Research On The Application Of PSO-RBF And CFD In MBR

MBR(Membrane Bio-Reactor)is a new type of water treatment technology that combines activated sludge process and membrane separation technology.Based on the unique advantages of membrane technology,MBR has more and more applications in the treatment of urban domestic sewage and industrial wastewater,and the scale of application is also growing.However,membrane pollution is an important factor hindering the rapid development of MBR,so it is necessary to carry out research on MBR membrane pollution.After studying the working mechanism of the membrane bioreactor,this paper analyzes the causes of membrane pollution and the main factors affecting membrane fouling.Some problems in the research of MBR membrane pollution prediction for traditional neural network models,for example,in the BP neural network model,there are shortcomings such as slow convergence rate,easy to fall into local minimum values,so this method is not good.After consulting a large number of literatures,it is found that the radial basis(RBF)neural network has the advantages of optimal approximation,simple structure,fast training speed,and there are mature applications of RBF in the prediction field,so RBF is applied in MBR simulation prediction simulation.The prediction effect of neural network has a great relationship with the parameter selection.In order to overcome the influence of RBF on the prediction result due to improper parameter selection,the particle swarm optimization algorithm(PSO)is used to optimize the RBF parameters(center value,width and weight)to improve The generalization ability of radial basis neural network algorithms.The results show that the average relative error is reduced by 2.9%compared with the prediction results of the BP algorithm model under the same conditions.Further research finds that the randomness of genetic algorithm(GA)can be used to increase the search range of RBF algorithm.Then the GA algorithm and PSO algorithm are combined to optimize the RBF input parameters,which not only preserves the efficiency of PSO algorithm but also improves the global optimal performance of the understanding.It was used to establish the GAPSO-RBF model to predict MBR membrane fouling.The results show that compared with the prediction results of the PSO-RBF algorithm model,the average relative error of the prediction results of the algorithm is reduced by 0.79%,which further improves the prediction accuracy of the membrane flux.Based on the study of membrane bioreactor(MBR)and computational fluid dynamics(CFD),this paper applies CFD to the field of MBR wastewater treatment research.The curtain type MBR not only has the excellent characteristics of traditional MBR,such as strong sewage treatment capacity,small floor space,less residual sludge and convenient operation.Moreover,the anti-pollution performance is good,and the bundle-shaped membrane unit can be arbitrarily arranged according to the amount of water,and the application field is wide.The multiphase flow mixture model in CFD can be used to deal with two-phase flow and multi-phase flow problems,and it occupies less resources.In this paper,the multiphase flow mixture(Mixture)model in CFD is used to simulate the flow state of sewage(solid-liquid two-phase flow)in a curtain membrane bioreactor.The amount of solid suspended particles(SS)at the outlet of the model is calculated as 0,which is in line with the actual flow phenomenon of the curtain MBR,and it has certain reference value for the sewage treatment process of curtain MBR.