A PMLSM time-varying system disturbance observation method, system and application
By separating and observing motor parameter mismatch and noise disturbance in PMLSM system, the problems of parameter dependence and noise influence in current prediction control are solved, and current control with high accuracy and high dynamic performance is achieved.
Patent Information
- Application Number
- CN202210311277.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-28
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2042-03-28
AI Technical Summary
The existing current prediction control has strong dependence on motor parameters and models. The mismatch of model parameters leads to deviation of the voltage reference value, and the delay of the digital control system reduces the control performance. The measurement noise and process noise are not considered during the iterative update process, which affects the control effect. .
A dynamic model of the PMLSM system including motor parameter mismatch and unmodeled dynamics was constructed, and measurement noise and iterative noise were introduced as time-varying disturbances, and slow-varying disturbances and time-varying disturbances were separated, and observing system disturbances were designed by expanding Kalman filters.
Real-time observation of disturbances of PMLSM system is achieved, current tracking errors are eliminated, dynamic performance and steady-state accuracy are improved, and is suitable for CNC machine tools, semiconductor manufacturing, logistics and transportation.
Smart Images

Figure CN114900097B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of servo system disturbance observation, and in particular relates to a PMLSM time-varying system disturbance observation method, system and application. Background Art
[0002] With the growing demand for high-speed and high-precision linear motion systems, permanent magnet linear synchronous motors (PMLSMs) have been adopted in numerous industrial applications, such as semiconductor manufacturing and CNC machine tools. In practical PMLSM control systems, current control performance is a key factor in determining control system performance. Therefore, to achieve high-precision positioning and high dynamic current response, numerous current control methods have been proposed. Hysteresis control, proportional-integral (PI) control, and predictive control are the most widely used.
[0003] Compared to traditional hysteresis control and PI control, predictive control offers superior steady-state and dynamic performance. However, current predictive control is highly dependent on motor parameters and models. Mismatches in model parameters can cause the calculated voltage reference to deviate from the desired value. Furthermore, control delays in digital control systems, such as sampling delay and duty cycle update delay, can significantly degrade system performance.
[0004] Currently, there are two main approaches to addressing the disturbance caused by parameter mismatch in current predictive control: one is to design corresponding observers, using system modeling to observe parameter disturbance errors online and compensate and adjust the model in real time; the other is to use identification algorithms to identify motor parameters online and use the identification results to correct the control model in real time. Both approaches have been extensively researched and have achieved some success. However, because predictive control requires iterative updates during each control cycle, and these approaches do not account for measurement and process noise during the iterative update process, the control effect still has room for improvement.
[0005] Through the above analysis, the problems and defects of the existing technology are as follows:
[0006] (1) The existing current prediction control is highly dependent on motor parameters and models. Mismatch of model parameters will cause the calculated voltage reference value to deviate from its expected value.
[0007] (2) The control delay of existing digital control systems, such as sampling delay and duty cycle refresh delay, will greatly reduce the control performance of the system.
[0008] (3) Since predictive control requires iterative update calculations in each control cycle, existing methods do not consider the measurement noise and process noise during the iterative update process, and there is room for improvement in the control effect. Summary of the Invention
[0009] In response to the problems existing in the prior art, the present invention provides a PMLSM time-varying system disturbance observation method, system and application, and in particular relates to a permanent magnet synchronous linear motor time-varying system disturbance observation technology based on extended Kalman filtering.
[0010] The present invention is implemented as follows: a PMLSM time-varying system disturbance observation method, the PMLSM time-varying system disturbance observation method comprising:
[0011] A PMLSM system dynamics model including motor parameter mismatch and unmodeled dynamics is constructed, and two time-varying disturbances, measurement noise and iteration noise, are introduced. Relative to the linear motor control period, the disturbance caused by parameter disturbance and unmodeled dynamics is a slow variable, while the measurement noise and process noise are a time-varying variable. All slowly varying terms are regarded as the system lumped disturbance, and the n-th order derivative is expanded to model the system state. Time-varying terms are introduced in the expanded state model, and the system disturbance is observed by designing a Kalman filter.
[0012] Furthermore, the PMLSM time-varying system disturbance observation method comprises the following steps:
[0013] Step 1: The system disturbance is divided into slow-varying terms and time-varying terms. Slow-varying terms such as motor parameter mismatch and unmodeled dynamics are regarded as the system lumped disturbance.
[0014] Step 2: Calculate the 1st to nth order derivatives of the lumped disturbance and model them as the system expansion state;
[0015] Step 3: introducing two time variables, measurement noise and iteration noise, into the system extended state model;
[0016] Step 4: Design a Kalman filter for the system extended state model with the time-varying disturbance term introduced, and determine the selection range of the gain parameter through simulation.
[0017] Furthermore, the system disturbance in step 1 is:
[0018] D(k)=f(ΔR s ,ΔL s ,Δψ f ,ε,ω,υ);
[0019] Where, ΔR s is the winding resistance error, ΔL s is the stator inductance error, Δψ fis the permanent magnet flux error, ε is the unmodeled dynamic error, ω is the iteration noise, and υ is the measurement noise. Taking the servo system control period as the benchmark, the first four items are motor parameter errors, which are slowly varying disturbances; the last two items are noise errors, which are time-varying disturbances.
[0020] The PMLSM system model including the slowly varying disturbance term is:
[0021]
[0022] Among them, u d / u q and i d / i q are the d / q axis stator voltage and current, R s and L s are the winding resistance and stator inductance, ψ f is the permanent magnet flux, ω e =πv / τ is the electrical angular velocity, where v is the linear velocity of the rotor and τ is the pole pitch; the subscript "o" indicates the nominal value of the parameter, f d ,f q represents the lumped perturbation due to slowly varying perturbations such as parameter mismatch and unmodeled dynamics and is calculated as:
[0023]
[0024] Among them, R s =R so +ΔR s , L s =L so +ΔL s , ψ f =ψ fo +Δψ f , ε d / ε q are the d / q axis lumped unmodeled dynamics, respectively.
[0025] Furthermore, in the step 2, the lumped disturbance is expanded by Taylor as follows:
[0026]
[0027] Among them, n is the order, a j (j=0,1,…,n-1) are the polynomial coefficients, and η(t) is the residual part with small and bounded continuous time derivative, which can be ignored.
[0028] According to the Taylor expansion of the lumped disturbance, the 1st to nth order derivatives of the lumped disturbance are expressed as:
[0029]
[0030] Therefore, the expansion state model is expressed as:
[0031]
[0032] Furthermore, in step 3, based on the zero-order hold characteristic of the inverter in the actual digital control system, the extended state model is rewritten into a discrete form:
[0033]
[0034] Among them, T s is the sampling time, and k is the discrete sampling moment.
[0035] The extended state model with time-varying disturbances is expressed in discrete time extended state space as:
[0036]
[0037] Among them, x is the state vector, u is the input vector, ω is the iteration noise, υ is the measurement noise, A d 、 B d 、C d is the coefficient vector.
[0038] Furthermore, in step 4, the Kalman filter is designed for the extended state model with time-varying disturbances, and the gain parameter matrix is determined by simulation for system disturbance observation, including:
[0039] (1) State prediction:
[0040] x p (k+1)=A d x e (k)+B d u(k);
[0041] Among them, x p (k+1) is the state prediction value at time (k+1), x e (k) is the estimated value of the state at time k.
[0042] (2) Error covariance matrix prediction:
[0043]
[0044] Among them, P p (k) is the prediction error covariance matrix at time k, P e (k-1) is the estimated error covariance matrix at time (k-1), and Q is the iterative noise covariance matrix.
[0045] (3) Kalman filter gain calculation:
[0046]
[0047] Where R is the measurement noise covariance matrix.
[0048] (4) Update the estimated value based on the measurement results:
[0049] x e (k+1)=x p (k+1)+K(k)(y(k)-C d x p (k));
[0050] (5) Update of the estimated error covariance matrix:
[0051] P e (k)=(IK(k)C d )P p (k);
[0052] (6) Return to step (1).
[0053] Another object of the present invention is to provide a PMLSM time-varying system disturbance observation system applying the PMLSM time-varying system disturbance observation method, the PMLSM time-varying system disturbance observation system comprising:
[0054] The system lumped disturbance establishment module is used to divide the system disturbance into slow-varying terms and time-varying terms, and treat slow-varying terms such as motor parameter mismatch and unmodeled dynamics as system lumped disturbances;
[0055] A model building module is used to obtain the 1st to nth order derivatives of the lumped disturbance and model them as the system expansion state;
[0056] A time variable introduction module is used to introduce two time variables, measurement noise and iteration noise, into the system extended state model;
[0057] The Kalman filter design module is used to design a Kalman filter for the system extended state model with the time-varying disturbance term introduced, and to determine the gain parameter selection range through simulation.
[0058] Another object of the present invention is to provide a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the following steps:
[0059] A PMLSM system dynamics model including motor parameter mismatch and unmodeled dynamics is constructed, and two time-varying disturbances, measurement noise and iteration noise, are introduced. Relative to the linear motor control period, the disturbance caused by parameter disturbance and unmodeled dynamics is a slow variable, while the measurement noise and process noise are a time-varying variable. All slowly varying terms are regarded as the system lumped disturbance, and the n-th order derivative is expanded to model the system state. Time-varying terms are introduced in the expanded state model, and the system disturbance is observed by designing a Kalman filter.
[0060] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor performs the following steps:
[0061] A PMLSM system dynamics model including motor parameter mismatch and unmodeled dynamics is constructed, and two time-varying disturbances, measurement noise and iteration noise, are introduced. Relative to the linear motor control period, the disturbance caused by parameter disturbance and unmodeled dynamics is a slow variable, while the measurement noise and process noise are a time-varying variable. All slowly varying terms are regarded as the system lumped disturbance, and the n-th order derivative is expanded to model the system state. Time-varying terms are introduced in the expanded state model, and the system disturbance is observed by designing a Kalman filter.
[0062] Another object of the present invention is to provide an information data processing terminal, which is used to implement the PMLSM time-varying system disturbance observation system.
[0063] In combination with the above technical solutions and the technical problems solved, please analyze the advantages and positive effects of the technical solutions to be protected by the present invention from the following aspects:
[0064] First, in view of the technical problems existing in the above-mentioned prior art and the difficulty of solving these problems, this paper closely combines the technical solutions to be protected by the present invention and the results and data during the research and development process, and analyzes in detail and in depth how the technical solutions of the present invention solve the technical problems and some creative technical effects brought about by solving the problems. The specific description is as follows:
[0065] The present invention provides a PMLSM time-varying system disturbance observation technology based on an extended Kalman filter. The PMSLM drive system based on current predictive control (PCC) has the advantages of good dynamic performance and high steady-state accuracy. However, this method relies heavily on the accuracy of the motor model parameters. In practical applications, there will be deviations between the actual parameters and the nominal parameters. At the same time, there are two time-varying disturbance terms, measurement noise and iteration noise, in the predictive control iterative calculation process. The above system disturbances will affect the quality of current control. To this end, the present invention regards slowly varying disturbances such as parameter errors and unmodeled errors as system lumped disturbances, expands their 1st to nth order derivatives to perform extended state modeling for the system state, and introduces time-varying disturbance terms. For the system extended state model after the introduction of the time-varying disturbance terms, a Kalman filter is designed, and the gain parameter selection range is determined by simulation. Finally, the Kalman filter is used to observe system disturbances.
[0066] Second, considering the technical solution as a whole or from the perspective of the product, the technical effects and advantages of the technical solution to be protected by the present invention are described in detail as follows:
[0067] This paper designs a Kalman filter based on the system's extended state under time-varying disturbances, enabling real-time observation of perturbations in a PMLSM system. Experimental results show that conventional methods exhibit current tracking errors across the entire range of motion, while the proposed method completely eliminates these tracking errors.
[0068] Third, as auxiliary evidence for the inventiveness of the claims of the present invention, it is also reflected in the following important aspects:
[0069] The expected benefits and commercial value of the technical solution of the present invention after transformation are as follows: The present invention addresses the two time-varying disturbances, measurement noise and process noise, in the current prediction and control of permanent magnet synchronous linear motors by establishing an extended state model of the system, introducing the two time-varying disturbances into the model, and constructing a Kalman filter to observe the disturbances. After transformation, the technical solution can be used in multiple fields and industries such as CNC machine tools, semiconductor manufacturing, and logistics and transportation. Specifically, when applied to CNC machine tools, especially gantry machine tools, it can meet the requirements of high-speed movement and high-precision positioning in the x, y, and z directions; when applied to the semiconductor industry, it can meet the requirements of rapid positioning and movement of crystals and chips; similarly, when applied to the field of logistics and transportation, it can meet the requirements of rapid movement and classification of transportation lines. BRIEF DESCRIPTION OF THE DRAWINGS
[0070] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.
[0071] Figure 1 Flowchart of a PMLSM time-varying system disturbance observation method provided by an embodiment of the present invention;
[0072] Figure 2 Schematic diagram of the disturbance observation method for a PMLSM time-varying system provided by an embodiment of the present invention;
[0073] Figure 3 This is a structural block diagram of a PMLSM time-varying system disturbance observation system provided by an embodiment of the present invention;
[0074] FIG4( a ) is a schematic diagram of q-axis current and disturbance observation simulation results provided by an embodiment of the present invention;
[0075] FIG4( b ) is a schematic diagram of KF gain update during the simulation process provided by an embodiment of the present invention;
[0076] FIG5( a ) is a schematic diagram of the results of a single-parameter fixed ratio mismatch experiment provided by an embodiment of the present invention;
[0077] FIG5( b ) is a schematic diagram of the KF gain parameter adjustment process during the experiment provided by an embodiment of the present invention;
[0078] FIG6( a ) is a schematic diagram of a single parameter time-varying mismatch experiment result without considering time-varying disturbances provided by an embodiment of the present invention;
[0079] FIG6( b ) is a schematic diagram of the experimental results of a single parameter time-varying mismatch after adding a time-varying disturbance provided by an embodiment of the present invention;
[0080] FIG7( a ) is a schematic diagram of the experimental results of dual-parameter time-varying mismatch without considering time-varying disturbances provided by an embodiment of the present invention;
[0081] FIG7( b ) is a schematic diagram of the experimental results of dual-parameter time-varying mismatch after adding time-varying disturbances according to an embodiment of the present invention;
[0082] FIG8( a ) is a schematic diagram of the experimental results of three-parameter time-varying mismatch when time-varying disturbance is not considered, provided by an embodiment of the present invention;
[0083] FIG8( b ) is a schematic diagram of the experimental results of time-varying mismatch of three parameters after adding time-varying disturbances according to an embodiment of the present invention;
[0084] FIG9( a ) is a schematic diagram of experimental results of time-varying mismatch of three parameters without considering time-varying disturbances in the position mode provided by an embodiment of the present invention;
[0085] FIG9( b ) is a schematic diagram of the experimental results of time-varying mismatch of three parameters after adding time-varying disturbances according to an embodiment of the present invention;
[0086] In the figure: 1. System lumped disturbance establishment module; 2. Model construction module; 3. Time variable introduction module; 4. Kalman filter design module. DETAILED DESCRIPTION
[0087] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0088] In view of the problems existing in the prior art, the present invention provides a PMLSM time-varying system disturbance observation method, system and application. The present invention is described in detail below with reference to the accompanying drawings.
[0089] 1. Explanatory Examples In order to enable those skilled in the art to fully understand how to implement the present invention, this section provides an illustrative example that expands upon the technical solutions of the claims.
[0090] like Figure 1 As shown, the PMLSM time-varying system disturbance observation method provided by the embodiment of the present invention includes the following steps:
[0091] S101, the system disturbance is divided into slowly varying terms and time varying terms, and the slowly varying terms such as motor parameter mismatch and unmodeled dynamics are regarded as the system lumped disturbance;
[0092] S102, obtaining 1st to nth order derivatives of the lumped disturbance and modeling them as the system expansion state;
[0093] S103, introducing two time-varying variables, measurement noise and iteration noise, into the system extended state model;
[0094] S104: Design a Kalman filter for the system extended state model with the time-varying disturbance term introduced, and determine a gain parameter selection range through simulation.
[0095] The principle diagram of the PMLSM time-varying system disturbance observation method provided by the embodiment of the present invention is as follows: Figure 2 shown.
[0096] like Figure 3 As shown, the PMLSM time-varying system disturbance observation system provided by the embodiment of the present invention includes:
[0097] System lumped disturbance establishment module 1 is used to divide the system disturbance into slow-varying terms and time-varying terms, and regard slow-varying terms such as motor parameter mismatch and unmodeled dynamics as system lumped disturbance;
[0098] Model building module 2, used for obtaining the 1st to nth order derivatives of the lumped disturbance and modeling them as the system expansion state;
[0099] A time variable introducing module 3 is used to introduce two time variables, measurement noise and iteration noise, into the system extended state model;
[0100] The Kalman filter design module 4 is used to design a Kalman filter for the system extended state model with the time-varying disturbance term introduced, and to determine the gain parameter selection range through simulation.
[0101] Example 1
[0102] The embodiment of the present invention provides a time-varying system disturbance observation technology for a permanent magnet synchronous linear motor based on an extended Kalman filter, which is characterized by comprising:
[0103] The system disturbance is divided into slowly varying terms and time varying terms, and slowly varying terms such as motor parameter mismatch and unmodeled dynamics are regarded as the system lumped disturbance;
[0104] Obtain the 1st to nth order derivatives of the above lumped disturbance and model them as the system expansion state;
[0105] Two time variables, measurement noise and iteration noise, are introduced into the above system extended state model;
[0106] According to the system extended state model with the time-varying disturbance term introduced above, a Kalman filter is designed, and the gain parameter selection range is determined through simulation.
[0107] The system disturbance is:
[0108] D(k)=f(ΔR s ,ΔL s ,Δψ f ,ε,ω,υ)
[0109] Where, ΔR s is the winding resistance error, ΔL s is the stator inductance error, Δψ f is the permanent magnet flux error, ε is the unmodeled dynamic error, ω is the iteration noise, and υ is the measurement noise. Based on the servo system control period, the first four terms are motor parameter errors, which are slowly varying disturbances; the last two terms are noise errors, which are time-varying disturbances.
[0110] The PMLSM system model including the slowly varying disturbance term is:
[0111]
[0112] Among them, u d / u q and i d / i q are the d / q axis stator voltage and current, R s and L s are the winding resistance and stator inductance, ψ f is the permanent magnet flux, ω e =πv / τ is the electrical angular velocity, where v is the linear velocity of the rotor and τ is the pole pitch. The subscript "o" indicates the nominal value of the parameter, f d ,f q represents the lumped perturbation due to slowly varying perturbations such as parameter mismatch and unmodeled dynamics, and can be calculated as:
[0113]
[0114] Among them, R s =R so +ΔR s , L s =L so +ΔL s , ψ f =ψ fo +Δψ f , ε d / ε q are the d / q axis lumped unmodeled dynamics, respectively.
[0115] The lumped perturbation is expanded by Taylor as:
[0116]
[0117] Among them, n is the order, a j (j=0,1,…,n-1) are the polynomial coefficients, and η(t) is the residual part with small and bounded continuous time derivative, which can be ignored.
[0118] According to formula (3), the 1st to nth order derivatives of the lumped disturbance can be expressed as:
[0119]
[0120] Therefore, the expansion state model can be expressed as:
[0121]
[0122] Considering the zero-order hold characteristic of the inverter in the actual digital control system, the extended state model (5) can be rewritten into a discrete form:
[0123]
[0124] Among them, T s is the sampling time, and k is the discrete sampling moment.
[0125] The extended state model with time-varying disturbances can be further expressed in discrete time extended state space as follows:
[0126]
[0127] Among them, x is the state vector, u is the input vector, ω is the iteration noise, υ is the measurement noise, A d 、 B d 、C d is the coefficient vector.
[0128] The method designs a Kalman filter for an extended state model with time-varying disturbances, and determines a gain parameter matrix through simulation for system disturbance observation, including:
[0129] 1) State prediction
[0130] x p (k+1)=A d x e (k)+B d u(k) (8)
[0131] Among them, x p (k+1) is the state prediction value at time (k+1), x e (k) is the state estimate at time k derived previously.
[0132] 2) Error covariance matrix prediction
[0133]
[0134] Among them, P p (k) is the prediction error covariance matrix at time k, P e (k-1) is the estimated error covariance matrix at time (k-1), and Q is the iterative noise covariance matrix.
[0135] 3) Kalman filter gain calculation
[0136]
[0137] Where R is the measurement noise covariance matrix.
[0138] 4) Update the estimated value based on the measurement results
[0139] x e (k+1)=x p(k+1)+K(k)(y(k)-C d x p (k)) (11)
[0140] 5) Estimation error covariance matrix update
[0141] P e (k)=(IK(k)C d )P p (k) (12)
[0142] 6) Return to 1).
[0143] Example 2
[0144] In practical control systems, accurate motor parameters are difficult to obtain due to the motor's need to operate in diverse environments and operating conditions. Furthermore, the speed and current values obtained through sensors or observation algorithms inevitably contain measurement noise, and the predictive control process also introduces iteration noise during its iterative calculations. These disturbances can severely impact system performance and stability.
[0145] Current research on current predictive control (CPC) primarily focuses on addressing stability issues caused by parameter mismatch, while relatively little research has focused on addressing the noise generated during the calculation and measurement processes. While the disturbance caused by parameter mismatch is a slowly varying term relative to the servo system's control cycle, the disturbance caused by this noise is a time-varying term, reflected in every current calculation cycle. Therefore, CPC also requires compensation for this time-varying disturbance.
[0146] Based on this, the embodiment of the present invention proposes an extended Kalman filter to observe the system lumped disturbance containing time-varying terms, such as Figure 2 As shown in the figure, the method includes: S1, constructing a PMLSM electrical model, treating slowly varying terms such as motor parameter mismatch and unmodeled dynamics as system lumped disturbances; S2, expanding the 1st to nth order derivatives of the lumped disturbance into system states, and performing expanded state modeling; S3, introducing time-varying terms into the expanded state model; S4, designing a Kalman filter, completing parameter tuning through simulation research, and observing system disturbances.
[0147] Among them, S1. The process of treating slowly varying terms such as motor parameter mismatch and unmodeled dynamics as the system lumped disturbance is as follows. The PMLSM electrical model can be expressed as:
[0148]
[0149] Among them, u d / u q and i d / i qare the d / q axis stator voltage and current respectively. s and L s are the winding resistance and stator inductance, ψ f is the permanent magnet flux. e =πv / τ is the electrical angular velocity, where v is the linear velocity of the rotor and τ is the pole pitch. The subscript "o" indicates the nominal value of the parameter, f d ,f q It represents the total disturbance of the system, which can be divided into two categories: slow-varying disturbance and time-varying disturbance. It can be calculated by the following formula:
[0150] F(k)=f(ΔR s ,ΔL s ,Δψ f ,ε,ω,υ) (14)
[0151] Where, ΔR s is the winding resistance error, ΔL s is the stator inductance error, Δψ f is the permanent magnet flux error, ε is the unmodeled dynamic error, ω is the iteration noise, and υ is the measurement noise. Based on the servo system control period, the first four terms are motor parameter errors, which are slowly varying disturbances; the last two terms are noise errors, which are time-varying disturbances.
[0152] For slowly varying disturbances, it can be calculated using the following formula:
[0153]
[0154] Among them, R s =R so +ΔR s , L s =L so +ΔL s , ψ f =ψ fo +Δψ f , ε d / ε q are the d / q axis lumped unmodeled dynamics, respectively.
[0155] S2. The process of expanding state modeling is as follows: all the above slowly varying disturbances are regarded as the system lumped disturbances. According to the Taylor expansion principle, any function can be expanded into the sum of polynomials and high-order functions, that is:
[0156]
[0157] Among them, n is the order, a j (j=0,1,…,n-1) are the polynomial coefficients, η(t) is the residual part with small and bounded continuous time derivative, and has the following characteristics:
[0158]
[0159] That is, the basic requirement for the disturbance is that the low-frequency part can be simply modeled as a finite-order polynomial, while the high-frequency part of the nonlinear disturbance can be ignored.
[0160] Therefore, the 1st to nth order derivatives of the above lumped disturbance can be expressed as:
[0161]
[0162] Therefore, the expansion state model can be expressed as:
[0163]
[0164] Considering the zero-order hold characteristic of the inverter in the actual digital control system, the extended state model (19) can be rewritten in discrete form:
[0165]
[0166] Among them, T s is the sampling time, and k is the discrete sampling moment.
[0167] S3. Introduce the time-varying disturbance term into the above discrete extended state model and write it in matrix form:
[0168]
[0169] Among them, the state vector Control input vector u=[u d u q -ω e ψ f0 ] T ; Iterative noise ω=[ω1 ω2 … ω 2n+2 ] T , its covariance matrix Q=E{ω(k)ω(k) T}>0;Measurement noise υ=[υ d υ q ] T Its covariance matrix R=E{υ(k)υ(k) T}≥0; measurement output matrix y(k)=[i d i q ] T The parameter matrices are defined as follows:
[0170]
[0171]
[0172]
[0173] Assuming that the speed dynamics are significantly slower than the current dynamics, which is generally reasonable, the matrix A between two consecutive current control cycles can be written as d As a constant, at time k there is A d (k+1)=A d (k), Therefore, the extended state model (21) can be regarded as a linear uncertain dynamic system, and then the Kalman filter is used to optimally estimate the system state and disturbance.
[0174] S4. The main process of designing the Kalman filter is as follows:
[0175] 1) State prediction
[0176] x p (k+1)=A d x e (k)+B d u(k) (22)
[0177] Among them, x p (k+1) is the state prediction value at time (k+1), x e (k) is the state estimate at time k derived previously.
[0178] 2) Error covariance matrix prediction
[0179]
[0180] Among them, P p (k) is the prediction error covariance matrix at time k, P e (k-1) is the estimated error covariance matrix at time (k-1), and Q is the iterative noise covariance matrix.
[0181] 3) Kalman filter gain calculation
[0182]
[0183] Where R is the measurement noise covariance matrix.
[0184] 4) Update the estimated value based on the measurement results
[0185] x e (k+1)=x p (k+1)+K(k)(y(k)-C d x p (k)) (25)
[0186] 5) Estimation error covariance matrix update
[0187] Pe (k)=(IK(k)C d )P p (k) (26)
[0188] 6) Return to 1).
[0189] It is worth noting that the design of the Kalman filter can be completed through equations (22) to (26), where the key point is the covariance matrix P e (0),Q,R tuning and initial state estimation x e The selection of (0) will directly affect the dynamic performance and convergence speed of KF. e The initial value of (0) primarily affects the transient amplitude of the state estimate, while the transient time and steady-state process are unaffected. Matrix Q represents the level of iterative noise and model uncertainty. Increasing the elements of matrix Q will increase the KF gain, thereby accelerating the filter dynamics. However, a large matrix Q will enhance the effect of measurement updates, thereby amplifying the side effects of noise. Matrix R represents the characteristics of measurement noise. Large values in matrix R mean low confidence in the measurement, which in turn reduces the dynamic performance of the filter. In practical applications, preliminary tuning can be performed based on test results, for example, directly using sampled current for calculation.
[0190] 2. Application Examples: In order to demonstrate the creativity and technical value of the technical solution of the present invention, this section provides application examples of the claimed technical solution on specific products or related technologies.
[0191] To meet the high-speed, high-precision control requirements of linear motors, current predictive control (CPC) offers superior steady-state and dynamic performance compared to traditional hysteresis control and PID control strategies. However, CPC relies on the motor model, which means it is highly dependent on the accuracy of motor parameters. While numerous solutions exist for this problem and have been widely applied in engineering, less research has been conducted on time-varying disturbances such as measurement noise and process noise in the iterative calculation of predicted current, compared to slowly varying system disturbances such as changes in motor parameters.
[0192] In specific technical applications, measurement noise includes position information measured by the scale, velocity information derived from the differential calculation of position information, and q-axis current / voltage information sampled by the driver. Process noise includes current and voltage calculations during prediction model calculations, and gain parameters and matrix update calculations during Kalman filter calculations.
[0193] In specific product applications, for the drive system, the current prediction control strategy does not change the hardware structure of the motor drive, but only replaces the innermost loop in the traditional cascade PI control: the current loop in the control algorithm.
[0194] For CNC, semiconductor manufacturing, logistics and transportation, and other fields where linear motors are most widely used, using current prediction control strategies to control permanent magnet synchronous linear motors can achieve better q-axis current steady-state and dynamic performance, thereby enabling the equipment to have higher response speed and high-precision positioning accuracy, thereby shortening the time of each work cycle and greatly improving processing, production and transportation efficiency.
[0195] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware portion can be implemented using dedicated logic; the software portion can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated hardware. Those skilled in the art will appreciate that the above-mentioned devices and methods can be implemented using computer-executable instructions and / or contained in processor control code, for example, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, or can be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.
[0196] 3. Evidence of the effects of the embodiments: The embodiments of the present invention have achieved some positive effects during the development or use process, and indeed have great advantages over the existing technology. The following content describes them with reference to the data, charts, etc. of the experimental process.
[0197] To verify the effectiveness of the proposed method, simulation and experimental tests were conducted. Under nominal conditions, the main parameters used in the simulation are the same as those of the prototype motor, as shown in Table 1.
[0198] Table 1 Main parameters of PMLSM
[0199]
[0200] The simulation conditions are as follows:
[0201] 1) Parameter perturbation: R s =2R so , L s =1.5L so , ψ f =1.5ψ fo ;
[0202] 2) And the q-axis current command is a square wave;
[0203] 3)x e (0)=0,P e (0)=0, R=diag(10,10), Q=diag(1,1,5000,5000).
[0204] Figure 4(a) shows the simulation results of the q-axis current response and disturbance observation. Figure 4(b) illustrates the Kalman filter gain update process during the calculation. The simulation results demonstrate that the proposed method can effectively observe q-axis disturbances when all three parameters are mismatched, ensuring that the q-axis current quickly and accurately tracks the command current, with the corresponding KF gain converging within 5ms.
[0205] Furthermore, experiments were conducted on the PMLSM for single-parameter fixed-ratio mismatch, single-parameter time-varying mismatch, dual-parameter time-varying mismatch, and triple-parameter time-varying mismatch. Figure 5(a) shows the results of the single-parameter fixed-ratio mismatch experiment, and Figure 5(b) shows the KF gain parameter adjustment process during the experiment. As can be seen from the figure, the static error of the q-axis current response is fully compensated, the q-axis lumped perturbation and current are accurately estimated, and the KF gain parameter converges within 2ms. Figure 6(a) shows the results of the single-parameter time-varying mismatch experiment without considering the time-varying perturbation, and Figure 6(b) shows the results of the single-parameter time-varying mismatch experiment after adding the time-varying perturbation. The experimental results show that the proposed method eliminates the influence of the linear change of resistance on the current response and ensures fast and accurate convergence. Similarly, Figure 7(a) shows the experimental results of dual-parameter time-varying mismatch when time-varying disturbance is not considered, and Figure 7(b) shows the experimental results of dual-parameter time-varying mismatch after adding time-varying disturbance; Figure 8(a) shows the experimental results of three-parameter time-varying mismatch when time-varying disturbance is not considered, and Figure 8(b) shows the experimental results of three-parameter time-varying mismatch after adding time-varying disturbance; Figure 9 shows the experimental results under position mode, among which Figure 9(a) shows the experimental results of three-parameter time-varying mismatch when time-varying disturbance is not considered, and Figure 9(b) shows the experimental results of three-parameter time-varying mismatch after adding time-varying disturbance. It can be seen from the experimental results that the traditional method has current tracking error in the entire motion range, and the method proposed in the present invention completely eliminates this tracking error.
[0206] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with this technical field within the technical scope disclosed by the present invention and within the spirit and principles of the present invention should be covered by the scope of protection of the present invention.
Claims
1. A PMLSM time-varying system disturbance observation method, characterized in that: The PMLSM time-varying system disturbance observation method comprises: A PMLSM system dynamics model including motor parameter mismatch and unmodeled dynamics is constructed, and two time-varying disturbances, measurement noise and iteration noise, are introduced. Relative to the linear motor control period, the disturbance caused by parameter disturbance and unmodeled dynamics are slow-varying variables, and the measurement noise and process noise are time-varying variables. All slow-varying terms are regarded as the system lumped disturbance, and the n-th order derivative is expanded to model the system state. Time-varying terms are introduced in the expanded state model, and the system disturbance is observed by designing a Kalman filter. The PMLSM time-varying system disturbance observation method comprises the following steps: Step 1: The system disturbance is divided into slow-varying terms and time-varying terms, and the slow-varying terms such as motor parameter mismatch and unmodeled dynamics are regarded as the system lumped disturbance; Step 2, obtaining the 1st to nth order derivatives of the lumped disturbance and modeling them as the system expansion state; Step 3, introduce two time variables, measurement noise and iteration noise, into the system expansion state model; Step 4: Design a Kalman filter for the system extended state model with time-varying disturbance terms, and determine the gain parameter selection range through simulation; The system disturbance in step 1 is: D(k)=f(ΔR s ,ΔL s ,Dp f ,e,oh,u); Among them, ΔR s is the winding resistance error, ΔL s is the stator inductance error, Δψ f is the permanent magnet flux error, ε is the unmodeled dynamic error, ω is the iteration noise, and υ is the measurement noise. Taking the servo system control cycle as the benchmark, the first four items are motor parameter errors, which are slowly varying disturbances; the last two items are noise errors, which are time-varying disturbances. The PMLSM system model including the slowly varying disturbance term is: Among them, u d / u q and i d / i q are the d / q axis stator voltage and current, R s and L s are the winding resistance and stator inductance, ψ f is the permanent magnet flux, ω e =πv / τ is the electrical angular velocity, where v is the linear velocity of the rotor and τ is the pole pitch; the subscript "o" indicates the nominal value of the parameter, f d ,f q represents the lumped perturbation due to slowly varying perturbations such as parameter mismatch and unmodeled dynamics and is calculated as: Among them, R s =R so +ΔR s , L s =L so +ΔL s , ψ f =ψ fo +Δψ f , ε d / ε q They are the d / q axis lumped unmodeled dynamics; In the step 2, the lumped disturbance is expanded by Taylor as follows: Among them, n is the order, a j (j=0,1,…,n-1) are the polynomial coefficients, η(t) is the residual part with small and bounded continuous time derivative, which is ignored; According to the Taylor expansion of the lumped disturbance, the 1st to nth order derivatives of the lumped disturbance are expressed as: Therefore, the expansion state model is expressed as: In step 3, based on the zero-order holding characteristic of the inverter in the actual digital control system, the extended state model is rewritten into a discrete form: Among them, T s is the sampling time, k is the discrete sampling moment; The extended state model with time-varying disturbance is expressed in discrete time extended state space as: Where x is the state vector, u is the input vector, ω is the iteration noise, υ is the measurement noise, A d , B d , C d is the coefficient vector.
2. The PMLSM time-varying system disturbance observation method according to claim 1, characterized in that: In the step 4, a Kalman filter is designed for the extended state model with time-varying disturbances, and a gain parameter matrix is determined by simulation for system disturbance observation, including: (1) Status prediction: x p (k+1)=A d x e (k)+B d u(k); Among them, x p (k+1) is the predicted value of the state at time (k+1), x e (k) is the estimated value of the state at time k; (2) Error covariance matrix prediction: Among them, P p (k) is the prediction error covariance matrix at time k, P e (k-1) is the estimated error covariance matrix at time (k-1), Q is the iterative noise covariance matrix; (3) Kalman filter gain calculation: Where R is the measurement noise covariance matrix; (4) Update the estimated value based on the measurement results: x e (k+1)=x p (k+1)+K(k)(y(k)-C d x p (k)); (5) Update of the estimated error covariance matrix: P e (k)=(I-K(k)C d )P p (k); (6) Return to step (1).
3. A PMLSM time-varying system disturbance observation system using the PMLSM time-varying system disturbance observation method according to any one of claims 1 to 2, characterized in that: The PMLSM time-varying system disturbance observation system comprises: The system lumped disturbance establishment module is used to divide the system disturbance into slow-varying terms and time-varying terms, and regard the slow-varying terms such as motor parameter mismatch and unmodeled dynamics as the system lumped disturbance; A model building module, used for obtaining the 1st to nth order derivatives of the lumped disturbance and modeling them as the system expansion state; A time variable introduction module is used to introduce two time variables, measurement noise and iteration noise, into the system extended state model; The Kalman filter design module is used to design a Kalman filter for the system extended state model with the time-varying disturbance term introduced, and to determine the gain parameter selection range through simulation.
4. A computer device, characterized in that: The computer device includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the PMLSM time-varying system disturbance observation method according to any one of claims 1 to 2.
5. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the steps of the PMLSM time-varying system disturbance observation method according to any one of claims 1 to 2.
6. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the PMLSM time-varying system disturbance observation system as described in claim 3.
Citation Information
Patent Citations
Permanent magnet synchronous motor adaptive sliding mode control method based on dynamic surface
CN108390606A
Inertia identification method of permanent magnet synchronous motor driving system
CN113708686A
Permanent magnet synchronous linear motor control method and system
CN114070145A
An adaptive sliding-mode control method of the permanent magnet synchronous motor based on dynamic surface technology
GB2572458A
