Permanent magnet motor predictive torque control method and apparatus with sine wave filter

Abstract

The application discloses a permanent magnet motor predictive torque control method and equipment with a sine wave filter, and belongs to the field of permanent magnet motor control. A discrete dynamic model is established according to the current state value of the permanent magnet motor; an electromagnetic torque and stator flux amplitude prediction model is constructed by using the discrete dynamic model; the electromagnetic torque given value is obtained, and the stator flux amplitude given value is obtained based on the maximum torque current ratio formula; the minimum beat predictive control is implemented on the stator flux amplitude and the electromagnetic torque respectively, and the inverter voltage given value is solved; then the optimal voltage vector and the action duty cycle are determined according to the inverter voltage given value, the control pulse signal is generated according to the symmetric 5-section pulse mode and is applied to the inverter, and the permanent magnet motor predictive torque control with the sine wave filter is realized. The application does not contain any to-be-designed parameters, avoids the complex parameter debugging process of the permanent magnet motor, and can effectively reduce the torque ripple and improve the steady-state control precision.

Description

Technical Field

[0001] This invention relates to a predictive torque control method and device for a permanent magnet motor with a sine wave filter, belonging to the field of permanent magnet motor control. Background Technology

[0002] In applications involving high-power permanent magnet motors driven by long cables, a sinusoidal wave filter is commonly incorporated into the inverter output to reduce overvoltage caused by traveling wave reflections from the long cable, lower harmonics, and extend motor lifespan. This creates a permanent magnet motor drive system with a sinusoidal wave filter. However, the filter inductor and capacitor of the introduced sinusoidal wave filter form a strong coupling relationship with the stator inductance of the permanent magnet motor, leading to an increased system order and complicating the design of the control strategy.

[0003] Traditional proportional-integral (PI) control schemes have significant limitations in multi-loop cascaded architectures: as the system order increases, the complexity of regulator parameter tuning grows exponentially, phase margin loss intensifies, and it becomes difficult to meet dynamic response requirements. Although model predictive control is considered an effective means of solving high-order system control due to its multi-objective optimization capabilities, existing multi-objective predictive control methods for permanent magnet motors with sinusoidal filters have the following drawbacks: First, multi-objective control mainly focuses on voltage and current, failing to achieve direct control of electromagnetic torque and flux linkage, resulting in low control accuracy of torque and flux linkage; second, the commonly used finite control set prediction strategy leads to large steady-state ripple in torque and flux linkage, and the non-fixed switching frequency makes sinusoidal filter design difficult; third, the inevitable introduction of multiple weighting coefficients makes adjustment difficult and fails to achieve optimal control performance.

[0004] The prior art disclosed in CN118826554A describes a torque prediction control method for a permanent magnet synchronous motor, specifically including: acquiring the three-phase current of the synchronous motor at time kTs and obtaining the current components through coordinate transformation; obtaining a given torque value based on the motor's electrical angle and the given speed value of the synchronous motor; obtaining the predicted flux linkage and torque values ​​corresponding to each voltage vector at the next time step based on the voltage equation and torque equation in the coordinate system; obtaining the flux linkage value function and torque value function corresponding to each voltage vector based on the predicted values ​​of flux linkage and torque, as well as the given torque value and given flux linkage value; selecting a first voltage vector from the voltage vectors based on the value function, and selecting the optimal voltage vector from the first voltage vector based on the comprehensive optimization principle and / or the value function numerical compensation principle. However, the torque flux linkage control accuracy and motor operation reliability are poor. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this paper provides a predictive torque control method and device for permanent magnet motors with a sinusoidal filter. By establishing predictive equations for electromagnetic torque and stator flux linkage amplitude, the method achieves minimum-step predictive control of stator flux linkage amplitude and electromagnetic torque using these equations, effectively improving the steady-state control accuracy of torque and flux linkage and reducing ripple. Furthermore, this method does not involve any weighting coefficients, eliminating the complex weighting coefficient adjustment process of existing predictive control methods. In addition, this method uses space vector modulation to generate switching signals with a fixed switching frequency, solving the problem of switching frequency variation in existing predictive control methods.

[0006] To achieve the above technical objectives, this invention discloses a predictive torque control method for a permanent magnet motor with a sinusoidal filter, comprising the following steps:

[0007] A predictive torque control method for a permanent magnet motor with a sinusoidal filter, characterized by the following steps:

[0008] S1. Collect the current state value of the permanent magnet motor system and convert the state value into a variable in the dq coordinate system;

[0009] S2. Establish a discrete dynamic model of the permanent magnet motor with a sinusoidal filter in the dq coordinate system with respect to the inverter current, capacitor voltage and stator current;

[0010] S3. Establish electromagnetic torque prediction model and stator flux amplitude prediction model;

[0011] S4. Obtain the electromagnetic torque setpoint based on the proportional-integral controller, and then obtain the stator flux amplitude setpoint based on the maximum torque-current ratio formula.

[0012] S5. Implement minimum-step predictive control for the stator flux amplitude and electromagnetic torque respectively, and obtain the inverter voltage setpoint.

[0013] S6. Determine the optimal voltage vector and duty cycle based on the inverter voltage setpoint, generate control pulse signals and apply them to the inverter to achieve predictive torque control of the permanent magnet motor with a sine wave filter.

[0014] Furthermore, the current state values ​​of the permanent magnet motor system include: the output voltage of the three-phase inverter. , , Three-phase inverter current , , Three-phase capacitor voltage , , 、 Three-phase stator current , , Rotor electrical angle The Parker transform is used to convert the collected three-phase inverter output voltage, three-phase inverter current, three-phase capacitor voltage, and three-phase stator current into variables in the dq coordinate system, expressed as: Inverter output voltage , Inverter current , capacitor voltage , and stator current , .

[0015] Furthermore, the discrete dynamic model of the permanent magnet motor with a sinusoidal filter in the dq coordinate system, established using the improved Euler method, with respect to the inverter current, capacitor voltage, and stator current, is as follows:

[0016] ,

[0017] In the formula, the subscript k and k +1 respectively represent k The collected values ​​at each moment and k The predicted value at time +1, Indicates the sampling period. Indicates the rotor's electrical angular velocity; , , , and These represent the filter inductance, filter capacitor, motor stator inductance, stator resistance, and rotor permanent magnet flux linkage of the sine wave filter, respectively.

[0018] Furthermore, based on the discrete dynamics model in the dq coordinate system, the electromagnetic torque prediction model is established as follows:

[0019] ,

[0020] , , ,

[0021] , ,

[0022] In the formula, This represents the predicted value of electromagnetic torque. This represents the number of pole pairs in a permanent magnet motor. , , , , All are coefficient matrices, specifically the discrete dynamics model. The coefficient matrix of each term in the specific expression obtained after substituting into the electromagnetic torque prediction model and expanding it.

[0023] Based on the discrete dynamic model of the dq coordinate system, the stator flux linkage amplitude prediction model is expressed as follows:

[0024] ,

[0025] , , , ,

[0026] ,

[0027] In the formula, This is the predicted value of the stator flux linkage amplitude. , The inverter output voltage is in the dq coordinate system. All are coefficient matrices, specifically representing the discrete dynamics model. and The coefficient matrix of each term in the specific expression obtained after substituting into the stator flux linkage amplitude prediction model.

[0028] Furthermore, the process of obtaining the electromagnetic torque setpoint based on the proportional-integral controller, and then obtaining the stator flux linkage amplitude setpoint based on the maximum torque-to-current ratio formula, is as follows:

[0029] Based on rotor electric angular velocity reference and electric angular velocity feedback The reference value of the electromagnetic torque is obtained by proportional-integral control of the inter-error. :

[0030] ,

[0031] In the formula, 1 / s represents the integral operation. and These are the proportional and integral parameters of the PI controller. and Determined based on the frequency domain response characteristics of a typical second-order system and the zero-pole cancellation principle;

[0032] Using the stator flux amplitude setpoint The reference value for stator flux linkage amplitude is derived from the maximum torque-current ratio formula. : .

[0033] Furthermore, minimum-step predictive control is implemented on the electromagnetic torque and stator flux linkage amplitude respectively to obtain the inverter voltage setpoint in the dq coordinate system, specifically including:

[0034] Based on the principle of minimum-step predictive control of electromagnetic torque, the reference value of the electromagnetic torque is set equal to its predicted value at the next moment. The q-axis component of the inverter voltage setpoint in the dq coordinate system is obtained using the electromagnetic torque prediction equation. :

[0035] ,

[0036] Based on the principle of minimum-step predictive control of stator flux linkage amplitude, let the stator flux linkage amplitude reference value... The d-axis component of the inverter voltage setpoint in the dq coordinate system is obtained using the stator flux linkage amplitude prediction equation. :

[0037] ,

[0038] Through the d-axis component expression and q-axis component expression The inverter voltage setpoint is formed in the dq coordinate system.

[0039] Furthermore, the specific expression for converting the inverter voltage setpoint in the dq coordinate system to the inverter voltage setpoint in the αβ coordinate system using the inverse Park transform is as follows:

[0040] ,

[0041] In the formula, and These are the inverter voltage setpoints in the αβ coordinate system, respectively. Is k The electric angle of the motor rotor is detected at any time, and the iPark matrix is ​​the inverse Parker transform matrix;

[0042] Based on the inverter voltage setpoint αβ axis components and And the spatial position angle is obtained from the arctangent function: ,judge The voltage vector sector is located, and the two adjacent non-zero voltage vectors within that sector are selected. , and zero vector As the voltage setpoint of the synthesized inverter The three optimal voltage vectors.

[0043] Furthermore, three optimal voltage vectors are defined. , and The corresponding duty cycles are respectively , and Then, the components of the combined voltage of the three optimal voltage vectors under these three duty cycles in the αβ coordinate system are:

[0044] ,

[0045] In the formula, and For the synthesized voltage α and β axis components; , and , These are the two selected non-zero voltage vectors. and αβ axis components; , Zero voltage vector αβ axis components;

[0046] Based on the principle of solving for the minimum value of a function, the synthesized voltage With predictive control reference instructions The sum of squares of the tracking errors of the α and β axis components affect the duty cycle. , By taking the partial derivatives, the three duty cycles corresponding to the three optimal voltage vectors are expressed as follows:

[0047] ,

[0048] Solving the above equation yields:

[0049] ,

[0050] In the formula, , and These represent the inverter voltage setpoints in the αβ coordinate system, respectively; , and These represent the three optimal voltage vectors respectively. , and The corresponding duty cycle.

[0051] Furthermore, the following symmetrical 5-segment pulse pattern is used to generate the control pulse signal for the inverter switching transistors to achieve a fixed switching frequency, as shown in the following expression:

[0052] ,

[0053] In the formula, the control pulse signal S consists of two optimal voltage vectors. , and It operates sequentially in a mirror-symmetric pattern. Represents the zero voltage vector. , , These represent the three optimal voltage vectors respectively. , and The corresponding duty cycle is half; the pulse signal S It is applied to inverter switching transistors to achieve predictive torque control of permanent magnet motors with sinusoidal wave filters.

[0054] A computer device includes a processor and a memory, the processor being electrically connected to the memory for storing instructions and data, and the processor being used to execute a predictive torque control method for a permanent magnet motor with a sinusoidal filter.

[0055] Beneficial Effects: The method provided by this invention achieves coordinated predictive control of electromagnetic torque and stator flux linkage amplitude, exhibiting better steady-state tracking accuracy compared to traditional predictive control methods. Furthermore, this invention's method does not contain any weighting factors, eliminating the complex parameter tuning process compared to traditional predictive control methods. Firstly, this invention establishes predictive equations for electromagnetic torque and stator flux linkage amplitude, and implements minimum-step predictive control of stator flux linkage amplitude and electromagnetic torque based on these equations, effectively improving control accuracy and reducing torque and flux linkage ripple. Secondly, this invention's method does not introduce any weighting coefficients, eliminating the cumbersome weighting coefficient adjustment process of existing multi-objective predictive control methods. Thirdly, this invention uses a symmetrical 5-segment pulse mode to generate control pulse signals for the inverter switching transistors, providing a fixed switching frequency and solving the problem of switching frequency variation in existing predictive control methods. Attached Figure Description

[0056] Figure 1 This is a schematic block diagram of the predictive torque control method for permanent magnet motors with a sine wave filter according to the present invention.

[0057] Figure 2 The waveforms of electromagnetic torque and stator flux linkage when the conventional model predictive control method in the embodiment reaches steady state are shown.

[0058] Figure 3 The diagram shows the electromagnetic torque and stator flux linkage waveforms when the method of the present invention reaches steady state in the embodiment.

[0059] Figure 4 The above is a dynamic waveform diagram of electromagnetic torque and stator flux linkage of the conventional model predictive control method in the embodiment when the system load changes abruptly.

[0060] Figure 5 The following is a dynamic waveform diagram of electromagnetic torque and stator flux linkage when the system load changes abruptly, as shown in the embodiment of the present invention. Detailed Implementation

[0061] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0062] Figure 1 This is a flowchart illustrating the permanent magnet motor system with a sinusoidal wave filter according to the present invention, and the corresponding predictive torque control method for the permanent magnet motor with a sinusoidal wave filter. The DC bus voltage is converted into an AC voltage square wave signal by a three-phase voltage source inverter, and then connected to the permanent magnet motor after passing through a sinusoidal wave filter. The three-phase inverter-side current, three-phase capacitor voltage, and three-phase stator current of the permanent magnet motor system with a sinusoidal wave filter are sampled. (See the diagram.) Indicates the DC bus voltage; This represents the output filter inductor of the inverter. This represents the filter capacitor. , , This indicates the current on the three-phase filter inverter side; , , Indicates the voltage of the three-phase capacitor; , , Indicates the three-phase stator current; Let k be the electrical angle of the permanent magnet motor rotor at the current time k. This is the reference value for the rotor's electric angular velocity at time k. This is the current electric angular velocity feedback value at time k. This is the reference value for the electromagnetic torque at time k. This is the reference value for the stator flux linkage at time k. , , They represent k Sampled values ​​of inverter current, capacitor voltage, and stator current in the αβ coordinate system at time 1; and They represent k Predicted values ​​of stator flux linkage and electromagnetic torque at time +1; and The inverter voltage setpoint; It is a spatial angular position angle; , and zero vector As the voltage setpoint of the synthesized inverter The three optimal voltage vectors; , and These are the three optimal voltage vectors. , and The function of duty cycle; This refers to the control pulse signal for the inverter's switching transistors.

[0063] A predictive torque control method for a permanent magnet motor with a sinusoidal filter includes the following steps:

[0064] The Parker transform is used to convert the collected three-phase inverter current, three-phase capacitor voltage, and three-phase stator current into variables in the dq coordinate system: inverter current. , capacitor voltage , and stator current , A discrete dynamic model of the permanent magnet motor in the dq coordinate system with a sinusoidal filter is established using the improved Euler method. The discrete dynamic model in the dq coordinate system is as follows:

[0065] ,

[0066] In the formula, the subscript k and k +1 respectively represent k The collected values ​​at each moment and k The predicted value at time +1, Indicates the sampling period. Indicates the rotor's electrical angular velocity; , , , and These represent the filter inductance, filter capacitor, motor stator inductance, stator resistance, and rotor permanent magnet flux linkage of the sine wave filter, respectively.

[0067] The specific steps for establishing the electromagnetic torque prediction equation and the stator flux linkage amplitude prediction equation include:

[0068] First, the electromagnetic torque prediction model is established based on the discrete dynamics model:

[0069] ,

[0070] In the formula, This represents the predicted value of electromagnetic torque. p Let be the number of pole pairs of the motor, and the expressions for each coefficient matrix are as follows:

[0071] , , ,

[0072] , ,

[0073] The matrices above represent: the discrete dynamics model The coefficient matrix of each term in the specific expression obtained after substituting into the electromagnetic torque prediction model;

[0074] Secondly, a stator flux linkage amplitude prediction model is established based on the discrete dynamics model:

[0075] ,

[0076] In the formula, This is the predicted value of the stator flux linkage amplitude. , The inverter output voltage is in the dq coordinate system.

[0077] The coefficient matrices are as follows: , , , ,

[0078] ,

[0079] The matrices above represent: the discrete dynamics model and The coefficient matrix of each term is obtained by substituting it into the stator flux linkage amplitude prediction model and expanding it.

[0080] The process of obtaining the electromagnetic torque setpoint from the proportional-integral controller and then obtaining the stator flux linkage amplitude setpoint based on the maximum torque-current ratio formula is as follows: Based on the rotor electric angular velocity reference... and electric angular velocity feedback PI control of the time error is used to obtain the reference value of the electromagnetic torque. :

[0081] ,

[0082] In the formula, 1 / s represents the integral operation. and These are the proportional and integral parameters of the PI controller. and Determined based on the frequency domain response characteristics of a typical second-order system and the zero-pole cancellation principle;

[0083] Stator flux amplitude setpoint The reference value for stator flux linkage amplitude is derived from the maximum torque-current ratio formula. : In the formula, For permanent magnet motors, the magnetic flux linkage is the permanent magnet. For the stator inductance of a permanent magnet motor, is the number of pole pairs of the permanent magnet motor. This is a reference value for the electromagnetic torque.

[0084] To achieve parallel minimum-step predictive control of electromagnetic torque and stator flux linkage amplitude, and to solve for the inverter voltage setpoint in the dq coordinate system, the specific steps include: First, based on the principle of minimum-step predictive control of electromagnetic torque, that is: setting the reference value of electromagnetic torque equal to its predicted value at the next moment. Substituting this into the electromagnetic torque prediction equation, the q-axis component of the inverter voltage setpoint in the dq coordinate system is obtained, and its expression is: ,

[0085] In the formula, This is the expression for the q-axis component of the inverter voltage setpoint in the dq coordinate system. This is a reference value for the electromagnetic torque, obtained from the output of the PI module in the speed loop. , , , , All are coefficient matrices, specifically the discrete dynamics model. The coefficient matrix of each term in the specific expression obtained after substituting into the electromagnetic torque prediction model.

[0086] Secondly, based on the principle of minimum step prediction control for stator flux linkage amplitude, that is: let the stator flux linkage amplitude reference value be... Substituting this into the stator flux linkage amplitude prediction equation, the expression for the d-axis component of the inverter voltage setpoint in the dq coordinate system can be obtained as follows:

[0087] ,

[0088] In the formula, and These are the d-axis and q-axis component expressions for the inverter voltage setpoint in the dq coordinate system, respectively. L s For the stator inductance of a permanent magnet motor, the stator flux linkage amplitude is given. It can be derived from the formula for the maximum torque-to-current ratio. To incorporate the discrete dynamics model and Substituting these values ​​into the stator flux linkage amplitude prediction model and expanding it, we obtain the coefficient matrix of each term in the specific expression. and This constitutes the inverter voltage setpoint in the dq coordinate system.

[0089] The specific expression for converting the inverter voltage setpoint from the dq coordinate system to the αβ coordinate system using the inverse Park transform is as follows:

[0090] ,

[0091] In the formula, and These are the inverter voltage setpoints in the αβ coordinate system, respectively. θ e,k Is k The electric angle of the motor rotor is detected at any time, and the iPark matrix is ​​the inverse Park transform matrix.

[0092] Based on the inverter voltage setpoint αβ axis components and And the spatial position angle is obtained from the arctangent function: ,judge The voltage vector sector is located, and the two adjacent non-zero voltage vectors within that sector are selected. , and zero vector As the voltage setpoint of the synthesized inverter The three optimal voltage vectors.

[0093] Define three optimal voltage vectors , and The corresponding duty cycles are respectively , and Then, the component form of the combined voltage of the three optimal voltage vectors under these three duty cycles in the αβ coordinate system is expressed as:

[0094] ,

[0095] In the formula, and For the synthesized voltage α and β axis components; , and , These are the two selected non-zero voltage vectors. and αβ axis components; , Zero voltage vector αβ axis components;

[0096] Based on the principle of solving for the minimum value of a function, the synthesized voltage With predictive control reference instructions The sum of squares of the tracking errors of the α and β axis components affect the duty cycle. , By taking the partial derivatives, the three duty cycles corresponding to the three optimal voltage vectors are expressed as follows:

[0097] ,

[0098] Solving the above equation yields:

[0099] ,

[0100] In the formula, , and These represent the inverter voltage setpoints in the αβ coordinate system, respectively; , and , These are the two selected non-zero voltage vectors. and αβ axis components; , Zero voltage vector αβ axis components; , and These represent the three optimal voltage vectors respectively. , and The corresponding duty cycle.

[0101] To achieve a fixed switching frequency, the following symmetrical 5-segment pulse pattern is used to generate the control pulse signal for the inverter switching transistors. The specific expression is as follows:

[0102] ,

[0103] In the formula, the control pulse signal S consists of two optimal voltage vectors. , and It operates sequentially in a mirror-symmetric pattern. Represents the zero voltage vector. , , These represent the three optimal voltage vectors respectively. , and The corresponding duty cycle is half; the pulse signal S It is applied to inverter switching transistors to achieve predictive torque control of permanent magnet motors with sinusoidal wave filters.

[0104] Example: To verify the predictive torque control method for permanent magnet motors with a sinusoidal filter provided by this invention, the method was applied to a permanent magnet motor system with a sinusoidal filter. The parameters of the permanent magnet motor system with a sinusoidal filter are shown in Table 1.

[0105] Table 1

[0106] .

[0107] Figure 2 , Figure 3 The figures show the electromagnetic torque and stator flux linkage waveforms at steady state for both the traditional model predictive control method and the method of this invention. Observe... Figure 2 and Figure 3 The comparison shows that the electromagnetic torque and stator flux linkage amplitude can reach the rated values ​​under both control methods. However, compared with the existing multi-objective predictive control method, the electromagnetic torque and stator flux linkage steady-state waveform ripple and pulsation under the method of this invention are smaller, and the steady-state control accuracy is higher.

[0108] Figure 4 and Figure 5 The figures show the dynamic waveforms of electromagnetic torque and stator flux linkage when the system load changes abruptly at a certain moment, representing the traditional model predictive control method and the method of this invention, respectively. Observation Figure 4 and Figure 5 It can be seen that the electromagnetic torque and stator flux linkage amplitude under both control methods can reach their rated values ​​after a sudden change in load. Nevertheless, compared with existing multi-objective predictive control methods, the method of this invention consistently exhibits lower torque flux linkage pulsation during the dynamic process of load change, thus achieving higher control accuracy.

[0109] The above description is merely one embodiment of the present invention and is not intended to limit the present invention. Any minor modifications, equivalent substitutions, and improvements made to the above embodiment based on the technical essence of the present invention should be included within the protection scope of the present invention.

Claims

1. A method of predictive torque control of a permanent magnet machine with a sinusoidal filter, characterized by, The steps are as follows: S1, collecting the state value of the permanent magnet motor system at the current time, and converting the state value into a variable in the dq coordinate system; S2, establishing a discrete dynamic model of the permanent magnet motor with a sine wave filter about the inverter current, capacitor voltage and stator current in the dq coordinate system; S3, establishing an electromagnetic torque prediction model and a stator flux amplitude prediction model; S4, obtaining the electromagnetic torque given value according to the proportional integral controller, and obtaining the stator flux amplitude given value based on the maximum torque current ratio formula; S5, implementing minimum beat prediction control on the stator flux amplitude and electromagnetic torque respectively, and obtaining the inverter voltage given value; S6, determining the optimal voltage vector and action duty cycle according to the inverter voltage given value, generating a control pulse signal applied to the inverter, and realizing the predictive torque control of the permanent magnet motor with a sine wave filter; The state values of the permanent magnet motor system at the current moment include: three-phase inverter output voltage , , , three-phase inverter current , , , three-phase capacitor voltage , , 、 three-phase stator current , , , detect the rotor electric angle ; the collected three-phase inverter output voltage, three-phase inverter current, three-phase capacitor voltage and three-phase stator current are transformed into dq coordinate system variables by using Park transformation, which are represented as: inverter output voltage , , inverter current , , capacitor voltage , and stator current , ; The discrete dynamic model of the permanent magnet motor with a sine wave filter about the inverter current, capacitor voltage and stator current in the dq coordinate system established by the improved Euler method is as follows: ， wherein the subscripts k and k +1 respectively represent k the acquisition value at the moment and k the predicted value at the moment +1, denotes the sampling period, denotes the rotor electric angular velocity; , , , and respectively represent the filter inductance, the filter capacitance, the motor stator inductance, the stator resistance and the rotor permanent magnet flux linkage of the sine wave filter. According to the dq coordinate system discrete dynamic model, the electromagnetic torque prediction model is established as follows: ， ， ， ， ， ， In the formula, This represents the predicted value of electromagnetic torque. This represents the number of pole pairs in a permanent magnet motor. , , , , All are coefficient matrices, specifically those derived from the discrete dynamics model. The coefficient matrix of each term in the specific expression obtained after substituting into the electromagnetic torque prediction model.

2. The permanent magnet machine predictive torque control method with sinusoidal filter according to claim 1, characterized in that, According to the dq coordinate system discrete dynamic model, the stator flux amplitude prediction model is expressed as: ， ， ， ， ， ， wherein is a predicted value of the stator flux linkage amplitude, , is the dq coordinate system inverter output voltage, are coefficient matrices, which are each a coefficient matrix in a specific expression obtained by substituting and expanding the dq coordinate system inverter output voltage and the dq coordinate system inverter input voltage in the stator flux linkage amplitude prediction model. and are each a coefficient matrix in a specific expression obtained by substituting and expanding the dq coordinate system inverter output voltage and the dq coordinate system inverter input voltage in the stator flux linkage amplitude prediction model.

3. The permanent magnet machine predictive torque control method with sinusoidal filter according to claim 2, characterized in that, The process of obtaining the electromagnetic torque given value according to the proportional integral controller, and obtaining the stator flux amplitude given value based on the maximum torque current ratio formula is as follows: According to the rotor electrical angular velocity reference and the electrical angular velocity feedback The proportional integral control of the error between the two values gives the reference value of the electromagnetic torque : ， where 1 / s represents an integral operation, and Kp and Ki are the proportional parameter and integral parameter of the PI controller, respectively, and are determined based on the frequency domain response characteristics of a typical second-order system and the zero-pole cancellation principle. Utilizing a stator flux linkage amplitude given value A reference value of the stator flux linkage amplitude is derived based on a maximum torque current ratio formula : .

4. The permanent magnet machine predictive torque control method with a sinusoidal filter according to claim 3, characterized by, Implementing minimum beat prediction control on the electromagnetic torque and the stator flux amplitude respectively, obtaining the dq coordinate system inverter voltage given value specifically includes: Based on the principle of electromagnetic torque minimum beat prediction control, the reference value of electromagnetic torque is equal to the predicted value of the next time , the q-axis component of the given value of the dq coordinate system inverter voltage is obtained by using the electromagnetic torque prediction equation : ， Based on the least-flap prediction control principle of the stator flux linkage amplitude, the stator flux linkage amplitude reference value is obtained by using the stator flux linkage amplitude prediction equation to obtain the d-axis component of the given value of the inverter voltage in the dq coordinate system : ， By the d-axis component expression and the q-axis component expression constitute the dq coordinate system inverter voltage command value.

5. The permanent magnet machine predictive torque control method with a sinusoidal filter according to claim 4, characterized by, The specific expression of converting the dq coordinate system inverter voltage given value to the αβ coordinate system inverter voltage given value by using the Park inverse transformation is as follows: ， wherein and are the αβ coordinate system inverter voltage command values, is the motor rotor electrical angle detected at k time, and iPark is the inverse Park transformation matrix. from the inverse tangent function and the αβ axis components and and the inverse tangent function gives the spatial angular position angle , the voltage vector sector in which lies is determined and the two adjacent non-zero voltage vectors , and the zero vector are selected as the three optimum voltage vectors for synthesizing the inverse converter voltage given value .

6. The permanent magnet machine predictive torque control method with a sinusoidal filter according to claim 5, characterized by, Three optimal voltage vectors are defined as , and The corresponding action duty cycles are , and The components of the resultant voltage of the three optimal voltage vectors under the three action duty cycles in the αβ coordinate system are: ， wherein and are the α and β axis components of the synthesized voltage ; , and , are the αβ axis components of the selected two non-zero voltage vectors and ; , are the αβ axis components of the zero voltage vector ; Based on the solution principle of function minimum value, the synthesized voltage and the square sum of αβ axis component tracking error of predictive control reference instruction occupy duty ratio , Partial derivative is obtained respectively, and three optimal voltage vectors corresponding to three action duty ratios are expressed as follows: ， Solving the above equation gives: ， wherein , , represent the inverter voltage reference values in the αβ coordinate system, respectively; , , represent the three optimal voltage vectors , , corresponding duty ratios.

7. The permanent magnet machine predictive torque control method with a sinusoidal filter according to claim 6, characterized by, The following symmetric 5-segment pulse mode is used to generate the control pulse signal of the inverter switching tube to realize fixed switching frequency, and the expression is as follows: ， In the formula, the control pulse signal S is acted by two optimal voltage vectors , and in turn according to the front and back mirror image symmetry mode, represents a zero voltage vector, , , respectively represent half of the action duty ratio of three optimal voltage vectors , and ; the pulse signal S is applied to the inverter switch tube, and the permanent magnet motor predictive torque control with a sine wave filter is realized.

8. A computer device, comprising: The device includes a processor and a memory, the processor is electrically connected with the memory, the memory is used for storing instructions and data, and the processor is used for executing the predictive torque control method of the permanent magnet motor with a sine wave filter according to any one of claims 1-7.

Images