On-line adaptive optimization method for maximum torque per ampere of interior permanent magnet synchronous motor

Abstract

The application discloses an online self-adaptive optimization method for maximum torque current ratio of an interior permanent magnet synchronous motor, and specifically comprises the following steps: step 1, establishing a maximum torque current ratio cost function of the interior permanent magnet synchronous motor; step 2, calculating an iteration step length of online self-adaptive optimization according to the maximum torque current ratio cost function obtained in step 1; and step 3, searching for a maximum torque current ratio working point by means of an adaptive optimization algorithm according to the iteration step length obtained in step 2. The method solves the problems that the existing gradient descent method has a very slow convergence speed when approaching the actual maximum torque current ratio working point, is difficult to reach the maximum torque current ratio working point, and is prone to oscillation in the process of searching for the maximum torque current ratio working point when the Jacobian matrix is a singular matrix.

Description

Technical Field

[0001] This invention belongs to the field of permanent magnet synchronous motor control technology, and relates to an online adaptive optimization method for the maximum torque-current ratio of a built-in permanent magnet synchronous motor. Background Technology

[0002] Under the development trend of advocating energy conservation and emission reduction, built-in permanent magnet synchronous motors (PMSMs) have been successfully applied in many fields and have achieved good energy-saving results due to their advantages such as high efficiency, high power density, high response, and salient pole structure. The direct-axis inductance and quadrature-axis inductance of a built-in PMSM are not equal, which can generate reluctance torque. By using maximum torque-to-current ratio control, the reluctance torque can be fully utilized to improve load capacity and working efficiency. The core challenge of maximum torque-to-current ratio control is finding the operating point of maximum torque-to-current ratio, that is, finding a suitable stator current vector angle that minimizes the amplitude of the electronic current while outputting the same electromagnetic torque.

[0003] The main methods for maximum torque-to-current ratio (MTR) control include: formula-based methods, formula calculation methods combined with parameter identification, table lookup methods, virtual signal injection methods, actual high-frequency signal injection methods, and search methods. Formula-based methods, virtual signal injection methods, and table lookup methods cannot overcome the problem of inaccurate operating points for the MTR caused by changes in motor parameters. Due to the rank deficiency problem, it is difficult to simultaneously identify the direct-axis inductance, quadrature-axis inductance, and permanent magnet flux linkage online. Furthermore, when the motor is running dynamically, online motor parameter identification methods struggle to accurately identify motor parameters, leading to inaccurate MTR operating points calculated by formula calculation methods combined with parameter identification. The actual high-frequency signal injection method causes torque and speed pulsations due to the injected high-frequency signal and generates additional high-frequency losses. Currently, the maximum torque-to-current ratio control method usually adopts a search method. Among the search methods, the gradient descent method is widely used due to its simplicity. However, the convergence speed of the gradient descent method is linear, and the convergence speed is very slow when approaching the actual maximum torque-to-current ratio operating point, making it difficult to reach the actual maximum torque-to-current ratio operating point. In addition, when the Jacobian matrix is ​​a singular matrix, the gradient descent method is prone to oscillation during the search for the maximum torque-to-current ratio operating point, leading to search failure. Summary of the Invention

[0004] The purpose of this invention is to provide an online adaptive optimization method for the maximum torque-current ratio of a built-in permanent magnet synchronous motor. This method solves the problems of slow convergence speed of the existing gradient descent method when approaching the actual maximum torque-current ratio operating point, making it difficult to reach the maximum torque-current ratio operating point, and the tendency of the gradient descent method to oscillate during the search for the maximum torque-current ratio operating point when the Jacobian matrix is ​​a singular matrix.

[0005] The technical solution adopted in this invention is an online adaptive optimization method for the maximum torque-to-current ratio of a built-in permanent magnet synchronous motor, which specifically includes the following steps:

[0006] Step 1: Establish the maximum torque-to-current ratio cost function for the built-in permanent magnet synchronous motor;

[0007] Step 2: Calculate the iteration step size for online adaptive optimization using the maximum torque-to-current ratio cost function obtained in Step 1;

[0008] Step 3: The maximum torque-to-current ratio operating point is found using an adaptive optimization method based on the iteration step size obtained in Step 2.

[0009] The invention is further characterized by:

[0010] The specific process of step 1 is as follows:

[0011] The steady-state voltage equation of the built-in permanent magnet synchronous motor is shown in the following formula (1):

[0012]

[0013] Among them, u d (k), u d (k) represent the components of the stator voltage on the d-axis and q-axis, respectively, for the k-th pulse. d (k), i q (k) represent the components of the stator current in the k-th cycle along the d-axis and q-axis, respectively, ω r (k) is the rotor electrical angular frequency of the kth cycle, L d It is the d-axis inductance, L q It is a q-axis inductor, L qd The mutual inductance along the d-axis is caused by the cross-coupling effect, L dq The mutual inductance along the q-axis is caused by the cross-coupling effect, R s It is the stator resistance, ψ f It is a permanent magnet flux chain;

[0014] The electromagnetic torque equation of the built-in permanent magnet synchronous motor is shown in the following formula (2):

[0015]

[0016] Among them, T e [k] is the electromagnetic torque at the k-th beat, n p It is an extreme logarithm;

[0017] Multiply both sides of the first equation in formula (1) by i d [k] and both sides of the second equation are multiplied by i. q [k] is shown in the following formula (3):

[0018]

[0019] The two equations in the addition formula (3) are rearranged as shown in the following formula (4):

[0020]

[0021] For a given current vector, the d-axis current and q-axis current are shown in the following formula (5):

[0022]

[0023] Among them, i s [k] is the given current vector output by the speed loop in the k-th cycle, and φ[k] is the given current vector i in the k-th cycle. s The angle between [k] and the q-axis;

[0024] Substituting formulas (2) and (5) into (4) yields... As shown in formula (6) below:

[0025]

[0026] In one switching cycle, the resistance R is considered to be... s and rotational speed ω r [k] is independent of the current vector angle φ[k]. The maximum value is equivalent to finding the maximum value of the function shown in formula (7) below:

[0027]

[0028] Multiplying both sides of formula (7) by -1 results in formula (8) as follows:

[0029] g(φ[k])=-f(φ[k]) (8);

[0030] Where g(φ[k]) is the maximum torque-to-current ratio cost function;

[0031] The maximum torque-to-current ratio operating point is to calculate T. e [k] / i s Find the maximum value of [k], and find T. e [k] / i s The maximum value of [k] is equivalent to finding Find the maximum value of . The maximum value of f(φ[k]) is equivalent to finding the maximum value of f(φ[k]), and finding the maximum value of f(φ[k]) is equivalent to finding the minimum value of g(φ[k]). Therefore, the maximum torque-current ratio operating point is equivalent to finding the minimum value of the maximum torque-current ratio cost function g(φ[k]).

[0032] The specific process of step 2 is as follows:

[0033] The adaptive step size adjustment factor τ of the adaptive optimization algorithm is calculated according to formula (8). i As shown in formula (9) below:

[0034] τ i =χ i |g(φ[i])| (9);

[0035] Where, τ i It is an adaptive step size adjustment factor, χ i is the adaptive coefficient, and i represents the i-th iteration.

[0036] The initial iteration step size of the adaptive optimization algorithm, calculated according to formulas (8) and (9), is shown in formula (10) below:

[0037]

[0038] Where, d i1 It is the initial iteration step size of the i-th iteration of the adaptive optimization algorithm. It is the Jacobian matrix of the i-th iteration;

[0039] The iteration step size of the adaptive optimization algorithm is calculated according to formulas (8) to (10) as shown in formula (11):

[0040]

[0041] Where, d i2 It is the iteration step size of the next iteration in the adaptive optimization algorithm;

[0042] The iteration step size of the tail iteration step of the adaptive optimization algorithm is calculated according to formulas (8) to (11) as shown in formula (12):

[0043]

[0044] Where, d i3 It is the iteration step size of the tail iteration step in the adaptive optimization algorithm;

[0045] The iteration step size of the adaptive optimization algorithm is calculated according to formulas (10) to (12) as shown in formula (13):

[0046]

[0047] Where, d i It is the iteration step size of the i-th iteration of the adaptive optimization algorithm. This is an adjustable parameter.

[0048] The specific process of step 3 is as follows:

[0049] To verify the effectiveness of the current iteration step, the effectiveness evaluation index of the current iteration step size is calculated as shown in the following formula (14):

[0050]

[0051] Among them, c i It is an evaluation index for the effectiveness of the current iteration step size;

[0052] The current vector angle for this iteration is calculated as shown in formula (15):

[0053]

[0054] Among them, h 01 h 02 h 03 It is the set constant threshold, and φ[i+1] is the current vector angle obtained in the i-th iteration;

[0055] Calculate adjustable parameters As shown in the following formula (16):

[0056]

[0057] The Jacobian matrix is ​​calculated as shown in formula (17):

[0058]

[0059] Where J(φ[i+1]) is the Jacobian matrix of the (i+1)th iteration;

[0060] The adaptive coefficient of the updated adaptive step size adjustment factor is shown in the following formula (18):

[0061]

[0062] Where h2 is the effectiveness evaluation index of the current iteration step size when the adaptive coefficient is used. i The lower limit threshold, h3 is the adaptive coefficient, and c is the effectiveness evaluation index of the current iteration step size. i The upper limit threshold;

[0063] The adaptive step size adjustment factor is updated as shown in formula (19):

[0064]

[0065] Where h1 is the effectiveness evaluation index of the current iteration step size when updating the adaptive step size adjustment factor. i The lower limit threshold.

[0066] In step 3, the process of finding the operating point with the maximum torque-current ratio using an online adaptive optimization method is as follows:

[0067] 1) Set the initial iteration value φ[1], and set the maximum number of iterations i max ≥1, given constant 0 < h 01 <h02 <h 03 <h2<h1<h3<1. Optimize the precision ε, χ1 and i = 1, calculate J(φ[1]), τ1 = χ1|g(φ[1])|;

[0068] 2) Calculate the step sizes d i1 , d i2 , d i3 and d i ;

[0069] 3) Calculate the evaluation index c of the current iteration step size according to formula (14) i ;

[0070] 4) Calculate the current vector angle of this iteration according to formula (15), and use the criterion |J(φ[i])g(φ[i])| < ε to judge whether the optimization precision is satisfied. If the optimization precision is satisfied, exit this optimization search and output the current vector angle φ[k] = φ[i] optimized in this beat, and calculate the d-axis current i d [k] and the q-axis current i d [k]. Otherwise, judge whether the iteration number i has reached the maximum iteration number i max . If the maximum iteration number i max has been reached, then exit this optimization search and output the current vector angle φ[k] = φ[k - 1], and calculate the d-axis current i d [k] and the q-axis current i d [k]. Otherwise, set i = i + 1 and continue to execute step 5);

[0071] 5) Calculate the adjustable parameter according to formula (16)

[0072] 6) Calculate the Jacobian matrix J(φ[i + 1]) according to formula (17);

[0073] 7) Update the adaptive coefficient χ according to formula (18) i+1 ;

[0074] 8) Update the adaptive step size adjustment factor τ according to formula (19) i+1 , and return to step 2).

[0075] The beneficial effects of this invention are that, compared with the traditional gradient descent method for searching the operating point of maximum torque-to-current ratio, this invention employs an online adaptive optimization method to search for the operating point of maximum torque-to-current ratio. By using an adaptive step size adjustment factor and multi-step step sizes, the convergence speed of searching for the operating point of maximum torque-to-current ratio is accelerated, reducing the search time. This solves the problem that the gradient descent method has a very slow convergence speed when approaching the actual operating point of maximum torque-to-current ratio, making it difficult to reach the actual operating point. Simultaneously, the online adaptive optimization method selects different step sizes based on the magnitude of the effectiveness evaluation index of the current iteration step size, ensuring convergence of the search for the operating point of maximum torque-to-current ratio even when the Jacobian matrix is ​​a singular matrix. Attached Figure Description

[0076] Figure 1 This is a block diagram of the vector control system used in the online adaptive optimization method for the maximum torque-current ratio of the built-in permanent magnet synchronous motor of the present invention.

[0077] Figure 2 This is a flowchart of the adaptive optimization method used in the online adaptive optimization method for the maximum torque-current ratio of the built-in permanent magnet synchronous motor of the present invention.

[0078] Figure 3 The q-axis current given i is optimized using the traditional gradient descent method. q [k] and q-axis feedback current i qf [k] Waveform;

[0079] Figure 4 The d-axis current given i is optimized using the traditional gradient descent method. d [k] and d-axis feedback current i df [k] Waveform;

[0080] Figure 5 The q-axis current given by i is optimized using the online adaptive optimization method for the maximum torque-to-current ratio of the built-in permanent magnet synchronous motor of this invention. q [k] and q-axis feedback current i qf [k] Waveform;

[0081] Figure 6 The d-axis current given by i is optimized using the online adaptive optimization method for the maximum torque-to-current ratio of the built-in permanent magnet synchronous motor of this invention. d [k] and d-axis feedback current i df [k] Waveform. Detailed Implementation

[0082] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0083] The present invention discloses an online adaptive optimization method for the maximum torque-to-current ratio of a built-in permanent magnet synchronous motor, wherein the block diagram of the vector control system used is as follows. Figure 1 As shown, the specific steps are as follows:

[0084] Step 1: Establish the maximum torque-to-current ratio cost function for the built-in permanent magnet synchronous motor, specifically as follows:

[0085] The steady-state voltage equation of the built-in permanent magnet synchronous motor is shown in the following formula (1):

[0086]

[0087] Among them, u d (k), u d (k) represent the components of the stator voltage on the d-axis and q-axis, respectively, for the k-th pulse. d (k), i q (k) represent the components of the stator current in the k-th cycle along the d-axis and q-axis, respectively, ω r (k) is the rotor electrical angular frequency of the kth cycle, L d It is the d-axis inductance, L q It is a q-axis inductor, L qd The mutual inductance along the d-axis is caused by the cross-coupling effect, L dq The mutual inductance along the q-axis is caused by the cross-coupling effect, R s It is the stator resistance, ψ f It is a permanent magnet flux linkage.

[0088] The electromagnetic torque equation of the built-in permanent magnet synchronous motor is shown in the following formula (2):

[0089]

[0090] Among them, T e [k] is the electromagnetic torque at the k-th beat, n p It is an extreme logarithm.

[0091] Multiply both sides of the first equation in formula (1) by i d [k] and both sides of the second equation are multiplied by i. q [k] is shown in the following formula (3):

[0092]

[0093] The two equations in the addition formula (3) are rearranged as shown in the following formula (4):

[0094]

[0095] For a given current vector, the d-axis current and q-axis current are shown in the following formula (5):

[0096]

[0097] Among them, i s [k] is the given current vector output by the speed loop in the k-th cycle, and φ[k] is the given current vector i in the k-th cycle. s [k] is the angle between the q-axis and the k-axis.

[0098] Substituting formulas (2) and (5) into (4) yields... As shown in formula (6) below:

[0099]

[0100] In one switching cycle, the resistance R can be considered as... s and rotational speed ω r [k] is independent of the current vector angle φ[k]. The maximum value is equivalent to finding the maximum value of the function shown in formula (7) below:

[0101]

[0102] Multiplying both sides of formula (7) by -1 results in formula (8) as follows:

[0103] g(φ[k])=-f(φ[k]) (8);

[0104] Where g(φ[k]) is the maximum torque-current ratio cost function.

[0105] The maximum torque-to-current ratio operating point is to calculate T. e [k] / i s Find the maximum value of [k], and find T. e [k] / i s The maximum value of [k] is equivalent to finding Find the maximum value of . The maximum value of f(φ[k]) is equivalent to finding the maximum value of f(φ[k]), and finding the maximum value of f(φ[k]) is equivalent to finding the minimum value of g(φ[k]). Therefore, the maximum torque-current ratio operating point is equivalent to finding the minimum value of the maximum torque-current ratio cost function g(φ[k]).

[0106] Step 2: Calculate the iteration step size for online adaptive optimization using the maximum torque-to-current ratio cost function obtained in Step 1;

[0107] The adaptive step size adjustment factor τ of the adaptive optimization algorithm is calculated according to formula (8). i As shown in formula (9) below:

[0108] τ i =χ i |g(φ[i])| (9);

[0109] Where, τ iIt is an adaptive step size adjustment factor, χ i is the adaptive coefficient, and i represents the i-th iteration.

[0110] The initial iteration step size of the adaptive optimization algorithm, calculated according to formulas (8) and (9), is shown in formula (10) below:

[0111]

[0112] Where, d i1 It is the initial iteration step size of the i-th iteration of the adaptive optimization algorithm. It is the Jacobian matrix of the i-th iteration.

[0113] The iteration step size of the next iteration of the adaptive optimization algorithm is calculated according to formulas (8) to (10) as shown in formula (11):

[0114]

[0115] Where, d i2 It is the iteration step size of the next iteration step in the adaptive optimization algorithm.

[0116] The iteration step size of the adaptive optimization algorithm is calculated according to formulas (8) to (11) as shown in formula (12):

[0117]

[0118] Where, d i3 It is the iteration step size of the tail iteration step in the adaptive optimization algorithm.

[0119] The iteration step size of the adaptive optimization algorithm is calculated according to formulas (10) to (12) as shown in formula (13):

[0120]

[0121] Where, d i It is the iteration step size of the i-th iteration of the adaptive optimization algorithm. This is an adjustable parameter.

[0122] Step 3, the iteration step size obtained in step 2 is obtained by... Figure 2 The adaptive optimization method shown seeks the operating point with the maximum torque-to-current ratio, specifically as follows:

[0123] To verify the effectiveness of the current iteration step, the effectiveness evaluation index of the current iteration step size is calculated as shown in the following formula (14):

[0124]

[0125] Among them, c i It is an evaluation index for the effectiveness of the current iteration step size.

[0126] The current vector angle for this iteration is calculated as shown in formula (15):

[0127]

[0128] Among them, h 01 h 02 h 03 φ[i+1] is the constant threshold set, and φ[i+1] is the current vector angle obtained in the i-th iteration.

[0129] Calculate adjustable parameters As shown in the following formula (16):

[0130]

[0131] The Jacobian matrix is ​​calculated as shown in formula (17):

[0132]

[0133] Where J(φ[i+1]) is the Jacobian matrix of the (i+1)th iteration.

[0134] The adaptive coefficient of the updated adaptive step size adjustment factor is shown in the following formula (18):

[0135]

[0136] Where h2 is the effectiveness evaluation index of the current iteration step size when the adaptive coefficient is used. i The lower limit threshold, h3 is the adaptive coefficient, and c is the effectiveness evaluation index of the current iteration step size. i The upper limit threshold.

[0137] The adaptive step size adjustment factor is updated as shown in formula (19):

[0138]

[0139] Where h1 is the effectiveness evaluation index of the current iteration step size when updating the adaptive step size adjustment factor. i The lower limit threshold.

[0140] Adopting such Figure 2 The process of the online adaptive optimization method shown to find the operating point with the maximum torque-to-current ratio is as follows:

[0141] 1) Set the initial iteration value φ[1], and set the maximum number of iterations i max ≥1, given constant 0 < h 01 <h 02 <h 03<h2><h1><h3>1. Optimize the precision ε, χ1 and i = 1, calculate J(φ[1]), τ1 = χ1|g(φ[1])|;

[0142] 2) Calculate the step size d according to formulas (10) to (13) i1 d i2 d i3 and d i ;

[0143] 3) Calculate the evaluation index c of the effectiveness of the current step iteration step according to formula (14) i ;

[0144] 4) Calculate the current vector angle of this iteration according to formula (15), and use the criterion |J(φ[i])g(φ[i])| < ε to judge whether the optimization precision is satisfied. If the optimization precision is satisfied, exit this optimization and output the current vector angle φ[k] = φ[i] optimized in this beat, and calculate the given d current i d [k] and q-axis current i d [k], otherwise judge whether the iteration number i has reached the maximum iteration number i max . If the maximum iteration number i max has been reached, then exit this optimization and output the current vector angle φ[k] = φ[k - 1], and calculate the given d current i d [k] and q-axis current i d [k], otherwise set i = i + 1 and continue to execute step 5);

[0145] 5) Calculate the adjustable parameter according to formula (16)

[0146] 6) Calculate the Jacobian matrix J(φ[i + 1]) according to formula (17);

[0147] 7) Update the adaptive coefficient χ according to formula (18) i+1 ;

[0148] 8) Update the adaptive step size adjustment factor τ according to formula (19) i+1 , and return to step 2).

[0149] The vector control system block diagram adopted by the online adaptive optimization method of the maximum torque current ratio of the built-in permanent magnet synchronous motor of the present invention is as Figure 1 shown. The system consists of 3 PI regulators to form a double-loop control of the speed loop and the current loop. The output of the speed loop PI regulator is the given current vector, and the given current vector passes through as Figure 2The online adaptive optimization shown yields the d-axis and q-axis current setpoints. These setpoints serve as inputs to the current loop PI regulator, whose output controls the power electronic converter.

[0150] The rotor speed ω is detected by installing an encoder on the rotor shaft of the built-in permanent magnet synchronous motor. r [k] and rotor position θ r [k] represents the given rotor speed of the speed loop. The rotor speed ω detected by the encoder r [k] is the difference, which, after passing through the speed loop PI controller, outputs the given stator current i. s [k]; Given stator current i s [k] via, as Figure 2 The online adaptive optimization method shown obtains the given d-axis current i d [k] and given q-axis current i q [k]; The stator current i of the built-in permanent magnet synchronous motor in the k-th phase of the three-phase stationary coordinate system is detected by a current Hall sensor. a [k]、i b [k]、i c [k]; Detected three-phase stator current i a [k]、i b [k]、i c [k] The k-th current value i is obtained by transforming the system to a two-phase stationary coordinate system using the abc / αβ transformation. α [k]、i β [k];i α [k]、i β [k] The k-th current value i is obtained by transforming the coordinate system to a two-phase synchronous rotating coordinate system using the αβ / dq transformation. df [k]、i qf [k]; Given d-axis current i d [k] and feedback current i df [k] is the difference, and the output d-axis voltage is obtained after passing through the current loop PI controller. Given q-axis current i q [k] and feedback current i qf [k] is the difference, and the output q-axis voltage is obtained after passing through the current loop PI controller. The voltage u in the two-phase stationary coordinate system is obtained after dq / αβ transformation. α [k]、u β [k], then through SVPWM modulation control of the three-phase inverter, finally driving the built-in permanent magnet synchronous motor to work.

[0151] Figure 3 Given i for the q-axis current optimized by the traditional gradient descent methodq [k] and q-axis feedback current i qf [k] Waveform; Figure 4 Given i for the d-axis current optimized by the traditional gradient descent method d [k] and d-axis feedback current i df [k] Waveform; Figure 5 The q-axis current is given by i in this invention. q [k] and q-axis feedback current i qf [k] Waveform;

[0152] Figure 6 The d-axis current is given by i in this invention. d [k] and d-axis feedback current i df [k] Waveform.

[0153] Figures 3-6 The parameters of the built-in permanent magnet synchronous motor used in the simulation are shown in Table 1. In the simulation results, the motor speed is set to the rated speed of 1000 rpm, and the load torque is set as follows: 0.8s-1s the motor runs under no-load; 1s-1.2s the load torque increases from 0 N.m to twice the rated torque; 1.2s-1.8s the motor runs under twice the rated torque; 1.8s-2s the load torque decreases from twice the rated torque to 0 N.m; 2s-2.2s the motor runs under no-load.

[0154] contrast Figure 3 and Figure 5 It can be observed that the method of this invention can significantly improve the optimization accuracy of the q-axis current. (Comparison) Figure 4 and Figure 6 It can be observed that the method of the present invention can significantly improve the optimization accuracy of the d-axis current. The simulation results above show that the online adaptive optimization method for the maximum torque-current ratio of the built-in permanent magnet synchronous motor of the present invention can significantly improve the accuracy of finding the operating point of the maximum torque-current ratio, thereby improving the working efficiency of the built-in permanent magnet synchronous motor drive system.

[0155] Table 1 Parameters of Built-in Permanent Magnet Synchronous Motor

[0156] Rated power 2kW Rated torque 19 N.m Extreme logarithm 4 Rated current 5.8A Rated speed 1000rpm Rated frequency 66.67Hz

Claims

1. An online adaptive optimization method for the maximum torque-to-current ratio of a built-in permanent magnet synchronous motor, characterized by: Specifically, the steps include the following: Step 1: Establish the maximum torque-to-current ratio cost function for the built-in permanent magnet synchronous motor; The specific process of Step 1 is as follows: The steady-state voltage equation of the built-in permanent magnet synchronous motor is shown in the following formula (1): （1）； in, , They are the first k Stator voltage at d shaft and q Components of the axis, , They are the first k Stator current at d shaft and q Components of the axis, ω r ( k ) is the first k The rotor's electrical angular frequency, L d yes d Shaft inductor, L q yes q Shaft inductor, It is caused by cross-coupling effect d Shaft mutual inductance, It is caused by cross-coupling effect q Shaft mutual inductance, It is the stator resistance. It is a permanent magnet flux chain; The electromagnetic torque equation of the built-in permanent magnet synchronous motor is shown in the following formula (2): （2） in, It is the first k Electromagnetic torque of the shot, It is an extreme logarithm; Multiply both sides of the first equation in formula (1) by Multiply both sides of the second equation by As shown in the following formula (3): （3）； The two equations in the addition formula (3) are rearranged as shown in the following formula (4): （4）； For a given current vector, d shaft current and q The shaft current is shown in the following formula (5): （5）； in, It is the first k The given current vector output by the speed loop, It is the first k Given current vector and q The included angle of the axis; Substituting formulas (2) and (5) into (4), we get As shown in the following formula (6): （6）； The resistance is considered in one switching cycle. and rotational speed With current vector angle Independence, Seeking The maximum value is equivalent to finding the maximum value of the function shown in formula (7) below: （7）； Multiplying both sides of formula (7) by -1 results in formula (8) as follows: （8）； in, It is the maximum torque-to-current ratio cost function; The maximum torque-to-current ratio operating point is to find Find the maximum value of . The maximum value is equivalent to finding Find the maximum value of . The maximum value is equivalent to finding Find the maximum value of . The maximum value is equivalent to finding The minimum value of the maximum torque current ratio operating point is therefore equivalent to finding the cost function of the maximum torque current ratio. The minimum value; Step 2: Calculate the iteration step size for online adaptive optimization using the maximum torque-to-current ratio cost function obtained in Step 1; the specific process of Step 2 is as follows: The adaptive step size adjustment factor of the adaptive optimization algorithm is calculated according to formula (8). As shown in the following formula (9): （9）； in, It is an adaptive step size adjustment factor. It is an adaptive coefficient. i Indicates the first i The next iteration; The initial iteration step size of the adaptive optimization algorithm, calculated according to formulas (8) and (9), is shown in formula (10) below: （10）； in, It is the adaptive optimization algorithm. i The initial iteration step size of the next iteration It is the first i The Jacobian matrix of the next iteration; The step size of the next iteration of the adaptive optimization algorithm is calculated according to formulas (8) to (10) as shown in formula (11): （11）； in, It is the step size of the next iteration in the adaptive optimization algorithm; The tail iteration step size of the adaptive optimization algorithm is calculated according to formulas (8) to (11) as shown in formula (12): （12）； in, It is the tail iteration step size of the adaptive optimization algorithm; The iteration step size of the adaptive optimization algorithm is calculated according to formulas (10) to (12) as shown in formula (13): （13）； in, It is the adaptive optimization algorithm. i The iteration step size of the next iteration These are adjustable parameters; Step 3: The maximum torque-to-current ratio operating point is found using an adaptive optimization method based on the iteration step size obtained in Step 2.

2. The online adaptive optimization method for the maximum torque-to-current ratio of the built-in permanent magnet synchronous motor according to claim 1, characterized in that: The specific process of step 3 is as follows: To verify the effectiveness of the current iteration step, the effectiveness evaluation index of the current iteration step size is calculated as shown in the following formula (14): （14）； in, It is an evaluation index for the effectiveness of the current iteration step size; The current vector angle for this iteration is calculated as shown in formula (15): （15）； in, , , It is a set constant threshold. It is the first The current vector angle obtained through iteration; Calculate adjustable parameters As shown in the following formula (16): （16） ； The Jacobian matrix is ​​calculated as shown in formula (17): （17）； in, It is the first i The Jacobian matrix after +1 iterations; The adaptive coefficient of the updated adaptive step size adjustment factor is shown in the following formula (18): （18）； in, When the adaptive coefficient is used as an evaluation index for the effectiveness of the current iteration step size, The lower limit threshold, When the adaptive coefficient is used as an evaluation index for the effectiveness of the current iteration step size, The upper limit threshold; The adaptive step size adjustment factor is updated as shown in formula (19): （19）； in, It is an evaluation index for the effectiveness of the current iteration step size when updating the adaptive step size adjustment factor. The lower limit threshold.

3. The online adaptive optimization method for the maximum torque-to-current ratio of the built-in permanent magnet synchronous motor according to claim 2, characterized in that: In step 3, the process of finding the operating point with the maximum torque-current ratio using an online adaptive optimization method is as follows: 1) Set initial values ​​for iteration Set the maximum number of iterations. Given a constant Optimize accuracy ε , and i =1, calculate , ; 2) Calculate the step size according to formulas (10) to (13). , , and ; 3) Calculate the effectiveness evaluation index of the current iteration step size according to formula (14). ; 4) Calculate the current vector angle for this iteration according to formula (15), and use the criterion. Determine if the optimization accuracy is met. If it is, exit the current optimization and output the optimized current vector angle for this cycle. And calculate the first according to formula (5) k Shoot given d Current and q shaft current Otherwise, check the number of iterations. i Has the maximum number of iterations been reached? If the maximum number of iterations has been reached Then exit the current optimization and output the current vector angle. And calculate the first according to formula (5) k Shoot given d Current and q shaft current Otherwise set i = i +1, continue to step 5); 5) Calculate the adjustable parameters according to formula (16) ; 6) Calculate the Jacobian matrix according to formula (17). ; 7) Update the adaptive coefficients according to formula (18) ; 8) Update the adaptive step size adjustment factor according to formula (19) (Return to step 2).

Images