CN-121984394-A - Demagnetizing space distribution self-adaptive compensation method for fractional slot concentrated winding permanent magnet motor

CN121984394ACN 121984394 ACN121984394 ACN 121984394ACN-121984394-A

Abstract

The invention relates to a fractional slot concentrated winding permanent magnet motor demagnetizing space distribution self-adaptive compensation method which comprises the steps of collecting motor driving physical quantity, calculating a permanent magnet flux estimated value based on the motor driving physical quantity, generating a flux linkage residual sequence, carrying out frequency domain analysis on the flux linkage residual sequence, obtaining a target fault harmonic order, constructing an analysis model of a demagnetizing area to obtain an equivalent flux linkage loss function, carrying out Fourier series expansion on the equivalent flux linkage loss function to obtain a harmonic complex coefficient set, constructing an error function aiming at demagnetizing space distribution parameters, carrying out iterative solution on the demagnetizing space distribution parameters by using a parameter inversion model to obtain an identification result, constructing an equivalent permanent magnet flux function under the current rotor electric angle by using the identification result, correcting a d-axis current instruction, setting a compensation voltage vector orthogonal to the demagnetizing flux linkage harmonic wave, and superposing the compensation voltage vector to a PI output end in a feedforward mode to generate a final d-q voltage instruction.

Inventors

TAO JIE
CHENG ZHIYUAN
CHEN JIANDONG
XU PINGPING
XIAO ZEHUI
LU RENQUAN

Assignees

广东工业大学

Dates

Publication Date: 20260505
Application Date: 20260203

Claims (10)

1. The self-adaptive compensation method for the demagnetizing space distribution of the fractional slot concentrated winding permanent magnet motor is characterized by comprising the following steps: Collecting motor driving physical quantity, calculating a permanent magnet flux linkage estimation value based on the motor driving physical quantity, defining flux linkage residual by utilizing the permanent magnet flux linkage estimation value, extracting q-axis components of the flux linkage residual, generating a flux linkage residual sequence, carrying out frequency domain analysis on the flux linkage residual sequence to obtain a target fault harmonic order, constructing an analysis model of a demagnetizing region, and obtaining an equivalent flux linkage loss function, wherein the analysis model is used for representing the space range and the intensity of the demagnetizing region; Performing Fourier series expansion on the equivalent flux loss function, combining an initial value setting strategy to obtain a harmonic complex coefficient set, constructing an error function aiming at demagnetizing space distribution parameters, and performing iterative solution on the demagnetizing space distribution parameters by using a parameter inversion model with the error function as a target to obtain an identification result; And constructing an equivalent permanent magnet flux linkage function under the current rotor electric angle by utilizing the identification result, correcting the d-axis current instruction, setting a compensation voltage vector orthogonal to the demagnetizing flux linkage harmonic wave, and overlapping the compensation voltage vector to the PI output end in a feedforward mode to generate a final d-q voltage instruction.
2. The method for adaptively compensating for the demagnetizing space distribution of a fractional-slot concentrated winding permanent magnet motor of claim 1, wherein calculating the permanent magnet flux linkage estimation value comprises: Converting three-phase current in the motor driving physical quantity into d-q axis current under a synchronous rotation coordinate system, and obtaining d-q axis voltage through space vector modulation back-pushing or sampling reconstruction; setting constraint conditions of steady-state operation or current change rate, calculating a permanent magnet flux linkage estimated value to be compensated by combining the d-q axis voltage, and carrying out temperature compensation on the permanent magnet flux linkage estimated value to be compensated by adopting a sectional temperature compensation strategy to obtain the permanent magnet flux linkage estimated value.
3. The method for adaptively compensating for the demagnetizing space distribution of a fractional-slot concentrated winding permanent magnet motor of claim 2, wherein obtaining the permanent magnet flux linkage estimation value comprises: ; Wherein, the For the temperature compensated actual flux linkage estimation, The actual permanent magnet flux linkage estimated value, namely the permanent magnet flux linkage estimated value to be compensated, Is the reversible temperature coefficient of the permanent magnet, Is the irreversible demagnetization factor in the high temperature area, Is the temperature of the permanent magnet and is equal to the temperature of the permanent magnet, As a result of the reference temperature, Is the inflection point temperature of the temperature characteristic.
4. The adaptive compensation method for the demagnetizing space distribution of the fractional-slot concentrated winding permanent magnet motor according to claim 1, the method is characterized in that the step of defining the flux linkage residual by using the permanent magnet flux linkage estimated value comprises the following steps: ; Wherein, the As a residual of the flux linkage, Is the nominal permanent magnet flux linkage of the motor, Is an estimated value of the actual flux linkage after temperature compensation.
5. The adaptive compensation method for the demagnetizing space distribution of the fractional-slot concentrated winding permanent magnet motor according to claim 1, wherein obtaining the equivalent flux loss function comprises: ; Wherein, the As a function of the equivalent flux linkage loss, In order to achieve the demagnetizing strength, The initial electrical angle and the final electrical angle of the demagnetizing region.
6. The fractional-slot concentrated winding permanent magnet motor demagnetizing space distribution adaptive compensation method of claim 1, wherein obtaining the harmonic complex coefficient set comprises: Performing Fourier series expansion on the equivalent flux loss function to generate a harmonic component calculation model; Setting a demagnetizing strength initial value, a window center initial value and a window width initial value through the initial value setting strategy, and acquiring an initial demagnetizing window center electric angle by utilizing the demagnetizing strength initial value, the window center initial value and the window width initial value; Substituting the initial demagnetizing window center electric angle serving as a starting value into the harmonic component calculation model to obtain the harmonic complex coefficient set.
7. The method for adaptively compensating the demagnetizing space distribution of a fractional-slot concentrated winding permanent magnet motor of claim 1, wherein constructing the error function comprises: ; Wherein, the For a selected set of harmonic orders, As a set of harmonic complex coefficients, A set of harmonic complex coefficients is predicted for the model.
8. The method for adaptively compensating the demagnetizing space distribution of a fractional-slot concentrated winding permanent magnet motor according to claim 1, wherein constructing the equivalent permanent magnet flux linkage function comprises: ; Wherein, the As a function of the equivalent permanent magnet flux linkage, Is the nominal permanent magnet flux linkage of the motor, In order to achieve the demagnetizing strength, Is the current rotor electrical angle.
9. The adaptive compensation method for the demagnetizing space distribution of the fractional-slot concentrated winding permanent magnet motor according to claim 1, wherein correcting the d-axis current command includes: ; Wherein, the For the corrected d-axis current command, As a function of the equivalent permanent magnet flux linkage, For the d-axis and q-axis inductances, Is a set q-axis current command.
10. The fractional-slot concentrated winding permanent magnet motor demagnetizing space distribution adaptive compensation method of claim 1, characterized in that setting the compensation voltage vector comprises: ; wherein K is the gain coefficient, Is the dominant characteristic order, i.e., the odd integer multiple closest to the theoretical fault harmonic order.

Description

Demagnetizing space distribution self-adaptive compensation method for fractional slot concentrated winding permanent magnet motor Technical Field The invention relates to the technical field of demagnetizing space distribution compensation, in particular to a method for adaptively compensating the demagnetizing space distribution of a fractional slot concentrated winding permanent magnet motor. Background Permanent magnet synchronous motors have become a mainstream choice for high-performance electric drive systems such as new energy automobiles, industrial servo and the like due to high power density, high efficiency, wide speed regulation range and excellent dynamic response characteristics. However, under severe conditions such as long-term high-load operation, frequent start-stop, high Wen Huojiang vibration, etc., the rotor permanent magnets may undergo local irreversible demagnetization. The fault can destroy the spatial sine of the air gap magnetomotive force, so that the electromagnetic torque pulsation is obviously increased, the rotating speed/position control precision is reduced, the system efficiency is deteriorated, and the controller can be induced to be unstable even stopped when serious. Aiming at the problems, the prior fault-tolerant control technology mainly comprises the following categories: (1) And (3) establishing a current compensation table look-up on an experimental platform in advance according to different demagnetizing degrees and working conditions, and triggering table look-up correction through fault detection in the operation process. The method has the advantages of simple realization and low calculation cost, but seriously depends on a large amount of off-line calibration data, cannot cover unknown demagnetization positions or multi-region combination demagnetization, and rapidly deteriorates the compensation effect once the motor model or the operation condition deviates from the calibration range. (2) The resonance compensation method is to introduce 5 times, 7 times and other resonance controllers with specific orders into the current inner loop to inhibit torque harmonic caused by demagnetization. Furthermore, the partial improvement scheme adopts an adaptive resonance controller, and the resonance frequency is adjusted on line through a phase-locked loop or a frequency estimation algorithm so as to cope with harmonic frequency deviation caused by rotation speed change. The method has the advantages that high gain suppression of preset order harmonics can be maintained in a certain rotating speed range, but the self-adaption is limited to frequency tracking, the target harmonic orders still need to be preset manually, and the mapping relation between the orders and the motor body structure (such as stator slot number and pole pair number) is not established, so that universality is lacking when facing motors with different slot pole combinations. (3) Disturbance observer method-the demagnetizing effect is considered as a total disturbance together with other uncertainties, estimated by an Extended State Observer (ESO) or a sliding mode observer, and feedforward compensated in a control law. The method has the advantages of having certain robustness on parameter perturbation, only obtaining the equivalent total disturbance amount, being incapable of analyzing the space position and the intensity of demagnetization, and lacking definite physical significance in compensation. (4) Model predictive fault-tolerant control, namely embedding demagnetizing related state variables in a Model Predictive Control (MPC) framework, and solving an optimal voltage vector through rolling optimization. The method has the advantages that multiple targets and multiple constraints can be considered, but the calculation complexity is high, millisecond control period requirements are difficult to meet on a conventional Digital Signal Processor (DSP), and engineering feasibility is limited. Summary of common defects of the existing fault-tolerant control technology: the demagnetizing characteristic extraction lacks theoretical basis, most methods select 5 times or 7 times of harmonic waves empirically, and the mapping relation between the harmonic waves and the geometric topology of the motor is not established; the signal processing dimension is unreasonable, namely, the demagnetizing harmonic wave is directly analyzed in a three-phase static coordinate system, the characteristic energy is dispersed, and the cross coupling interference is easy to occur; diagnosis and control disjoint that even if demagnetization is recognized, it is difficult to generate a compensation instruction matching with the actual demagnetization spatial distribution; The generalization capability is weak, and the same algorithm is difficult to be used for PMSM with different slot pole combinations of 12s/10p, 9s/8p and the like. Therefore, a complete system closed-loop control method capable of automatically adapti