阅读说明：本技术 一种电压前馈补偿方法的无差拍电流预测控制方法 (Dead-beat current prediction control method of voltage feedforward compensation method ) 是由 黄晓艳 盛方 邱麟 吴立建 方攸同 张健 于 2020-04-30 设计创作，主要内容包括：本发明公开了一种电压前馈补偿方法的无差拍电流预测控制方法。本发明首先通过最小二乘递推算法辨识电机的实际电感参数,建立电压前馈补偿通道；然后将当前时刻的电流给定值与下一个时刻的电流实际值做差得到电流估计误差,将该误差作为输入应用在电压前馈补偿通道上,即可完成对三相逆变器的控制电压的补偿。本发明方法可以在电机参数失配的情况下,提高电流预测控制方法的鲁棒性以及准确性,并且只需要对电机电感进行单独辨识,易于实现且计算量很小。(The invention discloses a deadbeat current prediction control method of a voltage feedforward compensation method. Firstly, identifying actual inductance parameters of a motor through a least square recursive algorithm, and establishing a voltage feedforward compensation channel; and then, the current set value at the current moment is differed with the current actual value at the next moment to obtain a current estimation error, and the error is applied to a voltage feedforward compensation channel as input to finish the compensation of the control voltage of the three-phase inverter. The method can improve the robustness and accuracy of the current prediction control method under the condition of motor parameter mismatch, only needs to separately identify the motor inductance, is easy to realize and has small calculation amount.)

1. A deadbeat current prediction control method of a voltage feedforward compensation method is characterized by comprising the following steps of:

1) identifying actual inductance parameters of the motor through a least square recursive algorithm;

2) establishing a voltage feedforward compensation channel based on a calculation formula of the control voltage compensation quantity of the three-phase inverter, and inputting the current estimation error into the voltage feedforward compensation channel to obtain the control voltage compensation quantity;

3) calculating a control voltage according to a traditional deadbeat current prediction control model;

4) and (3) adding the control voltage compensation quantity obtained in the step 2) into the control voltage obtained by the calculation of the traditional deadbeat current prediction control model in the step 3), so as to realize the compensation of the control voltage of the three-phase inverter.

2. The deadbeat current prediction control method of the voltage feedforward compensation method according to claim 1, wherein the step 1) is specifically: adopt least square recurrence algorithm to distinguish alone the actual inductance parameter of motor among the model, it is:

wherein T represents a sampling time, u_dRepresenting d-axis stator voltage vector, i_d(k) And i_d(k +1) represents the d-axis stator current i at two moments before and after k and k +1, respectively_q(k) Representing the q-axis stator current, ω, at time k_eRepresenting the electrical angular velocity at which the motor is rotating,representing the actual inductance parameter to be identified.

3. The deadbeat current prediction control method of the voltage feedforward compensation method according to claim 1, wherein the step 2) is specifically:

1) establishing a voltage feedforward compensation channel based on a calculation formula of control voltage compensation quantity of the three-phase inverter; the calculation formula of the control voltage compensation quantity of the three-phase inverter is as follows:

wherein, T represents the sampling time,

2) the given value of the stator current at the current moment is differed with the actual value of the stator current at the next moment to obtain a current estimation error, and the current estimation error is input into a voltage feedforward compensation channel to obtain a control voltage compensation β_x。

4. The deadbeat current predictive control method of a voltage feed forward compensation method as claimed in claim 1, wherein the conventional deadbeat current predictive control model in step 3) is:

u(k)＝G^-1(x(k+1)-F(k)·x(k)-H(k))

the specific form of each matrix in the formula is as follows:

x(k)＝[i_d(k) i_q(k)]^T；u(k)＝[u_d(k) u_q(k)]^T

wherein i_d(k) And i_q(k) Respectively representing a d-axis component and a q-axis component of the stator current at the k-th moment; u. of_d(k) And u_q(k) Respectively representing a d-axis component and a q-axis component of a stator voltage vector at the kth moment, namely the control voltage of the three-phase inverter; omega_e(k) Electrical angular velocity at time k; t represents a sampling time; l isRepresenting a nominal value of the inductance of the stator winding; r represents a nominal value of the stator resistance; psi_fRepresenting the nominal value of the permanent magnet flux linkage.

Technical Field

The invention relates to a current control method of a motor, in particular to a deadbeat current prediction control method of a voltage feedforward compensation method.

Background

The CPC is a current loop control method of the motor, and obtains a control voltage vector of the three-phase inverter by acquiring stator current and motor rotating speed signals and calculating through a formula based on motor parameters. Compared with a common PID controller, the method can enable the current loop output of the motor to follow a given signal in a sampling period, improves the system bandwidth, does not need parameter setting, has simple calculation process and is beneficial to the realization of codes in engineering.

CPC is suitable for permanent magnet synchronous machines, but in practical applications, parameter mismatch can occur due to the fact that the nominal value of the machine is not completely accurate. Parameter mismatch can have many adverse effects on the practical control effect of CPC, such as output quietness, system stability degradation, etc. For this problem, the traditional solution is parameter identification, i.e. for three motor-related parameters involved in CPC: and identifying the stator resistance, the stator inductance and the permanent magnet flux linkage. However, this method introduces complicated parameter identification calculation, which puts a burden on the system.

Disclosure of Invention

In order to solve the problems in the background art, the invention provides a deadbeat Current prediction Control method for voltage feedforward compensation under the condition of motor system parameter mismatch, which is a Current Predictive Control (CPC) method for a permanent magnet synchronous motor, can optimize the Control effect when the motor parameters are mismatched, reduce the dependency on the motor parameters as much as possible, and realize the deadbeat Current prediction Control method on the basis.

The technical scheme of the invention is that under the condition that the parameter mismatch occurs in the system, the control voltage of the three-phase inverter is compensated by adopting the following steps:

1) identifying actual inductance parameters of the motor through a least square recursive algorithm;

2) establishing a voltage feedforward compensation channel based on a calculation formula of the control voltage compensation quantity of the three-phase inverter, and inputting the current estimation error into the voltage feedforward compensation channel to obtain the control voltage compensation quantity;

3) calculating a control voltage according to a traditional deadbeat current prediction control model;

4) and (3) adding the control voltage compensation quantity obtained in the step 2) into the control voltage obtained by the calculation of the traditional deadbeat current prediction control model in the step 3), so as to realize the compensation of the control voltage of the three-phase inverter.

The step 1) is specifically as follows: adopt least square recurrence algorithm to distinguish alone the actual inductance parameter of motor among the model, it is:

wherein T represents a sampling time, u_dRepresenting d-axis stator voltage vector, i_d(k) And i_d(k +1) represents the d-axis stator current i at two moments before and after k and k +1, respectively_q(k) Representing the q-axis stator current, ω, at time k_eRepresenting the electrical angular velocity at which the motor is rotating,

representing the actual inductance parameter to be identified.

Therefore, in the parameter identification process of the invention, other two motor parameters (stator resistance and permanent magnet flux linkage) are not involved, the identification process is simple, and the identification process has no coupling relation with other parameters.

The step 2) is specifically as follows:

1) establishing a voltage feedforward compensation channel based on a calculation formula of control voltage compensation quantity of the three-phase inverter; the calculation formula of the control voltage compensation quantity of the three-phase inverter is as follows:

wherein T represents a sampling time，Representing the stator winding inductance, i.e. the actual inductance parameter identified separately in step 1),representing d-axis or q-axis stator current set value i at time k_x(k +1) represents the actual value of the stator current of d-axis or q-axis at the time of k +1, β_xRepresenting the control voltage compensation quantity of the three-phase inverter, wherein d and q respectively represent the directions of a d axis and a q axis, namely the directions of a direct axis and a quadrature axis;

the voltage feedforward compensation channel established by the invention only comprises one transfer function

/T。

2) The given value of the stator current at the current moment is differed with the actual value of the stator current at the next moment to obtain a current estimation error, and the current estimation error is input into a voltage feedforward compensation channel to obtain a control voltage compensation β_x；

The traditional deadbeat current prediction control model in the step 3) is as follows:

u(k)＝G^-1(x(k+1)-F(k)·x(k)-H(k))

u (k) denotes an output matrix, x (k) denotes an input matrix, G, F (k), H (k) denotes a coefficient matrix.

The specific form of each matrix in the formula is as follows:

x(k)＝[i_d(k) i_q(k)]^T；u(k)＝[u_d(k) u_q(k)]^T

wherein i_d(k) And i_q(k) Respectively representing the actual value of the stator current at the k-th momentD-axis component and q-axis component of (a); u. of_d(k) And u_q(k) Respectively representing a d-axis component and a q-axis component of a stator voltage vector at the kth moment, namely the control voltage of the three-phase inverter; omega_e(k) Electrical angular velocity at time k; t represents a sampling time; l represents the nominal value of the inductance of the stator winding; r represents a nominal value of the stator resistance; psi_fRepresenting the nominal value of the permanent magnet flux linkage.

The invention has the beneficial effects that:

the method only needs to identify the inductance parameter of the motor, does not need to carry out complex identification calculation on other parameters, has a very simple operation mode on the voltage feedforward channel, and effectively solves the problems of current static error, stability reduction and the like caused by parameter mismatch. Meanwhile, the advantages of large system bandwidth and simple calculation process are kept, and the method is very beneficial to engineering implementation.

Drawings

Fig. 1 is a block diagram of a current predictive control system.

Fig. 2 is a diagram of a voltage feed forward compensation architecture.

Fig. 3 shows the result of the inductive parameter simulation.

Fig. 4 is a simulation result of the conventional deadbeat current prediction control at the time of parameter mismatch.

FIG. 5 is a simulation result of a deadbeat current prediction control based on voltage feed forward with parameter mismatch.

Detailed Description

The invention is further illustrated by the following figures and examples.

The embodiment of the invention and the working principle and process thereof are as follows:

because the nominal value of the motor is not accurate, and the motor parameters used in the current prediction control method are all estimated values, the motor current loop output has static error, and serious parameter mismatch even causes system instability. In order to achieve the goal that the current prediction control can still work normally under the condition of parameter mismatch, the control voltage vector of the three-phase inverter needs to be compensated. A non-salient pole permanent magnet synchronous motor and a control system thereof are selected for description. As shown in fig. 1, the control system mainly includes a speed regulator, a current prediction controller, a coordinate converter (PARK inverse transformation, PARK transformation, CLARK transformation), a sensor (speed and electrical angle calculation, etc.), a three-phase inverter, and a control circuit thereof.

The implementation steps of the deadbeat current prediction control based on the voltage feedforward compensation are as follows:

(1) and calculating the control voltage through a traditional deadbeat current prediction control model.

The traditional deadbeat current prediction control model is as follows:

u(k)＝G^-1(x(k+1)-F(k)·x(k)-H(k))

u (k) denotes an output matrix, x (k) denotes an input matrix, G, F (k), H (k) denotes a coefficient matrix.

The specific form of each matrix in the formula is as follows:

x(k)＝[i_d(k) i_q(k)]^T；u(k)＝[u_d(k) u_q(k)]^T

wherein i_d(k) And i_q(k) Respectively representing the dq-axis stator current at the k-th time_d(k) And u_q(k) Respectively represent dq-axis stator voltage vectors at the k-th time, i.e., control voltages, ω, of the three-phase inverter_e(k) Electrical angular velocity at time k, T represents sampling time, L represents stator winding inductance, R represents stator resistance, #_fRepresenting a permanent magnet flux linkage.

A dq axis coordinate system is established on a motor rotor, a d axis (a direct axis) is a rotor magnetic field direction, and a q axis (a quadrature axis) is perpendicular to the rotor magnetic field direction.

(2) And calculating the compensation quantity of the control voltage through a voltage feedforward link.

The voltage feedforward compensation structure shown in fig. 2 is established based on a calculation formula of the control voltage compensation amount of the three-phase inverter, and the calculation formula of the control voltage compensation amount of the three-phase inverter is as follows:

wherein, T represents the sampling time,

represents the actual inductance parameter of the motor,representing a given value of current, i_x(k +1) represents the actual value of the current, β_xRepresenting the voltage compensation amount, d, q represent the direct and quadrature directions, respectively.

The identification method is a least square recursion algorithm, and the model is as follows:

wherein T represents the sampling time of the control system, u_dRepresenting d-axis stator voltage vector, i_d(k) And i_d(k +1) represents the d-axis stator current i at two moments before and after_qRepresenting the q-axis stator current, ω_eRepresenting the electrical angular velocity at which the motor is rotating,representing the stator winding inductance.

The voltage feedforward path established according to the formula only comprises one transfer function

The input is the difference between the current set value at the current moment and the current actual value at the next moment to obtain a current estimation error α_x(k)。

(3) The control voltage compensation amount calculated in the voltage feedforward compensation link is added to the control voltage calculated in the traditional deadbeat current prediction control, as shown in fig. 2. Finally, compensation of the control voltage of the three-phase inverter under the current prediction control can be realized under the condition of parameter mismatch.

Fig. 3 is a process of identifying actual inductance parameters of the motor. The identification process is convergent, the convergence process is fast, the precision is high, and the real inductance value of the motor can be effectively identified.

FIG. 4 and FIG. 5 show the initial set value of q-axis current at a given rotation speed ω 10 π rad/sSimulation result, i, which became 1A at time 0.1s_qRepresenting the actual value of the current. The simulation assumes the change of a plurality of groups of parameter variables, and as a result, the traditional current prediction control can face the problems of output static error and poor stability under the condition of parameter mismatch, and after the voltage feedforward compensation, the current static error is eliminated, and the system stability is also improved, so that the excellent performance of the voltage feedforward compensation method is embodied.

Therefore, the invention only needs to identify the motor inductance parameter value, and the transfer function in the compensation channel is

The method has the advantages of being simple, simple in calculation process, free of parameter setting, capable of reducing the dependency of the current prediction control on the motor parameters on the basis of keeping the advantage of high bandwidth and quick response of the current prediction control, and eliminating current loop output static error and system instability caused by parameter mismatch.