The permutation entropy and its applications on fire tests data

TL;DR

This paper explores the use of permutation entropy to analyze temperature time series data from full-scale fire experiments, aiming to identify order/disorder characteristics and improve understanding of fire dynamics.

Contribution

It introduces a novel application of permutation entropy to fire test data and discusses methods for embedding dimension selection and analysis of fire phenomena.

Findings

Permutation entropy effectively characterizes fire temperature data.

The study identifies optimal parameters for entropy analysis in fire data.

Proposes new research directions for entropy-based fire analysis.

Abstract

Based on the data gained from a full-scale experiment, the order/disorder characteristics of the compartment fire temperatures are analyzed. Among the known permutation/encoding type entropies used to analyze time series, we look for those that fit better in the fire phenomena. The literature in its major part does not focus on time series with data collected during full-scale fire experiments, therefore we do not only perform our analysis and report the results, but also discuss methods, algorithms, the novelty of our entropic approach and details behind the scene. The embedding dimension selection in the complexity evaluation is also discussed. Finally, more research directions are proposed.