The Research On Ewage Treatment Control System Design And Soft-sensor Modeling

With the progress and development of computer science and technology, the levelof industrial control automation is increasing day by day. These advanced controltechnology is closely related to the application of equipment. However, some complexobjects, such as sewage treatment process, due to the characteristics of highly nonlinearparameter and great measurement random variation, results in a number of importantprocess variables which are still not mature and stable. So there is great practicalsignificance and social value to take the research and exploration on sewage processingsystem and modeling.Combined with the specific requirements of the sewage treatment project, thispaper design and develop an integrated, efficient sewage treatment control system. Thesystem can realize the accurate variable measurement of sewage treatment processsuch as temperature, liquid level, PH value, and so on. At the same time, complete acontinuous real-time control for the various implementing agencies, pump, valve, mixer,scraping mud, and so on. This system follows the design principles of decentralizedcontrol and centralized monitoring. Control form is divided three, including the fieldmanual control, the PLC linkage control and the centralized control.It is difficult to realize online measurement for water quality indicator in sewagetreatment. To solve the key problem, this paper introduces the theory about softmeasurement technology. With nonlinear and complex process modeling in sewagetreatment system, we select the least squares support vector machine to build a waterparameter COD (chemical oxygen demand) soft sensor model. We adopt new particleswarm optimization algorithm to make parameter optimization for train model,achieving a high rate of learning. In MATLAB environment, we select the actualmeasurement data in modeling. The simulation verifies the feasibility and effectiveness for this approach.Finally, use the generalized predictive control (GPC) strategy to take the controlmethod research on biological fluidized bed sewage treatment process.Regarding leastsquares support vector machine as a predictive model, the paper gives the generalizedpredictive control framework based on least squares support vector machine. And makefurther exploration with the control algorithm. For sewage biological fluidized bedprocess, design a predictive controller to achieve closed-loop control of sewagetreatment. The simulation results show that there is good control effect based on leastsquares support vector machine generalized predictive control method in sewagetreatment system.