QSPR METHOD FOR PREDICTING AUTOIGNITION TEMPERATURE OF ORGANIC COMPOUND

摘要

The present invention relates to a quantitative structure-property relationships (QSPR) method of predicting the autoignition temperature of organic compound, and more specifically, to a QSPR method of predicting the autoignition temperature of organic compound, which includes the steps of: inputting a 2-dimensional chemical structure of the compound after collecting an organic compound from which autoignition temperature experimental data is obtainable among a plurality of compounds; selecting a common molecule descriptor for describing physicochemical characteristics of the compound from the 2-dimensional chemical structure of the compound; and establishing a model predicting the autoignition temperature. According to the QSPR method of predicting the autoignition temperature of the organic compound, of the present invention, an autoignition temperature value is theoretically and highly accurately predicted using only 2-dimensional molecular structure data, thereby reducing costs and time for experiments, and estimating the autoignition temperature value of molecules from which is impossible to obtain the autoignition temperature value. Therefore, according to the present invention, stability is remarkably increased in the chemistry industry. Further, various analyses on features and reaction mechanisms with respect to the autoignition temperature of the organic compound are proceeded by using the molecule descriptor used for the estimation model. Statistical functions using various molecule descriptors as independent variables are expressed in a machine learning scheme used for the model, thus a relation between a substance to be estimated and a variable is configured out in combination, through interrelated expression between the independent variables.;COPYRIGHT KIPO 2016