Dimensionality Reduction Line Constraint EMPC Control Method for New Energy Vehicle Motors

Abstract

A dimensionality reduction line-constrained EMPC control method for new energy vehicle motors, which relates to the field of motor control. The present invention is to solve the problems of high dimension and long execution time of the existing line-constrained EMPC algorithm, and low portability of the algorithm when the parameters of the controlled object change. The dimensionality reduction line-constrained EMPC control method for new energy vehicle motors according to the present invention inherits the advantages of line-constrained EMPC, and through model improvement, further reduces the system dimension to 4 dimensions. This improvement not only reduces the memory occupancy rate, but also improves the execution efficiency. In addition, the present invention transfers the parameters related to the controlled object outside the offline part of the algorithm. When the controlled object changes, there is no need to regenerate the offline matrix of EMPC, which greatly improves the portability of the algorithm without increasing the computational burden of the algorithm.

Description

Technical Field

[0001] The present invention belongs to the field of motor control. Background Art

[0002] Interior Permanent Magnet Synchronous Motors (IPMSMs) have been widely used in modern industries and electric vehicles due to their high efficiency, high power density, and excellent dynamic performance. When controlling such motors, Proportional-Integral (PI) controllers are commonly adopted because of their simple design and strong stability. Nevertheless, when faced with high requirements for motor dynamic performance, especially in the environment of parameter changes and load disturbances, traditional PI controllers may not provide the best performance.

[0003] To address these limitations, an advanced control strategy called Explicit Model Predictive Control (EMPC) has been introduced. EMPC improves system performance by predicting the future behavior of the system and selecting the optimal control action in each control cycle. In the control application of IPMSM, EMPC can handle motor dynamics more accurately, achieving faster response speed and more effective disturbance rejection.

[0004] EMPC transforms part of the online calculation process of continuous set model prediction into offline calculation, thus reducing the computational load while inheriting its advantages. Its powerful constraint handling ability makes it particularly suitable for electric vehicle drive systems. However, EMPC faces the problems of high memory occupancy and long execution time. As the model dimension and constraint conditions increase, the number of Piecewise Affine (PWA) spaces grows exponentially, resulting in a large consumption of the storage space of the main control chip, and further leading to an overly long execution time, which is not suitable for motor drive algorithms.

[0005] To address this challenge, a line-constrained EMPC strategy has been proposed in recent years. This method introduces a constraint variable and uses the constraints of EMPC to limit the d-axis current, thus avoiding the optimization of the d-axis current following process in the cost function of EMPC. By transforming the constraints of EMPC from traditional plane constraints to line constraints, the number of PWAs is significantly reduced, thereby improving the execution efficiency of the algorithm.

[0006] Although the line-constrained EMPC has made significant progress in reducing algorithm complexity and improving efficiency, there are still some problems:

[0007] 1. Even though the line-constrained EMPC has significantly reduced the algorithm complexity and improved the efficiency, the algorithm still has a high dimension, a long execution time, and imposes an operational pressure on low-cost embedded devices.

[0008] 2. EMPC converts part of the online calculation of CCS-MPC into offline calculation. Although the amount of calculation is reduced, when the parameters of the controlled object change, it is necessary to regenerate the offline data of EMPC, which reduces the portability of the algorithm. Summary of the Invention

[0009] The present invention is to solve the problems of the existing line-constrained EMPC algorithm with high dimension, long execution time, and low portability when the parameters of the controlled object change. Now, a dimension-reduced line-constrained EMPC control method for a new energy vehicle motor is provided.

[0010] The dimension-reduced line-constrained EMPC control method for a new energy vehicle motor specifically includes:

[0011] Construct a super-local model of the controlled motor:

[0012]

[0013] where, i d '(k + 1) and i q '(k + 1) are the transient currents on the d-axis and q-axis at the (k + 1)-th moment respectively, i d (k) and i q (k) are the synchronous currents on the d-axis and q-axis at the k-th moment respectively, T s is the sampling period, u' d (k) and u' q (k) are the transient voltages on the d-axis and q-axis at the k-th moment respectively, α' is the transient input quantity coefficient, α' = 1 / L bas , L bas is the transient inductance;

[0014] Based on the super-local model, construct a dimension-reduced line-constrained EMPC model:

[0015]

[0016] s.t.G ac U' dq ≤W ac +E ac x(k)

[0017] where, J(k) is the control objective of the dimension-reduced line-constrained EMPC model,

[0018] x(k) is the state variable and there is:

[0019] i' dref(k) and i' qref (k) are the given values of the transient currents on the d-axis and q-axis at time k, U'dq is the transient voltage and there is u'(k) is the transient voltage of the motor control system at time k, N u is the time domain length of the motor control, Y is the quadratic coefficient matrix of the state variables, H is the quadratic coefficient matrix of the input variables, F is the linear coefficient matrix of the state variables, G ac is the effective constraint coefficient matrix of the input variables, W ac is the constant term matrix of the effective constraint conditions, E ac is the effective coefficient matrix of the state variables;

[0020] Solve the reduced-dimension line-constrained EMPC model using the KKT conditions to obtain the transient voltage U' dq ;

[0021]

[0022] where λ is the Lagrange coefficient;

[0023] Restore the transient voltage U' dq to the actual voltage U of the motor control system dq , and use U dq to achieve the reduced-dimension line-constrained EMPC control of the new energy vehicle motor.

[0024] Furthermore, the above construction of the super-local model of the controlled motor includes:

[0025] Construct the initial model of the controlled motor according to the Euler discretization method:

[0026]

[0027] where i d (k + 1) and i q (k + 1) are the synchronous currents on the d-axis and q-axis at time k + 1 respectively, i d (k) and i q (k) are the synchronous currents on the d-axis and q-axis at time k respectively, F d (k) and F q (k) are the observable disturbance terms on the d-axis and q-axis at time k respectively, α d and α q are the input variable coefficients on the d-axis and q-axis respectively, u d (k) and u q (k) are the synchronous voltages on the d-axis and q-axis at time k respectively;

[0028] Phase-shift the initial model to obtain:

[0029]

[0030] Introduce the transient currents \(i_d'(k + 1)\) and \(i_q'(k + 1)\) on the d-axis and q-axis at the \((k + 1)\) -th moment into the initial model after phase shift, and there are: d '\((k + 1)\) and \(i\) q '\((k + 1)\), and there is: \(i\) d '\((k + 1)=i\) d (k + 1)-T s F d (k), \(i\) q '\((k + 1)=i\) q (k + 1)-T s F q (k), so that the initial model after phase shift is rewritten as a model that does not contain disturbance term information:

[0031]

[0032] Introduce the transient input quantity coefficient \(\alpha'\) into the model that does not contain disturbance term information to obtain:

[0033]

[0034] Introduce the transient voltages \(u_d'(k)\) and \(u_q'(k)\) on the d-axis and q-axis into the model with the transient input quantity coefficient \(\alpha'\) d (k) and \(u'\) q (k), and there is: so that the model with the transient input quantity coefficient \(\alpha'\) is rewritten in the following form:

[0035]

[0036] Furthermore, the given values of the transient currents on the d-axis and q-axis at the k-th moment, \(i_d'(k)\) and \(i_q'(k)\), are obtained by the following formula: dref (k) and \(i'\) qref (k) are obtained by the following formula:

[0037]

[0038] where \(i\) dref (k) and \(i\) qref (k) are the given values of the synchronous currents on the d-axis and q-axis at the k-th moment respectively, and \(F_d(k)\) and \(F_q(k)\) are the observable disturbance terms on the d-axis and q-axis at the k-th moment respectively. d (k) and \(F_q(k)\) are the observable disturbance terms on the d-axis and q-axis at the k-th moment respectively.

[0039] Furthermore, the above observable disturbance terms are observed by an extended state observer, and the expression of the extended state observer is as follows:

[0040]

[0041] where err(k) is the observation error vector at time k, z 1 (k) and z 2 (k) are the current observation vector and the disturbance term observation vector at time k, respectively, and there are and are the current observation values of the d-axis and q-axis at time k, respectively, z 1 (k + 1) and z 2 (k + 1) are the current observation vector and the disturbance term observation vector at time k + 1, respectively, i dq (k) is the current vector at time k and there are β 1 and β 2 are both the loop pole coefficients of the extended state observer, is the input quantity gain of the extended state observer.

[0042] Furthermore, the loop pole coefficients β 1 and β 2 of the above extended state observer are respectively expressed as:

[0043] β 1 = 2ω o ,

[0044] where ω o is the observation loop pole of the extended state observer.

[0045] Furthermore, the input quantity gain

[0046] L dq of the above extended state observer is the dq-axis inductance vector, and there is L dq = [L d L q , L d and L q are the d-axis and q-axis inductances respectively.

[0047] Furthermore, the constraint conditions for the above ultra-local model to construct the reduced-dimensional line constraint EMPC model are:

[0048]

[0049] where I qmax is the maximum current value of the q-axis, U max is the maximum voltage value of the dq-axis, and ε is the correction coefficient of the voltage constraint.

[0050] Furthermore, the correction coefficient ε of the above voltage constraint = [ε d ε q , where εd and ε q are the d - axis and q - axis components of ε respectively.

[0051] Furthermore, the above-mentioned reduction of the transient voltage U' dq to the actual voltage U of the motor control system dq includes:

[0052] Since the transient voltage the actual voltage of the motor control system then:

[0053]

[0054] ε d and ε q are the d - axis and q - axis components of the correction coefficient ε of the voltage constraint respectively.

[0055] Furthermore, the above-mentioned Lagrange coefficient λ expression is as follows:

[0056] λ = M ac x(k)+m ac ,

[0057] where, M ac is the coefficient matrix output by the reduced - dimension line - constraint EMPC model and has:

[0058]

[0059] m ac is the constant matrix output by the reduced - dimension line - constraint EMPC model and has:

[0060]

[0061] The reduced - dimension line - constraint EMPC control method for the motor of new - energy vehicles described in the present invention inherits the advantages of the line - constraint EMPC and further reduces the system dimension to 4 - D through model improvement. This improvement not only reduces the memory occupancy rate but also improves the execution efficiency. In addition, the present invention transfers the parameters related to the controlled object outside the offline part of the algorithm. When the controlled object changes, there is no need to regenerate the offline matrix of the EMPC, which greatly improves the portability of the algorithm without increasing the computational burden of the algorithm. Description of the Drawings

[0062] Figure 1 is the simulation diagram of the verification result of portability, where (a) represents the DRLC EMPC simulation of the dq - axis inductance given from the external parameter interface, and (b) represents the line - constraint EMPC simulation after accurate modeling;

[0063] Figure 2It is the structure diagram of the dimensionality reduction line-constrained EMPC;

[0064] Figure 3 It is the structure diagram of the closed-loop system;

[0065] Figure 4 It is the flowchart of the dimensionality reduction line-constrained EMPC control method for the motor of new energy vehicles described in the specific implementation manner. Specific implementation manner

[0066] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention. It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0067] 1. Model of the controlled motor

[0068] According to the Euler discretization method, the initial model of the controlled motor can be obtained as follows:

[0069]

[0070] Where: F d and F q are the observable disturbance terms on the d-axis and q-axis respectively, k is the sampling time, and there are:

[0071]

[0072] α d and α q are the input quantity coefficients on the d-axis and q-axis respectively, and there are

[0073] R is the stator resistance, L d and L q are the synchronous inductances on the d-axis and q-axis respectively, i d and i q are the synchronous currents on the d-axis and q-axis respectively, u d and u q are the synchronous voltages on the d-axis and q-axis respectively, ω is the rotor electrical angular velocity, λ r is the rotor magnetic flux, T s is the sampling time.

[0074] The disturbance terms F d and F qIt can be observed by the observer and compensated to the original system. The observer in this embodiment adopts an extended state observer, and its expression is as follows:

[0075]

[0076] where: err is the observation error vector. i dq is the current vector and has z 1 is the current observation vector and has and are the d-axis and q-axis current observation values respectively; z 2 is the disturbance term observation vector and has and are the d-axis and q-axis disturbance term observation values respectively. is the observer input quantity gain and has L dq =[L d L q is the dq-axis inductance vector. β 1 and β 2 are both the loop pole coefficients of the observer and have:

[0077]

[0078] where, ω o is the observation loop pole of the observer.

[0079] Compensating the dq-axis disturbance terms observed by formula (2) into formula (1) can complete the construction of the initial model of the permanent magnet synchronous motor.

[0080] 2. Basic Principle of EMPC

[0081] The control objective of EMPC can be expressed in the following quadratic programming form with constraints:

[0082]

[0083] where, x is the state variable, u is the input quantity of the motor control system, y is the process output variable, r is the target control quantity, A and B are the system transfer matrix and input matrix respectively (only related to the constructed motor mathematical model), C is the motor control system output matrix, Q is the state variable parameter matrix, P is the input quantity parameter matrix, N p is the prediction time domain length.

[0084] Through iteration, formula (4) can be transformed into the following standard form:

[0085]

[0086] Among them, is the input matrix of the cost function, and N u is the length of the control time domain.

[0087] Y is the quadratic coefficient matrix of the state variable, H is the quadratic coefficient matrix of the input quantity, F is the linear coefficient matrix of the state variable, and Y, H, F are iterated from A, B; G, W, E are the coefficient matrices of the constraint conditions.

[0088] The constraint conditions can be divided into effective constraints and ineffective constraints according to whether they are triggered:

[0089]

[0090] In the formula, G ac is the effective constraint coefficient matrix of the input quantity, W ac is the constant term matrix of the effective constraint condition, E ac is the effective coefficient matrix of the state variable; G ia is the ineffective constraint coefficient matrix of the input quantity, W ia is the constant term matrix of the ineffective constraint condition, E ia is the ineffective coefficient matrix of the state variable.

[0091] Using the KKT conditions to solve, the following solutions can be obtained:

[0092]

[0093] Among them: λ is the Lagrange coefficient.

[0094] M ac is the coefficient matrix output by the EMPC and has

[0095] m ac is the constant matrix output by the EMPC and has

[0096] According to formula (7), it can be seen that as long as the model of the controlled object is determined, then F, G ac , λ are uniquely determined. Therefore, only after the EMPC determines the model of the controlled object, it discretely calculates the parameters of F, G ac , λ, avoiding online iterative calculation. The parameters calculated offline are saved in the affine space, which reduces the computational amount to a certain extent but increases the memory occupancy rate of the system.

[0097] 3. Reduced-dimensional line constraint EMPC

[0098] The Dimensionality-reduced Line-constrained EMPC (DRLC EMPC) proposed in this embodiment aims to construct the state model of EMPC with fewer dimensions and make it portable, which is not possessed by traditional EMPC.

[0099] First step, to achieve the dimensionality reduction of EMPC, by transposing formula (1), we can get:

[0100]

[0101] Introduce new variables: the transient currents i d '(k + 1) and i q '(k + 1) on the d-axis and q-axis, and satisfy:

[0102] i d '(k + 1) = i d (k + 1) - T s F d (k), i q '(k + 1) = i q (k + 1) - T s F q (k),

[0103] Then we can get:

[0104]

[0105] According to the discrete state model rule, the processing of formula (9) is actually equivalent to shifting the d-axis and q-axis current reference values i dref (k) and i qref (k) forward by -T s F d (k) and -T s F q (k). And formula (9) does not contain the information of the disturbance term, so the dimension of the system is smaller, achieving the dimensionality reduction of the system.

[0106] The second step is to achieve the portability of the system. According to formula (7), as long as the parameters of the controlled object of the system are uniquely determined, the parameter matrix of the solution is uniquely determined. Therefore, when the controlled object changes, the parameter matrix has to be recalculated offline, which also limits the portability of the algorithm. The method of this embodiment is to transfer the parameters of the controlled object outside the EMPC algorithm. The only terms in formula (9) that contain the parameters of the controlled object are α d and α q , so the following processing can be done to formula (9):

[0107]

[0108] where: α' is the transition input quantity coefficient, α' = 1 / L bas , L bas is the transition inductance, and its value can be at 1×10 -3 or even smaller.

[0109] Introduce the transition voltages u' d (k) and u' q (k) on the d-axis and q-axis, and their expressions are as follows:

[0110]

[0111]

[0112] where ε d and ε q are the correction coefficients of the transition voltages u' d (k) and u' q (k) on the d-axis and q-axis respectively.

[0113] Using the transition voltages u' d (k) and u' q (k), formula (10) can be transformed into the following expression:

[0114]

[0115] By introducing the transition voltages u' d (k) and u' q (k) on the d-axis and q-axis and the transition input quantity coefficient α', the parameters related to the controlled motor are transferred outside the prediction model of the EMPC, thus realizing the portability of the motor control system.

[0116] According to formula (12), the state variables x(k) of the system and the state equation matrices A, B, and C can be selected as follows:

[0117]

[0118]

[0119] where i' dref (k) and i' qref (k) are the given values of the transition currents on the d-axis and q-axis at time k respectively, and their expressions are as follows:

[0120]

[0121] i dref (k) and i qref (k) are the given values of the synchronous currents on the d-axis and q-axis at time k respectively.

[0122] The constraint conditions of the dimensionality reduction line constraint EMPC are constructed as follows:

[0123]

[0124] Among them, I qmax is the maximum q-axis current value; U max is the maximum dq-axis voltage value; is the transition voltage vector; and ε is the correction coefficient of the voltage constraint, and ε = [ε d ε q .

[0125] For the actual output of the system it is necessary to restore the transition voltage to the actual voltage, that is:

[0126]

[0127] The construction of the entire dimensionality reduction line constraint EMPC is completed above. The DRLC EMPC system structure diagram and the closed-loop system structure block diagram are as shown in Figure 2 and Figure 3 .

[0128] To verify the effectiveness of this embodiment, the Infineon XMC4500 embedded chip is used in this embodiment to prove the effectiveness of the algorithm.

[0129] First, it is a comparison of the execution efficiency. The comparison algorithm is the line constraint EMPC,

[0130] Table 1 Algorithm Comparison

[0131] DRLC EMPC Line-constrained EMPC System dimension 4 6 Minimum prediction horizon 2 2 Number of partitions 4 7 Memory occupancy rate 1.7 kB 4 kB Execution time 27 μs 61 μs Whether it has portability Yes No

[0132] It can be seen from Table 1 that the proposed DRLC EMPC in this embodiment has a significant improvement in both the memory occupancy rate and the execution time compared with the line constraint EMPC. Through Figure 1 it can be known that the response of the DRLC EMPC to give parameters through the parameter interface is exactly the same as the response of the EMPC after accurate modeling, which also proves the portability of the DRLC EMPC.

[0133] In summary, the DRLC EMPC proposed in this embodiment has the following advantages compared with the traditional line constraint EMPC algorithm:

[0134] 1. The system dimension proposed by the algorithm proposed in this embodiment is only 4D, and the resulting memory occupancy rate and execution time are both reduced to one-half, which is more suitable for the application of new energy vehicle embedded devices.

[0135] 2. By transferring the parameters related to the controlled system outside the system, the DLRC EMPC algorithm enables the prediction model of the EMPC not to include the parameters of the controlled object and provides a parameter interface, which is not available in traditional EMPC, greatly improving the portability of the algorithm.

[0136] Although the present invention has been described herein with reference to particular embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Accordingly, it should be understood that numerous modifications may be made to the exemplary embodiments, and other arrangements may be designed, without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that the features described herein may be combined in different ways than those described in the original claims. It should also be understood that the features described in connection with separate embodiments may be used in other described embodiments.

Claims

1. A dimensionality reduction line constraint EMPC control method for a new energy vehicle motor, Characterized in that, Construct a super-local model of the controlled motor: where, i d '(k + 1) and i q '(k + 1) are the transient currents on the d-axis and q-axis at the (k + 1)-th moment, respectively, i d (k) and i q (k) are the synchronous currents on the d-axis and q-axis at the k-th moment, respectively, T s is the sampling period, u' d (k) and u' q (k) are the transient voltages on the d-axis and q-axis at the k-th moment, respectively, α' is the transient input quantity coefficient, α' = 1 / L bas , L bas is the transient inductance; Construct a dimensionality reduction line constraint EMPC model based on the super-local model: Where J(k) is the control objective of the dimensionality reduction line constraint EMPC model, x(k) is a state variable and there is: i' dref (k) and i' qref (k) are the given values of the transient currents on the d-axis and q-axis at time k, respectively, and U' dq is the transient voltage and there is u'(k) is the transient voltage of the motor control system at time k, N u is the time domain length of motor control, Y is the quadratic coefficient matrix of the state variables, H is the quadratic coefficient matrix of the input variables, F is the linear coefficient matrix of the state variables, G ac is the effective constraint coefficient matrix of the input variables, W ac is the constant term matrix of the effective constraint conditions, E ac is the effective coefficient matrix of the state variables; Solve the dimensionality reduction line constraint EMPC model using the KKT conditions to obtain the transition voltage U'. dq ; Where λ is the Lagrange coefficient; Restore the transition voltage U' dq to the actual voltage U of the motor control system dq , and use U dq to achieve the reduced-dimension line-constrained EMPC control of the motor of new energy vehicles.

2. The dimensionality reduction line constraint EMPC control method for a new energy vehicle motor according to claim 1, Characterized in that, The construction of the super-local model of the controlled motor includes: Construct an initial model of the controlled motor according to the Euler discretization method: where, i d (k + 1) and i q (k + 1) are the synchronous currents of the d-axis and q-axis at the (k + 1)-th moment, respectively, and i d (k) and i q (k) are the synchronous currents of the d-axis and q-axis at the k-th moment, respectively. F d (k) and F q (k) are the observable disturbance terms of the d-axis and q-axis at the k-th moment, respectively. α d and α q are the input quantity coefficients of the d-axis and q-axis, respectively. u d (k) and u q (k) are the synchronous voltages of the d-axis and q-axis at the k-th moment; Phase shift the initial model to obtain: Introduce the transient currents \(i_d'(k + 1)\) and \(i_q'(k + 1)\) of the d-axis and q-axis at the \((k + 1)\)-th moment into the initial model after phase shift, and there is: \(i_d'(k + 1)=i_d(k + 1)-T_F(k)\), \(i_q'(k + 1)=i_q(k + 1)-T_F(k)\), so that the initial model after phase shift is rewritten as a model that does not contain disturbance term information: d '(k + 1) and i q '(k + 1), and there is: i d '(k + 1) = i d (k + 1)-T s F d (k), i q '(k + 1) = i q (k + 1)-T s F q (k), such that the initial model after phase shift is rewritten as a model that does not contain disturbance term information: Introduce a transition input quantity coefficient α' into the model without disturbance term information to obtain: In the model introducing the transition input quantity coefficient α', the transition voltages u' d (k) and u' q (k) of the d-axis and q-axis are introduced, and there is: such that the model introducing the transition input quantity coefficient α' is rewritten into the following form:

3. The dimensionality reduction line constraint EMPC control method for a new energy vehicle motor according to claim 1, Characterized in that, The given values of the transient currents on the d-axis and q-axis at the k-th moment, \(i'_{d}(k)\) and \(i'_{q}(k)\), are obtained by the following formula: dref \(i'_{d}(k)\) and \(i'_{q}(k)\) qref are obtained by the following formula: where, i dref (k) and i qref (k) are the d-axis and q-axis synchronous current reference values at time k respectively, and F d (k) and F q (k) are the observable disturbance terms of the d-axis and q-axis at time k respectively.

4. The dimensionality reduction line constraint EMPC control method for a new energy vehicle motor according to claim 2 or 3, Characterized in that, The observable disturbance term is obtained by observing through an extended state observer, and the expression of the extended state observer is as follows: where err(k) is the observation error vector at time k, z 1 (k) and z 2 (k) are the current observation vector and the disturbance term observation vector at time k, respectively, and there are and are the current observation values of the d-axis and q-axis at time k, respectively, z 1 (k + 1) and z 2 (k + 1) are the current observation vector and the disturbance term observation vector at time k + 1, respectively, i dq (k) is the current vector at time k and there are β 1 and β 2 are both the loop pole coefficients of the extended state observer, is the input gain of the extended state observer.

5. The dimensionality reduction line constraint EMPC control method for a new energy vehicle motor according to claim 4, Characterized in that, The loop pole coefficient β of the extended state observer 1 and β 2 are respectively expressed as: β 1 = 2ω o , where ω o is the observation loop pole of the extended state observer.

6. The dimensionality reduction line constraint EMPC control method for a new energy vehicle motor according to claim 5, Characterized in that, Expansion state observer input gain L dq is the dq-axis inductance vector, and there is L dq = [L d L q ], L d and L q are the d-axis and q-axis inductances respectively.

7. The dimensionality reduction line constraint EMPC control method for a new energy vehicle motor according to claim 1, Characterized in that, The constraint condition for the super-local model to construct the dimensionality reduction line constraint EMPC model is: Among them, I qmax is the maximum q-axis current value, U max is the maximum dq-axis voltage value, and ε is the correction coefficient of voltage constraint.

8. The dimensionality reduction line constraint EMPC control method for a new energy vehicle motor according to claim 7, Characterized in that, The correction coefficient ε of the voltage constraint = [ε d ε q , where ε d and ε q are the d-axis and q-axis components of ε, respectively.

9. The dimensionality reduction line constraint EMPC control method for a new energy vehicle motor according to claim 1, Characterized in that, The reduction of the transition voltage U' dq to the actual voltage U of the motor control system dq , includes: Due to the transition voltage The actual voltage of the motor control system Then: ε d and ε q are the d-axis and q-axis components of the correction coefficient ε for voltage constraint, respectively.

10. The dimensionality reduction line constraint EMPC control method for a new energy vehicle motor according to claim 1, Characterized in that, The expression of the Lagrange coefficient λ is as follows: λ = M ac x(k) + m ac , Among them, M ac is the coefficient matrix output by the dimensionality reduction line-constrained EMPC model and has: m ac is a constant matrix output by the dimensionality reduction line constraint EMPC model, and there is:

Images