Dual-motor torque synchronization model predictive control method based on quadratic value function

Abstract

The invention discloses a dual-motor torque synchronization model predictive control method based on a quadratic value function. The method comprising steps of establishing a unified prediction modelby taking a dual-motor system that drives the same heavy load through two motors as an object, by taking the synchronization performance between motor output torques as a main control target, and bytaking the current values of respective motors and the torque synchronization error between the motors as state variables; using the unified prediction model to obtain state variable predicted values,and enabling multiple control variables to be included in the same error vector by using the quadratic value function in order that only one weight coefficient matrix to be solved is in the value function; and calculating the optimal weight coefficient matrix by using a weight coefficient matrix offline solution algorithm, and using the optimal weight coefficient matrix in a finite set model predictive control process of a dual-motor torque synchronization system. The method ensures the convergence of the error vector and the asymptotic stability of the system, not only improves the torque synchronization performance of the dual-motor system, but also improves the dynamic and steady tracking performance of the two motors.

Description

technical field [0001] The invention relates to a dual motor system synchronous control method belonging to the field of power electronic circuit control, in particular to a quadratic value function finite set model predictive control method for dual motor torque coordinated control. Background technique [0002] Finite set model predictive control (FCS-MPC) has good dynamic response speed and steady-state control accuracy, and at the same time, the switching signal of the inverter is directly evaluated and selected through the value function, which is easier to realize the multi-objective control of the system, so it is suitable to be applied to the dual Motor torque synchronization system. The value of the value function directly determines the optimal voltage vector finally selected, and the effect of each variable in the value function is adjusted through the weight coefficient. It can be considered that the form of the value function and the selection of the weight coef...

AI Technical Summary

Problems solved by technology

However, the control variables of the dual-motor system include not only the tracking error of each motor, but also the synchronization error between the motors, so there are at least three weight coefficients in the value function, and the algorithm tuning method can only solve the tuning problem of a single weight coefficient, and cannot adjust three There are two interrelated weight coefficients, so the empirical setting method can only be used to adjust the three weight coefficients.

[0003] However, the empirical tuning of the weight coefficients in the weighted summation value function will bring the following problems: (1) In order to ensure the univariate nature of the adjustment of the weight coefficients, it is necessary to approximate the weight coefficients of the tracking errors of the two motors, but due to The system adopts the master-slave control structure. The status of the two motors is actually unequal. The slave motor indirectly tracks the torque given value of the master motor. To achieve the same tracking effect, the slave motor tracking error should be given in the value function The weight of the item is larger, that is, this approximate processing is actually costly; (2) Usually, the initial value of the synchronization error weight coefficient is selected to cover a large selection range, and these values ​​​​need to depend on the control requirements of the system. Selected by subjective experience, if the order of magnitude is large, the number of adjustments will increase accordingly; (3) There are many evaluation indicators defined, so when there is a problem that cannot take into account all indicators in the process of gradually narrowing the interval, you can only rely on (4) There are many comparative experiments, which increase the complexity of the experiment, but in the end only the approximate value range of the synchronization error weight coefficient can be obtained, and the optimal point of the control effect cannot be determined; (5) only in each In the control cycle, selecting a group that can make the cost function optimal from the alternative switch state combinations is used as the input of the inverter, which can only ensure that the voltage vector combination selected in each control cycle is optimal, but the error is continuous The control cycle may still tend to diverge, so there is no guarantee that the error terms are convergent

Images