A motor control method based on current prediction error

Abstract

The present invention discloses a motor control method based on current prediction error, comprising S1: obtaining current parameters, which include voltage, speed, d-axis current and q-axis current; S2: establishing a motor current prediction model to obtain the current prediction error PE(k) at time k; S3: calculating the proportional relationship P between the d-axis and q-axis current prediction errors based on the current prediction error PE(k) d 、P q ; S4: According to the proportional relationship P of the d-axis and q-axis current prediction errors d 、P q Calculate the current prediction error corresponding to all voltage vectors at time k+1; S5: Use the prediction error compensation coefficient to optimize the current prediction error in S4 to obtain the predicted current at time k+1; S6: Calculate the predicted current at time k+2 based on the predicted current at time k+1, and then input the predicted current at time k+2 into the cost function for solution, and output the minimum value to the voltage of the motor.

Description

Technical Field

[0001] The present invention relates to the technical field of motor control, and in particular to a motor control method based on current prediction error. Background Art

[0002] As an inductive load, the main variable parameters of the motor are motor resistance, magnetic flux, and quadrature and direct axis inductance. When these variable parameters are mismatched, motor control may be out of balance.

[0003] In existing technologies, the d- and q-axis current loops in vector control methods for motor controllers generally use PI controllers. This controller offers good stability but limited dynamic response. Therefore, model-predictive current control (MPC) is employed. However, this traditional MPC relies heavily on the accuracy of the motor model, resulting in insufficient interference immunity. Some researchers have also employed state observers and recursive least squares algorithms to compensate for system disturbances caused by parameter mismatches. However, this approach requires significant computational effort and places high demands on the hardware system. Summary of the Invention

[0004] In response to the problem of insufficient anti-interference ability of the model-predicted current-controlled motor in the existing technology, the present invention proposes a motor control method based on current prediction error, which increases the current prediction error and compensation coefficient, improves the anti-interference performance of the current prediction model, and makes the motor control more stable.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] A motor control method based on current prediction error specifically comprises the following steps:

[0007] S1: Obtain current parameters, including voltage, speed, d-axis current and q-axis current;

[0008] S2: Establish a motor current prediction model to obtain the current prediction error PE(k) at time k;

[0009] S3: Calculate the proportional relationship P between the d-axis and q-axis current prediction errors based on the current prediction error PE(k) d 、P q ;

[0010] S4: According to the proportional relationship P between the d-axis and q-axis current prediction errors d 、P q Calculate the current prediction error corresponding to all voltage vectors at time k+1;

[0011] S5: Optimize the current prediction error in S4 using the prediction error compensation coefficient to obtain the predicted current at time k+1;

[0012] S6: Calculate the predicted current at time k+2 based on the predicted current at time k+1, then input the predicted current at time k+2 into the cost function for solution, and output the minimum value to the voltage of the motor.

[0013] Preferably, said S2 comprises the following steps:

[0014] S2-1: The actual motor model and prediction model established are expressed as:

[0015]

[0016] In formula (1): i d (k+1) represents the actual d-axis current at time k+1, i q (k+1) represents the actual current of the q axis at time k+1; T s represents the sampling period; R s Indicates stator resistance; L d represents the d-axis stator inductance; ω e Indicates the rotor electrical angular velocity; L q represents the q-axis stator inductance; i d (k) represents the actual d-axis current at time k, i q (k) represents the actual current of the q axis at time k; u d (k) represents the actual voltage of the d-axis stator at time k; u q (k) represents the actual q-axis stator voltage at time k; ψ f represents the magnetic flux of the rotor permanent magnet;

[0017] represents the d-axis predicted current at time k+1; represents the q-axis predicted current at time k+1; PE represents the current prediction error; represents the predicted current at time k+1; i dq (k+1) represents the actual current at time k+1;

[0018] S2-2: The stator resistance, rotor permanent magnet flux, and d-axis and q-axis stator inductance used in the prediction model are denoted as R sp , ψ fp 、L dp 、L qp ; The four parameters in the actual model are denoted as R sn , ψ fn 、L dn 、L qn ; The relationship between the two sets of parameters in the prediction model and the actual model is expressed as:

[0019] R sp =R sn ×NR , ψ fp =ψ fn ×N ψ , L dp =L dn ×N d , L qp =L qn ×N q (2)

[0020] In formula (2), N R Indicates the stator resistance conversion coefficient; N ψ Indicates the permanent magnet flux conversion coefficient; N d Indicates the d-axis inductance conversion coefficient; N q Indicates the q-axis inductance conversion coefficient;

[0021] S2-3: The current prediction error PE(k) at time k can be expressed as:

[0022]

[0023] In formula (2), PE_i d represents the d-axis current prediction error; represents the d-axis predicted current at time k+1; i d (k+1) represents the actual d-axis current at time k+1; T s represents the sampling period; R sn represents the stator resistance; i d (k) represents the actual d-axis current at time k; N d Indicates the d-axis inductance conversion coefficient; N R Indicates the stator resistance conversion coefficient; L dn represents the d-axis stator inductance; ω e Indicates the rotor electrical angular velocity; L qn represents the q-axis stator inductance; i q (k) represents the actual current of the q axis at time k; N q Indicates the q-axis inductance conversion coefficient; u d (k) represents the actual voltage on the d-axis at time k;

[0024] PE_i q represents the q-axis current prediction error; represents the q-axis predicted current at time k+1; i q (k+1) represents the actual current of the q axis at time k+1; i q (k) represents the actual current of the q axis at time k; L qn represents the q-axis stator inductance; u q (k) represents the actual voltage on the q axis at time k; ψ fn Indicates the actual permanent magnet flux of the rotor; Nψ It represents the permanent magnet flux conversion coefficient;

[0025] S2-4: When only the influence of inductance mismatch on the prediction model is considered, the current prediction error PE_i d PE_i q Simplified and reformulated as:

[0026]

[0027] In formula (3), PE_i d (k) represents the d-axis current prediction error at time k; represents the d-axis predicted current at time k; i d (k) represents the actual d-axis current at time k; T s Indicates the sampling period; L dn represents the d-axis stator inductance; N d Indicates the d-axis inductance conversion coefficient; u d (k-1) represents the actual voltage of the d-axis at time k-1; R sn represents the stator resistance; i d (k-1) represents the actual d-axis current at time k-1; N q Indicates the q-axis inductance conversion coefficient; ω e Indicates the rotor electrical angular velocity; L qn represents the q-axis stator inductance; i q (k-1) represents the actual current on the q axis at time k-1;

[0028] PE_i q (k) represents the q-axis current prediction error at time k; represents the q-axis predicted current at time k; i q (k) represents the actual current of the q axis at time k; u q (k-1) represents the actual voltage on the q axis at time k-1; ψ fn Represents the actual permanent magnet flux of the rotor.

[0029] Preferably, the step S3 includes the following steps:

[0030] S3-1: The difference in prediction error between two adjacent control cycles, ΔPE, can be expressed as:

[0031] ΔPE=PE(k)-PE(k-1) (4)

[0032] In formula (4), PE(k) represents the current prediction error at time k, and PE(k-1) represents the current prediction error at time k-1;

[0033] S3-2: Assuming that the d-axis and q-axis currents are constant, ΔPE is expressed as:

[0034]

[0035] In formula (5), ΔPE_i d represents the d-axis current prediction error between two adjacent control cycles; T s Indicates the sampling period; L dn represents the d-axis stator inductance; N d Indicates the d-axis inductance conversion coefficient; u d (k-1) represents the actual voltage of the d-axis at time k-1; U d (k-2) represents the actual voltage of the d-axis at time k-2; P d Indicates the proportional relationship of the d-axis current prediction error; Δu d Indicates the d-axis voltage difference between two adjacent control cycles;

[0036] ΔPE_i q represents the q-axis current prediction error between two adjacent control cycles; L qn represents the q-axis stator inductance; N q Indicates the q-axis inductance conversion coefficient; u q (k-1) represents the actual voltage on the q axis at time k-1; u q (k-2) represents the actual voltage on the q axis at time k-2; P q Indicates the proportional relationship of the q-axis current prediction error; Δu q Indicates the q-axis voltage difference between two adjacent control cycles.

[0037] Preferably, in S4, the voltage vector includes 6 non-zero voltage vectors and 2 zero voltage vectors, and each voltage vector has a corresponding current prediction error, which is expressed as follows:

[0038]

[0039] In formula (6), represents the current prediction error of the d-axis at the y-th voltage vector k+1; PE_i d (k) represents the current prediction error of the d-axis at time k; P d Indicates the proportional relationship of the d-axis current prediction error; u d (k) represents the voltage at time k; u d (k-1) represents the voltage at time k-1;

[0040] represents the current prediction error of the q-axis at the y-th voltage vector k+1; PE_i q (k) represents the current prediction error of the q axis at time k; P q Indicates the proportional relationship of the q-axis current prediction error; u q(k) represents the voltage at time k; u q (k-1) represents the voltage at time k-1.

[0041] Preferably, in S5, the predicted current expression at time k+1 is:

[0042]

[0043] In formula (6), It represents the predicted current on the d-axis at the time k+1 of the y-th voltage vector after optimization; represents the predicted d-axis current at the k+1 moment of the y-th voltage vector before optimization; represents the current prediction error of the d-axis at the time k+1 of the y-th voltage vector; α represents the prediction error compensation coefficient;

[0044] It represents the predicted current on the q axis at the k+1 moment of the y-th voltage vector after optimization; represents the predicted current on the q axis at the k+1 moment of the y-th voltage vector before optimization; Represents the q-axis current prediction error at the y-th voltage vector k+1.

[0045] Preferably, the prediction error compensation coefficient α has a value range of 0-1.

[0046] Preferably, in S6, the cost function g is expressed as:

[0047]

[0048]

[0049] In formula (7), represents the predicted current of the d-axis at time k+2, represents the reference value of the d-axis current at time k+2; q represents the weight coefficient, represents the predicted current of the q axis at time k+2, represents the reference value of the q-axis current at time k+2; f(i d ,i q ) is the current constraint condition, ensuring that the stator current does not exceed the maximum allowable current I max .

[0050] In summary, due to the adoption of the above technical solution, compared with the prior art, the present invention has at least the following beneficial effects:

[0051] When the motor inductance changes, the current prediction error increases, thereby improving the current prediction model's anti-interference performance and making motor control more stable. Real-time updates of the prediction error over the previous two cycles improve accuracy. The cost function incorporates weighting coefficients for the d- and q-axis current errors, further enhancing control accuracy while minimizing hardware system design requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 Schematic diagram of a motor control method based on current prediction error according to an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0053] The present invention will be further described in detail below with reference to the examples and specific implementation methods. However, this should not be understood as limiting the scope of the present invention to the following examples, as all technologies implemented based on the present invention fall within the scope of the present invention.

[0054] In the description of the present invention, it should be understood that the terms "longitudinal", "transverse", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention.

[0055] As an inductive load, the motor's main variable parameters are motor resistance, flux, and quadrature and direct-axis inductance. When these variable parameters are mismatched, motor control may be misaligned. Since quadrature and direct-axis inductances mainly vary with current, while motor resistance and flux vary mainly with temperature, and the frequency of current variation is much greater than the frequency of temperature variation, and the mismatch between quadrature and direct-axis inductance parameters has the greatest impact on motor control accuracy, the present invention only considers the impact of inductance mismatch on the predicted current model, achieving a high degree of computational accuracy while adding a small amount of computation to the original model without reselecting hardware. In addition, the present invention sets different weight coefficients for the d-axis and q-axis current errors in the cost function, which can further improve computational accuracy.

[0056] Therefore, the present invention uses a current loop controller based on current prediction error and a traditional PI controller to control the current, and then outputs the corresponding current and voltage through the inverter to control the motor. The traditional PI controller is existing technology, while the current loop controller based on current prediction error is the innovative technology of the present invention.

[0057] like Figure 1As shown, the present invention provides a design method of a current loop controller based on current prediction error, which specifically includes the following steps:

[0058] S1: Get current parameters, including bus voltage, speed, AC and DC axis current.

[0059] In this embodiment, the bus voltage is collected by a voltage sensor, the rotational speed is collected by a resolver sensor, and the AC and DC axis currents are collected by a three-phase current sensor.

[0060] S2: Establish a current prediction model to obtain the current prediction error PE(k) at time k.

[0061] In this embodiment, the actual model and the prediction model of the permanent magnet synchronous motor can be expressed as follows (using the forward Euler discretization method):

[0062]

[0063] In formula (1): i d (k+1) represents the actual d-axis current at time k+1, i q (k+1) represents the actual current of the q axis at time k+1; T s represents the sampling period; R s Indicates stator resistance; L d represents the d-axis stator inductance; ω e Indicates the rotor electrical angular velocity; L q represents the q-axis stator inductance; i d (k) represents the actual d-axis current at time k, i q (k) represents the actual current of the q axis at time k; u d (k) represents the actual voltage of the d-axis stator at time k; u q (k) represents the actual q-axis stator voltage at time k; ψ f represents the magnetic flux of the rotor permanent magnet;

[0064] represents the d-axis predicted current at time k+1; represents the q-axis predicted current at time k+1; PE represents the current prediction error; represents the predicted current at time k+1; i dq (k+1) represents the actual current at time k+1;

[0065] In this embodiment, the stator resistance, rotor permanent magnet flux, and d-axis and q-axis stator inductance used in the prediction model are denoted as R sp , ψ fp , L dp , L qp ; The four parameters in the actual motor model are denoted as R sn , ψfn 、L dn 、L qn ; The relationship between the two sets of parameters in the prediction model and the actual model is expressed as:

[0066] R sp =R sn ×N R , ψ fp =ψ fn ×N ψ , L dp =L dn ×N d , L qp =L qn ×N q (2)

[0067] In formula (2), N R Indicates the stator resistance conversion coefficient; N ψ Indicates the permanent magnet flux conversion coefficient; N d Indicates the d-axis inductance conversion coefficient; N q Indicates the q-axis inductance conversion coefficient.

[0068] Then the current prediction error PE(k) at time k can be expressed as:

[0069]

[0070] In formula (2), PE_i d represents the d-axis current prediction error; represents the d-axis predicted current at time k+1; i d (k+1) represents the actual d-axis current at time k+1; T s represents the sampling period; R sn represents the stator resistance; i d (k) represents the actual d-axis current at time k; N d Indicates the d-axis inductance conversion coefficient; N R Indicates the stator resistance conversion coefficient; L dn represents the d-axis stator inductance; ω e Indicates the rotor electrical angular velocity; L qn represents the q-axis stator inductance; i q (k) represents the actual current of the q axis at time k; N q Indicates the q-axis inductance conversion coefficient; u d (k) represents the actual voltage on the d-axis at time k;

[0071] PE_i q represents the q-axis current prediction error; represents the q-axis predicted current at time k+1; i q(k+1) represents the actual current of the q axis at time k+1; i q (k) represents the actual current of the q axis at time k; L qn represents the q-axis stator inductance; u q (k) represents the actual voltage on the q axis at time k; ψ fn Indicates the actual permanent magnet flux of the rotor; N ψ It represents the permanent magnet flux conversion coefficient.

[0072] In this embodiment, since the quadrature and direct axis inductances mainly change with the current, the motor resistance and flux change mainly with the temperature, and the current change frequency is much greater than the temperature change frequency, and the quadrature and direct axis inductance parameter mismatch has the greatest impact on the motor control accuracy, the present invention only considers the impact of the inductance mismatch on the predicted current model, and the current prediction error PE_i d PE_i q can be simplified and reformulated as:

[0073]

[0074] In formula (3), PE_i d (k) represents the d-axis current prediction error at time k; represents the d-axis predicted current at time k; i d (k) represents the actual d-axis current at time k; T s Indicates the sampling period; L dn represents the d-axis stator inductance; N d Indicates the d-axis inductance conversion coefficient; u d (k-1) represents the actual voltage of the d-axis at time k-1; R sn represents the stator resistance; i d (k-1) represents the actual d-axis current at time k-1; N q Indicates the q-axis inductance conversion coefficient; ω e Indicates the rotor electrical angular velocity; L qn represents the q-axis stator inductance; i q (k-1) represents the actual current on the q axis at time k-1;

[0075] PE_i q (k) represents the q-axis current prediction error at time k; represents the q-axis predicted current at time k; i q (k) represents the actual current of the q axis at time k; u q (k-1) represents the actual voltage on the q axis at time k-1; ψ fn Represents the actual permanent magnet flux of the rotor.

[0076] S3: If the voltages applied in the two cycles are the same, the prediction error is considered unchanged; if the voltages applied in the two cycles are different, the proportional relationship P between the voltages applied in the two cycles and the prediction errors of the d-axis and q-axis currents is calculated. d 、P q .

[0077] In this embodiment, the prediction error difference ΔPE between two adjacent control cycles can be expressed as:

[0078] ΔPE=PE(k)-PE(k-1) (4)

[0079] In formula (4), PE(k) represents the current prediction error at time k, and PE(k-1) represents the current prediction error at time k-1.

[0080] Since the electrical change frequency of the permanent magnet synchronous motor is much greater than the mechanical change frequency, the speed of two adjacent cycles can be considered constant; the d-axis and q-axis currents of two adjacent cycles in steady state can also be considered constant. Even if the current suddenly changes, the magnitude of the change in the d-axis and q-axis voltages is much greater than the magnitude of the change in the current, so the d-axis and q-axis currents can also be considered constant.

[0081] Therefore, the prediction error difference ΔPE between two adjacent control cycles can be expressed as:

[0082]

[0083] In formula (5), ΔPE_i d represents the d-axis current prediction error between two adjacent control cycles; T s Indicates the sampling period; L dn represents the d-axis stator inductance; N d Indicates the d-axis inductance conversion coefficient; u d (k-1) represents the actual voltage of the d-axis at time k-1; u d (k-2) represents the actual voltage of the d-axis at time k-2; P d Indicates the proportional relationship of the d-axis current prediction error; Δu d Indicates the d-axis voltage difference between two adjacent control cycles;

[0084] ΔPE_i q represents the q-axis current prediction error between two adjacent control cycles; L qn represents the q-axis stator inductance; N q Indicates the q-axis inductance conversion coefficient; u q (k-1) represents the actual voltage on the q axis at time k-1; u q (k-2) represents the actual voltage on the q axis at time k-2; P q Indicates the proportional relationship of the q-axis current prediction error; Δu qIndicates the q-axis voltage difference between two adjacent control cycles.

[0085] Then the proportional relationship between the d-axis and q-axis current prediction errors is P d 、P q It can be expressed as:

[0086]

[0087] S4: According to the proportional relationship P between the d-axis and q-axis current prediction errors d 、P q Calculate the current prediction error corresponding to all voltage vectors at time k+1.

[0088] In this embodiment, the motor controller adopts a three-phase two-level voltage source inverter, whose six IGBTs can generate eight valid switching state combinations, including six non-zero voltage vectors and two zero voltage vectors, and each voltage vector has a corresponding current prediction error.

[0089] Then the expression of the current prediction error corresponding to the voltage vector at time k+1 is as follows:

[0090]

[0091] In formula (6), represents the current prediction error of the d-axis at the y-th voltage vector k+1; PE_i d (k) represents the current prediction error of the d-axis at time k; P d Indicates the proportional relationship of the d-axis current prediction error; u d (k) represents the voltage at time k; u d (k-1) represents the voltage at time k-1;

[0092] represents the current prediction error of the q-axis at the y-th voltage vector k+1; PE_i q (k) represents the current prediction error of the q axis at time k; P q Indicates the proportional relationship of the q-axis current prediction error; u q (k) represents the voltage at time k; u q (k-1) represents the voltage at time k-1.

[0093] S5: In this embodiment, since the motor is assumed to be in a stable operating condition and the influence of the speed is not considered, it is necessary to use a prediction error compensation coefficient to optimize the current prediction error, so as to obtain the predicted current at time k+1.

[0094]

[0095] In formula (6), It represents the predicted current on the d-axis at the time k+1 of the y-th voltage vector after optimization; represents the predicted d-axis current at the k+1 moment of the y-th voltage vector before optimization;

[0096] represents the current prediction error of the d-axis at the time k+1 of the y-th voltage vector; α represents the prediction error compensation coefficient;

[0097] It represents the predicted current on the q axis at the k+1 moment of the y-th voltage vector after optimization;

[0098] represents the predicted current on the q axis at the k+1 moment of the y-th voltage vector before optimization; Represents the q-axis current prediction error at the y-th voltage vector k+1.

[0099] In this embodiment, the prediction error compensation coefficient is an empirical coefficient and can be gradually increased and debugged according to the control effect between [0, 1]. If the resistance value of the controlled motor is too large, a smaller compensation coefficient should be selected as much as possible to reduce the impact of current fluctuations.

[0100] S6: Calculate the predicted current at time k+2 based on the predicted current at time k+1, then input the predicted current at time k+2 into the cost function for solution, and take the minimum value as the optimal output voltage, that is, the voltage output to the motor.

[0101] In this embodiment, the cost function g is expressed as:

[0102]

[0103]

[0104] In formula (7), represents the predicted current of the d-axis at time k+2, represents the reference value of the d-axis current at time k+2; q represents the weight coefficient, represents the predicted current of the q axis at time k+2, represents the reference value of the q-axis current at time k+2; f(i d ,i q ) is the current constraint condition, ensuring that the stator current does not exceed the maximum allowable current I max , the corresponding result is 0, which has no effect on the cost function. If the maximum current is exceeded, a very large value is output, causing the voltage vector to be too large and not selected.

[0105] In the present invention, when the motor inductance changes, the current prediction error increases, thereby improving the anti-interference performance of the current prediction model and making the control of the motor more stable.

[0106] The prediction error proportional coefficient of the first two cycles is updated in real time to improve the accuracy of the prediction error.

[0107] The cost function takes into account the weight coefficient λ of the d and q axis current errors q ,λ q When >1, iq has a higher priority than id. The weight coefficient mainly adopts the debugging method. You can first assign a value >1 to ensure the priority and importance of iq, and then adjust the specific value according to the debugging effect to further improve the control accuracy.

[0108] In summary, the present invention not only ensures accuracy but also reduces the design requirements of the hardware system as much as possible.

[0109] Those skilled in the art will appreciate that the above-mentioned embodiments are specific examples for implementing the present invention, and that in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present invention.

Claims

1. A motor control method based on current prediction error, characterized in that: The specific steps include: S1: Obtain current parameters, including voltage, speed, d-axis current and q-axis current; S2: Establish a motor current prediction model to obtain the current prediction error PE(k) at time k; S3: Calculate the proportional relationship P of the d-axis current prediction error based on the current prediction error PE(k) d The proportional relationship between the q-axis current prediction error and the q-axis current prediction error is P q , where the proportional relationship of the d-axis current prediction error P d The proportional relationship P of the q-axis current prediction error is calculated by the ratio of the d-axis current prediction error difference between two adjacent control cycles to the d-axis voltage difference between two adjacent control cycles. q It is calculated by the ratio of the difference in q-axis current prediction error between two adjacent control cycles to the difference in q-axis voltage between two adjacent control cycles; S4: According to the proportional relationship P between the d-axis and q-axis current prediction errors d 、P q Calculate the current prediction error corresponding to all voltage vectors at time k+1; S5: Optimizing the current prediction error in S4 using a prediction error compensation coefficient to obtain the predicted current at time k+1. The prediction error compensation coefficient is an empirical numerical range that reduces the impact of current fluctuations and has a value range of 0-1. S6: Calculate the predicted current at time k+2 based on the predicted current at time k+1, then input the predicted current at time k+2 into the cost function for solution, and output the minimum value to the voltage of the motor.

2. The motor control method based on current prediction error according to claim 1, characterized in that: The S2 comprises the following steps: S2-1: The actual motor model and prediction model established are expressed as: In formula (1): i d (k+1) represents the actual d-axis current at time k+1, i q (k+1) represents the actual current of the q axis at time k+1; T s represents the sampling period; R s Indicates stator resistance; L d represents the d-axis stator inductance; ω e Indicates the rotor electrical angular velocity; L q represents the q-axis stator inductance; i d (k) represents the actual d-axis current at time k, i q (k) represents the actual current of the q axis at time k; u d (k) represents the actual voltage of the d-axis stator at time k; u q (k) represents the actual q-axis stator voltage at time k; ψ f represents the magnetic flux of the rotor permanent magnet; represents the d-axis predicted current at time k+1; represents the q-axis predicted current at time k+1; PE represents the current prediction error; represents the predicted current at time k+1; i dq (k+1) represents the actual current at time k+1; S2-2: The stator resistance, rotor permanent magnet flux, and d-axis and q-axis stator inductance used in the prediction model are denoted as R sp , ψ fp , L dp , L qp ; The four parameters in the actual model are denoted as R sn , ψ fn , L dn , L qn ; The relationship between the two sets of parameters in the prediction model and the actual model is expressed as: R sp =R sn ×N R ,ψ fp =ψ fn ×N ψ ,L dp =L dn ×N d ,L qp =L qn ×N q (2) In formula (2), N R Indicates the stator resistance conversion coefficient; N ψ Indicates the permanent magnet flux conversion coefficient; N d Indicates the d-axis inductance conversion coefficient; N q Indicates the q-axis inductance conversion coefficient; S2-3: The current prediction error PE(k) at time k is expressed as: In formula (2), PE_i d represents the d-axis current prediction error; represents the d-axis predicted current at time k+1; i d (k+1) represents the actual d-axis current at time k+1; T s represents the sampling period; R sn Indicates the actual stator resistance; i d (k) represents the actual d-axis current at time k; N d Indicates the d-axis inductance conversion coefficient; N R Indicates the stator resistance conversion coefficient; L dn represents the actual d-axis stator inductance; ω e Indicates the rotor electrical angular velocity; L qn represents the actual q-axis stator inductance; i q (k) represents the actual current of the q axis at time k; N q Indicates the q-axis inductance conversion coefficient; u d (k) represents the actual d-axis stator voltage at time k; PE_i q represents the q-axis current prediction error; represents the q-axis predicted current at time k+1; i q (k+1) represents the actual current of the q axis at time k+1; i q (k) represents the actual current of the q axis at time k; L qn represents the actual q-axis stator inductance; u q (k) represents the actual q-axis stator voltage at time k; ψ fn Indicates the actual permanent magnet flux of the rotor; N ψ It represents the permanent magnet flux conversion coefficient; S2-4: When only the influence of inductance mismatch on the prediction model is considered, the current prediction error PE_i d PE_i q Simplified and reformulated as: In formula (3), PE_i d (k) represents the d-axis current prediction error at time k; represents the d-axis predicted current at time k; i d (k) represents the actual d-axis current at time k; T s Indicates the sampling period; L dn Indicates the actual d-axis stator inductance; N d Indicates the d-axis inductance conversion coefficient; u d (k-1) represents the actual d-axis stator voltage at time k-1; R sn Indicates the actual stator resistance; i d (k-1) represents the actual d-axis current at time k-1; N q Indicates the q-axis inductance conversion coefficient; ω e Indicates the rotor electrical angular velocity; L qn represents the actual q-axis stator inductance; i q (k-1) represents the actual current on the q axis at time k-1; PE_i q (k) represents the q-axis current prediction error at time k; represents the q-axis predicted current at time k; i q (k) represents the actual current of the q axis at time k; u q (k-1) represents the actual q-axis stator voltage at time k-1; ψ fn Represents the actual permanent magnet flux of the rotor.

3. The motor control method based on current prediction error according to claim 1, characterized in that: The S3 includes the following steps: S3-1: The difference in prediction error between two adjacent control cycles, ΔPE, is expressed as: ΔPE=PE(k)-PE(k-1) (4) In formula (4), PE(k) represents the current prediction error at time k, and PE(k-1) represents the current prediction error at time k-1; S3-2: Assuming that the d-axis and q-axis currents are constant, ΔPE is expressed as: In formula (5), ΔPE_i d represents the d-axis current prediction error between two adjacent control cycles; T s Indicates the sampling period; L dn Indicates the actual d-axis stator inductance; N d Indicates the d-axis inductance conversion coefficient; u d (k-1) represents the actual d-axis stator voltage at time k-1; u d (k-2) represents the actual d-axis stator voltage at time k-2; P d Indicates the proportional relationship of the d-axis current prediction error; Δu d Indicates the d-axis voltage difference between two adjacent control cycles; ΔPE_i q represents the q-axis current prediction error between two adjacent control cycles; L qn Indicates the actual q-axis stator inductance; N q Indicates the q-axis inductance conversion coefficient; u q (k-1) represents the actual q-axis stator voltage at time k-1; u q (k-2) represents the actual q-axis stator voltage at time k-2; P q Indicates the proportional relationship of the q-axis current prediction error; Δu q Indicates the q-axis voltage difference between two adjacent control cycles.

4. The motor control method based on current prediction error according to claim 1, characterized in that: In S4, the voltage vector includes 6 non-zero voltage vectors and 2 zero voltage vectors. Each voltage vector has a corresponding current prediction error, which is expressed as follows: In formula (6), represents the current prediction error of the d-axis at the y-th voltage vector k+1; PE_i d (k) represents the current prediction error of the d-axis at time k; P d Indicates the proportional relationship of the d-axis current prediction error; u d (k) represents the actual d-axis stator voltage at time k; u d (k-1) represents the actual d-axis stator voltage at time k-1; represents the current prediction error of the q-axis at the y-th voltage vector k+1; PE_i q (k) represents the current prediction error of the q axis at time k; P q Indicates the proportional relationship of the q-axis current prediction error; u q (k) represents the actual q-axis stator voltage at time k; u q (k-1) represents the actual q-axis stator voltage at time k-1.

5. The motor control method based on current prediction error according to claim 1, characterized in that: In S5, the predicted current expression at time k+1 is: In formula (6), It represents the predicted current on the d-axis at the time k+1 of the y-th voltage vector after optimization; represents the predicted d-axis current at the k+1 moment of the y-th voltage vector before optimization; represents the current prediction error of the d-axis at the time k+1 of the y-th voltage vector; α represents the prediction error compensation coefficient; It represents the predicted current on the q axis at the k+1 moment of the y-th voltage vector after optimization; represents the predicted current on the q axis at the k+1 moment of the y-th voltage vector before optimization; Represents the q-axis current prediction error at the y-th voltage vector k+1.

6. The motor control method based on current prediction error according to claim 1, characterized in that: In S6, the cost function g is expressed as: In formula (7), represents the predicted current of the d-axis at time k+2, represents the reference value of the d-axis current at time k+2; λ q represents the weight coefficient, represents the predicted current of the q axis at time k+2, represents the reference value of the q-axis current at time k+2; f(i d ,i q ) is the current constraint condition, ensuring that the stator current does not exceed the maximum allowable current I max .

Images