Our research includes quantifying the exposures from fire environments onto materials and systems. In this work, we use experimental and computational methods to determine the heat transfer and combustion products during extreme events. Some examples includ
- Probabilistic fire exposures
- Fires in nuclear power plants
- Jet fires
- Scaling fire environments
Probabilistic Fire Exposures
The actual type of fire that may occur in or around a structure is typically not known, making the selection of the worst-case fire for design subjective and difficult. An alternative approach is to create a method to generate probabilistic distributions of exposures from all potential types of fires in an application for engineers to select the appropriate exposure for their risk assessments.
Our group is part of a Society of Fire Protection Engineers (SFPE) standards development team that is using Monte-Carlo analysis to generate exposures from a range of different types of fires to quantify the distribution of exposures for specific base geometries. Methods are being developed to use these detailed distributions to inform how distributions can efficiently be created for other similar geometries.
Fires in Nuclear Power Plants
Probabilistic risk assessments in nuclear power plants frequently use historic design fire data to determine the impact of a fire plant systems. However, some of these data are old, inaccurate, and may not include all the necessary information.
We are also performing research to quantify the data needed to accurately predict fire exposures in nuclear power plant risk assessments. In this research, we are using statistical and machine learning techniques to identify the most important parameters that affect fire conditions and cross-correlating that with existing data to determine future test needs.
Local fire exposures from pressurized leaks can produce severe exposures that represent the most challenging environment for materials. In addition, these high velocity gas jets can be used to represent large-scale fire behavior (e.g., wildfires) that are difficult to generate in a laboratory setting.
We have characterized these exposures using advanced measurement techniques to provide exposures for a range of applications including material fire resistance, vegetation burning, and firebrand generation.
Scaling Fire Environments
Conducting full-scale experiments is costly and time consuming. As a result, being able to scale down the geometry while maintaining the same large-scale response is important to be able to efficiently evaluate material system behavior.
Our research group has developed scaling law for various fire applications. Most recently, we have been exploring the simultaneous scaling of thermal and structural response. The competing scaling laws for each individual behavior requires new solutions based on dimensionless parameters from the governing equations and boundary conditions.