A parameter identification method and system of a permanent magnet synchronous motor servo system

Abstract

The application discloses a parameter identification method and system for a permanent magnet synchronous motor servo system, and belongs to the field of permanent magnet synchronous motor control. The parameter identification method for the permanent magnet synchronous motor servo system comprises the following steps: firstly, in an information acquisition system, input cross-axis current, direct-axis current, cross-axis voltage instruction, direct-axis voltage instruction and rotating speed, and average values of nine intermediate variables are obtained through an average window sampling method. Secondly, in a full-rank identification equation set construction system, a full-rank identification equation set is constructed according to the average values of the nine intermediate variables. Finally, in a full-rank identification equation set solution system, a recursive least square method is used to iteratively solve the full-rank identification equation set, so that online identification of equivalent voltage constants of inverter nonlinearity, stator resistance, cross-axis inductance, direct-axis inductance and rotor flux linkage is realized.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of permanent magnet synchronous motor control, and more particularly, to a parameter identification method and system for a permanent magnet synchronous motor servo system. BACKGROUND

[0002] Permanent magnet synchronous motors have the advantages of large torque inertia ratio, high power density, and good control performance, and are widely used in servo systems. In order to improve the control performance of the permanent magnet synchronous motor servo system, it is necessary to obtain accurate motor parameters. Online parameter identification can track the changes of motor parameters in real time, and is of great significance for motor state detection and fault diagnosis. In addition, the nonlinearity of the inverter not only affects the control performance of the motor, but also introduces errors for the parameter identification of the motor. Considering the inverter nonlinearity as a parameter to be identified and performing real-time identification and compensation is a good solution. Therefore, an online parameter identification method that can identify the inverter nonlinearity and the motor parameters simultaneously is of great significance.

[0003] In the existing related technology, it is often assumed that the motor is running in a steady state, i.e., it is considered that the motor is running at a constant speed and constant torque most of the time. Therefore, the related identification algorithm is developed for steady state conditions, and the performance is poor in transient state conditions with variable torque and variable speed. However, in servo applications, the motor often works in a variable load or variable torque scenario, or even the torque or speed of the motor is always changing. In this scenario, the traditional identification algorithm is not applicable.

[0004] Therefore, it is necessary to design a parameter identification method for a permanent magnet synchronous motor servo system to overcome the above problems. SUMMARY

[0005] In view of the defects of the prior art, the purpose of the present application is to provide a parameter identification method and system for a permanent magnet synchronous motor servo system to meet the demand for online parameter identification of a permanent magnet synchronous motor servo system in a variable torque or variable speed application scenario.

[0006] To achieve the above object, the application provides a parameter identification method of a permanent magnet synchronous motor servo system, comprising the following steps: obtaining average values of nine intermediate variables through an average window sampling method based on real-time cross-axis current, direct-axis current, cross-axis voltage instruction, direct-axis voltage instruction and rotating speed; the nine intermediate variables include cross-axis current, direct-axis current, cross-axis voltage instruction, direct-axis voltage instruction, rotating speed, product of cross-axis current and rotating speed, product of direct-axis current and rotating speed, cross-axis inverter nonlinearity coefficient and direct-axis inverter nonlinearity coefficient; the cross-axis inverter nonlinearity coefficient is the quotient of the cross-axis component of the inverter nonlinearity voltage drop and the equivalent voltage constant; the direct-axis inverter nonlinearity coefficient is the quotient of the direct-axis component of the inverter nonlinearity voltage drop and the equivalent voltage constant; each determined average window can obtain a group of average value data, i.e. average values of the nine intermediate variables in the average window; the average window is a time interval containing several sampling periods; three or more average windows are selected to obtain three or more groups of average values of the intermediate variables to construct a full-rank identification equation group with a rank of 5.

[0007] The full-rank identification equation group is solved based on a recursive least square method, and the solution is the identification result of the to-be-identified parameters.

[0008] The average window sampling method is divided into four independent types according to the average window demarcation rules, which are demarcated based on information for tracking cross-axis current change, information for tracking rotating speed change, information based on direct-axis current disturbance injection and information based on time. The four average window demarcation rules are different, and they are in parallel relationship, can be independently operated, do not interfere with each other and share information collected from the control system, so that the information of each sampling period can be collected by the four sampling methods at the same time.

[0009] The first average window sampling method is based on tracking the change of cross-axis current, and an average window is demarcated every time the cross-axis current changes by a preset increment value; the average windows are connected end to end, i.e. the end point of the current average window is the starting point of the next average window; the sampling value of the cross-axis current in the first sampling period of the first average window is zero. The purpose of this average window sampling method is to make the average values of the cross-axis current in adjacent average windows different, and the average value of the cross-axis current in any average window non-zero.

[0010] The second average window sampling method is based on tracking the change of rotating speed, and an average window is demarcated every time the rotating speed changes by a preset increment value; the average windows are connected end to end, i.e. the end point of the current average window is the starting point of the next average window; the sampling value of the rotating speed in the first sampling period of the first average window is zero. The purpose of this average window sampling method is to make the average values of the rotating speed in adjacent average windows different, and the average value of the rotating speed in any average window non-zero.

[0011] The third average window sampling method is to inject a non-zero current disturbance pulse into the direct axis, and to define an average window in the pulse. The purpose of the average window sampling method is to make the average value of the direct axis current in the defined average window non-zero.

[0012] The fourth average window sampling method is to make the length of the average window the same; the average window is connected head to tail, that is, the end point of the current average window is the start point of the next average window. The purpose of the average window sampling method is to periodically obtain the average value of the nine intermediate variables.

[0013] The full-rank identification equation set is constructed based on the average value of the intermediate variable. The identification equation set constructed based on the average value of the intermediate variable of an average window is called a sub-identification equation set, and the number of equations in each sub-identification equation set is two, corresponding to the quadrature axis and the direct axis respectively. The full-rank identification equation set contains at least three sub-identification equation sets, so that the number of equations in the identification equation set is greater than or equal to the number of parameters to be identified.

[0014] The constituent elements of the sub-identification equation set include: the average value of the quadrature axis current value in an average window , the average value of the direct axis current value in the same average window , the average value of the quadrature axis voltage command value in the same average window , the average value of the direct axis voltage command value in the same average window , the average value of the speed value in the same average window , the average value of the product of the quadrature axis current value and the speed value in the same average window , the average value of the product of the direct axis current value and the speed value in the same average window , the average value of the quadrature axis inverter nonlinearity coefficient in the same average window , the average value of the direct axis inverter nonlinearity coefficient in the same average window , the change amount Δi of the quadrature axis current in the same average window q , the change amount Δi of the direct axis current in the same average window d , and the parameters to be identified.

[0015] The sub-identification equation set is:

[0016]

[0017] In the formula: V c o m is the equivalent voltage constant of inverter nonlinearity, R s is the stator resistance, L q is the quadrature axis inductance, L d is the direct axis inductance, ψ fis the rotor flux linkage, N is the number of sampling periods contained in the average window, T s is the sampling period.

[0018] The full rank identification equation set should contain 5 linearly independent equations, so that the identification equation set has a unique solution. The present application proposes two schemes to construct a full rank identification equation set containing 5 linearly independent equations:

[0019] Scheme one: the constructed identification equation set includes more than three sub-identification equation sets, one of which has a non-zero direct-axis current average value, and two of which have different cross-axis current average values.

[0020] Scheme two: the constructed identification equation set includes more than three sub-identification equation sets, one of which has a non-zero direct-axis current average value, and two of which have different average speed values.

[0021] The identification equation set solving system uses recursive least squares method to solve the full rank identification equation set. The basic model equation of the recursive least squares method is:

[0022]

[0023]

[0024]

[0025] θ=[A s L d L q ψ f V dom ] T

[0026] In the formula, subscript m represents the index of the identification equation set; subscript a represents the index of the sub-identification equation set, that is, a identification equation set contains a sub-identification equation set. Lambda is the forgetting factor, and its value is usually selected in the range of 0.9-1.

[0027] The present application also provides a parameter identification system of a permanent magnet synchronous motor servo system, comprising: a computer readable storage medium and a processor;

[0028] The computer readable storage medium is used to store executable instructions;

[0029] The processor is used to read the executable instructions stored in the computer readable storage medium, and execute the parameter identification method of the permanent magnet synchronous motor servo system described above.

[0030] Compared with the prior art, the above technical scheme conceived by the present application can achieve the following

[0031] Advantages:

[0032] 1. The present application is suitable for parameter identification of permanent magnet synchronous motor servo system in variable load or variable torque application scenarios. The existing permanent magnet synchronous motor parameter identification method is proposed under the condition that the motor works in steady state for a long time. Therefore, the existing identification algorithm can only accurately identify the parameters under steady state. For servo motor systems working in variable torque or variable speed conditions for a long time, the existing technology cannot achieve high-performance parameter identification. The present application can fill this technical gap and is particularly suitable for parameter identification of permanent magnet synchronous motor servo system in variable load or variable torque application scenarios.

[0033] 2. The present application can identify a large number of parameters simultaneously. Not only can the motor parameters be identified, but the equivalent voltage constant of the inverter nonlinearity can also be identified simultaneously. The present application can simultaneously identify the equivalent voltage constant of the inverter nonlinearity and all parameters of the permanent magnet synchronous motor, including the stator resistance, the quadrature axis inductance, the direct axis inductance, and the rotor flux linkage. Compared with the existing parameter identification technology that cannot simultaneously identify the above five parameters, the present application does not have the problem of affecting the identification accuracy of the identified parameters due to un-identified parameters.

[0034] 3. The present application can identify in real time. Since the information acquisition system can acquire and calculate the latest average values of each variable from the control system in real time, the identification equation set construction system can construct the latest identification equation set according to the latest average values of each variable, and the identification equation set solving system can solve the latest identification equation set to obtain the latest to-be-identified parameters. The three systems can be updated in real time, ensuring the real-time performance of parameter identification. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 is a parameter identification system diagram of the present application;

[0036] Figure 2 is a flowchart of the information acquisition system in one embodiment;

[0037] Figure 3 is an experimental effect diagram of parameter identification in one embodiment. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical scheme and advantages of the present application clearer and more understandable, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application. In addition, the technical features involved in each embodiment of the present application described below can be combined with each other as long as they do not conflict with each other.

[0039] The parameter identification system described in this invention is as follows: Figure 1 As shown, the input to this system is the quadrature-axis current value i obtained in real time from the control system. q Direct-axis current value i d Cross-axis voltage command value u q Direct-axis voltage command value u d And the rotational speed ω, the output is the equivalent voltage constant V of the inverter nonlinearity of the permanent magnet synchronous motor servo system. com Stator resistance R s quadrature axis inductance L q Direct-axis inductor L d and rotor flux ψ f .

[0040] The present invention solves its technical problem by adopting the following technical solution:

[0041] like Figure 1 As shown, the parameter identification system provided by this invention includes an information acquisition system, a full-rank identification equation system construction system, and a full-rank identification equation system solving system. The information acquisition system collects information from the control system; the collected information is used by the full-rank identification equation system construction system to construct a full-rank identification equation system; the identification equation system solving system is used to solve for the unique solution of the identification equation system, which is the real-time parameter identification value, including the equivalent voltage constant V of the inverter nonlinearity. com Stator resistance R s quadrature axis inductance L q Direct-axis inductor L d and rotor flux ψ f The information acquisition system takes into account the quadrature-axis current, direct-axis current, quadrature-axis voltage command, direct-axis voltage command, and rotational speed as inputs, and obtains the average value of nine intermediate variables through an average window sampling method. The full-rank identification equation system constructs a full-rank identification equation system based on the average value of the nine intermediate variables. The full-rank identification equation system solving system is used to find the unique solution to the full-rank identification equation system, i.e., the real-time parameter identification values, including the equivalent voltage constant V of the inverter nonlinearity. com Stator resistance R s quadrature axis inductance L q Direct-axis inductor L d and rotor flux ψ f .

[0042] The information acquisition system is connected to the control system via a data transmission line, and obtains the quadrature-axis current value i for each sampling period from the control system in real time. q Direct-axis current value i d Cross-axis voltage command value u q Direct-axis voltage command value ud and the rotational speed value ω.

[0043] The information acquisition system adopts the method of average window sampling. The average window is a time interval containing several sampling periods. In an average window, the 9 intermediate variables that need to be recorded are recorded in the form of accumulation, and at the end of the average window, the average value of each intermediate variable in an average window is calculated. The 9 intermediate variables that need to be recorded are respectively the quadrature axis current value, the direct axis current value, the quadrature axis voltage command value, the direct axis voltage command value, the rotational speed value, the product of the quadrature axis current value and the rotational speed value, the product of the direct axis current value and the rotational speed value, the quadrature axis inverter nonlinearity coefficient, and the direct axis inverter nonlinearity coefficient. The quadrature axis inverter nonlinearity coefficient is the quotient of the component of the inverter nonlinearity voltage drop on the quadrature axis and the equivalent voltage constant. The direct axis inverter nonlinearity coefficient is the quotient of the component of the inverter nonlinearity voltage drop on the direct axis and the equivalent voltage constant.

[0044] The information acquisition system contains four subsystems, which are the information acquisition subsystem for tracking the change of the quadrature axis current, the information acquisition subsystem for tracking the change of the rotational speed, the information acquisition subsystem based on the direct axis current disturbance injection, and the information acquisition subsystem based on time. The four information acquisition subsystems only have different average window demarcation rules, and they are in parallel relationship, can be independently operated, do not interfere with each other, and share the information collected from the control system, so that the information of each sampling period can be collected by the four information acquisition subsystems at the same time. The flowchart of the information acquisition system in an embodiment is shown in Figure 2 , wherein the average window corresponding to the information acquisition subsystem for tracking the change of the quadrature axis current is referred to as average window a, the average window corresponding to the information acquisition subsystem for tracking the change of the rotational speed is referred to as average window b, the average window corresponding to the information acquisition subsystem based on the direct axis current disturbance injection is referred to as average window c, and the average window corresponding to the information acquisition subsystem based on time is referred to as average window d.

[0045] The average window setting rule of the information acquisition subsystem for tracking the change of the quadrature axis current is: an average window is set based on the change of the quadrature axis current, and the quadrature axis current changes by a preset increment value to set an average window; the average windows are connected in a head-to-tail manner, i.e. the end point of the current average window is the start point of the next average window; the sampling value of the quadrature axis current in the first sampling period of the first average window is zero. The purpose of this average window sampling method is to make the average value of the quadrature axis current in adjacent average windows different, and the average value of the quadrature axis current in any average window non-zero. In an embodiment, the initial state of the motor is zero current state, and then the first average window starts from the first sampling period. The increment is set as one fifth of the rated current value, and the increment remains unchanged in the information acquisition subsystem. Then, when the quadrature axis current first changes by one fifth of the rated current value relative to the quadrature axis current value corresponding to the start point of the current average window, the end point of the current average window (also the start point of the next average window) is obtained. Therefore, the quadrature axis current in the sampling period corresponding to the end point of the first average window (also the start point of the second average window) is one fifth of the rated current value or negative one fifth of the rated current value. Similarly, if the quadrature axis current in the sampling period corresponding to the end point of the first average window (also the start point of the second average window) is one fifth of the rated current value, the quadrature axis current in the sampling period corresponding to the end point of the second average window (also the start point of the third average window) is two fifths of the rated current value or zero, and the determination method of the average window end point (also the start point of the next average window) is the same as above, and the same applies to the subsequent average windows.

[0046] The average window of the information acquisition subsystem based on the change of the tracking speed is determined based on the change of the tracking speed, and an average window is determined when the speed changes by a preset increment value. The average window is connected at both ends, i.e. the end of the current average window is the start of the next average window. The sampling value of the speed in the first sampling period of the first average window is zero. The purpose of this average window sampling method is to make the average value of the speed in adjacent average windows different, and the average value of the speed in any average window is non-zero. In an embodiment, the initial state of the motor is a static state, and the first average window starts from the first sampling period. The increment is set to one fifth of the rated speed value, and the increment remains unchanged in the information acquisition subsystem. When the speed changes by one fifth of the rated speed value relative to the speed value corresponding to the start of the current average window for the first time, the end of the current average window (which is also the start of the next average window) is obtained. Therefore, the speed of the sampling period corresponding to the end of the first average window (which is also the start of the second average window) is one fifth of the rated speed value or negative one fifth of the rated speed value. Similarly, if the speed of the sampling period corresponding to the end of the first average window (which is also the start of the second average window) is one fifth of the rated speed value, the speed of the sampling period corresponding to the end of the second average window (which is also the start of the third average window) is two fifths of the rated speed value or zero. The determination method of the end of the subsequent average window (which is also the start of the next average window) is the same as above, and the same applies to the subsequent average windows.

[0047] The average window of the information acquisition subsystem based on the change of the tracking speed is determined based on the change of the tracking speed, and an average window is determined when the speed changes by a preset increment value. The average window is connected at both ends, i.e. the end of the current average window is the start of the next average window. The sampling value of the speed in the first sampling period of the first average window is zero. The purpose of this average window sampling method is to make the average value of the speed in adjacent average windows different, and the average value of the speed in any average window is non-zero. In an embodiment, the initial state of the motor is a static state, and the first average window starts from the first sampling period. The increment is set to one fifth of the rated speed value, and the increment remains unchanged in the information acquisition subsystem. When the speed changes by one fifth of the rated speed value relative to the speed value corresponding to the start of the current average window for the first time, the end of the current average window (which is also the start of the next average window) is obtained. Therefore, the speed of the sampling period corresponding to the end of the first average window (which is also the start of the second average window) is one fifth of the rated speed value or negative one fifth of the rated speed value. Similarly, if the speed of the sampling period corresponding to the end of the first average window (which is also the start of the second average window) is one fifth of the rated speed value, the speed of the sampling period corresponding to the end of the second average window (which is also the start of the third average window) is two fifths of the rated speed value or zero. The determination method of the end of the subsequent average window (which is also the start of the next average window) is the same as above, and the same applies to the subsequent average windows. d

[0048] ​The average window demarcation rule of the time-based information acquisition subsystem is that the length of the average window is the same, and the average window is connected at the head and tail, that is, the end point of the current average window is the start point of the next average window. The purpose of this average window sampling method is to periodically obtain the average values of the nine intermediate variables. In an embodiment, one average window is demarcated every 1000 sampling periods, and then the first to 1000th sampling periods belong to the first average window, the 1001st to 2000th belong to the second average window, the 2001st to 3000th belong to the third average window, and so on.

[0049] The full-rank identification equation set construction system constructs the full-rank identification equation set based on the average values output by the information acquisition system. The identification equation set constructed based on the average value of the intermediate variables of one average window is referred to as a sub-identification equation set, and the number of equations in each sub-identification equation set is two, corresponding to the cross axis and the direct axis respectively. The identification equation set contains at least three sub-identification equation sets, so that the number of equations in the identification equation set is greater than or equal to the number of parameters to be identified

[0050] The constituent elements of the sub-identification equation set include: the average value of the cross-axis current value in one average window , the average value of the direct-axis current value in the same average window , the average value of the cross-axis voltage command value in the same average window , the average value of the direct-axis voltage command value in the same average window , the average value of the speed value in the same average window , the average value of the product of the cross-axis current value and the speed value in the same average window , the average value of the product of the direct-axis current value and the speed value in the same average window , the average value of the cross-axis inverter nonlinearity coefficient in the same average window , the average value of the direct-axis inverter nonlinearity coefficient in the same average window , the change amount Δi of the cross-axis current in the same average window q , the change amount Δi of the direct-axis current in the same average window d , and the parameters to be identified.

[0051] The sub-identification equation set is:

[0052]

[0053] In the formula, V com is the equivalent voltage constant of inverter nonlinearity, R s is the stator resistance, L q is the cross-axis inductance, L d is the direct-axis inductance, ψ fis the rotor flux linkage, N is the number of sampling periods contained in the average window, T s is the sampling period.

[0054] The full rank identification equation set should contain 5 linearly independent equations, so that the identification equation set has a unique solution. The present application proposes two schemes to construct a full rank identification equation set containing 5 linearly independent equations:

[0055] Scheme one: the constructed full rank identification equation set includes more than three sub-identification equation sets, one of which has a non-zero direct-axis current average value, and the other two have different cross-axis current average values. In one embodiment, the constructed full rank identification equation set includes three sub-identification equation sets. Among them, the first sub-identification equation set corresponds to the average window derived from the information acquisition subsystem based on direct-axis current disturbance injection, and is the latest average window of the information acquisition subsystem based on direct-axis current disturbance injection at the current time; the second and third sub-identification equation sets correspond to the average windows derived from the information acquisition subsystem tracking cross-axis current changes, and these two average windows are the latest two adjacent average windows of the information acquisition subsystem tracking cross-axis current changes at the current time.

[0056] Scheme two: the constructed full rank identification equation set includes more than three sub-identification equation sets, one of which has a non-zero direct-axis current average value, and the other two have different average values of rotational speed. In one embodiment, the constructed full rank identification equation set includes three sub-identification equation sets. Among them, the first sub-identification equation set corresponds to the average window derived from the information acquisition subsystem based on direct-axis current disturbance injection, and is the latest average window of the information acquisition subsystem based on direct-axis current disturbance injection at the current time; the second and third sub-identification equation sets correspond to the average windows derived from the information acquisition subsystem tracking rotational speed changes, and these two average windows are the latest two adjacent average windows of the information acquisition subsystem tracking rotational speed changes at the current time.

[0057] In one embodiment, the constructed full-rank identification equation set belongs to both scheme one and scheme two. Specifically, the full-rank identification equation set includes six sub-identification equation sets. The first sub-identification equation set corresponds to an average window derived from the information acquisition subsystem based on direct-axis current disturbance injection, and is the latest average window of the information acquisition subsystem based on direct-axis current disturbance injection at the current time; the second and third sub-identification equation sets correspond to average windows derived from the information acquisition subsystem for tracking change in quadrature-axis current, and the two average windows are the latest two adjacent average windows of the information acquisition subsystem for tracking change in quadrature-axis current at the current time; the fourth and fifth sub-identification equation sets correspond to average windows derived from the information acquisition subsystem for tracking change in speed, and the two average windows are the latest two adjacent average windows of the information acquisition subsystem for tracking change in speed at the current time; and the sixth sub-identification equation set corresponds to an average window derived from the information acquisition subsystem based on time, and is the latest average window of the information acquisition subsystem based on time at the current time.

[0058] The full-rank identification equation set solving system adopts a recursive least square method to solve the full-rank identification equation set. In one embodiment, the full-rank identification equation set includes six sub-identification equation sets. The first sub-identification equation set corresponds to an average window derived from the information acquisition subsystem based on direct-axis current disturbance injection, and is the latest average window of the information acquisition subsystem based on direct-axis current disturbance injection at the current time; the second and third sub-identification equation sets correspond to average windows derived from the information acquisition subsystem for tracking change in quadrature-axis current, and the two average windows are the latest two adjacent average windows of the information acquisition subsystem for tracking change in quadrature-axis current at the current time; the fourth and fifth sub-identification equation sets correspond to average windows derived from the information acquisition subsystem for tracking change in speed, and the two average windows are the latest two adjacent average windows of the information acquisition subsystem for tracking change in speed at the current time; and the sixth sub-identification equation set corresponds to an average window derived from the information acquisition subsystem based on time, and is the latest average window of the information acquisition subsystem based on time at the current time. The recursive least square method can be expressed as:

[0059]

[0060]

[0061]

[0062] θ=[R s L d L q ψ f V com ] T

[0063] In the formula, subscript n represents the average value corresponding to the current average window output by the information acquisition subsystem based on the direct-axis current disturbance injection; subscript i represents the average value corresponding to the current average window output by the information acquisition subsystem for tracking the change of the quadrature-axis current, i-1 represents the average value corresponding to the last average window; subscript j represents the average value corresponding to the current average window output by the information acquisition subsystem for tracking the change of the speed, j-1 represents the average value corresponding to the last average window; subscript t represents the average value corresponding to the current average window output by the information acquisition subsystem based on time; λ is a forgetting factor, and the value thereof is usually selected in the range of 0.9-1.

[0064] In one embodiment, the permanent magnet synchronous motor servo system parameter identification system provided by the application is used to identify parameters of an experimental permanent magnet synchronous motor servo system, and the experimental results are as shown in Figure 3 The experimental results show that the proposed method has good identification effect in both variable load application scenarios and variable speed application scenarios.

[0065] Those skilled in the art can easily understand that the above description is only a preferred embodiment of the application, and is not intended to limit the application, and any modification, equivalent replacement and improvement made within the spirit and principle of the application shall be included in the protection scope of the application.

Claims

1. A parameter identification method of a permanent magnet synchronous motor servo system, characterized by, It comprises the following steps: Based on real-time cross-axis current, direct-axis current, cross-axis voltage instruction, direct-axis voltage instruction and rotating speed, nine intermediate variables including cross-axis current, direct-axis current, cross-axis voltage instruction, direct-axis voltage instruction, rotating speed, product of cross-axis current and rotating speed, product of direct-axis current and rotating speed, cross-axis inverter nonlinear coefficient and direct-axis inverter nonlinear coefficient are obtained by average window sampling method, and the average values of the nine intermediate variables in each average window are obtained. Any three or more average windows are selected to obtain the average values of three or more groups of intermediate variables, and a full-rank identification equation group with rank 5 is constructed; the average values of each group of intermediate variables construct a sub-identification equation group; the equation number of each sub-identification equation group is two, corresponding to the cross-axis and the direct-axis respectively; The sub-identification equation group is: wherein is an average of the direct axis voltage command over an average window, is an average of the quadrature axis voltage command over the same average window, is an average of the direct axis current over the same average window, is an average of the quadrature axis current over the same average window, is an average of the speed over the same average window, is an average of the product of the direct axis current and the speed over the same average window, is an average of the product of the quadrature axis current and the speed over the same average window, is an average of the direct axis inverter nonlinearity coefficient over the same average window, is an average of the quadrature axis inverter nonlinearity coefficient over the same average window, Δi d is a variation of the direct axis current over the same average window, Δi q is a variation of the quadrature axis current over the same average window, R s is a stator resistance of the permanent magnet synchronous machine, L d is a direct axis inductance of the permanent magnet synchronous machine, L q is a quadrature axis inductance of the permanent magnet synchronous machine, ψ f is a rotor flux of the permanent magnet synchronous machine, V com is an equivalent voltage constant of the inverter nonlinearity; The method for constructing the full-rank identification equation group comprises two kinds: The first method for constructing the full-rank identification equation group comprises three or more sub-identification equation groups, one of which has a non-zero direct-axis current average value, and two of which have different cross-axis current average values; The second method for constructing the full-rank identification equation group comprises three or more sub-identification equation groups, one of which has a non-zero direct-axis current average value, and two of which have different rotating speed average values; The recursive least square method is used to solve the full-rank identification equation group, and the solution is the identification result of the to-be-identified parameters.

2. The method of claim 1, wherein, The average window sampling method is divided into four independent kinds according to the average window demarcation rules: The first average window sampling method demarcates the average window based on tracking the change of the cross-axis current, and an average window is demarcated every time the cross-axis current changes by a preset increment value; The average windows are connected head to tail, i.e. the end point of the current average window is the start point of the next average window; The sampling value of the cross-axis current in the first sampling period of the first average window is zero; The second average window sampling method demarcates the average window based on tracking the change of the rotating speed, and an average window is demarcated every time the rotating speed changes by a preset increment value; The average windows are connected head to tail, i.e. the end point of the current average window is the start point of the next average window; The sampling value of the rotating speed in the first sampling period of the first average window is zero; The third average window sampling method demarcates an average window in a non-zero current disturbance pulse injected in the direct axis; The fourth average window sampling method has the same length of average window; The average windows are connected head to tail, i.e. the end point of the current average window is the start point of the next average window.

3. The method of claim 1, wherein, The basic model equation of the recursive least square method is: wherein the subscript m denotes the index of the full-rank identification equation set; the subscript a denotes the index of the sub-identification equation set, i.e. one full-rank identification equation set contains a sub-identification equation sets; λ is a forgetting factor.

4. A parameter identification system of a permanent magnet synchronous motor servo system, characterized by, It comprises: A computer readable storage medium and a processor; The computer readable storage medium is used to store executable instructions; The processor is used to read the executable instructions stored in the computer readable storage medium, and execute the parameter identification method of the permanent magnet synchronous motor servo system according to any one of claims 1 to 3.

Images