活性污泥过程溶解氧浓度预测  

Prediction of Dissolved Oxygen Concentration in Activated Sludge Process

作　　者：胡瑛汉 HU Yinghan(College of Electrical and Information Engineering,Lanzhou University of Technology,Lanzhou 730050)

机构地区：[1]兰州理工大学电气工程与信息工程学院,兰州730050

出　　处：《计算机与数字工程》2024年第4期1228-1234,共7页Computer & Digital Engineering

摘　　要：溶解氧浓度是活性污泥法污水处理过程中的重要过程参数。准确的溶解氧浓度测量是保证出水水质达标以及节能生产的前提,对此,提出了一种基于优化神经网络的溶解氧浓度软测量模型。首先,将自适应步长策略和学习策略引入标准的麻雀搜索算法,提高了算法的搜索能力和搜索精度。其次,为了提高溶解氧浓度的预测精度和效率,采用改进麻雀搜索算法用于优化BP神经网络模型参数,并以自动获取的最佳参数组合构建溶解氧软测量模型。最后,利用该软测量模型对国际基准仿真模型BSM1和实际污水处理过程的溶解氧浓度进行预测。仿真结果表明:与BP、RBF、ELM、JS-BP和PSO-BP等预测模型相比,ISSA-BP预测模型的预测精度更高,收敛速度更快,具备更好的实践应用价值。Dissolved oxygen concentration is an important process parameter in activated sludge wastewater treatment.Accu-rate dissolved oxygen concentration measurement is the premise to ensure effluent quality to meet the standards and energy-saving production.Therefore,a soft sensor model of dissolved oxygen concentration based on an optimized neural network is proposed.Firstly,the adaptive step size strategy and learning strategy are introduced into the standard sparrow search algorithm to improve the search capability and search accuracy of the algorithm.Secondly,to improve the prediction accuracy and efficiency of dissolved oxy-gen,the improved sparrow search algorithm(ISSA)is used to optimize the BP neural network parameters,and the soft sensor mod-el of dissolved oxygen is constructed with the best combination of automatically selected parameters.Finally,the soft sensor model is used to predict the dissolved oxygen concentration of the benchmark simulation model No.1(BSM1)and the actual wastewater treatment process.The simulation results show that the ISSA-BP prediction model has higher prediction accuracy and faster conver-gence compared with BP,RBF,ELM,JS-BP and PSO-BP prediction models,and it is more suitable for practical application.

关 键 词：污水处理 溶解氧预测 改进麻雀搜索算法 神经网络 软测量 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]