Study On Grid-Connected Interface Circuit Power Quality Optimization Method Based On Data Model Compensation

When the linear system uses the frequency domain design method to design the controller,there is a significant problem of mutual limitation of the inner and outer loop control gains,so the grid-connected interface circuit must make a compromise in the control of the DC bus voltage index and the grid-connected power quality index.This paper intends to study the method to solve this contradiction from the perspective of inner loop compensation control,combined with the means of data-driven control,in order to ensure that the system has sufficient stability margin and improve the grid-connected power quality of the grid-connected interface circuit.Firstly,this article analyzes the mathematical model of the LCL type grid connected interface circuit,the design basis of filter parameters,and the contradiction of multiple control objectives in a single control loop,namely the problem of mutual gain limitation between the inner and outer loops.Secondly,traditional repetitive control was applied to current loop control to improve steady-state grid connected power quality.The range of controller gain values and optimization methods for repetitive controllers were studied from the perspectives of grid side current feedback and inverter side current feedback,respectively.The comprehensive challenges brought by the use of repetitive control in the inner loop to the dynamic response characteristics and stability of the system were analyzed through time-domain analysis.To address this challenge,repetitive controllers are used to train steady-state current loop compensation data under different operating conditions,and neural networks and 1-stop curve fitting methods are used to offline obtain the mathematical relationships between operating conditions,given and compensation data.This involves obtaining a regression model of the compensation data,which is then brought into the control system as the compensation basis for the inner loop compensation loop,thereby achieving online data compensation for power quality.To verify the effectiveness of this method,phase diagram analysis was first used to compare and analyze the overall dynamic and static characteristics of the system when repetitive control and data compensation control were applied to the inner loop,and preliminary evaluation indicators were formed.To verify the feasibility of this method,a simulation platform was built using PLECS simulation software,and the method was evaluated based on actual operating conditions.To verify the engineering feasibility and effectiveness of the method,a hardware platform for grid connected interface circuits was built,and it was verified that the data compensation method has better dynamic response ability compared to repetitive control,which can achieve stable control of bus voltage while ensuring the control goal of high grid connected current waveform quality.