用BP神经网络模型预测纳米二氧化钛光催化降解硝基甲苯类废水  

BP Neural Network in Predicting the Nano-Titanium Dioxide Photocatalytic Degradation of Nitrotoluene Wastewater

作　　者：尹艳华[1] 王春芳[1] 晏明杨 

机构地区：[1]北京理工大学化工与环境学院,北京100081 [2]中国民用航空西南地区空中交通管理局技术保障中心,四川成都610202

出　　处：《火炸药学报》2011年第3期86-90,共5页Chinese Journal of Explosives & Propellants

摘　　要：基于光催化降解硝基甲苯类废水的实验数据,采用反向传播(BP)神经网络训练并建立了硝基甲苯类废水处理过程的神经网络模型。用训练好的神经网络模型模拟光催化降解硝基甲苯废水过程,模型硝基甲苯浓度的模拟输出与实验数据的相关度为0.998。用神经网络模型对此光催化降解硝基甲苯废水过程进行预测,得到硝基甲苯浓度的预测数据与实验数据的相关度为0.976。采用神经网络模型预测得到光催化降解硝基甲苯废水过程的最佳降解条件为:TiO2的质量浓度为0.10 g/L、H2O2的体积浓度为0.10 mL/L、pH值为3。The Back Propagation（BP）network was trained with the data from the photocatalytic degaration nitrotoluene wastewater treatment experiment and a network model was built for this nitrotolune wastewater treatment process.The nitrotoluene photocatalytic process was stimulated with the trained network model.The correlation degree between network analog output and experimental data of nitrotoluene concentration is 0.998.The photocatalytic degradation nitrotoluence wastewater treatment was predicted with this neural network model.And the correlation degree between network predictive data of and experimental data of nitrotoluene concentration is 0.976.BP neural network model was established to predict the optimal reaction conditions of the photocatalytic degradation process,and the determined conditions are as follows： the mass concentration of TiO2 is 0.10 g/L,the concentration of H2O2 is 0.10 mL/L,and the value of pH is 3.

关 键 词：应用化学 BP神经网络 光催化降解 硝基甲苯 废水处理 

分 类 号：TJ55[兵器科学与技术—军事化学与烟火技术] X703[环境科学与工程—环境工程]