Model-free deadbeat predictive current control method based on adaptive current variation component

Abstract

The invention belongs to the technical field of motor driving and control application, and provides a model-free deadbeat predictive current control method based on an adaptive current change component, and the main scheme is as follows: firstly, designing a novel model-free deadbeat predictive mechanism for predicating by using a current change natural response component and unit voltage and current change; the mechanism only relates to sampling current, and no mechanical parameter is needed, so that the mechanism has good parameter robustness; secondly, in order to obtain a current change natural response component and unit voltage current change required by prediction, taking the current change natural response component as a self-adaptive variable, taking the measured current change as a constraint term, and designing a self-adaptive law of the current change natural response component on the premise of pursuing the overall prediction performance; when a prediction error converges to be zero, the optimal proportion of a current change component is obtained. And finally, combining an average filtering strategy with a model-free strategy to suppress the influence of sampling noise on the control performance.

Description

Technical Field

[0001] The present invention relates to the technical field of motor drive and control application, and in particular to a model-free deadbeat prediction current control method based on adaptive current variation components. Background Art

[0002] Permanent magnet synchronous motor (PMSM) is widely used in many fields such as transportation, national defense, agricultural production, industrial manufacturing, etc. due to its simple structure, high reliability and high power density. In recent years, with the rapid development of digital processors, model predictive control (MPC) has received more and more attention. MPC originated from industrial applications and has been applied to many fields such as transportation network control, power transmission system and motor drive. Especially in the field of motor drive, MPC is considered to be an effective alternative to field oriented control and direct torque control.

[0003] Deadbeat predictive current control (DPCC) is a common MPC algorithm. Its core idea is to predict the future current behavior based on the model, and then select the voltage vector that can make the future current closest to the reference current as the output. When there is a large error between the reference current and the actual current of the PMSM, DPCC will force the stator current to be close to the reference current with the maximum driving capacity of the system, so it has good dynamic performance. This control mechanism gives the DPCC algorithm good dynamic performance while making it extremely sensitive to motor parameters. However, in actual applications, affected by motor temperature and magnetic saturation, the model parameters are always fluctuating, and it is difficult to obtain accurate motor parameters through offline identification. In order to solve this problem, scholars have conducted a lot of research.

[0004] At present, common parameter robustness improvement methods include online parameter identification, disturbance compensation, and model-free control based on hyperlocal models. Since there are many parameters to be identified in the prediction model, the online parameter identification method often has the problem of large computational complexity. The disturbance compensation method and the model-free control method based on the hyperlocal model can effectively suppress the current tracking static error caused by the mismatch of inductance and flux parameters, but the current ripple suppression performance caused by inductance mismatch is limited. In contrast, the model-free prediction technology based on current change uses the intrinsic connection between current change and voltage vector to predict current, and has stronger parameter robustness. This strategy is often used in single-vector model prediction current control. For DPCC, the system output vector is synthesized by pulse width modulation of the control vector, and it is difficult to obtain the current change corresponding to the zero vector. Therefore, it is impossible to simply apply the model-free strategy based on current change to DPCC. Summary of the invention

[0005] The purpose of the present invention is to provide a model-free and deadbeat predictive current control method based on an adaptive current change component, aiming to solve the problems of poor robustness of the model predictive control algorithm and large current ripple of the traditional single-vector model-free predictive current control technology, and to ensure the prediction accuracy of the strategy, eliminate the influence of sampling noise on the control performance, and improve the performance of MFDPCC control.

[0006] The present invention solves the technical problem and adopts the following technical solution: The method for model-free deadbeat predictive current control based on adaptive current variation component comprises the following steps: Acquisition of phase current , , And bus voltage , and the sampled phase current is park transformed to obtain dq Axis rotation coordinate system ( k ) Periodic stator current , , read the predicted current And the inverter k -1) Voltage vector of periodic output , and the first ( k -1) The optimal reference voltage vector calculated in the period acts on the inverter and is recorded as ; The first ( k ) cycle sampling current , and the k -1) Cycle sampling current , Make a difference, get The corresponding current change ; Calculate the natural response component of current change by using the adaptive rate of current change component And the current change forced response component ; Calculate the unit voltage and current change considering the sampling error, and convert the adjacent M The current change forced response component within a cycle and its corresponding voltage vector are stored in the array and In the summation, in ( k ) cycle with a new current change to force the response component and its corresponding voltage vector Replace the old value in the array, update the summation result, and then further calculate the unit voltage and current changes; Considering one-beat delay compensation, the current change component is used to predict the future ( k+1) current value at the moment; With reference speed Subtract feedback speed , obtained after calculation by PI controller dq Reference current in axis rotation coordinate system , and set d Axis reference current , and then calculate the first ( k +1) Periodic reference current variation; According to the basic principle of model-free deadbeat predictive current control, the first k +1) cycle optimal reference voltage vector , prepare for the inverter output of the next cycle; Repeat the above steps in sequence to form a closed-loop control.

[0007] As a further optimization, the permanent magnet synchronous motor dq The mathematical model in the axis rotation coordinate system is: , In the formula , They are dq Stator current in the axis rotating coordinate system, , They are dq The stator voltage in the axis rotation coordinate system, L d , L q yes dq Inductance in the axis rotation coordinate system, , R s and They are the flux linkage of the permanent magnet, the stator resistance and the electrical angular velocity of the rotor.

[0008] As a further optimization, the forward Euler method is used to optimize the permanent magnet synchronous motor. dq The mathematical model is discretized in the axis rotation coordinate system, and the following is derived: , in, T is the control period, k Corresponding to ( k ) control cycles, represents the natural response component of the current change, Represents the forced response component of current change to an arbitrary voltage vector; and It is expressed as: , in represents the input voltage vector, and the matrix A( k )、B( k ) and C( k ) is related to the parameters of the motor in the model.

[0009] As a further optimization, the following reference current variation is defined to constrain the permanent magnet synchronous motor: , in, , is the reference current. When the actual current change in the next cycle is equal to the reference current change, the tracking control of the reference current is achieved.

[0010] As a further optimization, the current changes The calculation formula is: , In the formula, , For the first k ) control cycle sampling current, , For the first k -1) sampling current of the control cycle.

[0011] As a further optimization, the current changes naturally in response to the component It can be described in one-dimensional discrete form as: , In the formula, represents the integral of the current prediction error, represents the integral coefficient, Indicates unit voltage and current change; Calculate the unit voltage current change using the current change and its natural response component : .

[0012] As a further optimization, the calculation formula for the unit voltage and current change considering the sampling error is: , in, M Indicates the number of cycles to be summed.

[0013] As a further optimization, the future ( k +1) The current value at the moment is predicted to be: , in, and Indicates the voltage vector acting on the inverter in this cycle, with the superscript p Represents the predicted value.

[0014] As a further optimization, the k +1) The calculation formula for the reference current change during the cycle is: .

[0015] As a further optimization, the optimal reference voltage vector The calculation formula is: .

[0016] The beneficial effects of the present invention are: 1. The present invention uses the sampled current to calculate all the information needed to predict the optimal voltage vector, completely getting rid of the dependence of the traditional deadbeat control algorithm on the current model. At the same time, this method avoids the cumbersome motor parameter determination process, is simple to debug, and improves the reliability of the PMSM drive system.

[0017] 2. The present invention develops a model-free current control technology based on adaptive current change components, which expands the model-free technology based on sampled current to the field of zero-beat predictive current control. Compared with traditional model-free predictive current control technology, it greatly reduces current ripple and improves the quality of the three-phase current of the motor.

[0018] 3. The present invention takes into account the influence of sampling noise on the calculation of current change forced response, develops an effective noise suppression method, and improves the prediction accuracy of the optimal voltage vector.

[0019] 4. The present invention is obviously different from the DPCC algorithm based on parameter identification. The current change component adaptation rate developed by the present invention is designed based on the current prediction error, aiming to improve the overall prediction accuracy, rather than to accurately identify a certain parameter. Thanks to this, the method also has a certain inhibitory effect on the sixth harmonic of the current caused by the nonlinearity of the inverter. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a block diagram of the model-free deadbeat predictive current control principle in an embodiment of the present invention; Figure 2 The characteristic equation of the natural response adaptive rate of current change in the embodiment of the present invention is different β The distribution of extreme points under ; Figure 3 This is a process for implementing unit voltage and current change sampling noise suppression in an embodiment of the present invention; Figure 4 This is a flow chart of MFDPCC control in an embodiment of the present invention; Figure 5In the embodiment of the present invention, DPCC is 0.3 L s Experimental waveform under parameters; Figure 6 The DPCC based on the model in the embodiment of the present invention is 3 L s Experimental waveform under parameters; Figure 7 1 is an experimental waveform of a DPCC under nominal parameters based on a model in an embodiment of the present invention; Figure 8 This is the MFDPCC experimental waveform in the embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations. Example

[0022] Figure 1 The schematic diagram of the model-free deadbeat predictive current control principle in this embodiment includes a speed outer loop PI regulator, a current change natural response adaptive module, and adjacent M cycle voltage vector summation module, adjacent M The current change forced response summation module corresponding to the periodic voltage vector, the unit voltage and current change calculation module, the delay compensation and optimal voltage vector prediction module, the inverter, the current sensor, the voltage sensor and the permanent magnet synchronous motor. Therefore, the present embodiment provides a model-free deadbeat predictive current control method based on adaptive current change components, comprising the following steps: Step 1: Initialize the parameters related to algorithm execution as required.

[0023] Specifically, the following two points need special attention in the parameter initialization process: (1) Natural response component of current change and current change forced response component Cannot be initialized to zero at the same time. and There is no such restriction. This is because d The reference current and actual current of the axis are both zero at startup. and At the same time, when it is zero, the current prediction error is also zero, and the adaptive mechanism startup is prone to startup failure. Although the sampled current may not always remain at zero in actual implementation, this situation should be avoided to ensure reliable startup of the system.

[0024] (2) Array and All data in an array must be initialized to zero. and The data in each M All control cycles are updated once to ensure that only adjacent M The data in the cycle is added. If and If the initial data in is not zero, then the first 2 M The summation result of the cycles is inaccurate, which will cause the motor to fail to start.

[0025] Step 2: Collect phase current through current sensor and voltage sensor , , And bus voltage , and perform park transformation on the sampled phase current to obtain dq Axis rotation coordinate system ( k ) Periodic stator current , , read the predicted current And the inverter k -1) Voltage vector of periodic output , and the first ( k -1) The optimal reference voltage vector calculated in the period acts on the inverter and is recorded as ; Here, the three-phase current collected by the current sensor is transformed into dq Axis rotation coordinate system: , in, θ is the rotor electrical angle, and the voltage vectors are synthesized through space voltage vector pulse width modulation (SVPWM).

[0026] Step 3: k ) cycle sampling current , and the k -1) Cycle sampling current , Make a difference, get The corresponding current change , this process can be written as: , In the formula, , For the first k ) control cycle sampling current, , For the first k -1) sampling current of the control cycle.

[0027] It is worth noting that the current change includes two parts, as follows: , In the formula Represents the natural response of current changes, mainly caused by back EMF; It represents the forced response of current change to any voltage vector, which is mainly controlled by the input voltage. By separating the natural response and forced response in the current change, the future optimal voltage vector can be simply predicted based on the reference current change.

[0028] Step 4: Calculate the natural response component of current change using the adaptive rate of current change component And the current change forced response component .

[0029] For simplicity of description, the natural response component of current change is The one-dimensional discrete form of the adaptive rate can be expressed as: , In the formula, represents the integral of the current prediction error, represents the integral coefficient, Represents the unit voltage and current change; using this adaptive rate in the first ( k ) cycles to estimate the required and .

[0030] The unit voltage current change is then calculated using the current change and its natural response component : , Natural response component of current change is a free term that changes according to the system operating state, and the measured current changes yes and When the motor is running in steady state, the DPCC in each cycle There won’t be too much fluctuation, so Will follow changes and reaches a stable state.

[0031] The value range of the integral coefficient in the adaptive rate described in step 4 can be determined in the following way: first, the adaptive rate of the natural response component of the current change is z The transformed closed-loop transfer function is: , The characteristic equation of the transfer function is: , Further solving the system characteristic equation gives The extreme point: , different β Value z The distribution of the pole positions in the domain is as follows Figure 2 As shown. β When it is less than 0, at most one pole is inside the unit circle; when β When it is greater than 10000, both poles are outside the unit circle. According to the principle of automatic control, the necessary and sufficient condition for the stability of a discrete system is that all characteristic roots are distributed within the unit circle. Figure 2 It can be seen that when β When the value of is in the range of 0 to 10000, the system is stable.

[0032] The above adaptive strategy is implemented on the basis of sampling current. However, in the actual current sampling process, sampling noise is inevitable. This noise will be brought into the above adaptive strategy, thereby further deteriorating the prediction of delay compensation current and reference voltage vector. Therefore, it is necessary to suppress the sampling noise.

[0033] Step 5: Calculate the unit voltage and current change considering the sampling error. M The current change forced response component within a cycle and its corresponding voltage vector are stored in the array and When a new control cycle comes, a new current change is needed to force the response component and its corresponding voltage vector Replace the old values ​​in the array and update the sum result, such as Figure 3 As shown, the unit voltage and current change considering the sampling error is then further calculated.

[0034] Specifically, based on the principle of mean filtering, the expression of unit voltage and current change is improved as follows: , Among them M Indicates the number of cycles to be summed.

[0035] The above formula can be further rewritten as: , When the summation period is large enough, the following equation holds: , Will( k-2 ) cycle unit voltage and current change expression is substituted into ( k-1 ) cycle, we can get: , in It can be seen from the formula that the unit voltage and current change after improvement is the result of low-pass filtering the unit voltage and current change before improvement, so the influence of sampling noise can be weakened. M The larger the value of λ The smaller the value of, the better the noise suppression performance. The sampling noise suppression process will move forward in time, and the adjacent M Cycle and its corresponding voltage vector Will be stored in the array and In the summation, M The choice of DSP requires a comprehensive consideration of the DSP's memory usage and noise suppression performance.

[0036] Step 6: Consider one-beat delay compensation and use the current change component to predict the future ( k +1) current value at the moment.

[0037] Specifically, the ( k +1) The current value at the moment can be predicted as: , in, and Indicates the voltage vector acting on the inverter in this cycle, with the superscript p Represents the predicted value.

[0038] Step 7: Use reference speed Subtract feedback speed , obtained after calculation by PI controller dq Reference current in axis rotation coordinate system , and set d Axis reference current , and then calculate the first ( k +1) cycle reference current variation.

[0039] Specifically, the reference current change can be calculated by the following formula: , Because thek-2 ) cycle and the reference current of the first ( k ) cycle is approximately equal, so it can be used and Alternative.

[0040] Step 8: Calculate the first ( k +1) cycle optimal reference voltage vector , preparing for the inverter output of the next cycle.

[0041] Specifically, the optimal reference voltage vector calculation formula can be expressed as: , It is not difficult to see from the above formula that the calculation process of the optimal reference voltage vector does not involve any motor mechanical parameters, so it has strong parameter robustness.

[0042] Step 9: Repeat steps 2 to 8 in sequence to form a closed-loop control. The above MFDPCC control process is as follows Figure 4 As shown, it is only necessary to Make a non-zero judgment, because after startup, The value of must be much greater than zero, which is sufficient to meet the calculation requirements of unit voltage and current changes and will not cause the calculation error to be amplified.

[0043] Finally, the embodiment is verified by experiments. Experimental data are collected by oscilloscope, and then the collected data are imported into MATLAB workspace, and then the experimental waveform is obtained by using From Workspace module in Simulink. Figure 5-7 They are respectively the DPCC based on the model at a speed of 600r / min at 0.3 L s , 3 L s As well as the experimental results under nominal parameters. It can be seen that the DPCC of the model has good current control performance under nominal parameters, and the phase current THD is 1.76%. L s Under inductor mismatch, the model-based DPCC only has a slight tracking error, and the phase current THD increases slightly to 3.37%. However, under the condition of three times the rated inductance, dq The ripple content in the shaft current increases greatly, and the THD increases sharply to 18%. However, the MFDPCC proposed in this embodiment is not affected by the motor parameters, and always maintains the optimal current tracking performance and ripple control performance, with a THD of only 1.25%. Figure 8 shown.

[0044] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A model-free deadbeat predictive current control method based on adaptive current variation component, characterized in that: The steps include: Acquisition of phase current , , And bus voltage , and perform park transformation on the sampled phase current to obtain dq Axis rotation coordinate system ( k ) Periodic stator current , , read the predicted current And the inverter k -1) Voltage vector of periodic output , and the first ( k -1) The optimal reference voltage vector calculated in the period acts on the inverter and is recorded as ; The first ( k ) cycle sampling current , and the k -1) Cycle sampling current , Make a difference, get The corresponding current change ; Calculate the natural response component of current change by using the adaptive rate of current change component And the current change forced response component ; Calculate the unit voltage and current change considering the sampling error, and convert the adjacent M The current change forced response component within a cycle and its corresponding voltage vector are stored in the array and In the summation, in ( k ) cycle with a new current change to force the response component and its corresponding voltage vector Replace the old value in the array, update the summation result, and then further calculate the unit voltage and current changes; Considering one-beat delay compensation, the current change component is used to predict the future ( k +1) current value at the moment; With reference speed Subtract feedback speed , obtained after calculation by PI controller dq Reference current in axis rotation coordinate system , and set d Axis reference current , and then calculate the first ( k +1) Periodic reference current variation; According to the basic principle of model-free deadbeat predictive current control, the first k +1) cycle optimal reference voltage vector , prepare for the inverter output of the next cycle; Repeat the above steps in sequence to form a closed-loop control.

2. The method for model-free deadbeat predictive current control based on adaptive current variation component according to claim 1, characterized in that: Permanent magnet synchronous motor dq The mathematical model in the axis rotation coordinate system is: ， In the formula , They are dq Stator current in the axis rotating coordinate system, , They are dq The stator voltage in the axis rotation coordinate system, L d , L q yes dq Inductance in the axis rotation coordinate system, , R s and They are the flux linkage of the permanent magnet, the stator resistance and the electrical angular velocity of the rotor.

3. The method for model-free deadbeat predictive current control based on adaptive current variation component according to claim 2, characterized in that: The forward Euler method is used to calculate the permanent magnet synchronous motor dq The mathematical model is discretized in the axis rotation coordinate system, and the following is derived: ， in, T is the control period, k Corresponding to ( k ) control cycles, represents the natural response component of the current change, Represents the forced response component of current change to an arbitrary voltage vector; and It is expressed as: in represents the input voltage vector, and the matrix A( k )、B( k ) and C( k ) is related to the parameters of the motor in the model.

4. The method for model-free deadbeat predictive current control based on adaptive current variation component according to claim 2 or 3, characterized in that: The following reference current variation is defined to constrain the permanent magnet synchronous motor: , in, , is the reference current. When the actual current change in the next cycle is equal to the reference current change, the tracking control of the reference current is achieved.

5. The method for model-free deadbeat predictive current control based on adaptive current variation component according to claim 1, characterized in that: The current changes The calculation formula is: , In the formula, , For the first k ) control cycle sampling current, , For the first k -1) sampling current of the control cycle.

6. The method for model-free deadbeat predictive current control based on adaptive current variation component according to claim 1, characterized in that: The current variation natural response component It can be described in one-dimensional discrete form as: , In the formula, represents the integral of the current prediction error, represents the integral coefficient, Indicates unit voltage and current change; Calculate the unit voltage current change using the current change and its natural response component : 。 7. The method for model-free deadbeat predictive current control based on adaptive current variation component according to claim 1, characterized in that: The calculation formula for the unit voltage and current change considering the sampling error is: , in, M Indicates the number of cycles to be summed.

8. The method for model-free deadbeat predictive current control based on adaptive current variation component according to claim 1, characterized in that: The future ( k +1) The current value at the moment is predicted to be: , in, and Indicates the voltage vector acting on the inverter in this cycle, with the superscript p Represents the predicted value.

9. The method for model-free deadbeat predictive current control based on adaptive current variation component according to claim 1, characterized in that: The said k +1) The calculation formula for the reference current change during the cycle is: 。 10. The method for model-free deadbeat predictive current control based on adaptive current variation component according to claim 1, characterized in that: The optimal reference voltage vector The calculation formula is: 。

Images