The invention discloses a
model prediction-based torque control method for a flux-switching
permanent magnet synchronous machine. The
model prediction-based torque control method comprises the following steps: combining the switch tube states of a motor
inverter to obtain eight sets of switch vector signals; at the current moment K, under each set of vector signals of the
inverter switch, predicting a p-phase winding current (please see the specification for the formula) at the next moment K+1 to obtain current prediction values (please see the specification for the formula) of a shaft d and a shaft q; predicting
motor torque (please see the specification for the formula) and motor
flux linkage (please see the specification for the formula) at the next moment k+1; calculating a cost function (please see the specification for the formula) to obtain the vector
signal of the
inverter switch when the cost function is the smallest; defining m (please see the specification for the formula), controlling the effective acting time of the vector
signal of the inverter switch when the cost function is the smallest in a single sampling period according to a value of m to complete torque control of the flux-switching
permanent magnet synchronous machine. The method can accurately control change of the vectors of the inverter switch, enables the
motor torque fluctuation and the
flux linkage fluctuation to be the smallest, and the effective acting time of the inverter switch is adjusted through the duty ratio, so that the on and off frequency of the inverter can be effectively reduced.