响应曲面法和人工神经网络模型对电催化氧化磺胺甲恶唑条件的优化

Modeling and Optimization of Sulfamethoxazole Treatment by Electrocatalytic Oxidationusing Response Surface Methodology and Artificial Neural Network Approaches

为获得电催化氧化磺胺甲恶唑的最佳实验条件,通过中心组合设计(CCD)设计实验,建立响应曲面(RSM)和人工神经网络(ANN)两种模型,并评价了两种模型的准确性和预测能力。在此基础上分别通过响应曲面法和遗传算法对所得模型进行优化。结果表明:RSM和ANN的均方误差(MSE)、平均绝对误差(MAE)和决定系数(R^2)分别为2.62、1.13、0.976和0.59、0.62、0.994,说明ANN模型比RSM模型具有更高拟合度、精度和预测能力。响应曲面法和遗传算法的优化结果与实验所得结果的相对误差分别为1.48%和0.74%,说明遗传算法具有更强的优化能力。本研究结果可为优化电催化氧化抗生素废水的条件提供参考。

In order to obtain the best experimental conditions for electrocatalytic oxidation of sulfamethoxazole,two models of response surface(RSM)and artificial neural network(ANN)were established through CCD design experiments.In addition,the accuracy and predictability of the two models were evaluated.On this basis,the resulting model was optimized by response surface methodology and genetic algorithm.The results show that the MSE,MAE,and R^2 of RSM and ANN are 2.62,1.13,0.976,and 0.59,0.62,and 0.994,respectively,indicating that the ANN model has higherdegreeof fitting,precision,and predictability than the RSM model.The relative errors of the experimental results obtained by the response surface method and genetic algorithm are 1.48%and 0.74%,respectively,indicating that the genetic algorithm has a stronger optimization ability.The results of this study can provide reference for optimizing electrocatalytic oxidation of antibiotic wastewater.