Research On Key Technologies Of Modular Multilevel Converter With (2N+1) SHEPWM

In the medium-voltage application scenarios,compared with the Sine Wave Pulse Width Modulation and other technologies,the Selective Harmonic Elimination Pulse Width Modulation(SHEPWM)technology adopted by MMC can not only reduce the switching frequency and system power consumption of the device,but also improve the accuracy of the fundamental wave amplitude,effectively suppress the low-order harmonics that are difficult to filter at the back end,and reduce the harm to the power grid and its equipment.Further adopting the new SHEPWM Technology,i.e.(2N+1)SHEPWM,which can output N additional levels,has more obvious advantages.However,MMC operation of(2N+1)SHEPWM cannot be separated from three closely related organic components: optimization model construction,solution and corresponding MMC capacitor voltage balancing control.Due to the lack of quantitative design of the evaluation function,there are randomness and uncertainty in the accuracy of the fundamental wave amplitude and the amplitude of the lower-order harmonics;The accuracy of fundamental wave amplitude decreases seriously and the contents of low-order harmonics exceed the IEEE-519standard;It is impossible to apply a higher degree of suppression to lower-order harmonics that are difficult to be filtered at the back end;The output waveform is distorted due to the lack of constraint conditions and the requirement of switching angle.In order to make up for the above defects,this paper studies and analyzes the parameter significance of the quantitative evaluation function,selects the appropriate quantitative evaluation function,and adds the constraint condition of the switching angle anisotropy to improve the traditional optimization model.Compared with the traditional optimization model,the improved optimization model can more effectively suppress the lower-order harmonic components and prevent the distortion of the output waveform on the premise of ensuring the accuracy of the fundamental wave amplitude.In terms of solving the optimization model,since the number of output levels of(2N+1)SHEPWM is nearly twice as much as that of the traditional(N+1)SHEPWM,and it is necessary to solve the optimal switching angle solutions under various MMC conditions such as different modulation ratios and different arm submodules.The traditional harmony search(HS)algorithm is not effective.This paper presents a meta heuristic algorithm for differential harmony search(DHS).The HS optimal solution search path is optimized by moving away from the worst harmony and approaching the best harmony.The mutation operation with strong DE search ability is used for reference to optimize the performance of local exploration ability and overcome the dependence on BW parameters with high sensitivity to problems.Aiming at the parameter tuning problem of the proposed DHS algorithm,the optimal parameter optimization model is further proposed,and the efficient Bayesian optimization algorithm is applied to realize the efficient and automatic tuning of DHS parameters.The proposed parameter self-tuning DHS can obtain switching angle solutions with higher accuracy of fundamental wave amplitude,higher degree of lower-order harmonic suppression and more number of lower-order harmonic suppression,and can adapt to various MMC conditions.In terms of the MMC capacitor voltage balancing algorithm based on(2N+1)SHEPWM,due to the switching state characteristics of(2N+1)SHEPWM,the traditional two-layer capacitor voltage balancing control architecture is not only difficult to realize the capacitor voltage balancing in the MMC arm based on(2N+1)SHEPWM,but also has an unfavorable mutual restriction relationship with different switching angle solutions.Therefore,based on the double-layer capacitor voltage balancing framework of(2N+1)SHEPWM,the causes of the capacitor voltage imbalance based on(2N+1)SHEPWM and its unfavorable constraints with different switching angle solutions are studied.This paper presents an improved modular selection method.Based on the analysis of the instantaneous power of the arm,the average capacitance voltage fluctuation range is derived,and then the amplitude limit of the capacitance voltage and the voltage difference limit between the capacitors of the same arm are increased by taking advantage of the significant change of the capacitance voltage near the extreme points of the instantaneous arm power.Compared with the traditional modular selection method,the additional power consumption is less,which can better suppress the capacitor voltage ripple,ensure the low-order harmonic suppression ability of(2N+1)SHEPWM,and eliminate the influence of different switching angle solutions on the capacitor voltage balance.