Model Predictive Current Control Of Induction Motors

Induction motors are one of AC motors with a high market share at present and their high-performance control is a major direction for future development.The traditional field orientation control and direct torque control cannot take care of the excellent steady-state control performance and quick dynamic response at the same time.However,model predictive control technology is easy to deal with nonlinear multivariable constraints,has quick dynamic response and real-time rolling optimization feature.Considering the the two points above,MPC has gradually become a new generation solution to realize efficient control of induction motor.This dissertation aims at the two-level voltage inverter-induction motor system.Based on a basic single-boundary circle and single-vector strategy model predictive current control,from the perspectives of vector selection,control strategy and prediction range,three-aspect contents in the improvement of steady-state performance,the improvement of dynamic performance and the reduction of algorithm burden are theoretically anaylzed and researched deeply.A basic single-boundary circle and single-vector strategy model predictive current control is established and its algorithm realization is completed.Firstly,based on the indirect rotor field-oriented control framework,the model predictive current controller is used to replace the traditional current inner loop PI controller and modulation module,so as to simplify its control structure.Nextly,based on the mathematical model of the induction motor in the rotating coordinate system,its prediction model is deduced and constructed.Based on the current error model,the cost function is deduced and analyzed.Lastly,its implementation and application in stator current control and the selection process of the optimal vector are elaborated.Besides,its dynamic and steady-state performance verification are completed.Aimed at the problem of poor and unstable steady-state performance of single-boundary circle and single-vector strategy model predictive current control,a double-vector strategy model predictive current control method which introduces "vector 0/7" is proposed.Firstly,the technical limitations of the application of single-boundary circle and single-vector strategy model predictive current control at low sampling frequency are analyzed.Nextly,the realization of the vector combination of "arbitrary vector + 0/7 vector" in the selection process of the optimal vector is elaborated in detail,the performance advantages of this method in terms of induction motor low-speed performance and low sampling frequency are analyzed,and steady-state performance advantages are verified.Aiming at the large flunctuations problem of performance indicators in the dynamic process of single-boundary circle and single-vector strategy model predictive current control,a dual boundary circle strategy model predictive current control method in two-phase static coordinate system is proposed.Firstly,AC current component can be controlled directly and the influence of orientation angle on control performance is weaken in the two-phase static coordinate system.So mathematical model of the induction motor in the two-phase stationary coordinate system is used to deduce and construct the prediction model.Nextly,Considering the current tracking performance and switching frequency as the control target at the same time,the cost function of the single-boundary circle strategy is constructed;Based on the consideration and further improvement of the dynamic performance of the induction motor,the cost function of the double boundary circle strategy is constructed.In addition,the influence of the radius of boundary circle and the switching weight factor in the cost function on the current harmonics,torque ripple and switching frequency at different speeds are analyzed.Lastly,the dynamic performance advantages of the dual-boundary circle strategy model predictive current control are verified and analyzed.In order to further boost the steady-state performance of the motor,from the perspective of perdiction range,the research is carried out on multi-step model predictive current control algorithm realization and calculation burden drop.Firstly,based on the state equation of induction motor in two-phase static coordinate system,a multi-step prediction model of induction motor is deduced and constructed.Taking the sum of the multi-step current error and switching frequency within the predicted range as the control objective,the cost function in the form of quadratic function is constructed.Nextly,based on the spherical interpretation optimization algorithm,the cost function is derived,and the multi-dimensional optimization problem of the cost function is converted into multiple one-dimensional problems for processing to reduce the computational burden of the algorithm.In addition,in order to realize the goal of on-line adjustable switch weight factor,a new weight factor is introduced into the cost function and the switch weight factor in the matrix is separated,so as to facilitate on-line debugging;For the sake of further reducing the calculation burden of multi-step model predictive current control algorithm,the correlation matrix containing the motor speed is processed to realize the off-line calculation of the motor speed part.Lastly,compare and analyze the performance advantages of the multi-step predictive current control strategy,and complete the performance verification of the improved strategy.This dissertation builds an induction motor control system in Matlab/Simulink,and conducts a series of simulation analysis on the performance advantages of the control strategies under different dynamic and steady-state conditions;Based on the "DSP+FPGA" hardware experimental platform and the DSPACE hardware-in-the-loop simulation platform,the experimental verification of each algorithm strategy is completed.Lastly,summarize and analyze the characteristics and suitable application of the three control strategies,and provide references for the application of model predictive current control in induction motors.Figures: 97,Tables:13,References:138.