Surface-mounted permanent magnet synchronous motor parameter online identification system and method

Abstract

The present application relates to a kind of surface-mounted permanent magnet synchronous motor parameter online identification system and method, belong to synchronous motor technical field.The present application is under the rated speed of motor and load, to the d-axis of motor injection low-frequency pulse small current, make d-axis voltage equation present two kinds of working conditions: the first state under d-axis injection current is zero, inductance parameter is identified based on d-axis voltage equation, resistance parameter keeps the identification result of last period;Second state under d-axis injection current is not zero, inductance is identified on the basis of having been identified, and resistance parameter is identified using d-axis voltage equation, inductance parameter keeps the result of first state.Then, inductance, resistance parameter of real-time identification is substituted into q-axis voltage equation, and rotor flux linkage parameter is calculated online.On this basis, inductance, resistance and flux linkage are completed online identification by combining adaptive forgetting factor recursive least square method.The present application can realize the complete decoupling identification of resistance, inductance and flux linkage while motor is normally operated.

Description

Technical Field

[0001] This invention relates to the field of synchronous motor technology, and in particular to an online parameter identification system and method for surface-mounted permanent magnet synchronous motors. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) are widely used in electric vehicles, industrial servos, aerospace, and other fields due to their high power density, excellent efficiency, and good control performance. The performance of a motor control system largely depends on the matching accuracy between controller parameters and motor parameters. However, during motor operation, stator resistance changes with temperature rise, inductance may drift due to magnetic saturation, and flux linkage is also affected by temperature. The time-varying nature of these parameters can degrade the performance of controllers designed with fixed parameters, and even affect system stability. Therefore, achieving online real-time identification of key motor parameters is crucial for realizing high-performance adaptive control, condition monitoring, and fault diagnosis.

[0003] Existing online parameter identification technologies can be broadly categorized into two types. The first type directly constructs the identification model from the motor's steady-state or dynamic voltage equations. Its advantage lies in its independence from specific injection signals or operating points, and its model covers a wide operating range. However, this type of method typically faces the problem of underrank in the identification model, meaning the number of equations is insufficient to independently decouple all unknown parameters, leading to severe parameter coupling between resistance, inductance, and flux linkage. When multiple parameters change simultaneously, it is difficult to accurately track their individual changes, and the model is often complex and computationally intensive. The second type of method artificially constructs specific steady-state or dynamic operating points by injecting high-frequency or low-frequency voltage / current signals into the motor, attempting to decouple and identify different parameters. While this type of method can simplify the model at specific moments, its identification process often depends on the specific state of the motor, and the injected signals may cause additional torque ripples, vibrations, and noise. More importantly, many injection methods still fail to completely resolve the coupling problem between motor parameters, or can only identify some parameters, failing to achieve online tracking of all key parameters. Furthermore, most schemes using the recursive least squares method employ a fixed forgetting factor, which cannot achieve a good balance between tracking speed and steady-state accuracy during parameter identification when the motor operating conditions change drastically or parameters drift rapidly.

[0004] Therefore, a new online identification strategy is urgently needed to address the challenge of time-varying and mutually coupled parameters in surface-mounted permanent magnet synchronous motors during operation. This strategy must be able to achieve complete decoupled identification of resistance, inductance, and flux linkage while the motor is operating normally. Summary of the Invention

[0005] In view of this, it is necessary to provide a surface-mount permanent magnet synchronous motor parameter online identification system and method to achieve the purpose of complete decoupling identification of resistance, inductance and magnetic flux while the motor is working normally.

[0006] To achieve the above objectives, in a first aspect, the present invention provides an online parameter identification system for a surface-mount permanent magnet synchronous motor, used in a surface-mount permanent magnet synchronous motor, wherein the surface-mount permanent magnet synchronous motor is equipped with a controller, the controller being equipped with a current control loop, including: a pulse injection unit and an identification unit; wherein; The pulse injection unit is connected at the d-axis current setting point of the current control loop and is used to generate and inject low-frequency pulsed small current to the d-axis. The identification unit collects d-axis voltage, d-axis current, q-axis voltage, q-axis current and motor speed values ​​from the current control loop and speed loop, and performs online parameter identification based on the adaptive forgetting factor recursive least squares method through the corresponding identification model. The identification model includes inductance identification model, stator resistance identification model and rotor flux identification model. In the first state where the d-axis pulse current generated by the pulse injection unit is zero, the inductor identification model is based on the d-axis voltage equation and uses the adaptive forgetting factor recursive least squares method to output the inductor identification value based on the d-axis voltage value, q-axis current value, and motor speed value. In this first state, the resistance identification value is the resistance value obtained at the end of the previous second state. In the second state where the d-axis pulse current is not zero, the stator resistance identification model outputs a resistance identification value based on the d-axis voltage value, d-axis current value, q-axis current value, motor speed value, and the inductance value identified at the end of the previous first state. In this second state, the inductance identification value is the inductance value identified at the end of the previous first state. The flux linkage identification model inputs the inductance identification value and the resistance identification value into the preset q-axis voltage equation, and combines the q-axis voltage value, d-axis current value, q-axis current value and motor speed value to output the flux linkage identification value.

[0007] In one possible implementation, when the first state is in a steady state, the expression for the d-axis voltage equation is: When the first state is in a steady state , The inductor identification model, after discretization, is constructed using the following formula: Where k represents the sampling point, This represents the d-axis voltage value at sampling point k. This represents the d-axis current value at sampling point k. This represents the q-axis current value at sampling point k. This represents the motor speed value at sampling point k. , Represents the extreme logarithm. This represents the inductance to be identified at time k. This represents the output of the inductance identification model at time k. This represents the regression vector at time k of the inductance identification model. This represents the inductance parameter to be identified at time k. This indicates the noise in the inductance identification measurement.

[0008] In one possible implementation, when the second state is in a steady state, the expression for the d-axis voltage equation is: When the second state is in a steady state , The resistance identification model, after discretization, is constructed using the following formula: in, This represents the resistance value to be identified at time k. This represents the output of the resistance identification model at time k. Let represent the regression vector at time k of the resistance identification model. This represents the resistance parameter to be identified at time k. This indicates resistance identification measurement noise. This represents the inductance value that was identified and maintained at the end of the previous first state.

[0009] In one possible implementation, under steady-state conditions of the motor, the q-axis voltage equation... because The discretized magnetic flux linkage identification model is constructed using the following formula: in, This is the q-axis voltage value. Let K be the permanent magnet rotor flux linkage to be identified at time k. and For real-time updated resistance and inductance identification values, This is the output of the magnetic linkage identification model at time k. Let be the regression vector of the magnetic linkage identification model at time k. Let K be the flux linkage parameter to be identified at time k. To identify noise in magnetic flux linkage measurements.

[0010] Secondly, the present invention also provides an online parameter identification method for a surface-mounted permanent magnet synchronous motor, applied to the identification system described in any of the above claims, characterized by comprising the following steps: Step 1: Inject a low-frequency pulsed small current into the surface-mounted permanent magnet synchronous motor under rated speed and rated load conditions through the pulse injection unit, so that the system alternately goes through a first state where the injected current into the d-axis is zero and a second state where it is not zero. Step 2: In the first state, based on the d-axis voltage equation, the inductance value is identified online using the adaptive forgetting factor recursive least squares method, while keeping the resistance value identified at the end of the previous second state unchanged. Step 3: In the second state, based on the d-axis voltage equation and using the inductance value identified at the end of the first state as a known quantity, the resistance value is identified online using the adaptive forgetting factor recursive least squares method, while keeping the inductance identification value unchanged. Step 4: Substitute the inductance and resistance values ​​identified and updated in real time in Step 2 and Step 3 into the q-axis voltage equation, and use the adaptive forgetting factor recursive least squares method to identify the flux linkage value online. Step 5: At the instant the first state and the second state switch, the d-axis current is in an unstable state. For a short period of time after the state switch, the resistance and inductance identification maintain the output value before the switching point.

[0011] In one possible implementation, the d-axis injection current is determined based on a d-axis injection current function.

[0012] In one possible implementation, the expression for the d-axis injected current function is: in, Indicates the d-axis injected current. This indicates the magnitude of the injected current along the d-axis. Indicates the injection current period. Represents the frequency of the injected current along the d-axis, where , Indicates the period number ( ).

[0013] In one possible implementation, the adaptive forgetting factor least recursive squares method is set by the following formula: Where n is the error amplification factor. To Low-pass digital filter, For the filtered Amplify by power. and The preset lower and upper limits for the forgetting factor. For adjustment coefficients, For the gain vector, Let be the covariance matrix.

[0014] The beneficial effects of this invention are as follows: The online parameter identification system for surface-mounted permanent magnet synchronous motors provided by this invention injects low-frequency pulse current into the d-axis and designs a state-alternating decoupling strategy, which provides a foundation for the completely independent identification of stator resistance, inductance, and rotor flux linkage. Furthermore, based on the state-alternating decoupling strategy, the identification model can synchronously and in real-time update the identification results of resistance, inductance, and flux linkage during continuous motor operation, providing timely and accurate parameter support for motor adaptive control, fault diagnosis, and condition monitoring. Attached Figure Description

[0015] Figure 1 A schematic diagram of the control and identification system framework provided by the present invention; Figure 2 A schematic diagram of the identification system provided by this invention; Figure 3 The overall flowchart of the online parameter identification method for surface-mounted permanent magnet synchronous motors provided by the present invention is shown below. Figure 4 The diagram shows the dynamic response of the motor speed under sudden parameter changes, as provided in a specific embodiment of the present invention. Figure 5 The diagram shows the dynamic response of the electromagnetic torque of the motor under sudden parameter changes, as provided in a specific embodiment of the present invention. Figure 6 The diagram shows the dynamic response of the motor's d-axis current and q-axis current under sudden parameter changes, as provided in a specific embodiment of the present invention. Figure 7 This is a schematic diagram of the online identification results of motor resistance in a specific embodiment of the present invention; Figure 8 This is a schematic diagram of the forgetting factor during the online identification of motor resistance in a specific embodiment of the present invention; Figure 9 This is a schematic diagram of the online identification results of motor inductance in a specific embodiment of the present invention; Figure 10 This is a schematic diagram of the forgetting factor during the online identification of motor inductance in a specific embodiment of the present invention; Figure 11 This is a schematic diagram of the online identification results of the motor rotor flux linkage in a specific embodiment of the present invention; Figure 12 This is a schematic diagram of the forgetting factor during the online identification of motor rotor flux linkage in a specific embodiment of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0017] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0018] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0019] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0020] This invention provides an online parameter identification system and method for surface-mounted permanent magnet synchronous motors, which will be described below.

[0021] Please see Figure 1 , Figure 2 as well as Figure 3 An online parameter identification system for a surface-mount permanent magnet synchronous motor is disclosed. The surface-mount permanent magnet synchronous motor is equipped with a controller, which includes a current control loop. The system is characterized by comprising: a pulse injection unit and an identification unit; wherein... The pulse injection unit is connected at the d-axis current setting point of the current control loop and is used to generate and inject low-frequency pulsed small current to the d-axis. The identification unit collects d-axis voltage, d-axis current, q-axis voltage, q-axis current and motor speed values ​​from the current control loop and speed loop, and performs online parameter identification based on the adaptive forgetting factor recursive least squares method through the corresponding identification model. The identification model includes inductance identification model, stator resistance identification model and rotor flux identification model. In the first state where the d-axis pulse current generated by the pulse injection unit is zero, the inductor identification model is based on the d-axis voltage equation and uses the adaptive forgetting factor recursive least squares method to output the inductor identification value based on the d-axis voltage value, q-axis current value, and motor speed value. In this first state, the resistance identification value is the resistance value obtained at the end of the previous second state. In the second state where the d-axis pulse current is not zero, the stator resistance identification model outputs a resistance identification value based on the d-axis voltage value, d-axis current value, q-axis current value, motor speed value, and the inductance value identified at the end of the previous first state. In this second state, the inductance identification value is the inductance value identified at the end of the previous first state. The flux linkage identification model inputs the inductance identification value and the resistance identification value into the preset q-axis voltage equation, and combines the q-axis voltage value, d-axis current value, q-axis current value and motor speed value to output the flux linkage identification value.

[0022] Compared with existing technologies, the surface-mount permanent magnet synchronous motor parameter online identification system provided in this embodiment injects low-frequency pulse current into the d-axis and designs a state-alternating decoupling strategy, which provides a foundation for the completely independent identification of stator resistance, inductance and rotor flux linkage. Furthermore, based on the state-alternating decoupling strategy, the identification model can synchronously and in real time update the identification results of resistance, inductance and flux linkage during the continuous operation of the motor, providing timely and accurate parameter support for motor adaptive control, fault diagnosis and condition monitoring.

[0023] In a specific embodiment of the present invention, an online identification system is provided for integration into the controller of a surface-mount permanent magnet synchronous motor. The controller includes a conventional current control loop, coordinate transformation module, and PWM generator, etc. This identification system, as a newly added functional module, mainly includes a pulse injection unit and an identification unit.

[0024] The pulse injection unit is connected at the d-axis current setpoint of the current control loop. Its core function is to generate and inject a low-frequency, small-amplitude pulse current signal according to a preset timing sequence. To the d-axis current loop. The mathematical expression of this pulse signal, as described in the invention summary, is determined by the amplitude. ,cycle and frequency Definition. The purpose of injecting this signal is to artificially and periodically construct two distinguishable system operating states, providing conditions for parameter decoupling.

[0025] The identification unit is responsible for collecting motor operation data from the control system in real time, including d-axis voltage. d-axis current q-axis voltage q-axis current and electric angular velocity The identification unit comprises three core sub-modules: inductance identification module, resistance identification module, and flux linkage identification module. Each module invokes the corresponding identification model based on different system states and employs an adaptive forgetting factor recursive least squares method for parameter identification and updating.

[0026] The system identification process is briefly described as follows: State 1 (Inductor Identification State): The system enters this state when the injected d-axis pulse current is zero. At this time, during steady-state motor operation, neglecting the differential term, the d-axis voltage equation shows that: Resistance voltage drop term Constructing with inductors For a linear model of the parameters to be identified, the inductor identification values ​​are updated online using the Adaptive Forgetting Factor Recursive Least Squares (AF-RLS) algorithm. During this state, the resistance identification value Keep the output as the final identification result at the end of the previous "state two".

[0027] State Two (Resistor Identification State): The system enters this state when the injected d-axis pulse current is not zero. At this time, during steady-state motor operation, the differential term is ignored, the inductance value is considered known, and is maintained by the final identification result at the end of the previous "State One". The d-axis voltage equation can be reconstructed using the resistor... For the linear model of the parameters to be identified, the resistance identification values ​​are also updated online using the AF-RLS algorithm. During this state, the inductance identification value The output remains unchanged from the final identification result at the end of the previous "state one".

[0028] Magnetic flux linkage identification: Obtaining real-time updated resistance identification values Inductance identification value Then, substitute it into the q-axis voltage equation. During steady-state operation of the motor, neglecting the differential terms, the equation can be simplified to using the rotor flux linkage... The linear model of the parameters to be identified is used, and then the AF-RLS algorithm is used to realize the online identification of magnetic flux.

[0029] It is important to note that at the instant of the state transition, because the d-axis current is in a transient process, the differential terms in the voltage equation no longer satisfy the steady-state assumption. Therefore, for a brief period after the state transition, the identification system will pause parameter updates, and the output values ​​of the resistor and inductor will remain unchanged from the identification results before the transition point. Identification will resume once the current re-enters a steady state.

[0030] In some embodiments of the present invention, when the first state is in a steady state, the expression for the d-axis voltage equation is: When the first state is in a steady state , The inductor identification model, after discretization, is constructed using the following formula: Where k represents the sampling point, This represents the d-axis voltage value at sampling point k. This represents the d-axis current value at sampling point k. This represents the q-axis current value at sampling point k. This represents the motor speed value at sampling point k. , Represents the extreme logarithm. This represents the inductance to be identified at time k. This represents the output of the inductance identification model at time k. This represents the regression vector at time k of the inductance identification model. This represents the inductance parameter to be identified at time k. Indicates the noise in inductance identification measurement. This indicates that the d-axis current is approximately zero. This indicates that the product of the inductance and the differential of the d-axis current is approximately equal to 0.

[0031] In a specific embodiment of the present invention, the inductor identification model is such that, in state one, the d-axis injected current is zero, and the system is in a steady state. At this time, the d-axis voltage equation... Simplified to After discretization, it becomes: In this state, The online identification value remains the same as the previous state's two-terminal identification value. The equation remains unchanged. Rearranging this equation into standard least squares form: in, For output quantity, For regression parameters, Let k be the inductance parameters to be identified. Noise is measured for inductor identification.

[0032] In some embodiments of the present invention, when the second state is in a steady state, the expression for the d-axis voltage equation is: When the second state is in a steady state , The resistance identification model, after discretization, is constructed using the following formula: in, This represents the resistance value to be identified at time k. This represents the output of the resistance identification model at time k. Let represent the regression vector at time k of the resistance identification model. This represents the resistance parameter to be identified at time k. This indicates resistance identification measurement noise. This represents the inductance value identified and maintained at the end of the previous first state. This indicates that the d-axis current is not equal to 0. The product of the inductance and the differential of the d-axis current is approximately equal to 0.

[0033] In a specific embodiment of the present invention, the resistance identification model is such that, in state two, the d-axis injected current is not zero, and the system is also in a steady state. The d-axis voltage equation... Simplified to After discretization, it becomes: At this point, the inductance value is taken as the terminal identification value from the previous state. The equation remains unchanged. Rearranging this equation into standard least squares form: in, For output quantity, For regression parameters, Let k be the resistance parameter to be identified. To identify measurement noise for resistance.

[0034] In some embodiments of the present invention, under steady-state conditions of the motor, the q-axis voltage equation because The discretized magnetic flux linkage identification model is constructed using the following formula: in, This is the q-axis voltage value. Let K be the permanent magnet rotor flux linkage to be identified at time k. and For real-time updated resistance and inductance identification values, This is the output of the magnetic linkage identification model at time k. Let be the regression vector of the magnetic linkage identification model at time k. Let K be the flux linkage parameter to be identified at time k. To identify noise in magnetic flux linkage measurements.

[0035] In a specific embodiment of the present invention, the magnetic flux linkage identification model is obtained in real time. and Then, under steady-state conditions of the motor, the q-axis voltage equation Simplified to After discretizing the real-time identification values, the result is: Summarized as follows: in, For output quantity, For regression parameters, Let k be the flux linkage parameter to be identified. To identify noise in magnetic flux linkage measurements.

[0036] In a specific embodiment of the present invention, the inductance identification model, resistance identification model, and flux linkage identification model are all solved online recursively using a unified AF-RLS algorithm. The core recursive formula of the algorithm is shown below: in, For the gain vector, Let be the covariance matrix.

[0037] Forgetting factor The adaptive adjustment is the key to this invention. Its update formula is: Where n is the error amplification factor, LPF( () indicates a low-pass filter, used to smooth changes and extract trends. To Low-pass digital filter, For the filtered Amplify by power. and The preset lower and upper limits for the forgetting factor. This is the adjustment coefficient. When the parameter changes drastically, As the value increases, λ(k) automatically decreases, assigning higher weights to new data and accelerating the algorithm's tracking speed; when the parameters tend to stabilize, As it decreases, λ(k) automatically increases to approach its value. By increasing the weight of historical data, the accuracy of steady-state identification is improved, noise is effectively suppressed, and data saturation is avoided.

[0038] Secondly, the present invention also provides an online parameter identification method for a surface-mounted permanent magnet synchronous motor, applied to the identification system described in any of the above claims, characterized by comprising the following steps: Step 1: The method first injects a low-frequency pulsed small current into the surface-mounted permanent magnet synchronous motor under rated speed and rated load conditions into the motor's d-axis through the pulse injection unit, so that the system alternately enters a first state where the injected current into the d-axis is zero and a second state where it is not zero. Step 2: In the first state, based on the d-axis voltage equation, the inductance value is identified online using the adaptive forgetting factor recursive least squares method, while keeping the resistance value identified at the end of the previous second state unchanged. Step 3: In the second state, based on the d-axis voltage equation and using the inductance value identified at the end of the first state as a known quantity, the resistance value is identified online using the adaptive forgetting factor recursive least squares method, while keeping the inductance identification value unchanged. Step 4: Substitute the inductance and resistance values ​​identified and updated in real time in Step 2 and Step 3 into the q-axis voltage equation, and use the adaptive forgetting factor recursive least squares method to identify the flux linkage value online. Step 5: At the instant the first state and the second state switch, the d-axis current is in an unstable state. For a short period of time after the state switch, the resistance and inductance identification maintain the output value before the switching point.

[0039] In some embodiments of the present invention, the d-axis injection current is determined based on a d-axis injection current function.

[0040] In some embodiments of the present invention, the expression for the d-axis injected current function is: in, Indicates the d-axis injected current. This indicates the magnitude of the injected current along the d-axis. Indicates the injection current period. Represents the frequency of the injected current along the d-axis, where , Indicates the period number ( ).

[0041] In some embodiments of the present invention, the adaptive forgetting factor least recursive squares method is set by the following formula: Where n is the error amplification factor. To Low-pass digital filter, For the filtered Amplify by power. and The preset lower and upper limits for the forgetting factor. For adjustment coefficients, For the gain vector, Let be the covariance matrix.

[0042] In summary, the advantages of this embodiment include: 1) This embodiment performs parameter identification under the rated operating conditions of a surface-mounted permanent magnet synchronous motor; 2) This embodiment achieves completely independent identification of stator resistance, inductance, and rotor flux linkage by injecting low-frequency pulse current into the d-axis and designing a state-alternating decoupling strategy, thus completely solving the identification inaccuracy problem caused by parameter coupling underrank in traditional methods; 3) The forgetting factor used in the identification process of this embodiment can be adaptively adjusted in real time according to the motor operating conditions and parameter change rate, accelerating the tracking speed during parameter mutations and improving the identification accuracy in steady state, effectively overcoming the "data saturation" and tracking lag problems in the traditional recursive least squares method; 4) The identification model proposed in this invention has a clear structure, and the equations in each state are all in linear form. Combined with the recursive algorithm, it has low computational load and low storage requirements, making it easy to run in real time in embedded systems or low-cost controllers; 5) The identification system proposed in this embodiment has a small amplitude and low frequency of the d-axis current loop injection signal, which has minimal impact on the normal operation of the motor, while the identification accuracy can meet the actual control requirements, possessing good engineering and application value. 6) The identification model proposed in this embodiment can synchronously and in real time update the identification results of resistance, inductance and flux linkage during the continuous operation of the motor, providing timely and accurate parameter support for motor adaptive control, fault diagnosis and condition monitoring.

[0043] To better illustrate the effects of this invention, specific simulation experiments are conducted below.

[0044] A simulation model of a surface-mounted permanent magnet synchronous motor vector control system and its parameter identification system was built in the Matlab / Simulink environment. The motor's rated parameters are: stator resistance... ,inductance Rotor flux polar number .

[0045] Simulation conditions: The motor operates under rated conditions, with a speed of 3000 r / min and a load torque of 0.32 N·m. During the simulation, sudden changes in motor parameters were artificially introduced, as shown in Table 1. Table 1: Motor Parameter Variation Table Identify system parameter settings: Injected current amplitude For rated 26.7% of shaft current, frequency To verify the advantages of the AF-RLS algorithm over the traditional Fixed Forgetting Factor Recursive Least Squares (FF-RLS) method, the FF-RLS and AF-RLS algorithms were deployed for the above-mentioned online identification scheme of motor parameters, as shown in Table 2. Table 2: Configuration of the forgetting factor in the RLS algorithm Please refer to the appendix for simulation results. Figure 4 To be continued Figure 12 .

[0046] Figures 4 to 6 The dynamic response of the motor to d-axis current, q-axis current, speed and electromagnetic torque under sudden parameter changes is shown.

[0047] Figure 7 , 9 Figures 1 and 11 show the comparison between the online identification results of resistance, inductance, and flux linkage under the two algorithms and their true values. It can be seen that after a step change in the parameters, both algorithms can track the new true values ​​of the motor parameters and remain stable, achieving decoupling and dynamic tracking of the three parameters. This proves that the online motor parameter identification method proposed in this invention can effectively solve the problems of parameter coupling and underranked identification models. However, it is clear that the overall identification performance based on the AF-RLS algorithm is better than the other two recursive least squares methods with a fixed forgetting factor (FF-RLS). The AF-RLS algorithm can simultaneously balance parameter identification accuracy and dynamic tracking speed.

[0048] Figure 8 , 101 and 12 respectively illustrate the adaptive forgetting factor during the identification process of resistance, inductance, and magnetic flux. The curve shows the change of parameters. It can be clearly observed that at the moment of parameter abrupt change, The value drops rapidly to speed up tracking; once the parameter converges to a new steady state, The value gradually rebounded to improve the steady-state identification accuracy and smooth noise, which intuitively verified the effectiveness of the adaptive mechanism.

[0049] In summary, through the combination of specific implementation methods and simulation examples, this invention fully demonstrates that the online parameter identification method and system for surface-mounted permanent magnet synchronous motors based on AF-RLS provided by this invention can effectively solve the problems of parameter coupling and underranked identification models, and utilizes an adaptive forgetting factor mechanism to balance tracking speed and identification accuracy.

[0050] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. An online parameter identification system for a surface-mount permanent magnet synchronous motor, used in a surface-mount permanent magnet synchronous motor, wherein the surface-mount permanent magnet synchronous motor is equipped with a controller, and the controller is equipped with a current control loop, characterized in that, include: Pulse injection unit and identification unit; wherein, The pulse injection unit is connected at the d-axis current setting point of the current control loop and is used to generate and inject low-frequency pulsed small current to the d-axis. The identification unit collects d-axis voltage, d-axis current, q-axis voltage, q-axis current and motor speed values ​​from the current control loop and speed loop, and performs online parameter identification based on the adaptive forgetting factor recursive least squares method through the corresponding identification model. The identification model includes inductance identification model, stator resistance identification model and rotor flux identification model. In the first state where the d-axis pulse current generated by the pulse injection unit is zero, the inductor identification model is based on the d-axis voltage equation and uses the adaptive forgetting factor recursive least squares method to output the inductor identification value based on the d-axis voltage value, q-axis current value, and motor speed value. In this first state, the resistance identification value is the resistance value obtained at the end of the previous second state. In the second state where the d-axis pulse current is not zero, the stator resistance identification model outputs a resistance identification value based on the d-axis voltage value, d-axis current value, q-axis current value, motor speed value, and the inductance value identified at the end of the previous first state. In this second state, the inductance identification value is the inductance value identified at the end of the previous first state. The flux linkage identification model inputs the inductance identification value and the resistance identification value into the q-axis voltage equation, and combines the q-axis voltage value, d-axis current value, q-axis current value and motor speed value to output the flux linkage identification value.

2. The online parameter identification system for surface-mounted permanent magnet synchronous motors according to claim 1, characterized in that, When the first state is in a steady state, the expression for the d-axis voltage equation is: When the first state is in a steady state , The inductor identification model, after discretization, is constructed using the following formula: Where k represents the sampling point, This represents the d-axis voltage value at sampling point k. This represents the d-axis current value at sampling point k. This represents the q-axis current value at sampling point k. This represents the motor speed value at sampling point k. , Represents the extreme logarithm. This represents the inductance to be identified at time k. This represents the output of the inductance identification model at time k. This represents the regression vector at time k of the inductance identification model. This represents the inductance parameter to be identified at time k. This indicates the noise in the inductance identification measurement.

3. The online parameter identification system for surface-mounted permanent magnet synchronous motors according to claim 1, characterized in that, When the second state is in a steady state, the expression for the d-axis voltage equation is: When the second state is in a steady state , The resistance identification model, after discretization, is constructed using the following formula: in, This represents the resistance value to be identified at time k. This represents the output of the resistance identification model at time k. Let represent the regression vector at time k of the resistance identification model. This represents the resistance parameter to be identified at time k. This indicates resistance identification measurement noise. This represents the inductance value that was identified and maintained at the end of the previous first state.

4. The online parameter identification system for surface-mounted permanent magnet synchronous motors according to claim 1, characterized in that, Under steady-state conditions of the motor, the q-axis voltage equation because The discretized magnetic flux linkage identification model is constructed using the following formula: in, This is the q-axis voltage value. Let K be the permanent magnet rotor flux linkage to be identified at time k. and For real-time updated resistance and inductance identification values, This is the output of the magnetic linkage identification model at time k. Let be the regression vector of the magnetic linkage identification model at time k. Let K be the flux linkage parameter to be identified at time k. To identify noise in magnetic flux linkage measurements.

5. The online parameter identification method for surface-mounted permanent magnet synchronous motors according to claim 1, applied to the identification system according to any one of claims 1 to 4, characterized in that, Includes the following steps: Step 1: Inject a low-frequency pulsed small current into the surface-mounted permanent magnet synchronous motor under rated speed and rated load conditions through the pulse injection unit, so that the system alternately goes through a first state where the injected current into the d-axis is zero and a second state where it is not zero. Step 2: In the first state, based on the d-axis voltage equation, the inductance value is identified online using the adaptive forgetting factor recursive least squares method, while keeping the resistance value identified at the end of the previous second state unchanged. Step 3: In the second state, based on the d-axis voltage equation and using the inductance value identified at the end of the first state as a known quantity, the resistance value is identified online using the adaptive forgetting factor recursive least squares method, while keeping the inductance identification value unchanged. Step 4: Substitute the inductance and resistance values ​​identified and updated in real time in Step 2 and Step 3 into the q-axis voltage equation, and use the adaptive forgetting factor recursive least squares method to identify the flux linkage value online. Step 5: At the instant the first state and the second state switch, the d-axis current is in an unstable state. For a short period of time after the state switch, the resistance and inductance identification maintain the output value before the switching point.

6. The method for online parameter identification of a surface-mounted permanent magnet synchronous motor according to claim 5, characterized in that, The d-axis injection current is determined based on the d-axis injection current function.

7. The method for online parameter identification of a surface-mounted permanent magnet synchronous motor according to claim 6, characterized in that, The expression for the d-axis injected current function is: in, Indicates the d-axis injected current. This indicates the magnitude of the injected current along the d-axis. Indicates the injection current period. Represents the frequency of the injected current along the d-axis, where , Indicates the period number ( ).

8. The method for online parameter identification of a surface-mounted permanent magnet synchronous motor according to claim 5, characterized in that, The adaptive forgetting factor least recursive squares method is set by the following formula: Where n is the error amplification factor. To Low-pass digital filter, For the filtered Amplify by power. and The preset lower and upper limits for the forgetting factor. For adjustment coefficients, For the gain vector, Let be the covariance matrix.

Images