Arson investigators rely on computer models to help predict how materials burn or degrade in a pyrolytic, or oxygen-free, environment that is common in intense residential fires. Those models depend on the input of accurate materials properties to allow the simulation to approximate what would happen in an actual fire.
Researchers at Jensen Hughes, a fire safety company based in Baltimore, MD., supported by a National Institute of Justice grant, developed a framework that uses multi-objective optimization, which simultaneously determines several parameters that describe the physical and chemical changes to a material when it burns. “The research produced a validated methodology that practitioners can follow to determine material properties,” the final report said. It also provided a benchmark data set for future work.
In the report, “Determination of Material Property Input Data for Fire Modeling,” researcher Brian Lattimer noted that the results, “will not only provide a more standardized approach for material property determination for forensic applications, but also for the general fire protection engineering community.” Lattimer, a professor of mechanical engineering at Virginia Tech, and his research team, had five objectives:
- Measure all the fire-related properties for several different materials, including a plastic polymer (Polymethyl methacrylate), medium-density fiberboard, white pine, cardboard, strand board, and glass fiber-reinforced vinyl.
- Identify the properties that most strongly affect simulation results by performing sensitivity analyses on three different pyrolysis models.
- Formulate tests to quantify model properties through optimization. (Optimization is a numerical method to determine the maximum or minimum value of an equation and is constrained by the need to arrive at a realistic solution).
- Design benchmark tests to determine the predictive capabilities of the models.
- Conduct pyrolysis model benchmarking using the new methodology.
The focus of the research was to measure the key fire-related properties to better understand not only the properties of the material, but also the difficulty and cost in measuring those properties. The researchers examined the density, thermal diffusivity, heat transmission, heat of combustion, char oxidation, specific heat, porosity, permeability, and thermal expansion coefficient for the range of materials in the study.
These complex tests, each specific to the individual materials, resulted in the development of an overall methodology that is expected to “improve the accuracy of fire modeling when predictions need to include material burning and its feedback with the predicted environment,” the report said. “The results of this project address the need for the forensic discipline of Fire and Arson Investigations for more adequate materials property data inputs for accurate computer fire models.”
Jensen Hughes used the new method to estimate the material properties used in railcars for the US Department of Transportation, the report said, and several commercial projects have used it to improve their Fluid Dynamics Simulator fire models.
“It is expected that this new method will improve the accuracy of fire modeling in general, especially when solid material burning is involved,” the researchers concluded.
About This Article
The research in this article was funded by National Institute of Justice award 2016-DN-BX-0185, awarded to Jensen Hughes, Baltimore, MD. This article is based on the grantee report “Determination of Material Property Input Data for Fire Modeling” by Brian Lattimer.