Research On Control Strategy Of Photovoltaic Grid Connected Inverter Based On Improved BP Neural Network And PID

In order to achieve the goal of "carbon peaking and carbon neutralization" and solve the energy crisis and ecological environment problems,people turn their attention to the development of new clean energy.Photovoltaic power generation is an important power source to achieve "zero emission" of greenhouse gases.Due to the influence of unstable factors such as light intensity and ambient temperature,as well as a large number of applications of power electronic devices,photovoltaic power generation will cause power pollution,which is mainly reflected in the excessive harmonic voltage and harmonic current,and the degradation of power quality.The reason is that photovoltaic power generation needs to be "inverted",which produces a large number of harmonics in the process of inversion,thus affecting the normal operation of other loads in the system.Therefore,it is necessary to improve the filtering effect of grid connected inverter and the tracking performance of grid connected current.In this thesis,the single-phase two-stage photovoltaic grid connected system is taken as the object,and the system is divided into two parts: the former and the latter.The front stage circuit includes photovoltaic cell module,DC boost module,voltage outer loop control module and MPPT module,and the rear stage circuit includes inverter filter module and grid connected current inner loop control module.The double closed-loop control method is used to control the grid connected inverter.The outer voltage loop adopts PI controller and the inner current loop adopts PID controller.Aiming at the problem that the traditional PID control is difficult to achieve the ideal control effect for the nonlinear system,the back-propagation BP learning algorithm with momentum update is used to accelerate the convergence of error performance function,output appropriate PID parameters in real time,and speed up the tracking speed of grid connected current to grid connected voltage phase.In view of the problem that the system is easy to fall into local minimum and oscillate in the learning process due to the fixed learning rate and other parameters of BP neural network,this thesis proposes a control strategy of photovoltaic grid connected inverter based on improved bp-pid.The improved particle swarm optimization algorithm is used to optimize the bp-pid controller,adjust the parameters of BP neural network adaptively,and reduce the tracking error of grid connected current to the command current.In this thesis,a photovoltaic system simulation model based on this control strategy is built,and the simulation results verify its correctness and effectiveness.