Research On Control Strategy Of Microbial Fuel Cell Based On State Estimation

Microbial Fuel Cell(MFC)is a bioelectrochemical device that converts the chemical energy in biomass or organic compounds into electrical energy through the catalytic reaction of microorganisms under specific controllable conditions.The unique advantages of high efficiency,cleanliness and environmental protection have attracted wide attention from scientific researchers from all over the world,and have become a research hotspot in this field.But the problems of low power density and unstable voltage lead to a certain gap from practical application.The microbial fuel cell system is a non-linear,strongly coupled,time-varying system,which is often affected by unmeasured interference in the actual industrial production process,which brings challenges to the control.Starting from the actual situation,state estimation and intelligent optimization control methods are combined in this thesis to optimize the output voltage of the microbial fuel cell to improve the stable operation of its power generation system.The main research contents are as follows:(1)Firstly,the working principle and basic composition of the microbial fuel cell system are analyzed,and the simulation model of the dual-compartment microbial fuel cell system is built under the Matlab/Simulink environment according to the MFC mechanism and mathematical equation.Through dynamic analysis,the influence of various operational parameters on the power production performance of microbial fuel cell is deeply studied.The feed flow which has great influence on the output voltage and is easy to control is selected as the control variable and the output voltage is taken as the controlled variable.(2)Aiming at the non-linear and interference problems in the microbial fuel cell system,a BP neural network PID control strategy based on unscented Kalman filter is proposed,which mainly controls the anode feed flow.The BP neural network is used to fully approximate the complex nonlinear relationship,to learn and adapt to the dynamic characteristics of the MFC system to realize the automatic adjustment of the controller parameters and to establish the best combination mode.The optimal estimation of system state is realized by using unscented Kalman filter to improve the control performance of controller.The simulation results show that compared with the traditional BP neural network PID control,the BP neural network PID control with unscented Kalman filter has short control time,fast response speed,small steady-state error,better stability and anti-interference ability,the output voltage can reach the set value quickly and smoothly.(3)Considering the characteristics of microbial fuel cells operating at steady-state operating points,the mechanism model of microbial fuel cells is linearized at anode and cathode steady-state points.Aiming at the linear MFC model and its interference problems,a fractional PI~λD~μcontrol strategy based on Kalman filter is proposed.The Kalman filter and the fractional PI~λD~μcontroller are designed for the linearized anode and cathode models respectively,and the Kalman filter state estimator is used to obtain accurate MFC anode and cathode voltage values,thereby improving the performance of the fractional PI~λD~μcontroller.The simulation results show that compared with the traditional fractional PI~λD~μcontrol,the fractional PI~λD~μcontrol after Kalman filter correction can effectively compensate the influence of interference on the MFC system,improve the steady-state accuracy and the control process is smoother.(4)The above control methods do not take into account the actual input constraints.Aiming at the constraints and interference problems existing in microbial fuel cell system,a model predictive control strategy based on Kalman filter is proposed.The Kalman filter and model predictive controller are designed for anode and cathode models respectively.By introducing the model predictive controller,the performance of constrained optimization control is improved,combined with the Kalman filter,the influence of unmeasured interference on the predictive control is compensated,and the control accuracy and anti-interference ability of the system is improved.The simulation results show that compared with the traditional model prediction control,the model prediction control with Kalman filter can make the system meet the actual constraints and achieve the optimal cost.It effectively overcomes the influence of unexpected interference on the output of microbial fuel cell system,improves tracking accuracy and stability of the voltage,which significantly improves the control performance of the model predictive controller.