高级检索

    成睿, 孟广源, 殷瑶, 郑雨诺, 张芯婉, 李童, 陈鹏, 张乐华. 神经网络BPNN模型机器学习方法应用于电化学去除氨氮过程的预测与优化[J]. 华东理工大学学报(自然科学版), 2023, 49(2): 202-210. DOI: 10.14135/j.cnki.1006-3080.20220123003
    引用本文: 成睿, 孟广源, 殷瑶, 郑雨诺, 张芯婉, 李童, 陈鹏, 张乐华. 神经网络BPNN模型机器学习方法应用于电化学去除氨氮过程的预测与优化[J]. 华东理工大学学报(自然科学版), 2023, 49(2): 202-210. DOI: 10.14135/j.cnki.1006-3080.20220123003
    CHENG Rui, MENG Guangyuan, YIN Yao, ZHENG Yunuo, ZHANG Xinwan, LI Tong, CHEN Peng, ZHANG Lehua. BPNN Algorithm Model Application to Prediction and Optimization of Electrochemical Ammonia Removal[J]. Journal of East China University of Science and Technology, 2023, 49(2): 202-210. DOI: 10.14135/j.cnki.1006-3080.20220123003
    Citation: CHENG Rui, MENG Guangyuan, YIN Yao, ZHENG Yunuo, ZHANG Xinwan, LI Tong, CHEN Peng, ZHANG Lehua. BPNN Algorithm Model Application to Prediction and Optimization of Electrochemical Ammonia Removal[J]. Journal of East China University of Science and Technology, 2023, 49(2): 202-210. DOI: 10.14135/j.cnki.1006-3080.20220123003

    神经网络BPNN模型机器学习方法应用于电化学去除氨氮过程的预测与优化

    BPNN Algorithm Model Application to Prediction and Optimization of Electrochemical Ammonia Removal

    • 摘要: 利用反向传播神经网络(Back Propagation Neural Network,BPNN)建立氨氮去除效果预测模型和智能控制策略。模型由具有BPNN模型的预测模块和控制模块组成。首先,采用4层隐藏层(每层60个神经元)和负反馈调节机制开发BPNN算法,优化模型并预测氨氮去除率。参数分析及响应面模型对比结果表明所提出的BPNN模型具有更好的决定系数(0.9580)。根据水质变化和确定的氨氮去除率目标,通过BPNN模型获得电化学过程中电流智能调控策略,该智能控制策略减少了水质波动对氨氮去除的负面影响,并使能耗降低38%。

       

      Abstract: The electrochemical method is proved to be an effective method to remove ammonia. However, the research on the energy consumption control is neglected. This paper uses artificial intelligence and back propagation neural network (BPNN) to establish the ammonia removal rate prediction model and intelligent control strategy. The model consists of a prediction module and a control module with BPNN algorithm model. Firstly, four hidden layers (per 60 neurons) and a negative feedback adjustment mechanism are used to develop the BPNN algorithm, optimize the model and predict the ammonia removal rate. Through parameter analysis and comparison of response surface models, the BPNN model proposed in this paper has better coefficient of determination and lower mean square error. According to the water quality changes and the determined target of ammonia removal rate, the current control strategy in the electrochemical can be obtained through the BPNN model. Finally, the proposed intelligent control strategy is applied to the electrochemical system for ammonia removal, which can reduce the negative impact of water quality changes and reduce energy consumption by 38% compared with the original strategy.

       

    /

    返回文章
    返回