基于NSGA-II结合BP网络算法的乙醇制备C4烯烃优化模型

Optimization Model for Preparation of C4 Olefins from Ethanol Based on NSGA-II Combined with BP Network Algorithm

乙醇偶合制备C4烯烃工艺广泛应用于化工生产,探寻实现乙醇转化率与烯烃收率最大化的催化剂组合与温度条件具有重要意义。根据反应过程性能数据,分别运用多元回归与BP网络建立反应条件与烯烃收率的回归预测模型,结果表明BP网络模型对烯烃收率的预测值有较高的精度。建立多目标优化模型,基于BP网络良好的预测能力,采用NSGA-II算法进一步优化可得满足目标的催化剂与温度方案,此时乙醇转化率与烯烃收率分别为64.67%,35.921%。随后进一步分析比较NSGA-II算法和经典遗传算法分别优化BP网络模型获得的烯烃收率预测误差,验证了所用算法的精确性与有效性。

The ethanol coupling process for the production of C4 olefins is widely used in chemical production,and it is important to explore the catalyst combinations and temperature conditions to maximise the conversion of ethanol and olefin yield.Based on the performance data of the reaction process,a regression prediction model of the reaction conditions and olefin yield was developed using multiple regression and BP network respectively.The results showed that the back propagation(BP)network model has high accuracy in predicting the olefin yield.A multi-objective optimisation model was established,and based on the good predictive capability of the BP network,the non-dominated sorting genetic algorithm-II(NSGA-II)was used to further optimise the catalyst and temperature solutions to meet the objectives,which resulted in ethanol conversion and olefin yield of 64.67%and 35.921%respectively.Further analysis was then carried out to compare the prediction errors of olefin yield obtained by the NSGA-II and the classical genetic algorithm for optimising the BP network model respectively,which verified the accuracy and effectiveness of the algorithm used.