不同神经网络描述的乙酸-水共沸精馏分离流程多目标设计优化OA

The recovery of acetic acid from industrial waste acid in acetic acid production is of great importance for the sustainable development of enterprises.This study used azeotropic distillation technology to separate acetic acid-water solutions.A simulation model of the process was first established,and the corresponding simulation results were used to construct a bootstrap aggregated neural network(BANN)and a random subspace stacked neural network(RSSNN)for classification and prediction of the acetic acid-water azeotropic separation process.Based on this,a multi-objective optimization model for the acetic acid-water separation process was established using BANN and RSSNN.The NSGA-Ⅱ algorithm was then applied to solve the multi-objective optimization problem.A case study based on the treatment of waste acid from an actual acetic acid industry was used to demonstrate the effectiveness of the proposed method.The results of the case study showed that BANN and RSSNN exhibited good classification and prediction capabilities in the acetic acid-water separation process.However,RSSNN was more accurate in predicting economic and environmental performance indicators.The multi-objective optimization results obtained by BANN and RSSNN were similar to those obtained by the simulation model,and the speed of multi-objective optimization based on BANN and RSSNN was respectively 7.3 times and 9.8 times faster than that based on the simulation model.