Research On Model Predictive Control For Anaerobic Digestion Process Based On ADM1

Biogas is a kind of renewable and secondary energy,which can produce heat and electricity.Biogas production is an important part of biogas projects.The input organic matter,as the substrate fermented into biogas,whose characteristics and flow rate is the key factor to high efficiency and stability of the whole anaerobic digestion process.Therefore,investigation of flow rate control strategy for the anaerobic digestion process is of significance to the biogas projects.The anaerobic digestion is related with various microorganism populations.It is including not only biochemical process but also physical process,which lead the anaerobic digestion process to a complex biological process that is multivariate,constrained,strongly coupled,nonlinear and with perturbation.Model predictive control(MPC)is an advanced process control which can deal with these complex system.MPC focus on the complicated industrial process and has achieved many successful and mature applications.Thus MPC is introduced to anaerobic digestion process in this dissertation.Based on ADM1 system analysis of input-output and state variables characteristics,the simplified model is obtained.And then the MPC strategy is investigated and simulated for the system.The main work of this dissertation can be summarized as the following parts:1)Development of the anaerobic digestion system modeling and simulation.Based on the mechanism model of the anaerobic digestion process,the modular simulation environment is built on the Matlab/Simulink platform.Since there existing a large number of parameters in ADM1 which have important effluence to the simulation result,parameter identifiability is studied and some key parameters are identified.To meet the demand of MPC,model simplifying methods are raised based on state sensitivity and PCA respectively.2)Research of model predictive control for anerobic digestion process.Though the analysis of actual condition in project,optimization objective and manipulated variables are determined.Considering the effect of temperature to the AD process,an upper decision maker is designed.A nonlinear state space model and an unscented kalman filter compounded structure are chosen for MPC.The feasibility and effectiveness of the MPC strategy is verified by comparing with open-loop control via simulation.3)Because of low-efficiency,high requirements for computing power and limits on treating disturbance when applying the nonlinear model predictive control strategy,the linear model predictive control based on extended state observer is introduced to anaerobic digestion system.The extended state observer is designed based on ARIMA which deal with the nonlinearity and uncertainty.Using the linear system model after adding extended state as predictive model,the model predictive control algorithm is deduced.Through designing simulation experiments,the effectiveness on setpoint tracking and anti-disturbance of the control stategy are proven.