Research On Active Front-end Converter System Design And Control Strategy

With the gradual improvement of electric power electronization in power system,the use of power electronic equipment to realize high-quality electric power conversion in future energy system will gradually occupy a dominant position.As an active front-end device connecting power grid and electronic devices,the high performance control of active front-end converter plays a key role in power quality.The active front-end converter is studied from the following aspects.1.Based on the three-phase two-level topology of the active front-end converter,the mathematical models of the active front-end converter in different coordinate systems are established.This thesis introduces the traditional current control method(VOC),and gives the parameter design and hardware realization method of active front-end converter systematically.The dynamic and steady-state performances are analyzed by simulation and experimental results.The results show that the response speed of the inner current loop is slow and the dynamic performance of the system is poor.The dc bias and amplitude error in the measured values of the input voltage and current will lead to the voltage fluctuation of the output on the dc side and the deterioration of the quality of the current on the grid side.2.Aiming at the problem of large amount of code and long calculation time of the traditional enumeration model predictive control,an improved simplified single vector model predictive control is proposed in this thesis.Based on the advantages of simple principle,easy implementation,good dynamic performance and no PI parameter setting,the improved simplified single vector model predictive control strategy can effectively reduce the online optimization time,improve the dynamic performance of the system,and facilitate the engineering implementation.3.To solve the problem that the steady state performance of the single vector model predictive control strategy is insufficient under the same switching frequency,a simplified multi-vector model predictive control is proposed in this thesis.The so-called multi-vector refers to the multiple voltage vectors participating in the action in a control cycle,which can ensure that the optimal two non-zero vectors in the 8 voltage vectors participate in the action at the same time to achieve the current beat free control.By introducing the method of vector sector judgment,and giving the calculation method of the action time of each vector,the steady state performance of the system is improved and the current ripple and total harmonic distortion rate(THD)are significantly reduced without changing the dynamic performance of the single vector predictive control.4.As for the dc bias and amplitude error contained in the measurement error of voltage and current,this thesis firstly makes a theoretical analysis on the influence of the error on various indexes of the system.Then an online error compensation scheme is proposed.The errors in the current measurement can be extracted by combining the fundamental frequency and double frequency fluctuations in the current in d-q axis with the bandpass filter and the low pass filter and compensated online in the sampling link.For voltage measurement error,an improved current controller is applied in this thesis.PI and double resonance controller are introduced to filter out the influence of voltage measurement error.At the same time,the phase-locked error caused by the voltage measurement error is eliminated by using a phase-locked loop based on a double second-order generalized integrator.On the basis of theoretical research,this thesis uses MATLAB/SIMULINK to build the model,and the designed prediction controller and measurement error compensation strategy are analyzed,the simulation results verify the correctness of the theoretical research.