Background
49 CFR 179.18 specifies that, to be qualified for use, a thermal protection system of a rail tank car has to pass a 30 min torch fire test. Fabricating the 1.22 x 1.22-m specimens needed to conduct a series of torch fire tests on different formulations of a novel thermal protection system is very expensive and presents major logistical challenges because of the quantity of material needed. Consequently, design by trial and error of a novel system that meets the requirements of the torch fire test is inefficient and cost prohibitive. The purpose of the targeted IR&D project was to provide a proof-of-concept (TRL3) for a cost-effective computational-experimental method that will greatly facilitate the design and qualification of innovative thermal protection systems for rail tank cars.
Approach
The method that was developed in the IR&D project consists of the following components:
A small-scale torch fire test that is used to obtain the thermal properties of the constituent materials of a new thermal protection system, needed for input into a numerical model of the full-scale torch fire test described in the CFR
A validated numerical model of the small-scale torch fire test coupled with an optimization algorithm to estimate the aforementioned thermal properties
A validated numerical model of the full-scale torch fire test that is used to perform virtual qualification tests and determine the likelihood of that a novel system will pass the test.
The experimental part of the project involved four types of tests:
Full-scale torch fire calibrations with a steel plate according to the procedure described in 49 CFR 179, Appendix B.
Full-scale torch fire tests on specimens of calcium silicate board, according to the procedure described in 49 CFR 179, Appendix B.
Small-scale torch fire calibrations with the same type of steel plate used in the full-scale torch fire calibrations.
Small-scale torch fire tests on specimens of the same calcium silicate board tested in full-scale.
The results of the calibrations in conjunction with one-dimensional inverse heat transfer calculations were used to quantify the thermal exposure conditions in the full- and small-scale tests. Subsequently, the same heat conduction model was used to calculate the unexposed surface temperatures in the tests on specimens of the calcium silicate board.
Accomplishments
Reasonable agreement was obtained between the measured and calculated unexposed surface temperatures, which provides a proof-of-concept for the proposed method. However, the heat transfer model needs to be extended to three dimensions so that the non-uniformity of the thermal exposure, lateral heat transfer and heat losses at the edges of the specimen can be accounted for in the calculations.
Furthermore, additional research is needed to extend and validate the three-dimensional approach to thermal protection systems that intumesce, ablate or exhibit specific behavior that needs to be accounted for in the heat transfer calculations.