Measuring the flash points (FP) of multicomponent organic mixtures, which are widely used in the chemical industry, is a time-consuming and laborious process. Although numerous FP prediction models have been proposed, most of them focus on predicting the FP of pure components and binary mixtures, and few studies have addressed ternary or higher component mixtures. In this study, we measured the FPs of 341 compositions from six ternary aqueous–organic mixtures and developed a FP prediction model based on the quantitative structure–property relationship (QSPR) principle using the obtained data. Molecular descriptors for each component were generated using Dragon software, and the mixture descriptors were computed based on logarithmic mixing rules. During the model construction process, a multistage approach utilizing random forest regression (RFR) was employed to develop a highly precise model for predicting the FP values of ternary miscible organic mixtures. The results demonstrate that our proposed model is robust and predictive, achieving coefficient of determination (), root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) values of 0.9641, 3.4086, 2.1961, and 0.6955%, respectively, on the test set. Furthermore, we compared our proposed model with other existing ternary mixture FP prediction methods from the literature to confirm its superiority. The results obtained her ein can be applied to assessments of fire and explosion hazards in the chemical industry and engineering.
扫码关注我们
求助内容:
应助结果提醒方式:
