Research On Energy-saving Optimal Control Method For Wastewater Treatment Process

Wastewater treatment plants (WWTPs) are dynamic non-linear systems whichhave characteristics of large time-varying and large time-delay with various variables,hence, the effective sewage treatment becomes one of the important issues to solvewater environmental problems in today’s world. In order to improve the operatingefficiency of WWTP devices, ensure effluent quality (EQ) and reduce operating costs(OC), it is a develop trend to study a new intelligent optimal control strategy toachieve energy-saving in current sewage treatment industry.In the thesis, a Benchmark Simulation Model No.1(BSM1)-based multivariateloop control strategy for WWTP is proposed, an energy consumption feature modelwhich can express the direct relation between EQ and energy consumption (EC) isbuilt, and then an energy-saving predictive optimal control system is designed. Finally,the feasibility and effectiveness of all researches are verified through the experiment.The main innovation points are summarized as follows:1. On the basis of thorough analysis in Activated sludge model No.1(ASM1)and BSM1of the activated sludge process (ASP), the simulation platform of BSM1isset up in MATLAB to be visualized. Then a WWTP multivariate loop control strategyis proposed, which controller design is based on recurrent neural network, and appliedon BSM1platform. The results show that this control method has a betterself-adaptability, robustness and stability compared to conventional PID control andBP neural network control.2. Considering there is no model representation between EQ and EC for WWTP,an energy consumption feature model (ECM) is put forward. By analyzing therelationship between EQ/EC indexes and13components in ASM1systematically, theinternal relation of EQ and EC can be digged out. Then the extended Elman neuralnetwork (EENN)-based modeling method is designed to build ECM. The simulationresults illustrate that the WWTP EENN-ECM can express the direct relationship ofEQ/EC, and also fill the gaps of energy common model in WWTP.3. To solve a higher EC problem, an EENN-ECM–based energy-saving optimalcontrol strategy is proposed. The mechanism of WWTP biological nitrogen removal isanalyzed in specific aiming at A/O nitrogen removal process so that the dissolvedoxygen (DO), the mixed liquor suspended solids (MLSS) and the nitrate nitrogen (SNO) concentrations are chosen as control optimal variables. Then a predictive controlmethod is designed to integrate a set of complete WWTP optimal control system.Experimental results indicate that this optimal control method can optimize theunderlying controller’s setpoints dynamically and lower OC successfully.