Evaluation Of Operation Status Of Constructed Wetlands Based On AHP And BP Neural Network

As a new sewage treatment technology,constructed wetlands need to be further improved.It is necessary to further study the operating conditions and monitoring data of artificial wetlands in different seasons,different regions and different processing technologies,and to give a correct evaluation of the operating conditions of constructed wetlands.It is beneficial to optimize the operating parameters of constructed wetlands,improve the management level of constructed wetlands,improve the evaluation of wetland operation status,and provide guidance and suggestions for the construction,management and evaluation of artificial wetland in the future.In this paper,4 typical constructed wetlands in Shandong province were selected for field investigation and water sample monitoring.The water quality of the 4 constructed wetlands and the changes in the water quality were analyzed.The AHP method was used to evaluate the operation state of the constructed wetland.The results of the evaluation of the operation state of the constructed wetland were evaluated by the AHP method,and the evaluation model of the operation state of the constructed wetland based on the BP neural network was constructed.Type.This article mainly carried out the following aspects of the work:(1)analysis of the water quality of four constructed wetlands,the removal rate of pollutants in artificial wetland,and the water quality of part of the surface water grade III standard.The average removal rates of COD,NH3-N,TN and TP were 38.87%,47.25%,66.68% and 75.66%respectively.The average removal rates of COD,NH3-N,TN and TP in the artificial wetland of the pig dragon river were 20.04%,50.95%,66.56% and 55.76% respectively.The average removal rates of COD,NH3-N,TN and TP in Xingjia wetland were 22.91%,32.67%,14.81%and 5.86%,respectively.The average removal rates of COD,NH3-N,TN and TP in the constructed wetland of Xuecheng wastewater treatment plant were 20.72%,28.07%,38.12% and38.21%,respectively.(2)analysis of the purification effect of four artificial wetlands in different seasons on theremoval of pollutants.Season has little effect on the removal of COD.The removal efficiency of TN and NH3-N is greatly influenced by season.With the increase of temperature,the removal rate of TN and NH3-N decreases.The removal rate of TP varies little with season.The removal of TP depends mainly on the adsorption,precipitation and exchange of fillers.(3)establish the evaluation index system of the operation state of constructed wetland,establish hierarchical structure of every index of wetland based on AHP method,establish three basic frame,target layer A,element layer B,index layer C.Then,according to the principle of index design,4 indexes of B "wetland index,operation index,water quality index and management index" are established,and 16 indexes,such as "pool design,packing design,plant design",are determined in the index layer,and so on.The weight of evaluation index was calculated by AHP method to determine the weight of B of factor layer relative to target layer A:water quality index > operation index > wetland index > management index;index layer C relative to target layer A: effluent COD> hydraulic retention time > packing design,plant design >TN,TP> water quality > hydraulic load > plant management >DO>pH> temperature >blocking Management > online monitoring.(4)comprehensive evaluation of the operation state of three constructed wetlands.Through the monitoring of 2 times per month for three constructed wetlands,the qualitative indexes are scored according to the scoring standard.According to the laboratory monitoring data,the quantitative indexes are obtained from four artificial wetland data samples,and then the data samples are pre processed to get the index of each index and AHP.The weight of each wetland is multiplied to get the comprehensive score of every wetland operation effect every month.(5)establish the BP neural network model of the AHP based wetland operation evaluation and establish the three layer BP neural network structure.The number of the input layer neural nodes is taken as the number of the input layer neural nodes,which are the main influencing factors of the operation state of the constructed wetland "pool body design,packing design,plant design,etc.",and the number of nodes of the operation state of the constructed wetland each month is the node number of the output layer based on the AHP algorithm.According to the literature and repeated calculation,the number of neural nodes in the hidden layer is determined to be 20.The evaluation model of artificial wetland operation status is established as 16-20-1.Then the training and prediction samples are constructed with the data of table 3.15."Trainlm" isselected as the training function and the learning rate is 0.05.Thus the model of the evaluation of the operation state of the constructed wetland based on BP neural network is established.The prediction score is divided into five grades: "excellent,good,medium,bad,extreme."(6)the BP neural network model was used for the 12 month data input of the PORON River into the artificial wetland in Hukou,and the comprehensive evaluation and evaluation grade of the operation status of the constructed wetland were obtained each month.