Research And Implemention On Algorithm Parallelization Of Inverter Predictive Control

The inverter plays an important role in the field of ac motor speed regulation,uninterrupted power supply system,new energy grid power generation and so on,and develops in the direction of digitalization and high frequency.With the rapid development of digital signal processing(DSP)technology,many new digital control strategies are widely used in inverters.Among them,the finite control set model predictive control(FCS-MPC)is a hotspot in current research because it has the advantages of simple concept,multivariable,nonlinear and constraint.Because the FCS-MPC is a kind of control strategy that uses the traversal optimization,the problem of high computation and high processing speed is found in the practical application.In the case that the system main frequency cannot be further improved,the traditional single-core DSP will be restricted to the implementation efficiency of the model predictive control program,which will affect the control effect and output performance.In view of the above problems,this paper adopts the multi-core DSP as the controller and combines the parallel processing technology to realize the model predictive control algorithm in a parallel way,and the research is carried out.In this paper,the basic principle of FCS-MPC algorithm is analyzed in detail,and its application is applied to the off-grid dual Buck full bridge inverter.Firstly,the working principle of dual Buck full bridge inverter is analyzed and the state space equation of the circuit is established,and the prediction model of output voltage is obtained.Then the model predictive voltage control(MPVC)algorithm based on FCS-MPC principle is designed.The correctness and feasibility of the control algorithm are verified by simulation experiments.In addition,the simulation experiment also compared the influence of different sampling frequency on the control effect of the algorithm,and the results showed that the model predictive voltage control had better control effect when the sampling frequency was above 100kHz.This paper introduces the commonly used parallel algorithm design techniques as well as the evaluation index.Then the parallelism of FCS-MPC algorithm is analyzed,and it is concluded that FCS-MPC algorithm suitable for partition strategy is adopted to improve the parallel processing,and through the modeling of time after the analysis of the parallel algorithm of time optimization effect.Then TMS320F28377D dual-core DSP communication mechanism between nuclear was studied,in consideration of program execution sequence,synchronous communication delay between nuclear and algorithm,designed the parallel MPVC algorithm as the core of double CPU control program flow.Finally,a lkW dual Buck full bridge inverter is produced.The feasibility and correctness of the parallel implementation of MPVC algorithm are verified by steady-state.and load switching experiments.The comparison and analysis of the execution time of series and parallel MPVC algorithm proves that the parallel processing of MPVC algorithm can effectively improve the efficiency of the algorithm.