U.S. flag

An official website of the United States government, Department of Justice.

Improving Fire Pattern Analysis Using Non-Invasive Sensor Technologies

Award Information

Award #
15PNIJ-24-GG-03845-RESS
Funding Category
Competitive Discretionary
Location
Awardee County
Montgomery
Congressional District
Status
Awarded, but not yet accepted
Funding First Awarded
2024
Total funding (to date)
$630,560

Description of original award (Fiscal Year 2024, $630,560)

Fire investigators use the guidelines in NFPA 921 Guide for Fire and Explosion Investigators when investigating fires. Fire investigators create and test hypotheses for the origin of a fire using data gathered from witness accounts, observations, measurements, and knowledge of fire dynamics. Fire effects are the observable or measurable changes in or on a material as a result of a fire, and fire patterns are the physical changes or identifiable shapes formed by a single or group of fire effects. Due to the limited availability of measurable evidence, fire investigators often rely on qualitative observations to determine fire patterns. These observations are typically based on identifying lines of demarcation which are apparent in imagery of the fire scene. However, during a fire, patterns can change and begin to overlap one another, and full room involvement can generate large amounts of smoke and soot that obscure pre-existing patterns. New approaches which can identify objectively lines of demarcation and obscured patterns will improve the accuracy and reliability of post-fire investigations.

The goal of this research is to improve the identification of patterns generated from fire effects using advanced sensing and statistical analysis/machine learning. Preliminary testing has shown that terahertz time domain spectroscopy and ultrasonic pulse velocity measurements have the potential to characterize the extent of thermal damage and soot deposition from fire exposures. The key question that his research seeks to answer is to what extent fire patterns which overlap in the visible spectrum are distinct when viewed with these advanced sensors. The research will take a systematic approach to characterize the fire effects from individual thermal exposure and soot deposition experiments. Models will be developed to use the data from multiple sensors to deconvolve the overlapping patterns generated during testing.

The data, algorithms, and models from this research will be made publicly available for fire investigators to use in future analyses. The findings will be published in a peer-reviewed journal as well as articles in various trade journals, such as the IAAI Fire & Arson Investigator Journal. The researchers will also do technical presentations of this information in conferences, such as International Association of Forensic Sciences (IAFS), International Symposium on Fire Investigation, and Forensic Research & Technology. In addition, as members of the NFPA 921 Technical Committee, the results of this study will be submitted as appropriate proposals for inclusion in future editions of NFPA 921. CA/NCF

Date Created: September 27, 2024