Journal Article Reports on Conducting an Uncertainty Analysis of Pyrolysis Models
A new peer-reviewed journal article on uncertainty quantification and propagation using pyrolysis models has been published in the Fire Safety Journal. The paper was authored by FSRI research engineer Mark McKinnon as part of the Materials and Products Database program.
Fire safety engineers, material and product designers, and building code authorities often rely on pyrolysis models to inform decisions and designs that improve the fire safety of the built environment. While accurate models are essential, there’s also an inherent degree of uncertainty to them. The type of material, the influence of external conditions, and a range of other characteristics make it challenging to correctly numerically represent how fires grow and the effect of materials on fire dynamics.
Understanding the uncertainty in measured data and how that uncertainty propagates to model predictions helps increase the reliability of these models and confidence in the conclusions drawn from the models. This study aimed to present and validate a method to quickly quantify the uncertainty of the representation of solid phase reactions in fire models. An additional objective was to understand the factors that have the greatest effect on predicting how heat breaks down materials into flammable gasses that contribute to fire growth.
Understanding Uncertainty: Methods and Results
To quantify the uncertainty in a material burning model, McKinnon used the generalized polynomial chaos expansion method, which has not been used previously in fire modeling, and systematically burned materials in a series of experiments. He discovered that the experimental data fell within the uncertainty bounds in the model predictions, which indicated the model accurately described the material through the burning process. McKinnon also conducted a sensitivity analysis to understand which factors contributed most to uncertainty in the predictions over the course of the simulated experiments. The results of the sensitivity analysis emphasized the need to reduce uncertainty in the temperature dependence of the material burning rate and the model representation of heat absorbed by the material for higher confidence in model predictions.
“Understanding how uncertainty propagates through fire models is vital for our confidence in these models. Our work introduces techniques to measure uncertainty in models, showcasing the ability of these techniques to efficiently handle input and output uncertainties. This work also allows us to conduct analyses that pinpoint exactly where we should focus our attention to improve the fidelity of our predictions.”
-Mark McKinnon, Research Engineer, FSRI
The methods presented in this work allow researchers to quickly quantify the uncertainty of pyrolysis models. This approach establishes the foundation for conducting similar analyses where many factors influence the results, such as in more complicated scenarios and larger-scale models.
This work was funded in part by the National Institute of Justice (US) through grant 2019-DU-BX-0018.
About Fire Safety Journal:
Fire Safety Journal is the leading publication dealing with all aspects of fire safety engineering. Its scope is purposefully wide, as it is deemed important to encourage papers from all sources within this multidisciplinary subject, thus providing a forum for its further development as a distinct engineering discipline. This is an essential step towards gaining a status equal to that enjoyed by the other engineering disciplines.
