A numerical model is developed to simulate the thermal behavior of a solid particle as it falls through air and interacts with a fuel layer. This scenario is significant to ignition competency in industrial safety and wildland fires. A Global Sensitivity Analysis identifies key input parameters that influence the maximum fuel temperature. A 16-dimensional parameter space, encompassing particle properties, fuel properties, and environmental conditions, is utilized to develop a model suitable for a broad range of particles and fuels. To address the computational limitations of performing large-scale probabilistic risk assessments, an Artificial Neural Network (ANN) surrogate model is developed and trained on over four million simulations to predict the maximum fuel layer thermal response. The ANN achieves a Mean Absolute Percentage Error of 0.094% on a held-out test set and has been validated against available experimental measurements. The surrogate model reduces computational cost by several orders of magnitude compared to conventional numerical models, enabling over 100 million predictions to map the parametric space. This study demonstrates the potential of ANN surrogate modeling to create rapid, probabilistic tools that allow for the evaluation of ignition potential across a wide range of real-world conditions, where traditional ignition tables or models are insufficient.

The activation status of a smoke alarm in a fire becomes a central question in forensic fire investigations when there are injuries or deaths. Forensic techniques have been developed to answer this question through the evaluation of the presence of enhanced soot deposition on specific locations of a smoke alarm. Testing was conducted to evaluate the thresholds where these observational indicators may be obscured or destroyed. Fifty-four smoke alarms were exposed to simulated fire conditions to create enhanced soot deposition observational indicators. A series of post exposure tests were conducted to determine the threshold(s) of thermal damage to determine when those indicators begin to be obscured. All powered smoke alarms presented observational indicators in locations like those identified in previous research, as well as four additional locations for this smoke alarm geometry. Enhanced soot deposition indicators were found to persist in most cases when the alarm was exposed to high temperatures (up to 325°C to 450°C), when the alarms were exposed to high temperatures and dropped to a floor surface, and when the alarm was exposed to carpet contamination.

This chapter provides an introduction for the fundamentals of fire through the tenets of the fire triangle and fire tetrahedron. The four sides of the fire tetrahedron are dissected and examined, focusing on the basic physics and chemistry of fuels, heat, oxidizers, and the chemical chain reaction. Concepts related to Ignition, fire spread, and extinguishment principles are provided for context into the overall environmental impact of fire. An introduction to fluid dynamics, products of combustion, and the principles of heat release rates are also covered in this chapter. Furthermore, a few essential calculations for heat release rates are described. Finally, combustion and fire development within compartments and in the open are discussed in order to establish a fundamental understanding of the physics of interior and exterior fires. Many references are cited throughout this chapter as a starting point for those interested in delving deeper into each of the topics introduced.

The science behind the formation of fire patterns and their ensuing use in the forensic analysis of fire scenes has been questioned since their introduction in the 1940s. This paper provides an overview of a prototype method for determining the area of origin based on fire patterns analysis, named the process for origin determination (POD). The POD is a seven step reasoning process for evaluating fire damage, which starts by identifying the value in further analysis of each surface and compartment of a structure and then procedurally evaluates each surface for use within the overall determination. This paper outlines the application of the POD with test subjects and presents an analysis of the outcomes showing its benefits. To facilitate testing the POD, numerical simulations and physical experiments were employed. The numerical simulations were completed through the use of fire dynamics simulator simulating a single compartment measuring 3.66 m × 3.66 m × 2.44 m with a single ventilation opening. The physical experiments were tests conducted specifically for fire patterns where accuracy rates had been previously identified in the literature. Sixty test subjects participated in the evaluation of thirty-two different origin scenarios. A decrease in variability, which indicates an increase in reliability, was noted in 21 of the 32 scenarios (66%) when participants used the POD. Three accuracy measurements were employed, all three of which illustrated an increase in accuracy when participants used the POD. The accuracy was shown to increase between 50% and 94% when participants used the POD.

Fire investigators have historically relied upon damage as a means to conclude where a fire originated. This review evaluates the historical and current literature on the topic, with a specific emphasis towards the research conducted over the past 80 years related to fire patterns. The concept of fire patterns for this review has been broken into four components that better assist in evaluating their effectiveness in determining an area of origin. The first component evaluated is the ability to assess the varying degree of fire damage along the surfaces of the compartment and contents. Next, the ability to identify clusters of damage was evaluated. Interpretation of the causal factors for the generation of the fire patterns was next appraised. Finally, the availability of processes using fire patterns in determining an area of origin was assessed. This deconstruction of the problem provides a gap analysis of the current processes and identifies areas where future work is needed. A seven step reasoning process for evaluating damage for determining the area of origin, along with a new definition for the term fire pattern is proposed.

A new method to characterize the degree of fire damage to gypsum wallboard is introduced, implemented, and tested to determine the efficacy of its application among novices. The method was evaluated by comparing degree of fire damage assessments of novices with and without the method. Thirty-nine "novice" raters assessed damage to a gypsum wallboard surface, completing 66 ratings, first without the method, and then again using the method. The inter-rater reliability was evaluated for ratings of damage without and with the method. For novice fire investigators rating degree of damage without the aid of the method, ICC(1,2) = 0.277 with 95% CI (0.211, 0.365), and with the method, ICC(2,1) = 0.593 with 95% CI (0.509, 0.684). Results indicate that the raters were more reliable in their analysis of the degree of fire damage when using the method, which support the use of standardized processes to decrease the variability in data collection and interpretation.

Full-scale research burns into the nature of patterns in compartment fires were conducted at the new fire research facility of Eastern Kentucky University. Key questions to be addressed by the research burns were: (a) patterns persistence through flashover and full room involvement, (b) reproducibility of patterns geometry in minimal variable testing methods, and (c) reaffirmation of standard patterns analysis methodologies, such as heat and flame vector analysis, depth of calcination measurement, and truncated cone patterns formation and analysis. As an added value these research burns were designed to test the validity of content of the National Fire Code © component document, NFPA 921 – Guide for Fire and Explosion Investigations chapters on Fire Patterns, and Origin Determination. These tests demonstrate a remarkable resemblance of patterns in minimal variable testing methods. Patterns persistence through flashover and full room involvement was observed, as well as the reproducibility of specific fire patterns, heat and flame vector analysis results, and depth of calcination measurements. In addition, several ancillary fire effects, fire patterns, and post-fire analysis issues were successfully examined.

Fire patterns, as defined by NFPA-921, are the visible or measurable physical effects that remain after a fire. Fire pattern analysis has been a key factor in the determination of the origin and cause of fires for the more than 50 years. In 1985 the Advanced Fire Patterns Research Project (AFPRP) was formed by the National Association of Fire Investigators (NAFI) and the Fire and Safety Engineering Technology Program of Eastern Kentucky University to complete research into the development of fire patterns. Since the creation of the project, considerable work has been undertaken to study fire growth and spread in both full and ¼ scale test burns. Much of the data generated has been utilized to supplement instruction in the area of Fire Pattern Analysis. The purpose of this paper is to describe the results of the most recent full scale test burns that were conducted at Eastern Kentucky University and sponsored by the AFPRP in order to give the reader an understanding of the wider research which is currently underway. For the past 3 years a series of ten full scale tests were conducted in identically constructed, finished, and furnished compartments. In each of the tests (with one exception) all fires progressed through flashover, to full room involvement. These full-scale test burns provided a considerable amount of data concerning fire pattern development and evolution during fire growth and spread. These test burns demonstrated fire pattern persistence and predictability during pre- and post-full room involvement fires. The full scale tests demonstrated that the fire patterns described in current peer- reviewed literature are correct and when used properly can assist in the determination of the origin of a fire. Due to the necessary brevity of this research report (10 pages), only a selected few of the more significant tests are reported here.

This research project is a continuation of a previous study (Hicks, et al., 2006), which analyzed fire patterns produced from wood cribs. The current study continued this fire patterns research by burning ten commercially available polyurethane (PU) foam chairs and documenting the fire patterns. The reproducibility of fire patterns was analyzed to compare one PU foam chair test to the next, as well as in association to those produced by burning wood cribs. Two aspects of fire pattern production were examined. The first aspect focuses on the reproducibility of a conical shaped fire pattern formed on standard gypsum wallboard surfaces. Second, this study analyzed the effects of the upper layer and its role in the production of a conical shaped fire pattern. This study showed that although the time to reach the fire pattern differed, a duplicate fire pattern was reproduced from a similar loss of mass. The results of this study illustrates that similar fuel packages will reproduce a similar conical shaped fire pattern. Additionally, lowering of the upper layer was found to affect the resulting conical shaped fire pattern. A subsequent aspect of this research is the implication that these patterns can be utilized by fire investigators in determining an area of origin.