摘要
为了提升污水排放总磷浓度的预测精度,保证污水达到排放标准,本文提出一种基于自组织神经网络的污水总磷浓度预测模型。通过建立基于粒子群优化的自组织神经网络,运用浮点数编码方式为粒子编码,并对其适应度进行评估,采用近邻粒子群方法对权重失真标准实施优化,对污水总磷数据实现精准挖掘;在自组织神经网络的基础上引入小波分析,利用进水指标和出水总磷浓度的映射关联构建小波神经网络预测模型。仿真结果表明,上述模型具备较高的收敛速率,能够及时有效对污水总磷浓度进行精准预测,为污水处理的稳定运行提供理论支撑和技术保障。
In order to improve the prediction accuracy for total phosphorus concentration in sewage discharge and ensure that the sewage meets the emission standards,this article puts forward a model of predicting total phosphorus concentration in sewage based on self-organizing neural network.Through the self-organizing neural network based on particle swarm optimization,the float-encoding method was used to encode the particle,and then its adaptability was evaluated.Moreover,the nearest neighbor particle swarm was used to optimize the weight distortion standard and accurately mine the total phosphorus data of sewage.Based on the self-organizing neural network,the wavelet analysis was adopted.Finally,the mapping relation between the inflow index and the total phosphorus concentration of effluent water was used to build the prediction model of wavelet neural network.Simulation results show that the designed model has high convergence rate,so it can predict the total phosphorus concentration of sewage seasonably and effectively.In addition,this model can provide theoretical support and technical support for the stable operation of sewage treatment.
作者
赵岚
刘瑛
李俊叶
ZHAO Lan;LIU Ying;LI Jun-ye(Institute of Technology,East China Jiaotong University,Nanchang Jiangxi 330000,China)
出处
《计算机仿真》
北大核心
2020年第11期465-469,共5页
Computer Simulation
基金
江西省教育厅科技课题(GJJ181486)
江西省2018年教育规划课题(18YB403)。
关键词
自组织神经网络
总磷浓度
预测模型
粒子群优化
Self-organizing neural network
Total phosphorus(TP)concentration
Prediction model
Particle swarm optimization(PSO)
