Permanent magnet motor local demagnetization fault-tolerant control method and system based on parameter robust model predictive control

Abstract

The application provides a permanent magnet motor local demagnetization fault tolerance control method and system based on a parameter robust model predictive control, and is characterized in that the permanent magnet motor model after local demagnetization fault is subjected to super-localization processing to obtain a super-local model; a generalized proportional integral observer is designed to estimate and feed forward compensate the lumped disturbance caused by demagnetization in real time; the voltage and current data of adjacent sampling periods are used as real-time updates of the super-local model gain, the estimated values of the current and disturbance are obtained from the generalized proportional integral observer, and input gain online self-tuning is realized. The motor model after the fault is subjected to super-localization, the disturbance and high-order derivatives introduced by the fault are unbiasedly estimated through the generalized proportional integral demagnetization disturbance observer, and through the input gain self-tuning algorithm, the input gain in the super-local model can be calculated in real time only by using the voltage and current data of two continuous periods, so that the motor operation quality and the robust performance after the fault are improved.

Description

Technical Field

[0001] This invention belongs to the field of fault-tolerant control of permanent magnet motors, and in particular relates to a fault-tolerant control method and system for local demagnetization faults of permanent magnet motors based on parameter robust model predictive control. Background Technology

[0002] After prolonged operation, the magnetic properties of the rotor permanent magnets in a permanent magnet synchronous motor are affected by various factors, including operating temperature, external magnetic field, and service time, with temperature having a particularly significant impact. When the operating temperature exceeds the Curie temperature, the permanent magnets will experience irreversible loss of magnetism, ultimately leading to localized or uniform demagnetization faults. Compared to uniform demagnetization faults, localized demagnetization faults are more severe. They not only cause harmonic disturbances in the magnetic flux linkage but also induce perturbations in all motor parameters, inevitably reducing the system's robustness and control accuracy. In severe cases, this can even lead to system instability or damage.

[0003] Current classical demagnetization fault-tolerant control algorithms primarily focus on uniform demagnetization, with less attention paid to localized demagnetization faults. Furthermore, existing algorithms cannot eliminate their dependence on motor parameters, thus failing to effectively address parameter perturbations caused by localized demagnetization faults, and consequently, parameter robustness is difficult to guarantee. Summary of the Invention

[0004] This invention proposes a fault-tolerant control method and system for local demagnetization faults of permanent magnet motors based on parametric robust model predictive control. The robust deadbeat predictive current control strategy based on a hyperlocal model optimizes the operating quality of the motor after a local demagnetization fault.

[0005] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0006] A fault-tolerant control method for local demagnetization faults in permanent magnet motors based on parametric robust model predictive control includes:

[0007] S1. Perform hyperlocalization processing on the permanent magnet motor model after local demagnetization fault to obtain a hyperlocal model;

[0008] S2. Design a generalized proportional-integral observer to estimate and feedforward compensation for lumped disturbances caused by demagnetization in real time.

[0009] S3. Utilize voltage and current data from adjacent sampling periods as real-time updates to the gain of the hyperlocal model, obtain estimates of current and disturbance from the generalized proportional-integral observer, and achieve online self-tuning of the input gain.

[0010] Furthermore, step S1 specifically includes:

[0011] S101. Analyze the no-load air gap magnetic flux density of the permanent magnet motor prototype to obtain the three-phase magnetic flux density after a local demagnetization fault. A - B - C Motor flux linkage equations in natural coordinate system;

[0012] S102. Reduce the fractional fault harmonics in the air gap flux density to the 7th / 5th order, and retain the 3rd and 5th harmonics with larger amplitudes for the integer harmonics. Then, transform the motor flux linkage equation from step S101 to... dq In the coordinate system, after obtaining the local demagnetization fault d - q The motor model in the coordinate system is then transformed into state-space equations.

[0013] S103. Perform hyperlocalization on the state-space equations to obtain a hyperlocal model. ;in The state vector represents the state-space equation. The input vector represents the state-space equation. This represents the input gain of the hyperlocal model. F This represents the unknown perturbation in the hyperlocal model.

[0014] Furthermore, step S2 specifically includes:

[0015] Constructing a generalized proportional-integral observer pair F Estimation and current prediction are performed to achieve model-free, deadbeat-free predictive current control under local demagnetization faults;

[0016] ;

[0017] In the formula, ,in for The estimated value, Unknown perturbations of the hyperlocal model F and m The estimated value of the second derivative, is the gain coefficient of the generalized proportional-integral observer.

[0018] Preferably, the gain coefficient of the generalized proportional-integral observer is tuned according to the pole placement strategy.

[0019] Furthermore, step S3 specifically includes:

[0020] S301. Discretize the hyperlocal model and assume a perturbation. F The input gain of the hyperlocal model remains constant over two consecutive control cycles. Perform real-time estimation;

[0021] S302, Consider one-shot delay compensation. d - q Shaft reference voltage Represented as:

[0022] ;

[0023] In the formula, For the first k Reference current at time , This is the estimated state vector value at time k+1. This is the estimated disturbance value at time k+1. and These are obtained from the generalized proportional-integral observer, respectively. T s For the system's control cycle, Input gain The estimated value at time k;

[0024] S303. The calculated reference voltage is input into the space vector pulse width modulation. The modulation signal is constructed by different combinations of voltage vectors and driven in a pre-designed order to drive the inverter power devices, thereby obtaining a sinusoidal current in the permanent magnet motor stator.

[0025] In another aspect, this invention proposes a fault-tolerant control system for local demagnetization faults in permanent magnet motors based on parametric robust model predictive control, comprising:

[0026] Hyperlocal model module: Performs hyperlocalization processing on the permanent magnet motor model after local demagnetization fault to obtain a hyperlocal model;

[0027] Observer module: Design a generalized proportional-integral observer to estimate and feedforward compensation for lumped disturbances caused by demagnetization in real time;

[0028] Control module: Utilizes voltage and current data from adjacent sampling periods as real-time updates to the gain of the hyperlocal model, obtains estimates of current and disturbance from the generalized proportional-integral observer, and achieves online self-tuning of the input gain.

[0029] Furthermore, the hyperlocal model module includes:

[0030] Motor flux linkage equation unit: Analyze the no-load air gap magnetic flux density of a permanent magnet motor prototype to obtain the three-phase flux density after a local demagnetization fault. A - B - C Motor flux linkage equations in natural coordinate system;

[0031] State-space equation unit: The fractional fault harmonics in the air gap flux density are reduced to the 7th / 5th order, while the integer harmonics retain the 3rd and 5th harmonics with larger amplitudes. The motor flux linkage equation from step S101 is then transformed to... dq In the coordinate system, after obtaining the local demagnetization fault d - q The motor model in the coordinate system is then transformed into state-space equations.

[0032] Model unit: The state-space equations are hyperlocalized to obtain a hyperlocal model. ;in The state vector represents the state-space equation. The input vector represents the state-space equation. This represents the input gain of the hyperlocal model. F This represents the unknown perturbation in the hyperlocal model.

[0033] Furthermore, the observer module includes:

[0034] Constructing a generalized proportional-integral observer pair F Estimation and current prediction are performed to achieve model-free, deadbeat-free predictive current control under local demagnetization faults;

[0035] ;

[0036] In the formula, ,in for The estimated value, Unknown perturbations of the hyperlocal model F and m The estimated value of the second derivative, is the gain coefficient of the generalized proportional-integral observer.

[0037] Preferably, the gain coefficient of the generalized proportional-integral observer is tuned according to the pole placement strategy.

[0038] Furthermore, the control module includes:

[0039] The hyperlocal model is discretized, and a perturbation is assumed. F The input gain of the hyperlocal model remains constant over two consecutive control cycles. Perform real-time estimation;

[0040] Consider one-shot delay compensation, d - q Shaft reference voltage Represented as:

[0041] ;

[0042] In the formula, For the first k Reference current at time , This is the estimated state vector value at time k+1. This is the estimated disturbance value at time k+1. and These are obtained from the generalized proportional-integral observer, respectively. T s For the system's control cycle, Input gain The estimated value at time k;

[0043] The calculated reference voltage is input into the space vector pulse width modulation, and a modulation signal is constructed by different combinations of voltage vectors. The signal is then used to drive the inverter power devices in a pre-designed sequence, thereby obtaining a sinusoidal current in the permanent magnet motor stator.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] This invention addresses the changes in motor characteristics following a local demagnetization fault. First, it performs hyperlocalization on the motor model after the fault, thereby eliminating the control algorithm's dependence on motor parameters. Second, a generalized proportional-integral demagnetization disturbance observer is designed to perform unbiased estimation of the disturbance introduced by the fault and its higher-order derivatives, thus improving the motor's operational quality. Finally, an input gain self-tuning algorithm is designed, which can calculate the input gain in the hyperlocal model in real time using only two consecutive cycles of voltage and current data, further enhancing the system's robustness after the fault. Attached Figure Description

[0046] Figure 1 This is an analysis diagram of the unloaded air gap magnetic flux density of Embodiment 1 of the present invention;

[0047] in Figure 1 (a) is a waveform diagram of the air gap magnetic flux density of a healthy motor. Figure 1 (b) is the air gap flux density spectrum analysis diagram of the healthy motor. Figure 1 (c) is a waveform diagram of the air gap flux density of a motor with a partial demagnetization fault. Figure 1 (d) is a spectrum analysis diagram of the air gap magnetic flux density of a motor with a partial demagnetization fault;

[0048] Figure 2 This is the demagnetization disturbance sensing based on a generalized proportional-integral observer in Embodiment 1 of the present invention;

[0049] Figure 3 This is a general block diagram of the method described in Embodiment 1 of the present invention;

[0050] Figure 4 This is a schematic diagram of the system structure of Embodiment 2 of the present invention. Detailed Implementation

[0051] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0052] Example 1:

[0053] The fault-tolerant control method for local demagnetization faults of permanent magnet motors based on parametric robust model predictive control proposed in this embodiment includes the following steps:

[0054] Step 1: Analyze the unloaded air gap magnetic flux density of the prototype, such as... Figure 1 (a)- Figure 1 As shown in (d), Figure 1 (a) is a waveform diagram of the air gap magnetic flux density of a healthy motor. Figure 1 (b) is the air gap flux density spectrum analysis diagram of the healthy motor. Figure 1 (c) is a waveform diagram of the air gap flux density of a motor with a partial demagnetization fault. Figure 1 (d) is the air gap flux density spectrum analysis diagram of the motor with a partial demagnetization fault. The analysis results show that once a partial demagnetization fault occurs in the rotor permanent magnet, the amplitude of the air gap flux density at the corresponding location will significantly decrease, and its original symmetry will be disrupted. Furthermore, the partial demagnetization fault will also introduce [a negative effect] into the spectrum of the air gap flux density. ( (where is a positive integer) fractional harmonic components, where The changes in fractional harmonics within the range are particularly pronounced, while the changes in integer harmonics are not significant. These harmonic characteristics of the air gap flux density will be directly reflected in the motor flux linkage, and can be expressed as:

[0055] (1)

[0056] In the formula, For the first magnetic linkage Second harmonic amplitude and These are the effective length of the motor stator core and the pole pitch, respectively. For the first Harmonic amplitude of secondary air gap magnetic flux density N The number of turns in series per phase of the stator winding. For the first Winding coefficients of the subharmonics.

[0057] Furthermore, after a localized demagnetization fault, the three phases A - B - C The flux linkage equation in the natural coordinate system can be described as:

[0058] (2)

[0059] In the formula, , and This refers to the flux linkage harmonic disturbance generated in the natural coordinate system after a local demagnetization fault. For electrical angle, These represent the harmonic amplitudes of the three-phase flux linkage. This represents the number of pole pairs of the motor.

[0060] To balance the accuracy and complexity of the model, this embodiment focuses on the fractional fault harmonics in the air gap flux density when establishing the local demagnetization model, and takes them up to the 7th / 5th order. Furthermore, given that the amplitude changes of integer harmonics are not significant after the fault, only the 3rd and 5th harmonics with larger amplitudes are retained in the modeling. Based on coordinate transformation theory, equation (2) is transformed to... d - q In the coordinate system, that is:

[0061]

[0062] (3)

[0063] In the formula, and After a localized demagnetization fault d - q The total harmonic disturbance component generated in the coordinate system. This represents the fundamental component of the magnetic flux linkage in a permanent magnet. and These are caused by local demagnetization faults. d - q The shaft flux linkage fractional harmonic disturbance component. In practical engineering systems, since the time constant of the electrical system is much smaller than that of the mechanical system, it is usually assumed that... , .

[0064] Furthermore, local demagnetization faults caused by high-temperature environments can also lead to perturbations in inductance and resistance. Taking these factors into account, and based on equation (3), the post-fault conditions can be obtained. d - q The motor model in the coordinate system, namely:

[0065] (4)

[0066] In the formula, , , , , and After each failure d - q Voltage, current, and flux linkage components in the coordinate system. ω is the electric angular velocity. and These are the measured values ​​of inductance and resistance, respectively. L s0 and R s0 These are the nominal values ​​of inductance and resistance, respectively, ∆ L s and ∆ R s These are the inductance and resistance perturbations, respectively. , is a permanent magnet flux linkage.

[0067] Rearranging equation (4) into a state-space equation, it can be described as follows:

[0068] (5)

[0069] In the formula, , and These are the state vector, input vector, and output vector of the permanent magnet motor control system, respectively. The state vector is the current, and the input vector is the voltage. State vector The derivative; is the flux linkage vector of the permanent magnet. C It is a 2×2 identity matrix. A , B and D They are respectively:

[0070] , , .

[0071] Caused by local demagnetization fault dq Shaft disturbance can be described as:

[0072] ;

[0073] According to equation (4), the electromagnetic torque of the motor after the fault is T e It can be represented as:

[0074] (6)

[0075] Step 2): Perform hyperlocalization on equation (5) to obtain:

[0076] (7)

[0077] In the formula, The input gain in a hyperlocal model is typically set to 1. . The unknown perturbation in the hyperlocal model can be represented as:

[0078] (8)

[0079] In the formula, This is the nominal value of the magnetic flux linkage of the permanent magnet.

[0080] Step 3: Based on equations (7) and (8), construct the generalized proportional integral observer pair. F Estimation and current prediction are performed to achieve model-free, deadbeat-free predictive current control under local demagnetization faults, i.e.:

[0081] (9)

[0082] The superscript in the formula " represents the derivative, .in, for The estimated value, Disturbance F The estimated value of and the estimated values ​​of its first to m-1th derivatives, i.e. g 2 For disturbance F The estimated value, g 3 For disturbance F The estimated value of the first derivative, g 4 For disturbance F The estimated value of the second derivative, ... g m+1 For disturbance F The estimated value of the m-1th derivative. is the gain coefficient of the generalized proportional-integral observer.

[0083] Furthermore, based on the pole configuration strategy, it can be... Adjusted to:

[0084] (10)

[0085] In the formula, This represents the bandwidth of the generalized proportional-integral (PII) observer. The block diagram of the generalized proportional-integral (PII) observer is shown below. Figure 2 As shown in the figure. 1 / s This indicates integration.

[0086] Step 4: After a partial demagnetization fault occurs in the motor, the faulty permanent magnet and the corresponding air gap magnetic field will weaken, thus disrupting the uniform distribution of the air gap magnetic field. This non-uniformity will cause changes in magnetic flux density, and given the linear relationship between magnetic flux density and reluctance, it will also induce reluctance fluctuations, ultimately causing a change in the average inductance value. Therefore, it is necessary to adjust the input gain. Real-time estimation is performed to further improve the motor's operating performance after a fault.

[0087] Discretizing equation (7) yields:

[0088] (11)

[0089] In the formula, T s For the system's control cycle, and The first k The state vector (current) and disturbance at any given time. and The first The state vector (current) and input vector (voltage) at each moment.

[0090] Given that the system's sampling period is very short, the disturbance can be considered as F It remains unchanged over two consecutive control cycles, therefore the two consecutive control cycles' d - q Subtracting the shaft currents yields:

[0091] (12)

[0092] In the formula, , , For the first The input vector (voltage) at time t. Equation (12) can be further expressed as:

[0093] (13)

[0094] In the formula, Input gain In the The estimated value of the time. Subtracting equation (12) from equation (13), we get:

[0095] (14)

[0096] From equation (14), it can be seen that if ,but ;like ,but Therefore, it can be Using symbolic functions Alternative. However, when When the change is minute, the sign function switches frequently between positive and negative signs, causing the estimated input gain to fluctuate wildly and preventing it from converging to the true value. Therefore, to overcome this problem, this embodiment uses the hyperbolic tangent function tanh instead. At this time, the input gain... The estimated value at time k It can be represented as:

[0097] (15)

[0098] The superscript "ˆ" in the formula represents the estimated value. In order to be in Gain coefficient designed during the observation process. The larger the value, the faster the dynamic response, but the greater the impact of noise on the estimation accuracy. Conversely, The smaller the value, the higher the estimation accuracy, but the slower the dynamic response speed will be.

[0099] Step 5: After considering the one-shot delay compensation, d - q Shaft reference voltage It can be represented as:

[0100] (16)

[0101] In the formula, For the first k Reference current at time , This is the estimated value of the state vector (current) at time k+1. This is the estimated disturbance value at time k+1. and These values ​​can be obtained from a generalized proportional-integral observer. Subsequently, the calculated reference voltage is input into a space vector pulse width modulation (SVPWM) system. Different combinations of voltage vectors are used to construct a modulation signal, which drives the inverter power devices in a pre-designed sequence, thereby obtaining a sinusoidal current in the motor stator.

[0102] The overall block diagram of the method described in this embodiment is as follows: Figure 3 As shown. Figure 3 In this context, PI represents a proportional-integral controller. V dc Indicates the DC bus voltage. e Electric angular velocity, nThe values ​​indicate the motor speed. The superscripts "ˆ" and "*" represent estimated and reference values, respectively. The table below shows... dq "", " ABC "Represents a two-phase rotating coordinate system, a two-phase stationary coordinate system, and a three-phase natural coordinate system, respectively."

[0103] As described above, this embodiment proposes a fault-tolerant control method for local demagnetization faults of permanent magnet motors based on parameter robust model predictive control. This method can achieve disturbance-free operation under local demagnetization faults and effectively suppress parameter perturbations, thereby improving the system control accuracy under fault conditions.

[0104] Example 2:

[0105] This embodiment proposes a fault-tolerant control system for local demagnetization faults in permanent magnet motors based on parametric robust model predictive control, such as... Figure 4 As shown, it includes:

[0106] Hyperlocal model module: Performs hyperlocalization processing on the permanent magnet motor model after local demagnetization fault to obtain a hyperlocal model;

[0107] Observer module: Design a generalized proportional-integral observer to estimate and feedforward compensation for lumped disturbances caused by demagnetization in real time;

[0108] Control module: Utilizes voltage and current data from adjacent sampling periods as real-time updates to the gain of the hyperlocal model, obtains estimates of current and disturbance from the generalized proportional-integral observer, and achieves online self-tuning of the input gain.

[0109] The hyperlocal model module includes:

[0110] Motor flux linkage equation unit: Analyze the no-load air gap magnetic flux density of a permanent magnet motor prototype to obtain the three-phase flux density after a local demagnetization fault. A - B - C Motor flux linkage equations in natural coordinate system;

[0111] State-space equation unit: The fractional fault harmonics in the air gap flux density are reduced to the 7th / 5th order, while the integer harmonics retain the 3rd and 5th harmonics with larger amplitudes. The motor flux linkage equation from step S101 is then transformed to... dq In the coordinate system, after obtaining the local demagnetization fault d - q The motor model in the coordinate system is then transformed into state-space equations.

[0112] Model unit: The state-space equations are hyperlocalized to obtain a hyperlocal model. ;in The state vector represents the state-space equation. The input vector represents the state-space equation. This represents the input gain of the hyperlocal model. F This represents the unknown perturbation in the hyperlocal model.

[0113] The observer module includes:

[0114] Constructing a generalized proportional-integral observer pair F Estimation and current prediction are performed to achieve model-free, deadbeat-free predictive current control under local demagnetization faults;

[0115] ;

[0116] In the formula, ,in for The estimated value, Unknown perturbations of the hyperlocal model F The estimated value and F Estimates of the first to m-1th derivatives, The gain coefficient of the generalized proportional-integral observer is indicated by the superscript " " in the formula. "" represents the derivative. The gain coefficient of the generalized proportional-integral observer is tuned according to the pole placement strategy.

[0117] The control module includes:

[0118] The hyperlocal model is discretized, and a perturbation is assumed. F The input gain of the hyperlocal model remains constant over two consecutive control cycles. Perform real-time estimation;

[0119] Consider one-shot delay compensation, d - q Shaft reference voltage Represented as:

[0120] ;

[0121] In the formula, For the first k Reference current at time , This is the estimated state vector value at time k+1. This is the estimated disturbance value at time k+1. and These are obtained from the generalized proportional-integral observer, respectively. T s For the system's control cycle, Input gain The estimated value at time k;

[0122] The calculated reference voltage is input into the space vector pulse width modulation, and a modulation signal is constructed by different combinations of voltage vectors. The signal is then used to drive the inverter power devices in a pre-designed sequence, thereby obtaining a sinusoidal current in the permanent magnet motor stator.

[0123] The fault-tolerant control system for partial demagnetization faults of permanent magnet motors based on parameter robust model predictive control proposed in this embodiment can achieve the fault-tolerant control method for partial demagnetization faults of permanent magnet motors based on parameter robust model predictive control as described in Embodiment 1, and has the same technical effect as Embodiment 1.

[0124] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A fault-tolerant control method for local demagnetization faults in permanent magnet motors based on parametric robust model predictive control, characterized in that, include: S1. Perform hyperlocalization processing on the permanent magnet motor model after local demagnetization fault to obtain a hyperlocal model; specifically including: S101. Analyze the no-load air gap magnetic flux density of the permanent magnet motor prototype to obtain the three-phase magnetic flux density after a local demagnetization fault. A - B - C Motor flux linkage equations in natural coordinate system; S102. Reduce the fractional fault harmonics in the air gap flux density to the 7th / 5th order, and retain the 3rd and 5th harmonics with larger amplitudes for the integer harmonics. Transform the motor flux linkage equation from step S101 to... dq In the coordinate system, after obtaining the local demagnetization fault d - q The motor model in the coordinate system is then transformed into state-space equations. S103. Perform hyperlocalization on the state-space equations to obtain a hyperlocal model. ;in The state vector represents the state-space equation. The input vector represents the state-space equation. This represents the input gain of the hyperlocal model. F This represents the unknown perturbation in the hyperlocal model; the state vector refers to the current, and the input vector refers to the voltage. S2. Design a generalized proportional-integral observer to estimate and feedforward compensate for lumped disturbances caused by demagnetization in real time; specifically, this includes: constructing a generalized proportional-integral observer to handle unknown disturbances in the hyperlocal model. F Estimation and current prediction are performed to achieve model-free, deadbeat-free predictive current control under local demagnetization faults; ； In the formula, ,in for The estimated value, Unknown perturbations of the hyperlocal model F The estimated value and F Estimates of the first to m-1th derivatives, The gain coefficient of the generalized proportional-integral observer is indicated by the superscript "" in the formula. " denotes the derivative; S3. Utilize voltage and current data from adjacent sampling periods as real-time updates to the gain of the hyperlocal model, obtain estimates of current and disturbance from the generalized proportional-integral observer, and achieve online self-tuning of the input gain.

2. The fault-tolerant control method for local demagnetization faults of permanent magnet motors based on parametric robust model predictive control according to claim 1, characterized in that, The gain coefficient of the generalized proportional-integral observer is tuned according to the pole placement strategy.

3. The fault-tolerant control method for local demagnetization faults of permanent magnet motors based on parametric robust model predictive control according to claim 1, characterized in that, Step S3 specifically includes: S301. Discretize the hyperlocal model and assume a perturbation. F The input gain of the hyperlocal model remains constant over two consecutive control cycles. Perform real-time estimation; S302, Consider one-shot delay compensation. d - q Shaft reference voltage Represented as: ； In the formula, For the first k Reference current at time , This is the estimated state vector value at time k+1. This is the estimated disturbance value at time k+1. and These are obtained from the generalized proportional-integral observer, respectively. T s For the system's control cycle, Input gain The estimated value at time k; S303. The calculated reference voltage is input into the space vector pulse width modulation. The modulation signal is constructed by different combinations of voltage vectors and driven in a pre-designed order to drive the inverter power devices, thereby obtaining a sinusoidal current in the permanent magnet motor stator.

4. A fault-tolerant control system for partial demagnetization faults of a permanent magnet motor based on parametric robust model predictive control, characterized in that, include: Hyperlocal model module: Performs hyperlocalization processing on the permanent magnet motor model after local demagnetization fault to obtain a hyperlocal model; The hyperlocal model module includes: a motor flux linkage equation unit: analyzing the no-load air gap flux density of a permanent magnet motor prototype to obtain the three-phase values ​​after a local demagnetization fault. A - B - C Motor flux linkage equations in natural coordinate system; State space equation unit: fractional fault harmonics in the air gap flux density are reduced to 7 / 5, and integer harmonics are retained, with larger amplitudes in the 3rd and 5th harmonics. The motor flux linkage equations in step S101 are then transformed to... dq In the coordinate system, after obtaining the local demagnetization fault d - q The motor model in the coordinate system is arranged into a state-space equation; Model element: The state-space equation is hyperlocalized to obtain a hyperlocal model. ;in The state vector represents the state-space equation. The input vector represents the state-space equation. This represents the input gain of the hyperlocal model. F This represents the unknown perturbation in the hyperlocal model; the state vector refers to the current, and the input vector refers to the voltage. Observer Module: Designs a generalized proportional-integral observer to estimate and feedforward compensate for lumped perturbations caused by demagnetization in real time; specifically including: constructing the generalized proportional-integral observer for unknown perturbations in the hyperlocal model. F Estimation and current prediction are performed to achieve model-free, deadbeat-free predictive current control under local demagnetization faults; ； In the formula, ,in for The estimated value, Unknown perturbations of the hyperlocal model F The estimated value and F Estimates of the first to m-1th derivatives, The gain coefficient of the generalized proportional-integral observer is indicated by the superscript "" in the formula. " denotes the derivative; Control module: Utilizes voltage and current data from adjacent sampling periods as real-time updates to the gain of the hyperlocal model, obtains estimates of current and disturbance from the generalized proportional-integral observer, and achieves online self-tuning of the input gain.

5. The fault-tolerant control system for partial demagnetization faults of permanent magnet motors based on parametric robust model predictive control according to claim 4, characterized in that, The gain coefficient of the generalized proportional-integral observer is tuned according to the pole placement strategy.

6. The fault-tolerant control system for partial demagnetization faults of permanent magnet motors based on parametric robust model predictive control according to claim 4, characterized in that, The control module includes: The hyperlocal model is discretized, and a perturbation is assumed. F The input gain of the hyperlocal model remains constant over two consecutive control cycles. Perform real-time estimation; Consider one-shot delay compensation, d - q Shaft reference voltage Represented as: ； In the formula, For the first k Reference current at time , This is the estimated state vector value at time k+1. This is the estimated disturbance value at time k+1. and These are obtained from the generalized proportional-integral observer, respectively. T s For the system's control cycle, Input gain The estimated value at time k; The calculated reference voltage is input into the space vector pulse width modulation, and a modulation signal is constructed by different combinations of voltage vectors. The signal is then used to drive the inverter power devices in a pre-designed sequence, thereby obtaining a sinusoidal current in the permanent magnet motor stator.

Images