to note that at no time did any of the monitored temperatures on the inner wall exceed the design criterion of 400°F. In fact, these temperatures never exceeded 300°F, which is approximately 100°F less than those experienced with the 1 in foam insulation. Although the temperatures of the

Fire endurance tests: A. N. Montestruc et al. interior of the 1 in insulated pipe with an intumescent coating were higher than for the 2 in insulated pipe with no intumescent coating, the temperature was significantly lower than that seen in the 1 in insulated pipe with no intumescent coating. In addition, the intumescent-coated pipes never caught fire in this study, while those with no coating did, even when not directly exposed to the flame. Therefore, from a fire-safety point of view, the results obtained for 2 in insulation and no intumescent coating are not considered better than those for the pipe with 1 in insulation and an intumescent coating. Finally, in comparing Figures 3, and 7, we conclude that similar results are obtained from a coated dual-wall pipe having 1 in of insulation and an uncoated dual-wall pipe with 2 in of insulation. It is interesting to note that the temperature gradient across the inner pipe was only 40°F towards the end of the experiment. This shows that the temperature difference across the pipe is quite small. At the end of the test (approximately 1 h), the pipe temperature was still rising in a uniformly linear fashion.

ACKNOWLEDGEMENTS This research was partially funded by the US Navy through research contract number N61533-89-C-0045; by the Louisiana Board of Regents under contract numbers LEQSF (1994-97)-RD-B-02 and LEQSF (1995-98)-RDB-05; Exxon Education Foundation; and Mobil Technology Company. We appreciate the invaluable assistance of the staff and management of Specialty Plastics, Inc. (Baton Rouge, LA). The second author would like to acknowledge the support from the Louisiana Board of Regents' Fellowship.




A systematic fire-resistance methodology that provides a considerably greater rigor than those currently used has been developed. The technique involves testing a benchscale pipe sample by exposure to a flame. Detailed temperature profiles and history are obtained. By using this methodology, we have studied the fire resistance of several dual-wall pipe samples. In particular, the effects of insulation thickness and the use of an intumescent coating were examined. These tests indicate that both the intumescent coating and the polyphosphazene foam insulation can significantly improve the fire resistance of composite pipe. The intumescent coating prevents the outer shell from burning, and therefore, it also prevents the heat transfer that results from direct flame contact. Similarly, the use of insulation can have the same effect. The challenge for the designer will be to find the optimal combination of these parameters in dual-wall piping systems. One possibility might be the following: use a high-strength resin for the inner pipe, covered by a layer of insulation, and a fire-resistant resin for the outer pipe, covered by an intumescent coating. Obviously there are a number of other combinations that will need to be investigated, such as modifying the chemical composition of the composite pipe and, in some cases, adding an additional structural component to the overall pipe. The results of this study will allow for development and verification of analytical models that might be used to make these design decisions.


5 6


8 9 10



13 14 15

Putnam, D.R. and Wilhelmi, G.F. Analysis of shipboard application limits for glass-reinforced epoxy-resin piping material based on emergency fire conditions, Commander, Naval Sea Systems Command, Washington D.C., 1979. International Labour Office 'Occupational Exposure Limits for Airborne Toxic Substances: A Tabular Compilation of Values from Selected Countries,' 2nd Edn, International Labour Office, Geneva, Switzerland, 1977. National Materials Advisory Board (NMAB) "Fire Safety Aspects of Polymeric Materials Vol. 2: Test Methods, Specifications, and Standards', Technomic Publishing Co. Inc., Westport, CT, 1978. American Society For Testing and Materials Designation F-117395, A4, Standard Specification for Thermosetting Resin Fiberglass Pipe and Fittings to be Used for Marine Applications, Philadelphia, PA, 1995. International Maritime Organization Resolution A.753(18) Guidelines for the application of plastic pipes on ships, 1993. Lea, R.H., Stubblefield, M.A. and Pang, S.S. The qualification of advanced composite pipe for use on open type offshore oil platforms. In 'Proc. ASME 15th Int. Conf. on Offshore Mechanics and Arctic Engineering (OMAE)', Materials Engineering, Vol. Ill (eds M. M. Salame et aL) 1996, pp. 265-268. Levin, B.C. The development of a new small-scale smoke toxicity test method and its comparison with real-scale fire tests. Toxicol. Lett. 1992, 64/65, 257-264. Sharma, S.K. Measurement of smoke from fires: the present trends. J. Sci. Ind. Res. 1995, 54, 98-107. Babraukas, V. Effective measurement techniques for heat, smoke and toxic gases. J. Fire Safety 1991, 17(1), 13-26. Briggs, P.J. Smoke generation--developments in international test methods and use of data for selection of materials and products. J. Fire Safe~' 1993, 20(4), 341-351. Pang, S.S., Lea, R.H., Cundy, V.A. and Griffin, S.A. Composite piping systems--phase II (SBIR Program), Final Report, US Navy, Contract No. N61533-89-C-0045, 15 August 1991. Montestruc, A.N., Cundy, V.A., Pang, S.S. and Lea, R.H. Smoke and toxicity tests of fiber glass pipe samples. In 'Design and Analysis of Pressure Vessels, Piping, and Components', ASME PVP Vol. 235 (eds C. Becht IV, et al.), American Society for Mechanical Engineers, New York, 1992, pp. 99-104. Freeman Chemicals Ltd, Polymers Division Low smoke emission composite materials. Aircraft Engng 1988, 60(4-6). Landrock, A.H. 'Handbook of Plastics Flammability and Combustion Toxicology', Noyes Publications, Park Ridge, NJ, 1983. Hirschler, M.M. Heat release from plastic materials. In 'Heat Release in Fires' (eds V. Babraukas and S.J. Grayson), Elsevier Applied Science, London, 1992, Chapter 12(a)