Research On Model Predictive Control Strategy Of Three-level Rectifier

To address the environmental problems caused by traditional fossil fuels,the demand for installation capacity of new energy sources such as photovoltaic and wind power is increasing year by year.As a traditional electrical energy conversion device,rectifiers have received widespread attention in the fields of new energy grid connection and high-power electric drive.With the continuous increase in voltage level requirements,three-level PWM converters are significantly superior to two-level PWM converters in terms of voltage level,power density,and energy quality.However,they also have problems such as multiple switch vectors and complex control objectives.The traditional vector control scheme is complex and has a high switching frequency,while the model predictive control strategy is simple to implement,has a fast dynamic response,and is advantageous for realizing multivariable control.This thesis focuses on the three-level neutral point clamped(NPC)PWM rectifier,and investigates the Finite Control Set Model Predictive Control(FCS-MPC)strategy for its control.In this thesis,the topology and working principle of the NPC-type three-level rectifier are first introduced,and its continuous mathematical model in different coordinate systems is analyzed.Based on the working state of the three-level topology power switch devices,the impact of different switch vectors on the midpoint potential in the three-level rectifier is analyzed.System control objectives are set as grid-side current tracking,midpoint potential balance,and switch switching restrictions,and a dual-weight value function is established.To address the drawbacks of high computational burden and high switching frequency in traditional rectifier model predictive control,this thesis proposes an optimized finite control set model predictive control strategy.This method sets the control objectives within a bounded invariant set by defining boundaries for the gridside current error and midpoint potential deviation error,which maintains system stability.When the system control conditions are satisfied,the switch state is maintained unchanged to reduce switching losses.Otherwise,the predictive vector partitioning method is used to optimize and reduce computational load within small sectors.Compared with traditional weight coefficient methods,this algorithm effectively reduces the switching frequency and computational load of the rectifier,and has certain application value in high-voltage,high-power and other working contexts.A novel control scheme is proposed for the multi-step model predictive control strategy of the three-level PWM rectifier.Traditional control schemes use iterative algorithms,which require large computational resources.This thesis investigates a lowcomplexity gradient descent solution with backtracking iteration approach for finite control set predictive current control.Firstly,FCS-MPC is reformulated as a quadratic programming problem from a geometric perspective.Then,a two-layer decision simplification algorithm is used to reduce the computational load,followed by the selection of the optimal solution through the reformulated objective function.Based on this research,a finite control set predictive current control method based on a two-layer decision tree is proposed for the control system of the three-level rectifier,combined with midpoint voltage judgment to avoid weight coefficient tuning.The multi-step length MPC algorithm is optimized to solve the problem of the explosion of computational complexity.Finally,on the three-level PWM rectifier experimental platform,the experimental verification of the algorithm proposed in this thesis is completed.There are 68 figures,8 tables and 84 references in this thesis.