The separation of ethyl acetate-methanol-water via the extractive distillation process as a chemical process safety simulation experiment is proposed in this paper by combining the chemical process simulation and intrinsic safety. The suitable entrainer with good separation performance is determined via the thermodynamic topology theory. The triple-column extractive distillation process model is constructed for separating the ternary azeotropic mixture ethyl acetate-methanol-water based on the process simulation software Aspen Plus. The developed process is optimized via the multi-objective particle swarm optimization algorithm with economic and safety performances as the objective functions. This simulation training experiment integrates professional courses such as chemical thermodynamics, chemical principles and thermal risk assessment of chemical processes. At the same time, it combines professional theoretical knowledge, cutting-edge disciplinary knowledge and engineering practice application to strengthen basic knowledge, motivate students' research interest and comprehensively improves their innovation ability.

This study aims to significantly improve existing quantitative structure-property relationship (QSPR) models for predicting the octanol-water partition coefficient (K_OW). This is because accurate predictions of K_OW are crucial for assessing the environmental behavior and bioaccumulation potential of chemicals. Previous models have reported determination coefficient (R²) values between 0.9451 and 0.9681, and this research seeks to exceed these benchmarks. Three machine learning (ML) models are explored, i.e., feed-forward neural networks (FNN), extreme gradient boosting (XGBoost), and random forest (RF). Using a dataset of 14,610 solvents (14,580 after data cleaning) and 21 molecular descriptors derived from SMILES representations, we rigorously evaluate these models based on R², mean absolute error (MAE), root mean squared error (RMSE), and mean relative error (MRE). Notably, the best model developed, the XGBoost-based QSPR, demonstrated exceptional performance, exhibiting an impressive R² value of 0.9772, surpassing benchmarks set by prior research models. Additionally, shapley additive explanation (SHAP) analysis is also employed for model interpretation, and it is revealed that the top five influential input features include SMR_VSA8, SMR_VSA3, Kappa2, HeavyAtomCount, and fr_furan. This study not only sets a new benchmark for K_OW prediction accuracy but also enhances the interpretability of QSPR models.