Modeling And Optimal Control For Aeration Process Based On LSSVM

With the development of industrial automation, the intensifying ofwater pollution has become the atopic of focus. In recent years, althoughthe automation level of domestic sewage treatment plants has beenrapidly improved, with the application of automation technologybeing late starters and compared with abroad, the level of automationcontrol and the efficiency of sewage treatment is relative low. Aerationprocess being as a key link in sewage treatment craft, the aeration rate isnot only has an important influence on water quality, but also itplays a decisive role to the sewage treatment plant operating costs.Therefore, the establishment of an accurate mathematical model for theaeration process and the introduction of advanced automatic controltechnology are the basical measure to improve the efficiency of sewagetreatment and reduce the energy consumption.It is the best choice to choose DO concentration as the controlledobject because it is the most important water quality parameters to reflectthe aeration rate of aeration process. Considering the complexity of thesewage treatment’s mechanism process, it is difficult to deduce themechanism model of the DO concentration accurately by the presentmathematics knowledge. So the DO concentration is as the dominantvariable and some water quality parameters impacting the dominantvariable are chosen as auxiliary variables in this paper, then the clusteranalysis algorithm and LSSVM algorithm are used to realize optimalestimation of dominant variable by assistant variables; For the problem ofkernel function’s choice restricting the modeling accuracy, the idea of themulti-core is introduced into the LSSVM algorithm to improve themodeling accuracy as much as possible; Moreover, in order to solve thesparseness of LS-SVM results, an improved multi-core LSSVM isdesigned. It need not pay attention to the mechanism of physical andchemical processes during the activated sludge process by this modelingapproach, So there is no the pressure to solve a large number ofdifferential equations based on the conservation rules, and it is easy tobuild accurate mathematical model of the DO concentration. For the unstable problem of conventional PID controlling the DOconcentration, this paper constructs the variable parameter PSO-PIDcontrol strategy. Its parameters can be self-tuning with the size of thedeviation e on line to deal with the changing work environment better.The variable parameter PSO-PID control system based on aerationprocess was established, the simulation results prove the robustness andeffectiveness of the variable parameter PSO-PID control strategy, andalso show its popularizing significance.