Modeling of autoignition temperature of organic energetic compounds using hybrid intelligent method

Abstract

Abstract Autoignition temperature (AIT) plays a significant role while characterizing the potential hazard of energetic chemical compounds and occurrence of fire disasters can be easily managed and controlled through adequate knowledge of autoignition temperature of the compounds. However, the experimental determination of autoignition temperature is laborious and consumes appreciable time and resources. This present work addresses the challenges using hybrid support vector regression (SVR) and gravitational search algorithm (GSA) for precise modeling of autoignition temperature of organic energetic compounds using only molecular weight, as well as the number of hydrogen, oxygen and carbon atoms as descriptors. Apart from the superior performance of the proposed hybrid model (SVR-GSA) as compared with the existing models, the absence of specific functional groups of the energetic compound as descriptor further eases the applicability of the model. On the basis of mean absolute error between the estimated autoignition temperatures and the experimentally measured values, the proposed SVR-GSA model outperforms Mohammad et al and Chen et al model with performance improvement of 37.34% and 79.05%, respectively. Implementation of the proposed model would definitely ease autoignition temperature determination of energetic compounds and ultimately reduces the potential risk associated with these compounds while handling.