Permanent magnet synchronous motor multi-parameter identification method based on harmonic coupling model

Abstract

The invention discloses a permanent magnet synchronous motor multi-parameter identification method based on a harmonic coupling model, and belongs to the technical field of permanent magnet synchronous motor parameter identification, and the method comprises the steps: decomposing a voltage equation of a permanent magnet synchronous motor harmonic coupling model into a fundamental wave voltage equation and a harmonic wave voltage equation; constructing two extended Kalman filter observers according to the fundamental wave voltage equation and the harmonic wave voltage equation, and distributing six key motor parameters under the condition of considering magnetic field saturation; on the basis of considering controller delay and inverter nonlinearity, the influence of motor parameter harmonic waves on fundamental waves and harmonic voltage is further considered, and real fundamental wave and harmonic voltage values are extracted from the output quantity of a current controller; and constructing a single interconnection type extended Kalman filtering parameter identification system according to the parameters identified by the fundamental wave and harmonic wave extended Kalman filtering observers. According to the method, the under-rank problem of parameter identification of the permanent magnet synchronous motor can be solved, and multiple motor parameters can be accurately identified.

Description

Technical Field

[0001] The present invention relates to the technical field of permanent magnet synchronous motor parameter identification, and in particular to a permanent magnet synchronous motor multi-parameter identification method based on a harmonic coupling model. Background Art

[0002] Since the 21st century, with the development of industrialization and urbanization, energy shortage and environmental pollution have become two major problems that need to be solved urgently worldwide. Traditional fuel vehicles not only consume a large amount of oil resources, but also pollute the environment with automobile exhaust (see reference [1]). The development of new energy vehicles can reduce the consumption of fossil resources and help solve the problems of energy shortage and environmental pollution. Permanent magnet synchronous motors have high efficiency (see reference [2]) and high power density (see reference [3]), and are widely used in new energy vehicles. The performance of permanent magnet synchronous motors is mainly affected by motor materials, structures and motor control methods. At present, due to the development of control theory, many high-performance control algorithms have been born. Control algorithms often rely on the accuracy of motor equations and motor parameters. Motor parameters are often affected by factors such as temperature and magnetic field saturation, and will change to varying degrees under different working conditions and environments. Therefore, online identification of permanent magnet synchronous motor parameters is of great significance.

[0003] Online parameter identification refers to the use of appropriate algorithms to estimate the parameters of a specific analysis model based on variables such as voltage, current and speed collected during motor operation. PMSM (permanent magnet synchronous motor) online parameter identification algorithms can be divided into two categories: one is numerical algorithms represented by recursive least squares (RLS), extended Kalman filter (EKF) and model reference adaptive method (MRAS); the other is various emerging intelligent algorithms in recent years, such as artificial neural network (ANN) and particle swarm optimization (PSO). Recursive least squares method and model reference adaptive method have simple structures and are easy to implement, but have poor anti-interference ability and are easily affected by noise (see reference [4]). Intelligent algorithms such as particle swarm optimization have good identification effects, but are relatively complex to implement and have high hardware performance requirements. They are currently difficult to apply in practice. The extended Kalman filter method takes system noise and measurement noise into account in its model, has good anti-interference ability and high accuracy (see reference [5]).

[0004] Since there are only two voltage equations for permanent magnet synchronous motors in the dq coordinate system, there will be a lack of rank problem when performing parameter identification on the PMSM. One way to solve the lack of rank problem is to reduce the number of parameters to be identified. Reference [6] considers the magnetic flux as a fixed constant and only identifies resistance and inductance. Reference [7] studies the law of stator resistance changing with temperature. The temperature of the current stator winding is measured by a temperature sensor to obtain the corresponding stator resistance value, which is then used as a known quantity in the voltage equation to identify other parameters. If all parameters are to be identified, the number of identification equations needs to be increased. The d-axis current injection method (see references [8] and [9]) is a commonly used method. By injecting a short pulse current i d Two sets of steady-state dq-axis voltage equations are obtained, doubling the number of equations and solving the lack-rank problem. However, this method of injecting short pulse current may affect the normal working state of the motor, and what is actually obtained are two steady-state equations under different conditions.

[0005] Online parameter identification of PMSM often requires real-time acquisition of voltage and current. Current can be measured directly by current sensors, while voltage is usually a pulse width modulated (PWM) wave, which is difficult to obtain directly by sampling. Therefore, we often use the output voltage of the current controller instead. The delay characteristics in the controller and the nonlinearity of the inverter will cause an error between the controller output voltage and the actual voltage (see references

[10] and

[11] ), which affects the accuracy of parameter identification. Reference

[12] uses the disturbance voltage caused by the nonlinearity of the voltage source inverter (VSI) as an identification parameter. Reference

[13] estimates the rotor permanent magnet temperature by extracting the first-order slot harmonics on the q-axis. Since the nonlinearity of the inverter mainly produces 6 th Voltage harmonics and therefore have less impact on the permanent magnet temperature estimation.

[0006] In view of the existing problems in parameter identification, the present invention aims to construct a multi-parameter identification system for permanent magnet synchronous motors based on the harmonic coupling model (see reference

[14] ) to solve the problem of lack of rank in multi-parameter identification.

[0007] References:

[0008] [1] Wang Qinghe. Automobile emission hazards, influencing factors and improvement measures[J]. Modern Industrial Economy and Informatization, 2020, 10(9): 48-49.

[0009] [2]F.Betin,G.Capolino,D.Casadei,et al.,“Trends in electrical machinescontrol:Samples for classical,sensorless,and fault-tolerant techniques,”IEEEInd.Electron.Mag.,vol.8,no.2,pp.43-55,Jun.2014,doi:10.1109 / MIE.2014.2313752.

[0010] [3]Y.Song,J.Lu,Y.Hu,et al.,“Expanding limit of minimum sampling timeusing auxiliary vectors for PMSM drives with single DC-link current sensor,”IEEE Trans.Ind.Electron.,vol.70,no.4,pp.3437-3448,Apr.2023,doi:10.1109 / TIE.2022.3174279.

[0011] [4]Li X,Kennel R.Comparison of state-of-the-art estimators forelectrical parameter identification of PMSM[C].2019IEEE InternationalSymposium on Predictive Control of Electrical Drives and Power Electronics(PRECEDE),Quanzhou,China,2019:1-6.

[0012] [5]Zerdali E.Adaptive extended Kalman filter for speed-sensorlesscontrol of induction motors[J].IEEE Transactions on Energy Conversion,2018,34(2):789-800.

[0013] [6]T.Boileau,N.Leboeuf,B.Nahid-Mobarakeh and F.Meibody-Tabar,“Onlineidentification of PMSM parameters:Parameter identifiability and estimatorcomparative study,”IEEE Trans.Ind.Appl.,vol.47,no.4,pp.1944-1957,Aug.2011,doi:10.1109 / TIA.2011.2155010.

[0014] [7]K.Liu and Z.Zhu,“Online estimation of the rotor flux linkage andvoltage-source inverter nonlinearity in permanent magnet synchronous machinedrives,”IEEE Trans.Power Electro.,vol.29,no.1,pp.418-427,Jan.2014,doi:10.1109 / TPEL.2013.2252024.

[0015] [8]Z.Liu,H.Wei,Q.Zhong,et al.,“GPU implementation of DPSO-REalgorithm for parameters identification of surface PMSM considering VSInonlinearity,”IEEE J.Emerg.Sel.Topics Power Electron.,vol.5,no.3,pp.1334-1345,Sep.2017,doi:10.1109 / JESTPE.2017.2690688.

[0016] [9]Y.Zhang,M.Zhou,C.Zhang,et al.,“Identification of PMSM parameterswith time-error compensated based on contractile factor antipredator PSO,”IEEE Trans.Transp.Electrif.,vol.10,no.2,pp.4006-4017,Jun.2024,doi:10.1109 / TTE.2023.3306872.

[0017]

[10] S.H.Hwang and J.M.Kim,“Dead time compensation method for voltage-fed PWM inverter,”IEEE Trans.Energy Convers.,vol.25,no.1,pp.1-10,Mar.2010,doi:10.1109 / TEC.2009.2031811.

[0018]

[11] Y.Liao,J.Yao and S.Yang,“Analysis and elimination method ofharmonics produced by forward voltage drop of ACEG excitation power source,”Proceedings of the Csee.,vol.24,no.4,pp.151-156,2004,doi:10.13334 / j.0258-8013.pcsee.2004.04.028.

[0019]

[12] Y.Wang,Y.Xu and J.Zou,“Online multiparameter identificationmethod for sensorless control of SPMSM,”IEEE Trans.Power Electron.,vol.35,no.10,pp.10601-10613,Oct.2020,doi:10.1109 / TPEL.2020.2974870.

[0020]

[13] B.Wen, K.Liu, Z.Zhu and R.Ding, “Estimation of magnet temperature for integer-slot permanent magnet synchronous machines via extraction of first-order slot harmonic in back EMF,” IEEE Trans.Transp.Electrif., vol.10, no.3, pp.7203-7213, Sep.2024, doi:10.1109 / TTE.2023.3333570.

[0021]

[14] P.Yi, Z.Sun, and Summary of the invention

[0022] The technical problem to be solved by the present invention is to provide a multi-parameter identification method for a permanent magnet synchronous motor based on a harmonic coupling model. On the one hand, the present method addresses the problem of under-rank in current parameter identification, and utilizes a harmonic coupling model to construct a multi-parameter identification system based on a fundamental / harmonic extended Kalman filter. The system adopts a single interconnection structure to avoid iterative operations and reduce the amount of calculation. On the other hand, the present method quantitatively evaluates the influence of controller delay, inverter nonlinearity, and motor parameter harmonics on parameter identification, establishes a voltage compensation formula, and can extract the actual voltage from the output voltage of the current controller, thereby greatly improving the accuracy of parameter identification.

[0023] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0024] A multi-parameter identification method for a permanent magnet synchronous motor based on a harmonic coupling model, comprising:

[0025] Step (a): by separating the voltage equation of the permanent magnet synchronous motor harmonic coupling model into two equations: a fundamental voltage equation and a harmonic voltage equation, wherein the required permanent magnet synchronous motor inductance harmonic parameters are obtained by a finite element analysis method;

[0026] Step (b): construct two extended Kalman filter observers based on the fundamental voltage equation and the harmonic voltage equation to allocate the six key motor parameters considering the magnetic field saturation;

[0027] Step (c): On the basis of considering the controller delay and the nonlinearity of the inverter, the influence of the motor parameter harmonics on the fundamental and harmonic voltages is further considered, and the real fundamental and harmonic voltage values ​​are extracted from the output of the current controller by quantitatively establishing the compensation voltage equation;

[0028] Step (d): Based on the parameters identified by the fundamental and harmonic extended Kalman filter observers, a single interconnected extended Kalman filter parameter identification system is constructed.

[0029] The present invention has the following beneficial effects:

[0030] (1) The harmonic coupling model can be used to solve the lack of rank problem in multi-parameter identification;

[0031] (2) The extended Kalman filter algorithm takes into account system noise and measurement noise, has good anti-interference performance and high accuracy. The single interconnected structure can avoid the iterative convergence of the parameters identified by the two extended Kalman filters in a single step, reducing the amount of calculation and accelerating the convergence speed.

[0032] (3) The influence of controller delay, inverter nonlinearity, and motor parameter harmonics on parameter identification is taken into account to improve the parameter identification accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a structural schematic diagram of a single interconnected extended Kalman filter in the present invention;

[0034] Figure 2 It is a schematic diagram of the framework of the multi-parameter identification system based on the harmonic coupling model;

[0035] Figure 3 It is a schematic diagram of the harmonic injection structure framework based on multi-reference frame PI control;

[0036] Figure 4 It is a schematic diagram of the electromagnetic structure of the permanent magnet synchronous motor under test;

[0037] Figure 5(a) shows the parameter l t0 Identification waveform of

[0038] Figure 5(b) shows the parameter l c2 Identification waveform of

[0039] Figure 5(c) shows the parameter l s2 Identification waveform of

[0040] Figure 5(d) is the identification waveform of parameter R;

[0041] Figure 5(e) shows the parameter ψ d1 Identification waveform of

[0042] Figure 5(f) shows the parameter ψ q1 Identification waveform of

[0043] Figure 6 It is a flow chart of the method for multi-parameter identification of a permanent magnet synchronous motor based on a harmonic coupling model of the present invention. DETAILED DESCRIPTION

[0044] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0045] The parameter accuracy of a permanent magnet synchronous motor directly affects the performance of motor control. The present invention aims to solve the under-rank problem in parameter identification based on a harmonic coupling model of a permanent magnet synchronous motor and realize the identification of six key parameters.

[0046] The present invention provides a method for multi-parameter identification of a permanent magnet synchronous motor based on a harmonic coupling model. Figure 6 As shown, including:

[0047] Step (a): by separating the voltage equation of the permanent magnet synchronous motor harmonic coupling model into two equations: a fundamental voltage equation and a harmonic voltage equation, wherein the required harmonic parameters of the permanent magnet synchronous motor inductance can be obtained by a finite element analysis method;

[0048] In this step, the voltage equation in the harmonic coupling model of the permanent magnet synchronous motor is used as the basis. th and 13 th After the current harmonics are injected, six voltage-current frequency coupling equations can be obtained, thereby solving the under-rank problem of multi-parameter identification; that is, by decomposing the voltage equation of the permanent magnet synchronous motor harmonic coupling model into two equations: the fundamental voltage equation and the harmonic voltage equation, where the required permanent magnet synchronous motor inductance harmonic parameters can be obtained through the finite element analysis method.

[0049] In the specific implementation, according to reference

[14] , the harmonic coupling model of permanent magnet synchronous motor studies the frequency coupling relationship between flux-current, voltage-current and torque-current. The voltage-current frequency coupling equation in the harmonic coupling model is used in the present invention. st and (f±1) th (f∈6N + )In the dq coordinate system, the frequency coupling relationship between voltage and current is shown in the following equation.

[0050]

[0051] Among them, u dqn ,i dqn and ψ dqn n th Voltage, current and permanent magnet flux in dq coordinate system; ω e is the electrical angular velocity; R = diag(R, R, R, R, R, R, R), R is the phase resistance; γ = [0 -1; 1 0], Γ = diag(y, (f-1)γ, (f+1)γ); is the coupled inductor matrix, L0, L2 - 、L2 + , L f-2 + , L f - , L f + and L f+2 - The specific form of can be found in reference

[14] , τ = diag(1, -1), It only depends on the amplitude and phase of the current fundamental wave and has nothing to do with the rotor position.

[0052] The present invention performs parameter identification according to formula (I) and selects injection 11 th and 13 th Current harmonics, that is, f = 12 in formula (I), the coupled inductance matrix can be expressed as:

[0053] The specific forms of the elements in the matrix can be referred to in reference

[14] , which are related to the self-inductance and mutual inductance of the motor. Preferably, formula (I) is divided into a fundamental wave part and a harmonic wave part, which are expressed as formula (II) and formula (III) respectively.

[0054]

[0055] Among them, Γ1=γ; Γ2=diag(11y, 13γ); R1=diag(R, R); R2=diag(R, R, R, R); u dq1113 =[u dq11 T u dq13 T ] T ;i dq1113 =[i dq11 T i dq13 T ] T ψ dq1113 =[ψ dq11 T ψ dq13 T ]T The high-order components of the inductance in formula (III) are and the higher-order components of the permanent magnet flux ψ dq1113 It can be obtained through finite element analysis (FEA) and is used as a known quantity in the parameter identification process of the present invention.

[0056] It can be seen that formula (II) contains two equations, and formula (III) contains four equations. When the parameters to be identified are phase resistance R, permanent magnet flux fundamental component (ψ d1 , ψ q1 ), the DC component and the second-order component of the inductance (l t0 , l c2 , l s2 ), there is no lack of rank problem when using the harmonic coupling model for multi-parameter identification.

[0057] Step (b): construct two extended Kalman filter observers based on the fundamental voltage equation and the harmonic voltage equation to allocate the six key motor parameters considering the magnetic field saturation;

[0058] In this step, the extended Kalman filter algorithm is selected as the estimation algorithm, which has good anti-interference performance, high accuracy and simple implementation. The six key parameters of the permanent magnet synchronous motor after magnetic field saturation (inductance parameter l t0 , l c2 , l s2 , resistance parameter R, permanent magnet flux ψ d1 , q1 ) is allocated, where the fundamental wave model identification parameters (ψ d1 , q1 ), harmonic model identification parameters (l t0 , l c2 , l s2 , R).

[0059] In specific implementation, the present invention realizes parameter identification by constructing two extended Kalman filter observers. According to formula (II) and formula (III), the fundamental wave identification model ∑1 and the harmonic wave identification model ∑2 can be obtained, as shown in formula (IV) and formula (V).

[0060]

[0061] Among them, x1 and x2 are state variables; y1 and y2 are output variables; u1 and u2 are input variables; f1(g) and f2(g) are state equations; h1(g) and h2(g) are output equations. The specific form is as follows:

[0062] x1=[i d1 i q1 ψ d1 ψ q1 ]T , x2=[i d11 i q11 i d13 i q13 l t0 l c2 l s2 R] T ;

[0063] y1=[i d1 i q1 ] T , y2=[i d11 i q11 i d13 i q13 ] T ;

[0064] u1=u dq1 =[u d1 u q1 ] T , u2=u dq1113 =[u d11 u q11 u d13 u q13 ] T ;

[0065]

[0066] It can be seen from equations (IV) and (V) that the fundamental wave identification model contains all six parameters to be identified, and the harmonic identification model contains l t0 , l c2 , l s2 In order to reduce the amount of calculation, the harmonic identification model is used to construct the extended Kalman filter observer EKF2 identification l t0 , l c2 , l s2 , R, using the fundamental wave coupling model to construct the extended Kalman filter observer EKF1 to identify ψ d1 and ψ q1 .

[0067] The state equations f1(g) and f2(g) are discretized using the first-order forward Euler method to obtain equations (VI) and (VII).

[0068]

[0069] in, and is the predicted value of the state variable at step k, and is the optimal estimate of the state variable at the k-1th step, and is the discrete expression of the state equation. The optimal estimate of the state variable at the kth step can be expressed by equation (VIII) and equation (IX).

[0070]

[0071] Among them, y 1，k and 2，k is the measurement value at step k, K k is the Kalman gain matrix of the kth step. The constructed single interconnected extended Kalman filter structure is as follows Figure 1 shown.

[0072] For EKF1 and EKF2, the Kalman filter gain matrix K k The calculation process is similar to that of EKF2, which is described here as follows: The predicted value of the error covariance matrix at step k is It can be deduced from formula (X).

[0073]

[0074] Among them, F 2，k-1 It is the k-1th step The Jacobian matrix of is the optimal estimate of the error covariance matrix at step k-1, Q 2，k-1 is the system noise covariance matrix of the k-1th step. The Kalman gain matrix K of the kth step can be obtained 2，k The expression is:

[0075]

[0076] Among them, R 2，k is the k-th step measurement noise covariance matrix. Correspondingly, the optimal estimate of the k-th step error covariance matrix is It can be calculated according to formula (XII).

[0077]

[0078] Step (c): comprehensively evaluate the influence of controller delay, inverter nonlinearity, and motor parameter harmonics on fundamental and harmonic voltages, and extract the true fundamental and harmonic voltage values ​​from the output of the current controller by quantitatively establishing a compensation voltage equation;

[0079] In this step, the effects of inverter nonlinearity, controller delay, and motor parameter harmonics on the fundamental and harmonic voltage equations are considered. By quantitatively establishing compensation equations, the actual voltage value can be extracted from the output signal of the current controller. The inverter nonlinear factors include the dead zone and the conduction voltage drop of the switch tube and the freewheeling diode. The controller delay includes the delay generated during the SVPWM on-time calculation process (one controller cycle) and the delay during the PWM voltage generation process (similar to the sampling and holding process, 0.5 switching cycles). The motor parameter harmonics take into account the inductance and flux harmonics caused by magnetic field saturation, tooth slot structure, and rotor structure.

[0080] In specific implementation, in digital control systems, the controller delay caused by discretization mainly includes two parts: first, the control voltage calculated based on the current sampling is used for the next control cycle, so there is a delay of one control cycle; second, the PWM voltage output is similar to the sample-and-hold process, which will produce a delay of half a switching cycle. Since the discrete frequency of the controller is usually consistent with the PWM switching frequency, the total controller delay time generated is 1.5T s , where T s is the inverter switching cycle.

[0081] Preferably, in multiple reference coordinate systems, define is the controller output n th Voltage harmonics. Definition α n n th The controller delay angle of voltage harmonics is α n =1.5nω e T s The voltage after delay when considering the controller and It can be obtained from formula (XIII).

[0082]

[0083] The nonlinear factors of the inverter mainly include dead time and tube voltage drop. The switching device in the voltage-type inverter analyzed in the present invention is IGBT.

[0084] Preferably, define u dc is the DC bus voltage, u fi is the forward conduction voltage drop of IGBT, u fd is the forward conduction voltage drop of the diode, K s is the switching frequency of the IGBT, T d is the dead time of the inverter. The voltage after considering the nonlinearity of the inverter can be obtained by formula (XIV).

[0085]

[0086] in,

[0087] In addition, the voltage equation (III) and the extended Kalman filter observer already include the influence of motor parameter harmonics, which are difficult to obtain through identification means. In the present invention, the motor parameter harmonics obtained by finite element analysis are calculated as known quantities.

[0088] Preferably, according to formula (XIII) and formula (XIV), a voltage compensation equation can be constructed to obtain the actual voltage on the motor according to the controller output voltage.

[0089] Step (d): Based on the parameters identified by the fundamental and harmonic extended Kalman filter observers, a single interconnected extended Kalman filter parameter identification system is constructed.

[0090] Preferably, a single interconnected parameter identification system consisting of two extended Kalman filter observers (fundamental and harmonic) is constructed, and the EKF2 identification parameter l t0 , l c2 , l s2 , R is passed to EKF1 to identify ψ d1 and ψ q1 , in order to avoid the iterative convergence process of EKF1 and EKF2 in single-step identification, and effectively improve the parameter convergence speed. By separating the fundamental and harmonic extended Kalman filter observers, the amount of calculation can be greatly reduced. The use of a single interconnected structure can avoid the iterative convergence of parameters in a single-step operation, and effectively improve the parameter convergence speed.

[0091] In the specific implementation, the above three steps are combined to build a single interconnection identification system based on the parameters identified by the fundamental and harmonic extended Kalman filter observers, such as Figure 2 shown. Figure 2 It can be seen from the parameter transfer structure in that during the parameter identification process, the two extended Kalman filters in a single-step operation do not need to iterate each other when calculating parameters, so the convergence speed of parameter identification can be greatly accelerated.

[0092] In order to verify the parameter identification effect of the system, a simulation model of the multi-parameter identification system was established in Matlab. First, a permanent magnet synchronous motor current harmonic control system based on multi-reference frame PI control was built. Figure 3 As shown, the electromagnetic structure of the permanent magnet synchronous motor is as follows Figure 4 As shown, Figure 4 In the figure, 1 is the core, 2 is the winding, and 3 is the permanent magnet. The key parameters of the motor electromagnetic structure are shown in Table 1, and the settings of the voltage source inverter used are shown in Table 2.

[0093] Table 1 Permanent magnet synchronous motor parameters

[0094]

[0095] Table 2 Parameters of voltage source inverter

[0096]

[0097] Secondly, the controller delay compensation module and the inverter nonlinear compensation module were built to obtain the actual voltage applied to the permanent magnet synchronous motor according to the controller output voltage. Finally, the single interconnected extended Kalman filter observer module was written through the S function, and the parameter values ​​of the two extended Kalman filter observers are shown in Table 3.

[0098] Table 3 Extended Kalman filter parameters

[0099]

[0100] The simulation working condition information and current harmonic setting values ​​are shown in Table 4.

[0101] Table 4 Key control parameters for harmonic injection simulation of permanent magnet synchronous motor

[0102]

[0103]

[0104] The identification waveforms of each parameter are as follows: Figure 5(a)-Figure 5(f) As shown, Figure 5(a) is the parameter l t0 The identification waveform of parameter l is shown in Figure 5(b). c2 The identification waveform of parameter l is shown in Figure 5(c). s2 Figure 5(d) is the identification waveform of parameter R, and Figure 5(e) is the identification waveform of parameter ψ d1 The identification waveform of the parameter ψ is shown in Figure 5(f). q1 From the identification waveform, it can be seen that each identification parameter can quickly converge to the true value. The true value of the parameter and the identification result are shown in Table 5. It can be seen that the identification error of each parameter is within 2% (l s2 Except, s2 represents the average value of the dq axis mutual inductance, which is caused by the motor magnetic field saturation, and the amplitude is very small, so the error is relatively large). Therefore, the above simulation proves that the present invention can realize accurate permanent magnet synchronous motor multi-parameter identification.

[0105] Table 5 True values ​​and identification results of each parameter

[0106]

[0107] Since the number of voltage equations of the traditional permanent magnet synchronous motor (PMSM) model in the dq coordinate system is 2, the online parameter identification of the permanent magnet synchronous motor often has a lack of rank problem. This paper proposes a multi-parameter identification technology of permanent magnet synchronous motor based on harmonic coupling model. th and 13 th Current harmonics are used to realize online identification of multiple parameters. By introducing the voltage-current harmonic coupling equation, the voltage equation of the permanent magnet synchronous motor is expanded to the fundamental voltage-current model and the harmonic voltage-current coupling equation. By constructing the fundamental model and the harmonic model, a single interconnected extended Kalman filter observation system is formed, which can simultaneously realize the identification of 6 motor parameters without the problem of lack of rank. In addition, in order to further improve the parameter identification accuracy, this technology also considers the influence of controller delay, inverter nonlinearity, and motor parameter harmonics on the parameter identification method.

[0108] In summary, the permanent magnet synchronous motor multi-parameter identification method based on the harmonic coupling model of the present invention has the following advantages compared with the prior art:

[0109] (1) The harmonic coupling model can solve the problem of lack of rank in multi-parameter identification;

[0110] (2) The extended Kalman filter algorithm takes into account system noise and measurement noise, has good anti-interference performance and high accuracy. The single interconnected structure can avoid the iterative convergence process between the parameters identified by the two extended Kalman filters in a single-step operation, reducing the amount of calculation and accelerating the convergence speed.

[0111] (3) The influence of controller delay, inverter nonlinearity and motor parameter harmonics on parameter identification is considered.

[0112] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A multi-parameter identification method for a permanent magnet synchronous motor based on a harmonic coupling model, characterized in that: include: Step (a): by separating the voltage equation of the permanent magnet synchronous motor harmonic coupling model into two equations: a fundamental voltage equation and a harmonic voltage equation, wherein the required permanent magnet synchronous motor inductance harmonic parameters are obtained by a finite element analysis method; Step (b): construct two extended Kalman filter observers based on the fundamental voltage equation and the harmonic voltage equation to allocate the six key motor parameters considering the magnetic field saturation; Step (c): On the basis of considering the controller delay and the nonlinearity of the inverter, the influence of the motor parameter harmonics on the fundamental and harmonic voltages is further considered, and the real fundamental and harmonic voltage values ​​are extracted from the output of the current controller by quantitatively establishing the compensation voltage equation; Step (d): Based on the parameters identified by the fundamental and harmonic extended Kalman filter observers, a single interconnected extended Kalman filter parameter identification system is constructed.

2. The method according to claim 1, characterized in that In the step (a), the fundamental voltage equation and the harmonic voltage equation are respectively: Among them, Γ1=γ; Γ2=diag(11y, 13γ); R1=diag(R, R); R2=diag(R, R, R, R); u dq1113 =[u dq11 T u dq13 T ] T ;i dq1113 =[i dq11 T i dq13 T ] T ψ dq1113 =[ψ dq11 T ψ dq13 T ] T ;u dqn ,i dqn and ψ dqn n th Voltage, current and permanent magnet flux in dq coordinate system; ω e is the electrical angular velocity; R is the phase resistance; γ=[0 -1; 1 0]; The coupled inductance matrix is ​​expressed as: Where, τ = diag(1, -1); l t0 Calculated from the DC components of self-inductance and mutual inductance; l ck , l sk Calculated from the kth harmonic components of self-inductance and mutual inductance; High frequency components of coupled inductors The higher-order components of the permanent magnet flux ψdq1113 are obtained through finite element analysis.

3. The method according to claim 1, characterized in that: The step (b) comprises: According to formula (II) and formula (III), the fundamental wave identification model ∑1 and the harmonic wave identification model ∑2 are obtained, as shown in formula (IV) and formula (V): Among them, x1 and x2 are state variables; y1 and y2 are output variables; u1 and u2 are input variables; f1(g) and f2(g) are state equations; h1(g) and h2(g) are output equations; Using the harmonic identification model Σ2 to construct the extended Kalman filter observer EKF2 identification l t0 , l c2 , l s2 , R, and use the fundamental wave coupling model Σ1 to construct the extended Kalman filter observer EKF1 to identify ψd1 and ψ q1 , where R is the phase resistance, ψ d1 and ψ q1 is the fundamental wave component of permanent magnet flux, l t0 , l c2 and l s2 Calculated from the DC component and second-order component of the inductance respectively.

4. The method according to claim 1, characterized in that The step (c) comprises: In multiple reference coordinate systems, define is the controller output n th Voltage Harmonics; Definition α n n th The controller delay angle of voltage harmonics is α n =1.5nω e T s ; Delayed voltage when considering the controller and From formula (XIII), we get: Among them, T s is the inverter switching cycle; Define u dc is the DC bus voltage, u fi is the forward conduction voltage drop of IGBT, u fd is the forward conduction voltage drop of the diode, K s is the switching frequency of the IGBT, T d is the dead time of the inverter; the voltage after considering the nonlinearity of the inverter is obtained by formula (XIV): in, Formula (III) itself includes the influence of motor parameter harmonics, where the motor parameter harmonics obtained from finite element analysis are calculated as known quantities; According to equations (XIII) and (XIV), a voltage compensation equation is constructed, and the actual voltage on the motor is obtained according to the output voltage of the current controller.

5. The method according to claim 1, characterized in that: The step (d) comprises: A single interconnected parameter identification system consisting of two extended Kalman observers is constructed, and the EKF2 identification parameter l t0 , l c2 , l s2 , R is passed to EKF1 to identify ψ d1 and ψ q1 , to avoid the iterative convergence process of EKF1 and EKF2 in single-step identification, and effectively improve the parameter convergence speed.

Images