Foc anti-disturbance control method fusing fuzzy pi and kalman filter

Abstract

The application discloses a FOC anti-disturbance control method fusing fuzzy PI and Kalman filtering, which comprises the following steps: d , q-axis actual current; estimated currents are obtained respectively through Kalman filtering; speed deviation is output through a first fuzzy PI controller q , d-axis current given value; q , the deviation between the d-axis current given value and the estimated q , d-axis current is output through a second fuzzy PI controller q , d-axis voltage, d , the deviation between the q-axis current given value and the estimated d , q-axis current is output through a third fuzzy PI controller d , q-axis voltage; and the motor is driven through inverse Park transformation and space vector modulation. The application restrains measurement noise through Kalman filtering, provides pure current feedback for the fuzzy PI controller, realizes cooperative optimization of state estimation and adaptive control, and improves dynamic response, steady-state accuracy and anti-disturbance robustness of a permanent magnet synchronous motor FOC system.

Description

Technical Field

[0001] This invention relates to the field of motor control technology, and specifically to a FOC disturbance rejection control method that integrates fuzzy PI and Kalman filtering. Background Technology

[0002] Vector control, as a core technology for high-performance AC motor drives, achieves DC motor-like control characteristics by decoupling excitation and torque components. Traditional vector control systems rely on linear PI controllers, whose fixed gain parameters are ill-suited to complex operating conditions such as sudden load changes and motor parameter perturbations, leading to lag in dynamic response and decreased steady-state control accuracy. Furthermore, the nonlinear distortion characteristics of the three-phase power inverter and measurement noise from the current sensor further exacerbate the observation errors in the current and speed loops, severely restricting the overall robustness of the system.

[0003] In existing technologies, some studies have attempted to introduce fuzzy logic control into motor control systems, dynamically adjusting PI controller parameters through fuzzy rules to adapt to changes in operating conditions. However, these methods typically use raw sensor measurement signals directly as inputs to the fuzzy controller. In actual systems, current and speed signals inevitably contain high-frequency noise, which interferes with the fuzzy inference process, causing unnecessary high-frequency fluctuations in PI parameters. This may degrade control performance or induce system oscillations.

[0004] Other studies have applied Kalman filters to motor state estimation, utilizing their optimal estimation characteristics to suppress measurement noise. However, these approaches often simply combine the Kalman filter with a fixed-parameter PI controller, failing to achieve dynamic co-optimization of state estimation and control parameters. When motor operating conditions deviate from the design conditions, the fixed PI parameters ( K p , K i The system cannot provide optimal dynamic response, resulting in significant overshoot, prolonged settling time, and even stability problems under disturbances such as sudden load increases.

[0005] Comprehensive analysis shows that existing technologies have significant limitations in addressing the disturbance rejection capabilities of motor control systems: fuzzy control strategies are susceptible to noise interference and fail, Kalman filtering technology has not been deeply integrated with adaptive control mechanisms, and the advantages of the two have not been effectively combined to form a synergistic effect, making it difficult to simultaneously meet the comprehensive performance requirements of the system in terms of dynamic response speed, steady-state control accuracy, and resistance to external disturbances. Summary of the Invention

[0006] The purpose of this invention is to provide a FOC disturbance rejection control method that integrates fuzzy PI and Kalman filtering, which can effectively suppress the interference of measurement noise on the control process, realize the synergistic optimization of Kalman state estimation and fuzzy adaptive PI control, and improve the dynamic response speed, steady-state control accuracy and external disturbance rejection capability of the permanent magnet synchronous motor FOC control system, thereby enhancing the overall robustness of the system.

[0007] This invention provides a FOC disturbance rejection control method that integrates fuzzy PI and Kalman filtering, comprising the following steps: S1, the three-phase current of the permanent magnet synchronous motor i a , i b , i c After passing through the Clarke converter, the motor's position in the two-phase stationary coordinate system is obtained. α shaft current I α and β shaft current I β ; S2, the aforementioned α shaft current I α and β shaft current I β After passing through the Park transformer, the actual result is obtained. d shaft current i d and q shaft current i q ; S3, the actual d shaft current i d and q shaft current i q The estimates are obtained by passing the first Kalman filter and the second Kalman filter, respectively. d shaft current and estimates q shaft current ; S4, the actual speed output of the permanent magnet synchronous motor. N r With expected speed The speed deviation is obtained by subtraction, and the speed deviation is input to the first fuzzy PI controller, which outputs... q Shaft current setpoint ; S5, the above q Shaft current setpoint With the estimate q shaft current By doing the difference, we get q The shaft current deviation is input to the second fuzzy PI controller, and the output is... q Shaft voltage setpoint U q At the same time, d Shaft current setpoint With the estimate d shaft current By doing the difference, we get d The shaft current deviation is input to the third fuzzy PI controller, and the output is... d Shaft voltage setpoint U d ; wherein, the d Shaft current setpoint Set to 0; S6, the above q Shaft voltage setpoint U and d Shaft voltage setpoint U d The input is fed into the inverse Park transformer to obtain... α Shaft voltage control quantity U α and β Shaft voltage control quantity U β The α Shaft voltage control quantity U α and β Shaft voltage control quantity U β The signal is generated by a space vector modulator to control the switching of the three-phase power inverter transistors; the three-phase power inverter converts DC power... V dc Inverting to three-phase current i a , i b , i c This drives the permanent magnet synchronous motor to operate.

[0008] As can be seen from the above, the FOC disturbance rejection control method integrating fuzzy PI and Kalman filtering provided in this application filters and reduces noise in the acquired current signal through Kalman filtering, and then inputs the estimated stable signal into the fuzzy PI controller. At the same time, the PI control parameters are adaptively adjusted through fuzzy rules, realizing the deep integration of state estimation and adaptive control. This effectively solves the problems of fuzzy control being susceptible to noise interference, Kalman filtering failing to be combined with the adaptive control mechanism, and insufficient disturbance rejection capability in the prior art. It has the advantages of effectively suppressing the interference of measurement noise on the control process, realizing the synergistic optimization of Kalman state estimation and fuzzy adaptive PI control, and improving the dynamic response speed, steady-state control accuracy, and external disturbance rejection capability of the permanent magnet synchronous motor FOC control system, thereby enhancing the overall robustness of the system. Attached Figure Description

[0009] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0010] Figure 1 This is a control principle diagram of an FOC anti-disturbance control method that integrates fuzzy PI and Kalman filtering according to an embodiment of the present invention; Figure 2 This is a control principle diagram of conventional method 1 in an embodiment of the present invention; Figure 3 This is a control principle diagram of conventional method 2 in an embodiment of the present invention; Figure 4 The conventional method 1 of the present invention i d Feedback signal diagram; Figure 5 The conventional method 1 of the present invention i q Feedback signal diagram; Figure 6 This is the conventional method 2 of the present invention. i d Feedback signal diagram; Figure 7 This is the conventional method 2 of the present invention. i q Feedback signal diagram; Figure 8 The method of the present invention i d Feedback signal diagram; Figure 9 The method of the present inventioni q Feedback signal diagram; Figure 10 This is a velocity loop curve diagram of conventional method 1 in an embodiment of the present invention; Figure 11 This is a current loop curve diagram of conventional method 1 in this embodiment of the invention; Figure 12 This is a velocity loop curve diagram after a sudden load is applied according to the conventional method 1 of this invention. Figure 13 This is a current loop curve after a sudden load is applied according to the conventional method 1 of this invention. Figure 14 This is a velocity loop curve diagram of conventional method 2 in an embodiment of the present invention; Figure 15 This is a current loop curve diagram of conventional method 2 in this embodiment of the invention; Figure 16 This is a velocity loop curve diagram after a sudden load is applied according to the conventional method 2 of this invention. Figure 17 This is a current loop curve diagram after a sudden load is applied in the conventional method 2 of this invention embodiment; Figure 18 This is a velocity loop curve diagram of the method of the present invention; Figure 19 This is a current loop curve diagram of the method of the present invention; Figure 20 This is a velocity loop curve diagram after a sudden load is applied according to the method of the present invention; Figure 21 This is a current loop curve after a sudden load is applied according to the method of the present invention. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] like Figure 1 As shown, the present invention provides a FOC disturbance rejection control method that integrates fuzzy PI and Kalman filtering. The control system applied by the method includes: a permanent magnet synchronous motor (PMSM), a first Kalman filter, a second Kalman filter, a first fuzzy PI controller, a second fuzzy PI controller, a third fuzzy PI controller, a Clarke converter, a Park converter, an inverse Park converter, a space vector modulator (SWPWM), and a three-phase power inverter.

[0013] The method includes the following steps: Step S1, Three-phase current of permanent magnet synchronous motor i a , i b , i c After passing through the Clarke converter, the motor's position in the two-phase stationary coordinate system is obtained. α shaft current I α and β shaft current I β ; Step S2 α shaft current I α and β shaft current I β After passing through the Park transformer, the actual result is obtained. d shaft current i d and q shaft current i q ; Step S3, the actual d shaft current i d and q shaft current i q The estimates are obtained by passing the first Kalman filter and the second Kalman filter, respectively. d shaft current and estimates q shaft current ; Step S4: Actual speed output by the permanent magnet synchronous motor N r With expected speed The speed deviation is obtained by subtraction, and the speed deviation is input to the first fuzzy PI controller, which outputs... q Shaft current setpoint .

[0014] Among them, actual speed N r Provided by the "Speed" signal output from the permanent magnet synchronous motor.

[0015] Step S5, q Shaft current setpoint With estimation q shaft current By doing the difference, we get q The shaft current deviation is input to the second fuzzy PI controller, and the output is...q Shaft voltage setpoint U q At the same time, d Shaft current setpoint With estimation d shaft current By doing the difference, we get d The shaft current deviation is input to the third fuzzy PI controller, and the output is... d Shaft voltage setpoint U d ;in, d Shaft current setpoint Set it to 0.

[0016] Step S6, q Shaft voltage setpoint U and d Shaft voltage setpoint U d The input is fed into the inverse Park transformer to obtain... α Shaft voltage control quantity U α and β Shaft voltage control quantity U β ; α Shaft voltage control quantity U α and β Shaft voltage control quantity U β The signal is generated by a space vector modulator to control the switching on and off of the three-phase power inverter's transistors; the three-phase power inverter converts DC power... V dc Inverting to three-phase current i a , i b , i c To drive the permanent magnet synchronous motor.

[0017] Specifically, the control method of this embodiment is implemented as follows: First, the three-phase current of the permanent magnet synchronous motor i a , i b , i c The data was collected and converted to a two-phase stationary coordinate system using a Clarke transformer. α shaft current I α and β shaft current I β The three-phase current ia , i b , i c Measurements can be taken using current sensors. The Clarke converter can be implemented using software algorithm modules in a digital signal processor, projecting the three-phase AC current onto orthogonal circuits. α - β Stationary coordinate system.

[0018] Furthermore, α shaft current I α and β shaft current I β The input is fed into the Park converter and converted into a two-phase rotating phase. d - q Actual coordinate system d shaft current i d and reality q shaft current i q Parker converters are typically implemented using software algorithms that utilize the motor's rotor position information to rotate the current component in the stationary coordinate system to a rotation synchronized with the rotor. d - q Coordinate system.

[0019] Subsequently, in reality d shaft current i d and reality q shaft current i q After processing by the first Kalman filter and the second Kalman filter respectively, the estimated values ​​are obtained. d shaft current and estimates q shaft current Each Kalman filter can be configured to perform state estimation based on a dynamic model of the current signal to filter out measurement noise and provide a smooth current estimate. For example, the filter can employ a linear state-space model to extract an estimate from noisy measurements through prediction and correction steps. d shaft current and estimates q shaft current .

[0020] Meanwhile, the actual speed of the motor N r Acquired by sensors and compared with the expected rotational speed. The difference is calculated to obtain the speed deviation. This speed deviation is input to the first fuzzy PI controller. The first fuzzy PI controller is configured to receive the deviation signal and process it according to a preset fuzzy rule.

[0021] Therefore, the first fuzzy PI controller receives the speed deviation as input, calculates and outputs the result through its internal fuzzy logic processing unit and PI control structure. q Shaft current setpoint .Should q Shaft current setpoint This represents the amount of torque required to achieve the desired speed.

[0022] then, q Shaft current setpoint The estimate from the output of the second Kalman filter q shaft current By comparing, we can obtain q Shaft current deviation. q The shaft current deviation is input to the second fuzzy PI controller. Simultaneously, d Shaft current setpoint It is typically set to zero to achieve magnetic field orientation control, which is compared with the estimate output by the first Kalman filter. d shaft current By comparing, we can obtain d Shaft current deviation. d The shaft current deviation is input to the third fuzzy PI controller.

[0023] Finally, the second fuzzy PI controller receives... q Shaft current deviation and output q Shaft voltage setpoint U q The third fuzzy PI controller receives d Shaft current deviation and output d Shaft voltage setpoint U d These voltage given values U q and U d It is fed into the inverse Park transformer and converted to α Shaft voltage control quantity U α and β Shaft voltage control quantity U β Subsequently, α Shaft voltage control quantity U α and β Shaft voltage control quantity U βThe signal is input to a space vector modulator, generating pulse signals to control the on / off switching of the three-phase power inverter's transistors. The three-phase power inverter converts DC power... V dc Converted to three-phase alternating current i a , i b , i c Ultimately, this drives the permanent magnet synchronous motor to run.

[0024] The control method proposed in this embodiment effectively solves the problems of slow dynamic response, insufficient steady-state accuracy, and parameter jitter caused by noise interference in traditional FOC (Focus-Oriented Control) systems by applying a fuzzy PI controller to the speed and current loops and introducing a Kalman filter to estimate the current signal. This method provides permanent magnet synchronous motors with more accurate and robust disturbance rejection capabilities in applications such as electric vehicle power systems and industrial servo motors, where high dynamic response and stability requirements are required.

[0025] In one optional implementation, the first fuzzy PI controller is a speed-loop fuzzy PI controller, used to adjust PI parameters according to fuzzy rules based on the speed deviation as input, and output... q Shaft current setpoint .

[0026] Specifically, the first fuzzy PI controller is designed as a speed loop fuzzy PI controller. This speed loop fuzzy PI controller is specifically designed for the outer speed loop in the FOC control system of a permanent magnet synchronous motor, and its core function is to receive the actual speed output by the permanent magnet synchronous motor. N r With expected speed The speed deviation obtained by subtraction is used as input. Unlike traditional fixed-parameter PI controllers, this speed-loop fuzzy PI controller can adaptively adjust its internal proportional gain in real time using a preset fuzzy rule set. K p ) and integrals ( K i These fuzzy rules are typically based on an understanding of the motor's dynamic characteristics and control experience. For example, when the speed deviation is large, the controller adjusts parameters to provide stronger control and accelerate speed convergence; when the speed deviation is small, it adjusts parameters to reduce overshoot and improve steady-state accuracy. Through this dynamic parameter adjustment mechanism, the controller can output more accurate and adaptive parameters. q Shaft current setpoint This drives the inner current loop to generate corresponding torque.

[0027] In one optional implementation, both the second and third fuzzy PI controllers are current-loop fuzzy PI controllers, respectively used for... d Shaft current deviation and q The shaft current deviation is the input; the PI parameter is adaptively adjusted, and the corresponding output is adjusted accordingly. d Shaft voltage setpoint U d and q Shaft voltage setpoint U q .

[0028] Among them, the current-loop fuzzy PI controller is a fuzzy PI controller specifically designed for regulating motor current. In the field-oriented control (FOC) of a permanent magnet synchronous motor, the estimate obtained after passing through the first Kalman filter and the second Kalman filter... d shaft current and estimates q shaft current They are decoupled, corresponding to the flux linkage and torque components respectively. Therefore, independent controllers are needed to precisely regulate these two current components. The current-loop fuzzy PI controller introduces fuzzy logic, enabling it to adaptively adjust the proportional coefficient of the PI controller based on the current deviation and the rate of change of the deviation. K p ) and integral coefficient ( K i This allows the current-loop fuzzy PI controller to maintain good control performance under different operating conditions. Compared with traditional fixed-parameter PI controllers, the current-loop fuzzy PI controller can better handle the nonlinearity and uncertainty of the system, improving the dynamic response speed and steady-state accuracy of current regulation.

[0029] Specifically, in the vector control of permanent magnet synchronous motors, estimation d shaft current and estimates q shaft current These are the key control variables. Estimation d shaft current It is typically used to control magnetic flux, and may not be zero in field weakening control, but is usually set to zero in the constant torque region. To achieve the maximum torque-to-current ratio. (Estimation) q shaft current This is directly related to the motor torque.

[0030] Therefore, by estimating d shaft current and d Shaft current setpoint The difference is obtained d Shaft current deviation, and estimation q shaft current and q Shaft current setpoint The difference is obtained q The shaft current deviation, as an input to the current-loop fuzzy PI controller, directly reflects the difference between the current and the desired current. The controller adjusts based on these deviation signals to eliminate the deviation and enable the actual current to quickly track the given current.

[0031] Adaptive adjustment of the PI parameter refers to the controller's ability to adjust its internal proportional coefficient in real time according to changes in the system's operating state. K p and integral coefficient K i In a current-loop fuzzy PI controller, this adaptability is achieved through fuzzy logic. The fuzzy controller takes the current deviation and the rate of change of deviation as input, and through a fuzzy inference mechanism, outputs a correction value for the PI parameters. K p and K i These adjustments are used to update the current... K p and K i This value allows the PI controller to automatically adjust its control characteristics under conditions such as changes in motor parameters, load disturbances, or speed variations, thereby maintaining excellent control performance. For example, when the current deviation is large, the controller may increase... K p To speed up response; when the deviation is small, adjustments may be made. K i To eliminate steady-state error.

[0032] d Shaft voltage setpoint U d and q Shaft voltage setpoint U q These are the outputs of the current-loop fuzzy PI controller; they represent the outputs of the current-loop fuzzy PI controller. d - q In a rotating coordinate system, the voltage components required for the actual current to track the given current are defined. These voltage setpoints are control quantities that directly act on the motor windings, and are converted to a two-phase stationary coordinate system using an inverse Park converter. α Shaft voltage control quantity U α and β Shaft voltage control quantity U βThen, a pulse signal is generated through a space vector modulator to drive a three-phase power inverter, ultimately producing a three-phase current to drive the permanent magnet synchronous motor. i a , i b , i c .therefore, U d and U q Precise output is key to achieving accurate current control and stable motor operation.

[0033] In one alternative implementation, the process by which each fuzzy PI controller (here representing any one of the first, second, and third fuzzy PI controllers) processes the input signal includes: Deviation E (This refers to the speed deviation) d Shaft current deviation and q (General term for shaft current deviation) and the corresponding rate of change of deviation EC As input, they are mapped to their respective fuzzy domains through quantization factors and transformed into fuzzy linguistic variables using membership functions; Fuzzy inference is performed on fuzzy linguistic variables based on a predefined fuzzy rule table to obtain fuzzy output; The fuzzy output is defuzzified using the centroid method, and then mapped to their respective fundamental universes using a scaling factor to obtain the precise PI parameter correction. The proportional and integral coefficients are updated in real time based on the PI parameter correction, and finally a control signal is generated. The parameter update expression for the fuzzy PI controller is as follows: ; in, K p This is a real-time scaling factor. K i For real-time integral coefficients, K p ' It is the initial scaling factor. K i ' These are the initial integration coefficients. K p This is the adjustment amount for the proportionality coefficient. K i This is the correction amount for the integral coefficient.

[0034] Specifically, in the input processing stage of the fuzzy PI controller, the system deviation is first obtained.E (e.g., speed deviation, d Shaft current deviation or q Shaft current deviation and the rate of change of deviation E over time EC .deviation E It reflects the degree of deviation between the current system state and the desired state, while the rate of change of deviation... EC This indicates the trend and speed of this deviation. To enable these precise input quantities to be processed by the fuzzy logic system, quantization mapping is required, converting them into discrete values ​​in the fuzzy domain through a quantization factor. Subsequently, these discrete values ​​are transformed into the degree of corresponding fuzzy linguistic variables (such as "negative large", "zero", "positive small", etc.) through a membership function. This embodiment uses 7 levels of fuzzy linguistic variables, namely... NB (Negative) NM (Negative) NS (Negative) Z (zero), PS (Small), PM (middle), PB (Zhengda). Membership functions can take many forms, such as triangular, trapezoidal, or Gaussian. Their function is to map precise input values ​​to membership degrees between 0 and 1, representing their degree of belonging in the fuzzy universe.

[0035] After receiving fuzzy input, the fuzzy PI controller performs fuzzy inference based on a pre-defined fuzzy rule table. The fuzzy rule table typically consists of a series of "IF-THEN" statements, as shown in Tables 1 and 2. Table 1 is... Fuzzy rule table, Table 2 is (Fuzzy rule table). For example: "If the deviation..." E for PB (Zhengda) and the rate of change of deviation EC for Z (Zero), then the proportional coefficient correction amount for NM (Negative), integral coefficient correction amount for PM (middle)".

[0036] Table 1 Table 2 These rules are derived from expert experience or system characteristics and are used to describe the relationship between input fuzzy quantities and output fuzzy quantities. The fuzzy inference process activates the corresponding fuzzy rules based on the current fuzzy input and combines them with fuzzy logic operations (such as the minimum value method, product method, etc.) to obtain a fuzzy output set.

[0037] To apply the fuzzy output to actual PI controller parameter adjustment, defuzzification is required. This application employs the centroid method to defuzzify the fuzzy output. The centroid method is a commonly used defuzzification method that obtains a precise value by calculating the centroid of the membership function of the fuzzy output. This precise value is then mapped to the fundamental universe of discourse through a scaling factor to obtain the PI parameter correction, including the scaling coefficient correction. K p and integral coefficient correction amount K i These adjustments reflect how the PI controller parameters should be adjusted based on the current system deviation and its trend.

[0038] Finally, based on the calculated proportional coefficient correction amount K p and integral coefficient correction amount K i Real-time updates of the proportional coefficient of the PI controller K p and integral coefficient K i In this way, the parameters of the PI controller can be dynamically adjusted according to the real-time operating status of the system, thereby generating control signals adapted to the current operating conditions (such as...). d Shaft voltage setpoint U d and stated q Shaft voltage setpoint U q ).

[0039] In one optional implementation, the quantization factor is 1 and the scaling factor is 0.833; deviation E and rate of change of deviation EC Both the fundamental and fuzzy universes of discourse are [-6,6]; the scaling factor correction is... K p and integral coefficient correction amount K i The basic universe of discourse is [-5,5] and the fuzzy universe of discourse is [-6,6], as shown in Table 3.

[0040] Table 3

[0041] In one alternative implementation, the membership function is a triangular membership function.

[0042] By employing a triangular membership function, the fuzzy PI controller of this application can achieve an efficient and intuitive mapping of input quantities to fuzzy sets during fuzzification. The concise linear structure of the triangular membership function not only reduces the computational complexity of the fuzzification process and ensures the real-time performance of fuzzy inference, but its clear boundary definition also helps improve the accuracy of fuzzy inference. This enables the fuzzy PI controller to generate proportional coefficient corrections more accurately. K p and integral coefficient correction amount K i This allows for real-time and effective updates of the scaling factor. K p and integral coefficient K i Ultimately, this helps improve the adaptive adjustment capabilities of the speed loop fuzzy PI controller, the second fuzzy PI controller, and the third fuzzy PI controller to the operating state of the permanent magnet synchronous motor, enhances the robustness and control accuracy of the entire FOC disturbance rejection control method, and ensures that the motor can operate stably and efficiently under various operating conditions.

[0043] To verify the technical effect of the present invention, a comparative verification was performed using the method of the present invention, conventional method 1, and conventional method 2, as follows: The specific control methods of traditional method 1 and traditional method 2 are explained below: Traditional Method 1 is a disturbance rejection control method for permanent magnet synchronous motors based on a traditional PI controller, such as... Figure 2 As shown, the control system involved includes a permanent magnet synchronous motor (PMSM), a first PI controller, a second PI controller, a third PI controller, a Clarke converter, a Park converter, an inverse Park converter, a space vector modulator (SWPWM), and a three-phase power inverter; it includes the following steps: Step S1, Three-phase current of permanent magnet synchronous motor i a , i b , i c The two-phase stationary coordinates of the motor are obtained after using the Clarke converter. α shaft current I α and β shaft current I β ; Step S2 α shaft current I α and β shaft current Iβ The actual result is obtained after passing through the Park transformer. d shaft current i d and q shaft current i q ; Step S3, Actual Motor Speed N r With expected speed The difference is used to obtain the speed deviation, which is then input to the first PI controller, and the output is... q Shaft current setpoint ; Step S4 q Shaft current setpoint With reality q shaft current i q The difference is calculated, and the difference is input to the second PI controller, which outputs... q Shaft voltage setpoint U q ; d Shaft current setpoint ( (Set to 0) and the actual d shaft current i d The difference is calculated, and the difference is input to the third PI controller, which outputs... d Shaft voltage setpoint U d ; Step S5 q Shaft voltage setpoint U and d Shaft voltage setpoint U d The input is fed into the inverse Park transformer to obtain... α Shaft voltage control quantity U α and β Shaft voltage control quantity U β ; α Shaft voltage control quantity U α and β Shaft voltage control quantity U β After passing through a space vector modulator, a pulse signal is obtained to control the switching of the three-phase power inverter's transistors. The three-phase power inverter then converts the DC power supply... V dc Inverting to three-phase current i a , i b , i c It drives the permanent magnet synchronous motor to run.

[0044] Traditional method 2 is a disturbance rejection control method for permanent magnet synchronous motors based on a fuzzy PI controller, such as... Figure 3 As shown, the control system involved includes a permanent magnet synchronous motor (PMSM), a first fuzzy PI controller, a second fuzzy PI controller, a third fuzzy PI controller, a Clarke converter, a Park converter, an inverse Park converter, a space vector modulator (SWPWM), and a three-phase power inverter; it includes the following steps: Step S1, Three-phase current of permanent magnet synchronous motor i a , i b , i c After passing through the Clarke converter, the motor's position in the two-phase stationary coordinate system is obtained. α shaft current I α and β shaft current I β ; Step S2 α shaft current I α and β shaft current I β After passing through the Park transformer, the actual result is obtained. d shaft current i d and q shaft current i q ; Step S3, Actual speed output by the permanent magnet synchronous motor N r With expected speed The speed deviation is obtained by subtraction, and the speed deviation is input to the first fuzzy PI controller, which outputs... q Shaft current setpoint ; Step S4 q Shaft current setpoint With reality q shaft current i q The difference is calculated, and the difference is input to the second fuzzy PI controller, which outputs... q Shaft voltage setpoint U q ; d Shaft current setpoint ( (Set to 0) and the actual d shaft current i dThe difference is calculated, and the difference is input to the third fuzzy PI controller, which outputs... d Shaft voltage setpoint U d ; Step S6, q Shaft voltage setpoint U and d Shaft voltage setpoint U d The input is fed into the inverse Park transformer to obtain... α Shaft voltage control quantity U α and β Shaft voltage control quantity U β ; α Shaft voltage control quantity U α and β Shaft voltage control quantity U β The signal is generated by a space vector modulator to control the switching on and off of the three-phase power inverter's transistors; the three-phase power inverter converts DC power... V dc Inverting to three-phase current i a , i b , i c To drive the permanent magnet synchronous motor.

[0045] The control algorithm of the present invention, conventional method 1, and conventional method 2 is verified using the Simulink platform to test the control algorithm of a motor-driven magnetic powder brake system. It should be noted that an analog oscilloscope was used to monitor the feedback during the verification process. i d and input and feedback i q and input A comparison was made. Experimental setup. i d =0, the simulation results are more intuitive. The FOC of the traditional PI controller in traditional method 1 can achieve vector control of the motor. First, observe the experiment... i d Feedback signals, such as Figure 4 As shown, the waveform exhibits high-frequency, irregular jitter, with a fluctuation amplitude of approximately ±0.5 N*m, and contains numerous random glitches. This indicates that high-frequency noise from the current measurement directly enters the control loop, leading to... d The shaft current tracking error is large and the torque fluctuation is severe.

[0046] Continuei q Feedback signal comparison, such as Figure 5 As shown. The electromagnetic torque fluctuation amplitude is approximately ±0.5 N*m, and the waveform contains numerous high-frequency random glitches, indicated by the red line ( The blue feedback waveform around () i q The severe jitter indicates that current measurement noise directly enters the control loop, causing... q The shaft current tracking error is large, and the torque fluctuation is irregular. The tracking error has no obvious convergence trend, random noise causes continuous interference, the current tracking accuracy is insufficient in steady state, and the system has weak anti-interference capability.

[0047] In traditional method 2, the FOC simulation verification after incorporating a fuzzy PI controller i d Feedback signals such as Figure 6 As shown. With Figure 4 In contrast, during the initial stage of the electromagnetic torque step response, the rise time was shortened to 0.02s, a reduction of 20%. High-frequency jitter was reduced, but a large number of random glitches still existed, indicating that the fuzzy PI only optimized the parameter response speed and did not suppress the current measurement noise at its source, and the random component of the tracking error was still significant.

[0048] Simulation verification after incorporating fuzzy adaptive PI algorithm i q Feedback signals such as Figure 7 As shown. With Figure 5 In comparison, the overshoot of the electromagnetic torque decreased from 30 to 25, representing a 5% performance improvement. The time to reach final steady state decreased from 2.1 seconds to 2 seconds, an improvement of 1%. This indicates that the adaptive adjustment of the fuzzy PI optimizes the parameter response speed, but does not fundamentally suppress current measurement noise. The steady-state tracking error is improved compared to the traditional PI, but the torque fluctuation caused by noise is still significant, indicating insufficient system robustness.

[0049] In the method of this application, the id feedback signal after the fusion of the fuzzy PI controller and the Kalman filter is as follows: Figure 8 As shown. With Figure 4 and Figure 6 In contrast, the waveform exhibits periodic and regular fluctuations without random high-frequency glitches. The fluctuation amplitude is concentrated within ±0.5 N*m, and noise components are significantly filtered out, retaining only the periodic error synchronized with the motor speed. The regularity of the error waveform is significantly improved, and the fuzzy PI can further optimize parameters for systematic errors, ultimately achieving higher steady-state tracking accuracy and torque stability.

[0050] i q Feedback signals such as Figure 9As shown, the amplitude of the fluctuation is minimal (approximately ±0.3 N*m), the waveform exhibits periodic and regular fluctuations, and there are no random high-frequency glitches. The curve becomes smooth and continuous, indicating that the Kalman filter effectively filters out high-frequency random noise in the current signal, providing a cleaner current feedback for the fuzzy PI. The regularity of the tracking error is significantly improved, with the q-axis current tracking accuracy being the highest. The fuzzy PI can further adaptively adjust parameters to address periodic fluctuations, ultimately achieving a more stable torque output.

[0051] Through the Figure 4 , 6 A comparison of 8 and 1 shows the FOC (Forward Opening) after fusing the fuzzy PI controller and the Kalman filter. d The random component of shaft current tracking error is significantly reduced, the steady-state performance of the system is more stable, and the torque fluctuation is smoother, making it particularly suitable for speed control scenarios of permanent magnet synchronous motors that are sensitive to noise and have high precision requirements.

[0052] Through the Figure 5 , 7 A comparison of 9 shows that after fusing the fuzzy PI controller with the Kalman filter, q It has the smallest axis current tracking error and the smoothest waveform, and its steady-state accuracy is significantly improved compared with traditional PI and pure fuzzy adaptive PI.

[0053] The control algorithm of the present invention, conventional method 1, and conventional method 2 is verified using the Matlab Simulink platform: First, the speed loop of the FOC control using the traditional PI controller in Method 1 is verified. Under unloaded conditions, the speed is set to 1000 rad / s. The speed loop image and a magnified view are shown below. Figure 10 As shown. The current loop is then further inspected; the current loop image and its magnified portion are shown below. Figure 11 As shown.

[0054] A sudden load is applied to the motor, with the speed set at 1000 rad / s. The speed loop and its enlarged local diagram after loading are shown below. Figure 12 As shown. The current loop image and its magnified portion are shown below. Figure 13 As shown.

[0055] Speed ​​loop verification was performed on the FOC (Functional Oscillator) after incorporating a fuzzy PI controller in traditional method 2. Under unloaded conditions, the rotational speed was also set to 1000 rad / s. The speed loop image and a magnified view are shown below. Figure 14 As shown. The current loop image and its magnified portion are shown below. Figure 15 As shown.

[0056] A sudden load is applied to the motor, with the speed set at 1000 rad / s. The speed loop and its enlarged local diagram after loading are shown below. Figure 16As shown, the current loop image and its magnified portion are as follows: Figure 17 As shown.

[0057] The FOC of the fuzzy PI controller and Kalman filter fusion in the method of this invention was verified. Under unloaded conditions, the rotational speed was set to 1000 rad / s. The velocity loop image and a magnified view are shown below. Figure 18 As shown. The current loop image and its magnified portion are shown below. Figure 19 As shown.

[0058] A sudden load is applied to the motor, with the speed set at 1000 rad / s. The speed loop and its enlarged local diagram after loading are shown below. Figure 20 As shown, the current loop image and its magnified portion are as follows: Figure 21 As shown.

[0059] (1) When the motor is suddenly unloaded, by Figure 10 , 14 Comparing the 18-speed loop diagrams, in traditional method 1, the speed fluctuation range is approximately ±0.5 rad / s, with a relatively large fluctuation range and obvious high-frequency random jitter in the waveform; in traditional method 2, the fluctuation range is slightly reduced (±0.4 rad / s), and the high-frequency jitter is reduced, but obvious periodic fluctuations still exist; the speed loop waveform in the method of this invention is the smoothest, with high-frequency random jitter almost disappearing and only weak periodic fluctuations remaining. The speed is closer to the given value, and the system's steady-state performance is superior.

[0060] pass Figure 11 , 15 A comparison of the current loop diagrams shows that in traditional method 1, the waveform exhibits high-frequency, irregular jitter, with the blue / red curve fluctuating by approximately ±0.04V and containing numerous random spikes. Traditional method 2 shows a slightly reduced fluctuation amplitude (approximately ±0.03V) and less high-frequency jitter, but random spikes still exist. The waveform obtained using the method described in this invention exhibits periodic, regular fluctuations without random high-frequency spikes, with a fluctuation amplitude of approximately ±0.025V. The current loop changes are significantly clearer, and the current more closely approximates a sinusoidal waveform.

[0061] (2) When the motor is suddenly subjected to a load, by Figure 12 , 16Comparing the 20-speed loop diagrams, in traditional method 1, the maximum speed drop after a sudden load application is approximately 870 rad / s (130 rad / s different from the given value of 1000 rad / s), a large drop; the recovery time is approximately 3 seconds, accompanied by significant high-frequency oscillations and overshoot, resulting in severe speed fluctuations after disturbance and weak anti-disturbance capability. Even after recovery, the speed still exhibits slight jitter, indicating insufficient steady-state accuracy. In traditional method 2, the maximum speed drop after a sudden load application is 750 rad / s, a larger drop than traditional PI, indicating that the adaptive adjustment of the fuzzy PI is insufficient in the initial response to sudden loads; the recovery time is approximately 2 seconds, the number of oscillations is reduced, and the anti-disturbance capability is better than traditional PI, but stability needs improvement. In the method of this invention, the maximum speed drop after a sudden load application is approximately 890 rad / s, the recovery time is approximately 0.8 seconds, there is no significant high-frequency oscillation, the overshoot is extremely small, and the speed quickly returns to the given value; after recovery, the speed is stable at around 1000 rad / s without jitter, achieving the highest steady-state accuracy.

[0062] pass Figure 13 , 17 A comparison of the 21 current loop diagrams reveals that in traditional method 1, the current fluctuation amplitude reaches ±0.2V when a sudden load is applied, exhibiting sharp positive and negative spikes accompanied by numerous high-frequency random glitches; the oscillation duration after disturbance is long (approximately 1 second or more), and the waveform shows no obvious convergence trend. In traditional method 2, the fluctuation amplitude is slightly reduced (approximately ±0.1V), and high-frequency glitches are reduced, but significant positive and negative spikes and secondary fluctuations still exist. The recovery time is shortened to approximately 0.5 seconds, but small-amplitude oscillations still exist, and the random fluctuations of the steady-state current are not completely eliminated. In the method of this invention, the waveform exhibits periodic and regular fluctuations, without random high-frequency glitches, the fluctuation amplitude is stable at approximately ±0.09V, and the spikes are completely suppressed, resulting in a faster recovery speed after disturbance.

[0063] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A FOC disturbance rejection control method integrating fuzzy PI and Kalman filtering, characterized in that, Includes the following steps: S1, the three-phase current of the permanent magnet synchronous motor i a , i b , i c After passing through the Clarke converter, the motor's position in the two-phase stationary coordinate system is obtained. α shaft current I α and β shaft current I β ; S2, the aforementioned α shaft current I α and β shaft current I β After passing through the Park transformer, the actual result is obtained. d shaft current i d and q shaft current i q ; S3, the actual d shaft current i d and q shaft current i q The estimates are obtained by passing the first Kalman filter and the second Kalman filter, respectively. d shaft current and estimates q shaft current ; S4, the actual speed output of the permanent magnet synchronous motor. N r With expected speed The speed deviation is obtained by subtraction, and the speed deviation is input to the first fuzzy PI controller, which outputs... q Shaft current setpoint ; S5, the above q Shaft current setpoint With the estimate q shaft current By doing the difference, we get q The shaft current deviation is input to the second fuzzy PI controller, and the output is... q Shaft voltage setpoint U q At the same time, d Shaft current setpoint With the estimate d shaft current By doing the difference, we get d The shaft current deviation is input to the third fuzzy PI controller, and the output is... d Shaft voltage setpoint U d ; wherein, the d Shaft current setpoint Set to 0; S6, the above q Shaft voltage setpoint U and d Shaft voltage setpoint U d The input is fed into the inverse Park transformer to obtain... α Shaft voltage control quantity U α and β Shaft voltage control quantity U β The α Shaft voltage control quantity U α and β Shaft voltage control quantity U β The space vector modulator generates pulse signals to control the switching on and off of the three-phase power inverter transistors. The three-phase power inverter converts DC power... V dc Inverting to three-phase current i a , i b , i c This drives the permanent magnet synchronous motor to operate.

2. The method according to claim 1, characterized in that, The first fuzzy PI controller is a speed-loop fuzzy PI controller, used to adjust PI parameters according to fuzzy rules based on the speed deviation as input, and output the... q Shaft current setpoint .

3. The method according to claim 1, characterized in that, Both the second and third fuzzy PI controllers are current-loop fuzzy PI controllers, respectively used for... d Shaft current deviation and the q The shaft current deviation is taken as input, the PI parameter is adaptively adjusted, and the corresponding output is given. d Shaft voltage setpoint U d and stated q Shaft voltage setpoint U q .

4. The method according to claim 2 or 3, characterized in that, Each of the fuzzy PI controllers processes the input signal as follows: Deviation E and the corresponding rate of change of deviation EC As input, they are mapped to their respective fuzzy domains through quantization factors and transformed into fuzzy linguistic variables using membership functions; Fuzzy inference is performed on fuzzy linguistic variables based on a predefined fuzzy rule table to obtain fuzzy output; The fuzzy output is defuzzified using the centroid method, and then mapped to their respective fundamental universes using a scaling factor to obtain the precise PI parameter correction. The proportional and integral coefficients are updated in real time based on the PI parameter correction amount, and finally a control signal is generated. The parameter update expression for the fuzzy PI controller is as follows: ； in, K p This is a real-time scaling factor. K i For real-time integral coefficients, K p ' It is the initial scaling factor. K i ' These are the initial integration coefficients. K p This is the adjustment amount for the proportionality coefficient. K i This is the correction amount for the integral coefficient.

5. The method according to claim 4, characterized in that, The quantization factor is 1, and the scaling factor is 0.833; deviation E and rate of change of deviation EC Both the fundamental and fuzzy domains are [-6,6]; proportional coefficient correction amount K p and integral coefficient correction amount K i The basic domains are all [-5,5] and the fuzzy domains are all [-6,6].

6. The method according to claim 4, characterized in that, The membership function is a triangular membership function.

Images