The simplest method for reliable prediction of autoignition temperature of organic hydroxyl compounds to assess their process safety in industrial applications

Abstract

The knowledge of autoignition temperature (AIT) is essential for the manufacture, processing, handling, transport, and storage of combustible materials but its measurement is very laborious. This work introduces a simple model for estimating AIT of different classes of organic hydroxyl compounds containing the other polar groups such as –O–, –S–, –CN, –C(=O)O–, –NH2 and > NH. The new correlation can predict the AIT of any organic compounds including the hydroxyl functional group through its molecular structure. It is based on the number of hydrogen atoms and two correcting functions under certain conditions. Various types of statistical parameters including internal and external validations are done to assess the reliability of the new model. The high reliability of the new model is compared with the outputs of two of the best available predictive methods. It is indicated that the values of Mean Absolute Percent Error (MAPE) of the new model for both training and test sets, corresponding to 109 organic hydroxyl compounds, is 3.90 K, which is much less than two comparative methods, i.e. 7.81 and 13.70 K.