The invention belongs to the technical field of ultra-high-speed
linear motor control, particularly relates to a self-learning time-varying parameter current stability control method,
system and equipment for an ultra-high-speed
linear motor, and aims to solve the problem that current stability control is difficult to realize by a constant
parameter control strategy. The method comprises the steps that
stator excitation current im in historical operation is collected, a
stator segment serial number n serves as input,
stator excitation
inductance Lnm serves as output, and a
data set training neural network is constructed; and obtaining im and n in real time, inputting the im and the n into a training model, outputting a predicted value of stator excitation
inductance Lnm, calculating to obtain a predicted value of mutual
inductance Lnr based on a
magnetic circuit relationship, inputting the predicted value of the stator excitation inductance Lnm and the predicted value of the mutual inductance Lnr into a time-varying parameter controller, and generating a control
voltage to realize stable current control. According to the method, current oscillation caused by nonlinear rapid time
variation of parameters is effectively suppressed, high-precision real-time prediction of excitation inductance is realized, and current tracking deviation is suppressed.