阅读说明：本技术 一种基于扰动观测的永磁同步电机无模型控制方法 (Permanent magnet synchronous motor model-free control method based on disturbance observation ) 是由 张硕 杨楠 张承宁 李雪荣 李雪萍 于 2021-08-13 设计创作，主要内容包括：本发明提供了一种基于扰动观测的永磁同步电机无模型控制方法,其在电流预测控制过程中仅需要调节两个观测器控制参数,而不依赖于任何电机参数,克服了现有技术易受温度、磁场饱和、运行状态等因素造成的参数漂移和模型失配影响的缺点,降低了电流预测控制在参数扰动时的谐波含量,避免了振荡和电流静差,提高了电机控制的鲁棒性,从而能够达到现有技术所不具备的诸多有益效果。(The invention provides a disturbance observation-based model-free control method for a permanent magnet synchronous motor, which only needs to adjust two observer control parameters in the current prediction control process without depending on any motor parameter, overcomes the defects that the prior art is easily influenced by parameter drift and model mismatch caused by factors such as temperature, magnetic field saturation, running state and the like, reduces the harmonic content of current prediction control during parameter disturbance, avoids oscillation and current static error, improves the robustness of motor control, and can achieve a plurality of beneficial effects which cannot be achieved by the prior art.)

1. A permanent magnet synchronous motor model-free control method based on disturbance observation is characterized in that: the method specifically comprises the following steps:

step 1, constructing a motor voltage equation containing system parameter disturbance;

step 2, obtaining a super-local model for model-free current predictive control of the permanent magnet synchronous motor based on the established motor voltage equation;

step 3, constructing a Longbeger disturbance observer and a system state equation without model parameters by taking the current and the system disturbance as state variables; analyzing the stability of the system to obtain the pole of the Roeberg disturbance observer;

and 4, observing system disturbance by using the Roeberg disturbance observer, substituting the system disturbance into the super-local model to calculate to obtain reference voltage, and using the reference voltage for model-free current prediction control.

2. The method of claim 1, wherein: the motor voltage equation constructed in the step 1 specifically adopts the following form:

wherein R is₀、L₀、ψ_f0Nominal parameters, ω, of stator resistance, inductance, and rotor permanent magnet flux linkage, respectively_eIs the electrical angular velocity, i, of the rotor of the machine_d、i_qD, q-axis currents, U, of the motor, respectively_d、U_qD and q-axis voltages, f, of the motor, respectively_d、f_qAre respectively the d-axis system disturbance and the q-axis system disturbance of the motor,representing the differential of the parameter.

3. The method of claim 2, wherein: in step 2, the super-local model is derived from the motor voltage equation to obtain the following form:

wherein, F_d、F_qThe disturbance caused by the unknown variation of the motor system of the d and q axes of the motor respectively, and epsilon is the gain value of the input voltage and is a constant related to the nominal value of the motor stator inductance.

4. The method of claim 3, wherein: the construction process of the Longbeger perturbation observer without the model parameters in the step 3 is as follows:

disturbance f based on d and q axis system_d、f_qThe rate of change being zero, i.e.Under the assumption that d-axis and q-axis currents and system disturbance are selectedAs a system state variable, constructing a system state equation of the Roeberg disturbance observer and performing discretization to obtain:

wherein k is₁、k₂For observer gain, T_sThe upper mark 'Λ' represents the estimated predicted value of the corresponding parameter as the sampling time, k is a certain time,

5. the method of claim 3, wherein: and 3, configuring poles to analyze the stability of the system, wherein a system state characteristic equation is specifically constructed in the following form:

|λI-G|＝-[λ²+(k₁-2)λ+1-k₁-T_sεk₂]²＝0

wherein, λ is the characteristic root of the characteristic equation, i.e. the pole of the system, and I is the identity matrix;

solving the characteristic equation can obtain the pole of the disturbance observer as follows:

distributing the poles within the z-domain unit circle to remain stable and determining k therefrom₁、k₂The respective value ranges.

6. The method of claim 5, wherein: in step 4, specifically, the system disturbance observed by the Longbeige disturbance observer is utilizedAndestimating a predicted value, substituting the predicted value into the super-local model and discretizing, and calculating to obtain the following reference voltages:

wherein the content of the first and second substances,respectively, are the reference voltages of the first and second transistors,respectively reference current, alpha is a constant related to the nominal value of the motor stator resistance.

Technical Field

The invention belongs to the technical field of permanent magnet synchronous motor control, and particularly relates to a method for realizing predictive control on current under the condition of motor parameter mismatch based on a Longbeige disturbance observer and a super-local model.

Background

In the driving of the permanent magnet synchronous motor, the current loop located at the innermost side of the control structure plays a very important role, and directly influences the dynamic and steady-state performance of the motor driving system. At present, predictive current control is gradually becoming the mainstream mode of permanent magnet synchronous motor control due to the advantages of being easy to handle multivariable conditions, fast in dynamic response, easy to include time variables and nonlinearity, and the like. In the dead-beat prediction current control, a d-q axis voltage vector is obtained by predicting a reference current value, feedback stator current and a rotor position, and voltage is added to the permanent magnet synchronous motor after being modulated by the inverter SVPWM, so that actual current can follow the reference current in a control period. However, the permanent magnet synchronous motor control system contains many non-linear factors, such as unavoidable interference and parameter variation under operating conditions, and the above-mentioned dead-beat predictive control also has the disadvantage that the accuracy is too dependent on the control model, the performance of the control system depends on the actual motor to a great extent, and when the motor operates, the motor parameters are changed, which results in parameter drift. At the same time, these parameters are also affected by internal and external unknown disturbances, which results in reduced motor control performance, poor disturbance rejection performance, and low robustness. Mismatch of the electromagnetic parameters may further cause the reference voltage calculated by the predictive current controller to deviate from the desired value.

Disclosure of Invention

Aiming at the technical problems in the prior art, the invention provides a permanent magnet synchronous motor model-free control method based on disturbance observation, which specifically comprises the following steps:

step 1, constructing a motor voltage equation containing system parameter disturbance;

step 2, obtaining a super-local model for model-free current predictive control of the permanent magnet synchronous motor based on the established motor voltage equation;

step 3, constructing a Longbeger disturbance observer and a system state equation without model parameters by taking the current and the system disturbance as state variables; analyzing the stability of the system to obtain the pole of the Roeberg disturbance observer;

and 4, observing system disturbance by using the Roeberg disturbance observer, substituting the system disturbance into the super-local model to calculate to obtain reference voltage, and using the reference voltage for model-free current prediction control.

Further, the motor voltage equation constructed in step 1 specifically takes the following form:

wherein R is₀、L₀、ψ_f0Nominal parameters, ω, of stator resistance, inductance, and rotor permanent magnet flux linkage, respectively_eIs the electrical angular velocity, i, of the rotor of the machine_d、i_qD, q-axis currents, U, of the motor, respectively_d、U_qD and q-axis voltages, f, of the motor, respectively_d、f_qAre respectively the d-axis system disturbance and the q-axis system disturbance of the motor,representing the differential of the parameter.

Further, the super-local model in step 2 is derived from the motor voltage equation to have the following form:

wherein, F_d、F_qThe disturbance caused by the unknown variation of the motor system of the d and q axes of the motor respectively, and epsilon is the gain value of the input voltage and is a constant related to the nominal value of the motor stator inductance.

Further, the construction process of the lobege disturbance observer without the model parameters in step 3 is as follows:

disturbance f based on d and q axis system_d、f_qThe rate of change being zero, i.e.The method comprises the following steps of (1) selecting d-axis current, q-axis current and system disturbance as system state variables, constructing a state equation of the Roeberg disturbance observer, and performing discretization to obtain:

wherein k is₁、k₂For observer gain, T_sThe upper mark 'Λ' represents the estimated predicted value of the corresponding parameter as the sampling time, k is a certain time,

further, the poles are configured in step 3, so that the stability of the system is analyzed, and the system state characteristic equation is specifically constructed in the following form:

|λI-G|＝-[λ²+(k₁-2)λ+1-k₁-T_sεk₂]²＝0

wherein λ is the characteristic root of the characteristic equation, i.e. the poles of the system, and I is the identity matrix.

Solving the characteristic equation can obtain the pole of the disturbance observer as follows:

distributing the poles within the z-domain unit circle to remain stable and determining k therefrom₁、k₂The respective value ranges.

Further, in step 4, specifically, the system disturbance observed by the longberg disturbance observer is usedAndestimating a predicted value, substituting the predicted value into the super-local model and discretizing, and calculating to obtain the following reference voltages:

wherein the content of the first and second substances,respectively, are the reference voltages of the first and second transistors,respectively reference current, alpha is a constant related to the nominal value of the motor stator resistance.

According to the method provided by the invention, only two observer control parameters need to be adjusted in the current prediction control process, and the method does not depend on any motor parameter, so that the defects that the prior art is easily influenced by parameter drift and model mismatch caused by factors such as temperature, magnetic field saturation, running state and the like are overcome, the harmonic content of the current prediction control during parameter disturbance is reduced, oscillation and current static error are avoided, the robustness of motor control is improved, and a plurality of beneficial effects which cannot be achieved by the prior art can be achieved.

Drawings

FIG. 1 is a frame of a PMSM current predictive control system constructed in accordance with the present invention;

FIG. 2 is a general flow chart of the method provided by the present invention;

FIG. 3 shows the results of dead-beat current prediction control performed without motor parameter disturbance at a sampling frequency of 20 kHz;

FIG. 4 is a current prediction control result for implementing the present invention under the condition of no motor parameter disturbance at a sampling frequency of 20 kHz;

FIG. 5 shows the results of dead-beat current prediction control performed at a 20kHz sampling frequency with motor parameter disturbances;

FIG. 6 shows the result of current predictive control implementing the present invention with motor parameter disturbances at a 20kHz sampling frequency.

Detailed Description

The technical solutions of the present invention will be described clearly and completely with reference to the accompanying drawings, and it should be understood that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Fig. 1 shows a block diagram of a current predictive control system that can be used to implement the method provided by the present invention, and the control system mainly comprises a PI speed loop, an improved current loop, a Clarke/Park conversion module, a permanent magnet synchronous motor module, an SVPWM model, and a three-phase voltage inverter. The abc three-phase current of the motor is detected by a current sensor and then is subjected to coordinate transformation to obtain dq-axis current, the rotating speed sensor provides real-time rotating speed signals and rotor position signals, a reference rotating speed is given, and the q-axis reference current can be obtained through PI control of a rotating speed ring. Due to the adoption of i_dAnd (3) performing current prediction control according to an improved current loop to obtain a dq-axis reference voltage acting in the next period, performing coordinate inverse transformation to obtain an alpha beta-axis voltage, outputting the alpha beta-axis voltage to an inverter power module as an input of an SVPWM (space vector modulation) module, and finally driving the motor to operate, wherein the d-axis reference current is 0. The improved current loop is composed of a LongBege disturbance observer and a super-local model-free current prediction controller.

The method applicable to the system provided by the present invention, as shown in fig. 2, specifically includes the following steps:

step 1, constructing a motor voltage equation containing system parameter disturbance;

step 2, obtaining a super-local model for model-free current predictive control of the permanent magnet synchronous motor based on the established motor voltage equation;

step 3, constructing a Longbeger disturbance observer and a system state equation without model parameters by taking the current and the system disturbance as state variables; analyzing the stability of the system to obtain the pole of the Roeberg disturbance observer;

and 4, observing system disturbance by using the Roeberg disturbance observer, substituting the system disturbance into the super-local model to calculate to obtain reference voltage, and using the reference voltage for model-free current prediction control.

Firstly, establishing a mathematical model of the permanent magnet synchronous motor under a synchronous rotating coordinate system:

because the permanent magnet synchronous motor is a nonlinear system, when the motor runs, the parameters of the motor inevitably change to cause disturbance, and the parameter disturbance item needs to be considered when the control robustness of the motor under various working conditions is to be improved:

in the above formulas, Δ R, Δ L, Δ ψ_fFor system parameter disturbances, R₀、L₀、ψ_f0Respectively stator resistance, inductance and rotor permanent magnet flux linkage as the nominal parameter on the motor nameplate, omega_eIs the electrical angular velocity, i, of the rotor of the machine_d、i_qD, q-axis currents, U, of the motor, respectively_d、U_qD and q-axis voltages, f, of the motor, respectively_d、f_qAre respectively the d-axis system disturbance and the q-axis system disturbance of the motor,representing the differential of a parameter, T_d、T_qD-axis noise and q-axis noise of the motor and other unknown interference terms are respectively.

Therefore, the motor voltage equation constructed in step 1 specifically takes the following form:

in step 2, the super-local model is derived from the motor voltage equation to obtain the following form:

wherein, F_d、F_qThe disturbance caused by the unknown variation of the motor system of the d and q axes of the motor respectively, and epsilon is the gain value of the input voltage and is a constant related to the nominal value of the motor stator inductance.

The invention introduces a Roeberg disturbance observer with excellent performance in parameter disturbance observation to observe system disturbance based on d-axis and q-axis system disturbance f_d、f_qRate of change of zeroI.e. byThe d-axis current, the q-axis current and the system disturbance are selected as system state variables to construct a state equation:

accordingly, the Longbeger perturbation observer is constructed as:

discretizing the Longbeige disturbance observer, and considering the sampling time T_sSufficiently small that T can be considered_sR₀/L₀0, and T_sω_eAnd (3) 0, so that the Longberg disturbance observer without model parameters can be constructed:

wherein k is₁、k₂For observer gain, T_sThe upper mark 'Λ' represents the estimated predicted value of the corresponding parameter as the sampling time, k is a certain time,

and 3, configuring poles so as to analyze the stability of the system, wherein a system state characteristic equation is specifically constructed in the following form:

|λI-G|＝-[λ²+(k₁-2)λ+1-k₁-T_sεk₂]²＝0

wherein λ is the characteristic root of the characteristic equation, i.e. the poles of the system, and I is the identity matrix.

Solving the characteristic equation can obtain the pole of the disturbance observer as follows:

distributing the poles within the z-domain unit circle to remain stable and determining k therefrom₁、k₂The respective value ranges.

In step 4, specifically, the system disturbance observed by the Longbeige disturbance observer is utilizedAndestimating a predicted value, substituting the predicted value into the super-local model and discretizing, and calculating to obtain the following reference voltages:

wherein the content of the first and second substances,respectively, are the reference voltages of the first and second transistors,respectively reference current, alpha is a constant related to the nominal value of the motor stator resistance.

Compared with the prior art, the beneficial effects of the invention can be more intuitively embodied by the specific embodiment based on the invention. Fig. 3 and 4 show the simulation results of the dead-beat current prediction control and model-free current prediction control based on the longberg disturbance observer under the condition of no motor parameter disturbance of 20kHz at the same sampling frequency, respectively. The test conditions were such that a speed reference speed value of 1000rpm was given and then the load torque was ramped from 5Nm to 10Nm at t-0.02 s. The first channel shows the reference and actual currents for the q-axis and the second channel shows the actual current for the d-axis. The result shows that the traditional dead-beat prediction current has good dynamic performance when no parameter disturbance exists, but has current static error and poor follow-up performance of q-axis current to reference current; the improved model-free current prediction control based on the Longbeige disturbance observer has no current static difference in the q axis when no parameter disturbance exists, the followability to the reference current is good, and the dynamic and static performances of motor control are improved.

FIGS. 5 and 6 show the perturbation of the motor parameter (L) at 20kHz under the same sampling frequency_s＝2L₀、ψ_f＝2ψ_f0) And then, simulation results of dead beat current prediction control and model-free current prediction control based on a Longbeige disturbance observer are obtained. The test conditions were such that a speed reference speed value of 1000rpm was given and then the load torque was ramped from 5Nm to 10Nm at t-0.02 s. The first channel shows the reference and actual currents for the q-axis and the second channel shows the actual current for the d-axis. The result shows that the traditional dead-beat prediction current is quite sensitive to parameter disturbance, the dq axis current cannot accurately track the reference value, obvious oscillation occurs, the harmonic content is large, and the control performance of the motor is greatly reduced; the improved model-free current prediction control based on the Longbege disturbance observer eliminates steady-state tracking errors of q-axis current under two conditions, has good tracking performance and low harmonic content, and shows strong robustness to changes of machine parameters.

It should be understood that, the sequence numbers of the steps in the embodiments of the present invention do not mean the execution sequence, and the execution sequence of each process should be determined by the function and the inherent logic of the process, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.