Testing Water Mist Systems Against Large Fires in Tunnels: Integrating Test Data with CFD Simulations

The high cost of conducting large, full-scale fire tests for the evaluation of suppression systems in tunnels tends to limit both the extent of the instrumentation provided and the number of tests that are conducted. Because of the variability of the large fires, performance criteria based on single point measurements derived from experience with smaller test fires were not reliable indicators of performance. Yet decisions about the acceptability of suppression systems must be based on the limited amount of performance information available. A means was sought to reduce the reliance on single point instrumentation readings, and to augment the value of the limited amount of test data by integrating the field testing with CFD modeling. In this study a computational fluid dynamic (CFD) model was used to simulate a series of full-scale fire tests of water mist systems conducted in 2006 in a highway test tunnel. The NIST Fire Dynamics Simulator version 4 (FDS4) was used to simulate five of the tunnel fire tests. The task was to confirm that the simulations could achieve a reasonable degree of agreement with the conditions measured in the tests. The model could then be used to evaluate the performance of the water mist system over a broader range of performance indicators than were measured. This paper illustrates what is unique about very large fire tests and presents highlights of the modeling. The level of agreement between simulation and test results is demonstrated for one test. Agreement was deemed to be good enough to establish confidence in applying the model to examine the conditions that would occur with an unsuppressed fire, which had not been tested. CFD modeling can be used to improve the understanding of the performance of the suppression system, and to augment the value of the test results. A second, complementary paper has been submitted to the SFPE Journal of Fire Protection Engineering to provide more detailed information about the FDS4 modeling than can be covered in this paper.