阅读说明：本技术 基于二阶滑模观测器的永磁同步电机参数辨识方法 (permanent magnet synchronous motor parameter identification method based on second-order sliding-mode observer ) 是由 叶益强 陈礼根 张玉北 于 2019-09-30 设计创作，主要内容包括：本发明涉及电机参数辨识技术领域,具体涉及一种基于二阶滑模观测器的永磁同步电机参数辨识方法,包括以下步骤：第一步,获取永磁同步电机在dq轴参考坐标系下的参数；第二步,以机械角速度和负载转矩为观测对象,构建可将扩展的滑模观测器；第三步,在第二步的基础上,以机械角速度和负载转矩为观测对象,构建二阶滑模观测器辨识永磁同步电机的负载转矩,第四步,在识别的负载转矩的基础上,采用直接计算法或PI调节器方法辨识永磁同步电机的转动惯量。本发明考虑了转动惯量、电磁转矩和粘性摩擦的变化,并在保证负载转矩精度的同时还可有效抑制抖振。(The invention relates to the technical field of motor parameter identification, in particular to a permanent magnet synchronous motor parameter identification method based on a second-order sliding-mode observer, which comprises the following steps of: firstly, acquiring parameters of a permanent magnet synchronous motor under a dq axis reference coordinate system; secondly, constructing a sliding-mode observer capable of expanding by taking the mechanical angular velocity and the load torque as observation objects; and a third step of constructing a second-order sliding mode observer to identify the load torque of the permanent magnet synchronous motor by taking the mechanical angular speed and the load torque as observation objects on the basis of the second step, and a fourth step of identifying the rotational inertia of the permanent magnet synchronous motor by adopting a direct calculation method or a PI regulator method on the basis of the identified load torque. The invention considers the changes of the rotational inertia, the electromagnetic torque and the viscous friction, and can effectively inhibit buffeting while ensuring the precision of the load torque.)

1. A permanent magnet synchronous motor parameter identification method based on a second-order sliding-mode observer is characterized by comprising the following steps:

Firstly, acquiring parameters of the permanent magnet synchronous motor under a dq-axis reference coordinate system, wherein the parameters comprise the number of pole pairs P and voltage components u of stator voltage on d and q axes_d、u_qcurrent component i of stator current in d and q axes_d、i_qinductance component L of stator inductance on d and q axes_d、L_qStator resistance R_scomponent ψ of flux linkage in d and q axes_d、ψ_qangular velocity ω of motor_m；

And secondly, taking the mechanical angular velocity and the load torque as observation objects, and describing the extended sliding mode observer as follows:

in formula (1), J, T_e、T_Lf and omega_mrespectively rotational inertia, electromagnetic torque, load torque, viscous friction coefficient and motor angular velocity,AndRespectively obtaining estimated values of rotational inertia, electromagnetic torque, load torque, viscous friction coefficient and motor angular speed; k is sliding mode gain, g is feedback gain;

and thirdly, on the basis of the second step, taking the mechanical angular velocity and the load torque as observation objects, and constructing a second-order sliding mode observer for identifying the load torque of the permanent magnet synchronous motor, wherein the second-order sliding mode observer is as follows:

in the formula_ωsis a control law of the sliding-mode observer, h is a feedback gain,AndEstimated values of electromagnetic torque, load torque, viscous friction coefficient and mechanical angular velocity, respectively;

Control law gamma of sliding-mode observer_ωsThe expression is as follows:

In the formula (3), the reaction mixture is,γ、l_gAnd c are sliding mode control rate design parameters and meet min (gamma, l)_gC) > 0, p, q are odd numbers and satisfy p/q>1，S_ωfor second-order slip-form switching surfaces, S_ωThe expression is as follows:

In the formula Is f_ωthe first derivative of (a).

2. the method for identifying the parameters of the permanent magnet synchronous motor based on the second-order sliding-mode observer according to claim 1, further comprising the following steps of:

And fourthly, on the basis of the identified load torque, identifying the rotational inertia of the permanent magnet synchronous motor by adopting a direct calculation method or a PI regulator method.

3. The method for identifying the parameters of the permanent magnet synchronous motor based on the second-order sliding-mode observer according to claim 2, wherein the direct calculation method is suitable for a driving system just before entering a steady state, and the specific algorithm is as follows:

When the drive system enters a steady state at T ═ x, the load torque T is loaded at that time_L(x) Can be expressed as:

T_L(x)＝T_e(x)-Fω_m(x) (5)

T_L(x) From t-x to t-x₁Is kept constant, the moment of inertia is directly calculated by the following expression:

In the formulaThe value of the moment of inertia at t x,Estimated load torque for initial moment of inertia value

4. the method for identifying the parameters of the permanent magnet synchronous motor based on the second-order sliding-mode observer according to claim 2, wherein the PI regulator method is suitable for the situation that a load torque observation error always exists in a dynamic state of a driving system when the driving system is just switched into the dynamic state from any stable state,load torque observation error f_TThe discrete form is expressed as follows:

The rule for identifying the moment of inertia using a PI regulator is as follows:

In the formula k_pand k_iProportional and integral gains, T, respectively_sis the sampling time, s_T(x) The expression of (a) is as follows:

s_T(x)＝sign[T_e(x)-T_L(x)-Fω_m(x)] (9)

J/[ T ] when the drive system goes to steady state_e(x)-T_L(x)-Fω_m(x)]The term tends to be infinite, possibly leading to recognition failure, so s is used_T(x) Alternative J/[ T ]_e(x)-T_L(x)-Fω_m(x)]。

5. the method for identifying the parameters of the permanent magnet synchronous motor based on the second-order sliding-mode observer according to claim 1, wherein a cut-off frequency is adoptedthe low-pass filter of (a) suppresses the buffeting signal of a second order sliding mode observer, wherein,k is the sliding mode gain, g is the feedback gain,is an estimate of the moment of inertia.

Technical Field

the invention relates to the technical field of motor parameter identification, in particular to a permanent magnet synchronous motor parameter identification method based on a second-order sliding-mode observer.

Background

Permanent Magnet Synchronous Motor (PMSM) drives have found widespread use in various electromechanical servo systems over the last several decades. Parameters of a driving system such as load torque, rotational inertia and the like have important significance for improving controller design in industrial application. For example, when the moment of inertia is used to design a speed loop controller, the drive system will have a higher speed tracking accuracy. In addition, when the load torque is used as a feedforward term of the reference torque, the load disturbance resistance of the driving system is remarkably improved. However, the load torque is generally unknown, and the moment of inertia can vary significantly with the shape and size of the mechanical load. In some practical permanent magnet synchronous motor drive systems, the moment of inertia and load torque are time-varying and difficult to obtain online. In this case, the unmatched moment of inertia or load torque may not be strong enough to ensure the speed control performance of the drive system. For this purpose, recognition algorithms for load torque and moment of inertia are proposed.

The load torque identification method mainly comprises model reference adaptive control, a Kalman filter, a sliding-mode observer and the like. In practical research and application, the load torque is estimated by constructing various types of state observers. The sliding-mode observer has the advantages of simplicity in implementation, strong uncertainty resistance, strong anti-interference capability and the like, so that more and more attention is paid to the sliding-mode observer.

The sliding-mode observer is used for enabling the system structure to change along with the current state, so that the system moves up and down according to a given track, namely the sliding-mode observer is used for realizing the sliding-mode motion. The control has certain switching characteristics and is in an intermittent switching working state, so that the control has certain robustness on parameter conversion and disturbance. The main problem of the application of sliding-mode observers in servo systems is the well-known phenomenon of flutter, i.e. the generation of high frequency harmonics. Therefore, scholars at home and abroad propose several methods for overcoming buffeting, including a classical method, an intelligent method and an observer method. The classical methods mainly include a quasi-sliding mode method, an approach method and the like. Intelligent methods include fuzzy control, artificial neural networks, and the like. An advanced control theory is applied to the buffeting reduction, but the calculation amount is large, and the quick real-time control is difficult to realize. Buffeting caused by uncertainty can be well eliminated by adopting an observer method.

The rotational inertia identification method can be divided into an off-line identification method and an on-line identification method. The off-line identification method mainly comprises a direct calculation method, an acceleration and deceleration method and a limited torque acceleration method. The off-line identification method has the advantages that the calculation conditions can be set according to the algorithm, and meanwhile, the problem that the change of the rotational inertia cannot be reflected in real time is also obvious. The online rotational inertia identification method mainly comprises a sliding-mode observer, a Kalman filter, model reference adaptive control and the like. The Kalman filtering method takes the rotational inertia as a state variable, and the identification value of the rotational inertia is directly output by a filtering algorithm. The model reference adaptive observation technology takes the deviation of the speed as feedback, and the deviation of the speed model tends to zero through the identification of the rotational inertia value.

in conventional sliding-mode observer based methods, the sign function will cause high frequency flutter and buffeting problems, which may cause system oscillation, performance degradation, and even system instability. Therefore, most of the existing load torque identification methods based on the sliding mode observer are mainly focused on inhibiting buffeting, and influence of parameter distortion on identification performance is ignored. Although useful information of the estimated parameters can be obtained by adding a low-pass filter, the introduction of the low-pass filter causes a phase delay, thereby affecting the estimation accuracy and the performance of the driving system. Therefore, appropriate compensation is needed to mitigate the effects of the low pass filter, especially when estimating a continuous phase delayed ac signal. Meanwhile, in some load torque identification methods, the rotational inertia is considered to be constant, and when the rotational inertia changes, no analysis is performed on the observation performance. When the transmission system is in a dynamic state, the rotational inertia has great influence on the observation precision of the load moment.

In the aspect of identifying the rotational inertia, in the off-line identification method, in order to simplify the problem, a direct calculation method only calculates the rotational inertia of the rotor of the servo motor and ignores the friction torque, and the influence of the friction torque is weakened by taking the average value of the speed and the acceleration and deceleration time of the rotor. The method has no error iterative convergence process, so that the online tracking cannot be realized and the identification result fluctuates to a certain extent. The acceleration and deceleration method and the limited torque acceleration method are widely applied to offline rotational inertia identification, but the methods are low in precision, long in identification time and large in storage space. The on-line identification method has a large calculation amount and requires a digital signal processor to have higher data processing speed and storage capacity. And when the load disturbance in the transient process is changed in a large range, the old load torque estimation value still participates in the recursion process, so that the identification distortion is caused.

Disclosure of Invention

To solve the above technical problems, the present invention aims to: the method for identifying the parameters of the permanent magnet synchronous motor based on the second-order sliding-mode observer considers the changes of rotational inertia, electromagnetic torque and viscous friction, and can effectively inhibit buffeting while ensuring the precision of load torque.

The technical scheme adopted by the invention for solving the technical problem is as follows:

the permanent magnet synchronous motor parameter identification method based on the second-order sliding-mode observer specifically comprises the following steps:

firstly, acquiring parameters of the permanent magnet synchronous motor under a dq-axis reference coordinate system, wherein the parameters comprise the number of pole pairs P and voltage components u of stator voltage on d and q axes_d、u_qCurrent component i of stator current in d and q axes_d、i_qinductance component L of stator inductance on d and q axes_d、L_qStator resistance R_scomponent ψ of flux linkage in d and q axes_d、ψ_qangular velocity ω of motor_m；

and secondly, taking the mechanical angular velocity and the load torque as observation objects, and describing the extended sliding mode observer as follows:

In formula (1), J, T_e、T_LF and omega_mRespectively rotational inertia, electromagnetic torque, load torque, viscous friction coefficient and motor angular velocity,andRespectively, moment of inertia, electromagnetic torqueEstimated values of load torque, viscous friction coefficient and motor angular velocity; k is sliding mode gain, g is feedback gain;

And thirdly, on the basis of the second step, taking the mechanical angular velocity and the load torque as observation objects, and constructing a second-order sliding mode observer for identifying the load torque of the permanent magnet synchronous motor, wherein the second-order sliding mode observer is as follows:

In the formula_ωsIs a control law of the sliding-mode observer, h is a feedback gain,AndEstimated values of electromagnetic torque, load torque, viscous friction coefficient and mechanical angular velocity, respectively;

control law gamma of sliding-mode observer_ωsthe expression is as follows:

in the formula (3), the reaction mixture is,γ、l_gand c are sliding mode control rate design parameters and meet min (gamma, l)_gC) > 0, p, q are odd numbers and satisfy p/q>1，S_ωfor second-order slip-form switching surfaces, S_ωThe expression is as follows:

In the formula Is f_ωthe first derivative of (a).

Preferably, the method for identifying parameters of the permanent magnet synchronous motor based on the second-order sliding-mode observer further includes the following steps:

And fourthly, on the basis of the identified load torque, identifying the rotational inertia of the permanent magnet synchronous motor by adopting a direct calculation method or a PI regulator method.

Preferably, the direct calculation method is suitable for the driving system just entering the steady state, and the specific algorithm is as follows:

When the drive system enters a steady state at T ═ x, the load torque T is loaded at that time_L(x) Can be expressed as:

T_L(x)＝T_e(x)-Fω_m(x) (5)

T_L(x) From t-x to t-x₁is kept constant, the moment of inertia is directly calculated by the following expression:

In the formulathe value of the moment of inertia at t x,Estimated load torque for initial moment of inertia value

Let Δ x be x-x₁If Δ x is too large, it is difficult to satisfy the assumption that the load torque is constant in the Δ x time. If Δ x is too small, there may be large computational errors due to sampling noise or other non-idealities.

preferably, the PI regulator method is suitable for the moment when the driving system is just switched into the dynamic state from any steady state, and the load torque observation error always exists in the drivingdynamic state of the system, load torque observation error f_TThe discrete form is expressed as follows:

The rule for identifying the moment of inertia using a PI regulator is as follows:

In the formula k_pand k_iproportional and integral gains, T, respectively_sis the sampling time, s_T(x) The expression of (a) is as follows:

s_T(x)＝sign[T_e(x)-T_L(x)-Fω_m(x)] (9)

J/[ T ] when the drive system goes to steady state_e(x)-T_L(x)-Fω_m(x)]the term tends to be infinite, possibly leading to recognition failure, so s is used_T(x) Alternative J/[ T ]_e(x)-T_L(x)-Fω_m(x)]。

Preferably, a cut-off frequency ofThe low-pass filter of (a) suppresses the buffeting signal of a second order sliding mode observer, wherein,k is the sliding mode gain, g is the feedback gain,is an estimate of the moment of inertia.

compared with the prior art, the invention has the following beneficial effects:

the invention provides a novel load torque second-order sliding mode observer, which considers the changes of rotational inertia, electromagnetic torque and viscous friction, can effectively inhibit buffeting while ensuring the precision of the load torque, has higher estimation precision and faster convergence speed, and provides two methods for estimating the rotational inertia.

Drawings

FIG. 1 is a block flow diagram of the present invention.

Detailed Description