| Activated sludge wastewater process has been widely applied in domestic wastewatertreatment plant. Activated sludge process is a kind of effective method,which consumes thepollutant in wastewater by microorganisms’ metabolism. Activated sludge process is amulti-variables system,which has characteristic of strong coupling,indetermination,highnon-linearity and big lag,and difficult to build process model.Traditional control methods(such as conventional PID control etal.) achieve ideal control effect with difficultly, so ithas especially essential realistic value and broad application prospect to research theintelligent control technology of wastewater treatment process. In the activated sludgewastewater treatment system, the percentage of Dissolved Oxygen (DO) in water is animportant parameter,the control of DO can directly affect the effect and energyconsumption of the system.In the paper, on the basis of thorough analyzing past achievements and literature, tosolve the problem existed in DO control, this work presents as follows:(1)The source and background of the subject is introduced in this paper andsignificance of the subject is discussed. Summarized the soft measuring technique and theautomatic control strategies in wastewater treatment.(2)Analyzed the characteristic of the wastewater treatment system, established a softsensor model based on least squares support vector machines, the estimation of somedifficultly but important parameters are gained by this model. And used particle swarmalgorithm to optimize the parameters of the least squares support vector machines.(3) On the basis of studying the active sludge model NO.1which was proposed by theinternational water association, related actual wastewater treatment situation, establishedsimplification active sludge mathematical model which reflects the relations amongdissolved oxygen, microorganism and substrate.(4) In order to solve the problem existed in present dissolved oxygen control method,designed the dissolved oxygen setting value optimization model and the dissolved oxygeninverse control system based on least squares support vector machines.①Designed the dissolved oxygen setting value optimization model: According to thewater quality, established a soft senor model based on least squares support vectormachines to predict the water leakage parameters with a variety of physical and chemicalparameters such as the influent conditions, then took the water discharge parameters as a feedback signal, and used fuzzy neural networks to optimize the DO settings.②Due to activated sludge process has characteristic of strong coupling,indetermination,high non-linearity and big lag,the inverse control system based on leastsquares support vector machine and PID was used to decouple and track the settings ofdissolved oxygen. Analyzed and designed the control algorithm in detail, and confirmed thissystem’s antijamming,the compatibility and robustness through the simulation experiment.The innovation of the paper is as follows:(1) This paper established a multi-input and multi-output LSSVM soft sensor model,this model achieved to predict several variables at the same time. In view of theshortcoming of PSO is currently only used for off-line optimization, this paper realized theprocess of PSO to optimize online, and using a sliding rectangular window form,realizing a rolling forecast function, so that the model has more accurate prediction.(2) In this paper, in view of the insufficiency of used the constant setting value tocontrol dissolved oxygen in mostly wastewater treatment,we established the optimizationmodel to dynamic renew dissolved oxygen setting value, and successful applied theLSSVM inverse control system in the dissolved oxygen follow-up control. This methodachieves dynamic optimize control to the aeration quantity, and realizes wastewatertreatment system’s energy conservation to fall consumes. |
PDF Full Text Request