阅读说明：本技术 交替执行采样与控制程序的双三相永磁同步电机控制方法 (Double three-phase permanent magnet synchronous motor control method for alternatively executing sampling and control program ) 是由 王政 顾珉睿 于 2021-01-12 设计创作，主要内容包括：本发明公开了交替执行采样与控制程序的双三相永磁同步电机控制方法,属于发电、变电或配电的技术领域。通过将电机两套绕组的采样时刻、装载矢量时刻和参考值追踪时刻相互交替一半采样周期,本发明在不改变单套三相绕组采样频率的前提下,使得电机控制系统的等效采样频率翻倍,控制延迟和预测时间跨度减半。此外,本发明通过双层模型预测控制策略,解决了交替执行采样与控制程序的双三相永磁同步电机控制方法导致的系统受控维度下降为二维而控制目标仍是四维的欠驱动问题。本发明提供的交替执行采样与控制程序的双三相永磁同步电机控制方法有效提高了双三相永磁同步电机系统的稳态和动态控制性能,并降低了控制算法的计算量。(The invention discloses a control method of a double three-phase permanent magnet synchronous motor for alternately executing sampling and control programs, and belongs to the technical field of power generation, power transformation or power distribution. By mutually alternating the sampling time, the loading vector time and the reference value tracking time of two sets of windings of the motor by a half sampling period, the equivalent sampling frequency of a motor control system is doubled on the premise of not changing the sampling frequency of a single set of three-phase windings, and the control delay and the prediction time span are halved. In addition, the invention solves the problem that the system controlled dimension is reduced to two dimensions and the control target is still four-dimensional under-actuation caused by the control method of the double three-phase permanent magnet synchronous motor which alternately executes the sampling and control program through a double-layer model predictive control strategy. The double three-phase permanent magnet synchronous motor control method for alternately executing the sampling and control program effectively improves the stable state and dynamic control performance of a double three-phase permanent magnet synchronous motor system and reduces the calculated amount of a control algorithm.)

1. The method for controlling the double three-phase permanent magnet synchronous motor alternately executing the sampling and control programs is characterized in that the method is used for controlling the double three-phase permanent magnet synchronous motor at kT_sThe phase current of the second set of windings is sampled at time (k +1/2) T_sPhase current of a first set of windings is sampled at a time and control vectors of a second set of windings are loaded, at (k +1) T_sLoad control vectors of the first set of windings at a time and complete the second setTracking of current reference by sheath winding at (k +3/2) T_sThe method comprises the steps of finishing tracking of a first set of windings on a current reference value constantly, determining a d-axis current reference value and a q-axis current reference value of each set of windings by adopting a double-layer model prediction control strategy which aims at d-axis current and q-axis current of the two sets of windings to equally divide a d-axis total current, q-axis total current and d-axis total current and q-axis total current following reference values, selecting candidate vectors for the set of windings from inverter control voltage vectors which never cause the inverter output level to jump between a P level and an N level under the constraint of the d-axis current reference value and the q-axis current reference value of each set of windings, calculating a value function of each candidate vector after the candidate vector acts on an inverter, and obtaining the candidate vector of a minimum value function as a final control vector, wherein k is a positive integer, T is a T, and_sis a switching cycle.

2. The dual three-phase pmsm control method of claim 1, wherein the method of selecting candidate vectors for each set of windings from among the inverter control voltage vectors that do not cause the inverter output level to jump between the P level and the N level under the constraints of the d-axis and q-axis current reference values for the set of windings is:

when at least one candidate vector is extrapolated, and each controlled quantity is within a control error range, deleting the candidate vectors which cannot be extrapolated, and taking the switching frequency calculated according to the extrapolation result as a value function;

and when all the candidate vectors cannot be extrapolated, taking the weighted result of the absolute value of the control error of each controlled quantity as a cost function.

3. The method of claim 1, wherein the vector of inverter control voltages that do not cause the inverter output level to jump between the P level and the N level is selected by: calculating d-axis reference voltage and q-axis reference voltage of one set of windings according to d-axis current reference values and q-axis current reference values of the two sets of windings, performing polar coordinate transformation on the d-axis reference voltage and the q-axis reference voltage of the set of windings, selecting a vector of a triangular sector where an inverter reference voltage vector corresponding to the set of windings is located as a candidate vector according to a polar coordinate transformation result, when at least one candidate vector does not cause the inverter output level to jump between a P level and an N level, all the candidate vectors are inverter control voltage vectors which do not cause the inverter output level to jump between the P level and the N level, and when all the candidate vectors cause the inverter output level to jump between the P level and the N level, excluding vectors in the triangular sector and the hexagonal sector simultaneously in a hexagonal area where the inverter reference voltage vector corresponding to the set of windings is located, the remaining vectors in the hexagonal area are selected as inverter control voltage vectors that do not cause the inverter output level to jump between the P-level and the N-level.

4. The method of claim 1, wherein the voltage reference values targeting d-axis current sharing of the two sets of windings to the d-axis total current and q-axis current sharing of the two sets of windings to the q-axis total current are:wherein u is_d1 ^*、u_q1 ^*、u_d2 ^*And u_q2 ^*Is a d-axis reference voltage, a q-axis reference voltage, i, of two windings respectively_d1 ^*＝i_d2 ^*＝(1/2)i_d ^*，i_q1 ^*＝i_q2 ^*＝(1/2)i_q ^*，i_d1 ^*、i_q1 ^*、i_d2 ^*、i_q2 ^*D-axis reference current and q-axis reference current i of two windings respectively_d ^*、i_q ^*Is d-axis total reference current, q-axis total reference current, i_d1(k+1)、i_q1(k+1)、i_d2(k +1) and i_q2(k +1) is d-axis current and q-axis current of the two windings in the (k +1) th switching period respectively,L_dand L_qAre d-axis and q-axis inductances, R_sIs stator resistance, L_dmAnd L_qmD-axis and q-axis mutual inductance, omega (k) being the electrical angular velocity, psi, of the kth switching period_fIs a permanent magnetic linkage.

5. The double three-phase permanent magnet synchronous motor control method alternately executing a sampling and control program according to claim 4,

the reference values of the first set of winding currents with the d-axis total current and the q-axis total current following reference values as targets are as follows:

the expression of the reference value of the second set of winding current with the d-axis total current and the q-axis total current as the target is as follows:

wherein i_d1(k+3/2)、i_q1(k +3/2) is d-axis current, q-axis current, i of the first set of windings in the (k +3/2) th switching period_d2(k+1)、i_q2(k +1) is the d-axis current, q-axis current of the second set of windings during the kth switching period.

6. The method as claimed in claim 2, wherein the switching frequency is calculated asg₁Is the switching frequency, n_sTo be from kT_sThe time-loaded vector switches to (k +1) T_sThe number of on-off switching times of the candidate vector at the moment, N being the extrapolation step,S_x(k+1)、S_x(k) is (k +1) T_sTime of day, kT_sThe switching function of phase x, x being a, b, c, d, e, f,

7. the method for controlling a double three-phase permanent magnet synchronous motor alternately executing a sampling and control program according to claim 2, wherein the expression for calculating the weighted result of the absolute value of the control error of each controlled quantity is:g₂for each controlled quantity controlling the weighted result of the absolute value of the error, g_xControlling the absolute value of the error, λ, for a controlled quantity, x_xIs g_xThe weight of (a) is determined,x (k +2) is (k +2) T_sPredicted value, x, of the time-dependent variable x_max、x_minMaximum and minimum values allowed for the controlled quantity x, x_max＝x_ref+Δx，x_min＝x_refΔ x, Δ x being the controlled quantity x at (k +2) T_sPredicted value and reference value x of time_refMaximum allowable absolute error of i_d1、i_q1D-axis current, q-axis current, i, for the first set of windings_d2、i_q2D-axis current, q-axis current, V, for the second set of windings_nIs the inverter DC bus midpoint voltage; when sampling and controlling the first set of windings, x ∈ { i ∈ } [ i ]_d1,i_q1,V_n}; when sampling and controlling the second set of windings, x ∈ { i }_d2,i_q2,V_n}。

8. The system for realizing the control method of the double three-phase permanent magnet synchronous motor according to claim 1 is characterized by comprising the following steps:

the position encoder is arranged on an output shaft of the double three-phase permanent magnet synchronous motor and used for detecting the position of the output shaft of the motor;

the input end of the rotating speed calculation module receives the angle position information and outputs the rotating speed;

the input end of the speed closed-loop PI regulator is connected with a rotating speed and a rotating speed reference value, and a torque reference value is generated according to a rotating speed difference value;

the input end of the maximum torque current ratio control module is connected with a torque reference value, and a d-axis current reference value and a q-axis current reference value of the first set of windings and a d-axis current reference value and a q-axis current reference value of the second set of windings are output;

the double-rotation coordinate transformation module is used for receiving six-phase current output by the double three-phase permanent magnet synchronous motor, performing rotation transformation on the six-phase current and outputting the current time kT_sD-axis current and q-axis current of two sets of windings at the moment;

a current prediction module with its input connected with the output of the dual-rotation coordinate transformation module for predicting (k +1) T_sD-axis current and q-axis current of the two sets of windings and neutral point potential of a direct current bus of the inverter are output at the moment;

one input end of the dead beat control module is connected with the output end of the current prediction module, the other input end of the dead beat control module is connected with the output end of the maximum torque-current ratio control module, and d-axis reference voltage and q-axis reference voltage of the first set of winding and d-axis reference voltage and q-axis reference voltage of the second set of winding are calculated and output according to d-axis current of the two sets of windings and reference values and predicted values of q-axis current;

the input end of the polar coordinate transformation module is connected with the output end of the dead beat control module based on the motor discrete model, and the polar coordinate transformation module is used for carrying out polar coordinate transformation on the d-axis reference voltage and the q-axis reference voltage of the first set of winding and the d-axis reference voltage and the q-axis reference voltage of the second set of winding;

multi-step model predictive control module at kT_sThe phase current of the second set of windings is sampled at time (k +1/2) T_sPhase current of a first set of windings is sampled at a time and control vectors of a second set of windings are loaded, at (k +1) T_sLoading the control vector of the first set of windings at a time and completing the tracking of the current reference value by the second winding, at (k +3/2) T_sThe tracking of the current reference value by the first set of windings is completed at any moment, and d-axis electricity with two sets of windings is adoptedThe method comprises the steps of determining a d-axis current reference value and a q-axis current reference value of each set of winding according to a double-layer model predictive control strategy with the objective of current-sharing d-axis total current, q-axis current sharing q-axis total current of two sets of windings and following reference values of the d-axis total current and the q-axis total current, selecting candidate vectors for the set of windings from inverter control voltage vectors which never cause the inverter output level to jump between a P level and an N level under the constraint of the d-axis current reference value and the q-axis current reference value of each set of windings, and calculating a cost function after each candidate vector acts on the inverter so as to obtain the candidate vector of the minimum cost function as a final control vector.

9. The system of the dual three-phase permanent magnet synchronous motor control method according to claim 8, wherein the multi-step model predictive control module stores a computer program, and the computer program when executed by a processor implements the dual three-phase permanent magnet synchronous motor control method of claim 1 that alternately executes a sampling and control program.

Technical Field

The invention relates to electrician, motor and power electronic technology, in particular to a control method of a double three-phase permanent magnet synchronous motor for alternatively executing a sampling and control program, belonging to the technical field of power generation, power transformation or power distribution.

Background

In recent years, in the field of medium-voltage high-power application, a multi-level driving system of a multi-phase motor draws more and more attention due to the advantages of low stress of power devices, high reliability, high power quality and the like. The double three-phase permanent magnet synchronous motor driving system has the advantages of simple structure, six-time torque ripple elimination, strong fault tolerance capability and the like, and becomes a novel motor driving system with great prospect. In order to improve the efficiency of the power electronic converter and reduce the capacity of the auxiliary heat dissipation system, the switching frequency of the power device of the conventional medium-voltage high-power motor driving system is generally lower, and is usually lower than 1 kHz. The lower switching frequency causes the following problems in the prior vector control based on Proportional Integral (PI) controller: (1) the influence of the delay problem in digital control is aggravated due to low switching frequency, the coupling effect of a motor dq axis component is aggravated, and a modeling mode of a traditional first-order inertia link has large errors, so that the steady-state and dynamic control performance of a motor system is obviously influenced; (2) the low switching frequency causes the output harmonic content of the power electronic converter to be increased, and the control performance is reduced; (3) the problem of large digital delay caused by low switching frequency reduces the bandwidth of the traditional vector control system and influences the stability of the control system. The double three-phase motor driving system is provided with two sets of three-phase motor windings and a power electronic converter, has higher control freedom degree compared with the traditional three-phase motor system, provides an optimized space for improving the control performance of the motor, and aims to overcome the defect of the existing control method under the working condition of low switching frequency by alternately executing the sampling and control program of the double three-phase permanent magnet synchronous motor.

Disclosure of Invention

The invention aims to provide a double three-phase permanent magnet synchronous motor control method for alternately executing a sampling and control program aiming at the defects of the background technology, and the integral equivalent sampling frequency of a motor control system is doubled on the premise of not changing the sampling frequency of a single set of three-phase winding by alternately executing the sampling and control program of double three-phase windings and corresponding power electronic converters, so that the steady-state and dynamic control performance of the motor control system under the working condition of low switching frequency is effectively improved, and the technical problems of control delay and control performance reduction under the working condition of low switching frequency of the existing control method are solved.

The invention adopts the following technical scheme for realizing the aim of the invention:

a control method of a double three-phase permanent magnet synchronous motor for alternately executing sampling and control programs comprises the following steps:

1) the motor speed value, the phase current value and the DC bus midpoint potential value of the calculator are acquired and calculated through a motor position encoder, a phase current sensor and a DC bus voltage sensor, and d is obtained through a double-rotation coordinate transformation module₁q₁Axial current component sum d₂q₂The value of the shaft current component;

2) performing first-order forward Euler discretization on a motor mathematical model, and obtaining d based on the discrete mathematical model₁q₁Axial current, d₂q₂The shaft current and the bus midpoint potential compensate for a sampling value of one-beat control delay;

3) according to a torque reference value obtained by a rotating speed closed-loop PI controller, a dq axis current reference value is obtained and averaged based on a vector control method, and then d is obtained₁q₁Shaft current sum d₂q₂A shaft current reference value;

4) based on the principle of dead-beat control, according to d₁q₁Shaft, d₂q₂Calculating the reference value and sampling value of shaft current₁q₁Shaft and d₂q₂An axis reference voltage vector;

5) d is selected by combining the distribution characteristics of voltage space vectors of the three-level inverter₁q₁Shaft and d₂q₂Four or five voltage vectors of a triangular sector where the axis reference voltage vector is located are used as candidate vectors for model prediction control;

6) predicting the control effect of the candidate vector in the step 5) based on a point potential discrete model of the direct current bus of the motor and the inverter, and selecting the candidate vector with the optimal control effect as a final loading vector according to a value function.

Furthermore, if the inverter can switch between the P level and the N level when switching from the current switching state to the switching state corresponding to the candidate vector in step 5), the candidate qualification of the candidate vector is cancelled; further, if all the candidate vectors in the step 5) are disqualified as candidates, selecting the voltage vector of the hexagonal large sector where the voltage reference vector is located as a new candidate vector, and screening out a vector which does not cause switching of the P level and the N level of the inverter from the new candidate vector as a final candidate vector;

further, the cost function described in step 6) is divided into two cases: in the first case, at least one candidate vector is present, under the action of which the predicted values of all controlled quantities are within the allowed error range, the current trajectory is subjected to a linear extrapolation, when the current trajectory reaches the allowed error boundary, the extrapolation is completed, and the average switching frequency is calculated as a cost function on the basis of the extrapolation step size and the number of switching actions resulting from switching to the candidate vector; in the second case, no candidate vector meets the condition of the first case, and the cost function is obtained by weighted summation of the control error absolute values of all controlled quantities;

furthermore, the candidate vector screening and evaluating processes described in step 4), step 5) and step 6) are performed alternately in two sets of windings, when the first set of windings is sampled, the second set of windings loads a new vector, and until the first set of windings is sampled next time, the switch state corresponding to the second set of windings remains unchanged, when the second set of windings is sampled, the first set of windings loads a new vector, and until the second set of windings is sampled next time, the switch state corresponding to the first set of windings remains unchanged; in order to solve the problem of under-actuation caused by the method of alternately executing the sampling and control program, a double-layer model predictive control strategy is adopted in the steps 4), 5) and 6), a four-dimensional control target is divided into two layers, and the steps 4) and 5) form first-layer model predictive control, wherein dead-beat control is performed by d₁q₁Shaft and d₂q₂The current reference value of the shaft current tracking average is taken as a control target; and step 6) is second-layer model prediction control, the control objects of the second-layer model prediction control are double-winding d-axis total current and q-axis total current, and the control target is to enable the double-winding d-axis total current and the double-winding q-axis total current to follow the double-winding d-axis total current reference value and the double-winding q-axis total current reference value obtained by the maximum torque-current ratio module.

By adopting the technical scheme, the invention has the following beneficial effects:

(1) according to the method, the equivalent sampling frequency of the motor control system is doubled and the digital delay is halved by alternately executing the sampling and control programs, meanwhile, the prediction time span of the candidate vector control effect in model prediction control is halved, the prediction precision is improved, and therefore the steady-state and dynamic performances of motor control are effectively improved;

(2) according to the invention, by means of alternately executing the sampling and control programs, the dimensionality of the motor control system is reduced from four dimensions to two dimensions, so that the calculated amount required by the adopted model prediction control is effectively reduced;

(3) according to the method, a double-layer model prediction control strategy is adopted, an equalized reference current value is taken as a control target in the first-layer control to determine a candidate vector so as to ensure that the candidate vector can ensure that the currents of two windings are basically uniformly divided, d-axis total current and q-axis total current are taken as the control target in the second-layer control, and a final loading vector is selected from the candidate vector, so that the control of the d-axis total current and the q-axis total current under the premise of ensuring that the currents of the two windings are basically uniformly divided is realized, the problem of underdrive that the controlled dimensionality of a system is reduced to two dimensions and the control target is still four dimensions caused by the method of alternately executing a sampling and control program is solved, and the control of the two-dimensional controlled quantity on the four-dimensional control;

(4) by comprehensively adopting the control method of the double three-phase permanent magnet synchronous motor for alternately executing the sampling and control programs, the equivalent sampling frequency of the motor driving system is higher, the control delay is shortened, and the problem of low control precision caused by high control delay under the working condition of low switching frequency in the traditional control method is effectively solved.

Drawings

FIG. 1 is a block diagram of a multi-step model predictive control of a dual three-phase PMSM with alternate sampling and control routines;

the system comprises a speed closed-loop PI regulator 1.1, a maximum torque-current ratio control module 1.2, a dead-beat control module based on a motor discrete model 1.3, a polar coordinate transformation module 1.4, a multi-step long model prediction control module 1.5, a direct-current bus 1.6, a double three-phase three-level inverter 1.7, a double three-phase permanent magnet synchronous motor 1.8, a position encoder 1.9, a double rotation coordinate transformation module 1.10, a current prediction module 1.11 and a rotating speed calculation module 1.12.

Fig. 2 is a flowchart of a multi-step long model predictive control method.

FIG. 3 is a schematic diagram of a voltage space vector distribution diagram of a first set of three-level inverters in a dual three-phase PMSM drive system and a candidate vector triangle region method selection;

wherein, V_ref1Is the reference voltage vector for the inverter for the first set of windings.

FIG. 4 is a schematic diagram of a voltage space vector distribution diagram and a candidate vector triangle region method selection manner of a second set of three-level inverters in a dual three-phase permanent magnet synchronous motor driving system;

wherein, V_ref2Is the reference voltage vector for the inverter for the second set of windings.

Fig. 5 is a schematic diagram of a conventional synchronous control dual three-phase inverter candidate vector intersection traversal.

Fig. 6 is a schematic diagram of a voltage space vector distribution diagram and a candidate vector hexagon region method selection mode of a first set of three-level inverters in a double three-phase permanent magnet synchronous motor driving system.

Fig. 7 is a schematic diagram of a voltage space vector distribution diagram and a candidate vector hexagonal region method selection mode of a second set of three-level inverters in a double three-phase permanent magnet synchronous motor driving system.

FIG. 8 is q₁An example graph of the axis current multi-step model predictive control.

Fig. 9 is a prediction control timing chart of the synchronous prediction control method.

Fig. 10 is a predictive control timing diagram of the method of alternately executing a sampling and control routine.

FIG. 11 is a waveform diagram of current experiment of motor A, B, D and E in a conventional synchronous control method, i_A,i_B,i_DAnd i_EThe experimental waveforms for motor A, B, D and E-phase current, respectively.

Fig. 12 is a waveform diagram of an experiment of phase currents of the motor A, B, D and E under the method of alternately executing the sampling and control program.

Fig. 13 is an experimental waveform diagram of the bus capacitor voltage and the bus midpoint potential obtained by the conventional synchronous control method.

Fig. 14 is an experimental waveform diagram of the bus capacitor voltage and the bus midpoint potential obtained by alternately executing the sampling and control program method.

Fig. 15 is a waveform of an experiment of the rotation speed and the torque obtained by the conventional synchronous control method.

Fig. 16 is a rotational speed and torque experimental waveform obtained by alternately executing the sampling and control program method.

Fig. 17 is a waveform of a rotational speed and torque dynamic response experiment obtained by a conventional synchronous control method.

FIG. 18 is an amplified waveform of the rotational speed and torque dynamic response experimentally obtained for the conventional synchronous control method.

FIG. 19 is a graph of experimental rotational speed and torque dynamic response waveforms obtained by alternately executing the sampling and control routine method.

FIG. 20 is an amplified waveform of the rotational speed and torque dynamic response obtained by alternately performing sampling and control routine method experiments.

Fig. 21 shows the program run time of the experiment of the conventional synchronous control and alternate execution sampling and control program method.

Detailed Description

The inventive concept will be further clarified by the following detailed description taken in conjunction with the accompanying drawings, which are intended to illustrate and not limit the scope of the invention, and various equivalent modifications made by those skilled in the art after reading this disclosure will fall within the scope defined by the appended claims.

The invention provides a double three-phase permanent magnet synchronous motor control method for alternately executing sampling and control programs, which aims at the problems of control delay and control performance reduction of the existing control method under the working condition of low switching frequency.

The invention provides a multi-step long model predictive control scheme for alternately executing sampling and control programs aiming at a control system of a double three-phase permanent magnet synchronous motor under the working condition of low switching frequency, which is realized by the control system shown in figure 1, and the control system comprises: the device comprises a speed closed-loop PI regulator 1.1, a maximum torque-current ratio control module 1.2, a motor discrete model-based dead-beat control module 1.3, a polar coordinate transformation module 1.4, a multi-step long model prediction control module 1.5, a direct current bus 1.6, a double three-phase three-level inverter 1.7, a double three-phase permanent magnet synchronous motor 1.8, a position encoder 1.9, a double rotation coordinate transformation module 1.10, a current prediction module 1.11 and a rotating speed calculation module 1.12. A position encoder 1.9 arranged on the output shaft of the double three-phase permanent magnet synchronous motor 1.8 is used for detecting the angular position theta of the output shaft of the motor_eThe rotation speed calculation module 1.12 calculates the rotation speed according to the angular position theta_eCalculating the rotation speed n, and comparing the rotation speed n with a reference value n^*Is fed into a speed closed-loop PI regulator 1.1, the speed closed-loop PI regulator 1.1 generating a torque reference valueThe maximum torque current ratio control module 1.2 then uses the torque reference value as a function of the torque reference valueGenerating and equally dividing dq-axis reference current to obtain a dq-axis current reference value of the first set of windings And dq-axis current reference value of the second set of windingsSix-phase current i obtained by sampling circuit_A、i_B、i_C、i_D、i_E、i_FObtaining the current time kT under the dual-rotation coordinate system through the dual-rotation coordinate transformation module 1.10_sDq axis current for two sets of windingsThen, the control delay is compensated through a current prediction module 1.11, and the next time (k +1) T is obtained through prediction_sDq-axis current values for two sets of windingsAnd DC bus midpoint potentialAccording to the current reference value and the current predicted value, a first set of winding voltage reference value is calculated by a dead-beat control module 1.3 based on a motor discrete modelAnd a second set of winding voltage reference valuesThe polar coordinate transformation module 1.4 and the multistep long model prediction control module 1.5 determine a loading vector, and control the direct current bus 1.6 and the double three-phase three-level inverter 1.7 to switch the on-off state, so that the double three-phase permanent magnet synchronous motor 1.8 is controlled.

In the method for controlling the prediction of the multi-step long model by alternately executing the sampling and control program, a specific flow of determining the loading vector by the multi-step long model prediction control module is shown in fig. 2, and the method comprises the following steps:

step 2.1, according to the reference value of the first set of winding voltageAnd a second set of winding voltage reference valuesSelecting a vector of a triangular sector where the two sets of windings correspond to the inverter reference voltage vector as a candidate vector according to the polar coordinate transformation result;

step 2.2, judging whether at least one candidate vector exists in the candidate vectors selected in the step 2.1 and does not cause switching between the P level and the N level of the inverter, if so, entering the step 2.4, otherwise, entering the step 2.3;

step 2.3, expanding the selection range from the triangular sector where the reference voltage vector is located to the hexagonal sector where the reference voltage vector is located, and executing the operation of step 2.4 on the candidate vector in the hexagonal sector;

step 2.4, directly canceling the candidate qualification of the candidate vector which can cause the switching of the P level and the N level of the inverter and reserving other candidate vectors;

step 2.5, controlling the double three-phase three-level inverter 1.7 based on the candidate vector determined in the step 2.4, collecting six-phase current at the current moment and obtaining dq-axis current of two sets of windings at the current momentPredicting the state variable of the motor at the next moment to obtain the current values of the dq axes of the two sets of windings at the next moment And DC bus midpoint potential

Step 2.6, it is then determined whether there is a candidate vector, under the action of which the controlled variable i is the next moment_d1，i_q1，i_d2，i_q2And V_nJudging whether candidate vectors can be extrapolated or not, if so, entering a step 2.7, and otherwise, entering a step 2.9;

step 2.7, deleting candidate vectors which cannot be extrapolated, and performing linear extrapolation on the rest candidate vectors;

step 2.8, calculating switching frequency based on the extrapolation result, evaluating the candidate vector by taking the calculated switching frequency as a value function, and entering step 2.10;

step 2.9, the result of the weighted summation of the absolute values of the controlled quantity control errors is used as a cost function to evaluate the candidate vectors;

step 2.10, based on the results of 2.8 or 2.9, the final loading vector is obtained.

The candidate vector triangle region method selection method according to the present invention is shown in fig. 3 and 4. Each phase bridge arm of the three-level inverter can output three levels of voltage U relative to the midpoint of the direct-current bus_dc/2、0、-U_dcAnd/2, respectively, as represented by reference numerals P, O, N. V_ref1And V_ref2Respectively, reference voltage vectors for the first and second sets of windings. V_ref1The triangular area comprises four vectors, namely PON, PNN, POO and ONN, V_ref2The triangle area includes four vectors, ONN, POO, PNN and PNO. The above selection process of candidate vectors corresponds to step 2.1.

The traditional vector crossing traversal process of the synchronous control double three-phase inverter candidate compared with the invention is shown in fig. 5. Starting from step 2.1, the first set of windings has four options, in each of which the second set of windings has four options corresponding to it. Therefore, finally, 16 candidate vector combinations are obtained by the conventional synchronization control method, which are { PNN, POO }, { PNN, ONN }, { PNN, PNO }, { PNN, PNN }, { POO, POO }, { POO, ONN }, { POO, PNO }, { POO, PNN }, { ONN, POO }, { ONN, ONN }, { ONN, PNO }, { ONN, PNN }, { PON, POO }, { PON, ONN }, { PON, PON }, { PON }, PON, PNN }, and { PON, PNN }.

FIG. 6 and FIG. 6 show the selection method of the candidate vector hexagonal region method according to the present inventionShown at 7. V_ref1And V_ref2Respectively, reference voltage vectors for the first and second sets of windings. Considering that the selection method using the hexagonal region method is premised on that the triangular region method does not satisfy the condition, vectors in both the hexagonal region and the triangular region should be excluded, and only vectors in the hexagonal region and no longer in the triangular region should be retained. For example, when V_ref1And V_ref2When in the positions shown in fig. 6 and 7, respectively, V_ref1The vectors in the hexagonal area are PPO, OON, OOO, POP, ONO, PNO, PON, PNN, POO and ONN, the total number of the vectors is 10, and V is excluded_ref1And the candidate vectors PON, PNN, POO and ONN in the triangular region are obtained, and the final candidate vectors determined by the hexagonal region method are 6 of PPO, OON, OOO, POP, ONO and PNO. V_ref2The vectors in the hexagonal area comprise PPO, OON, PON, OOO, POP, ONO, ONN, POO, PNN and PNO, the total number of the vectors is 10, and V is excluded_ref2And (3) the candidate vectors ONN, POO, PNN and PNO in the triangular region are obtained, and the final candidate vectors determined by the hexagonal region method are 6 PPO, OON, PON, OOO, POP and ONO.

The method for evaluating candidate vectors based on extrapolation according to the present invention is shown in fig. 8. Vector VC₃、VC₁、VC₂And VC₄The predicted and extrapolated trajectories of (a) are shown as the four-segment trajectories in fig. 8. The solid line part is the model-based trajectory; the dashed part is the linear extrapolation trajectory. Touch error upper boundary i of candidate vector trajectory_q1maxOr lower boundary i_q1minThe required step size is defined as the extrapolation step size of the vector with respect to the controlled quantity. For example, VC in FIG. 8₃、VC₁、VC₂And VC₄With respect to q₁The extrapolation step sizes for the shaft currents are 1, 2, 3 and 4, respectively. In order to ensure that all controlled quantities are within the error range, for each candidate vector, the minimum value of the extrapolation step sizes corresponding to all controlled quantities is selected as the final extrapolation step size N of the entire extrapolation process when the candidate vector executes the control program. To introduce an evaluation index for the switching frequency, a criterion from kT is defined_sNumber of times n of switching of a time-loaded vector to a candidate vector_sComprises the following steps:

wherein, T_sIs the sampling period, the variable with label (k) indicates that the variable is at kT_sValue of time, S_xIs a switching function defined as:

it is obvious that the larger the total extrapolation step size and the smaller the number of switching times, the better the control and the smaller the switching frequency. Thus, the cost function is defined as:

equation (3) is the cost function involved in step 2.8, and the candidate vector with the smallest value of the cost function is selected as the final loading vector.

The value function based on the weighted summation of the absolute values of the controlled quantity control errors is shown as a formula (4):

wherein λ is_xIs a weight coefficient of a variable x, x_minAnd x_maxRespectively, the minimum and maximum values allowed for the variable x, i.e. x_min＝x_ref-Δx，x_max＝x_ref+ Δ x, Δ x is the variable x at (k +2) T_sMaximum allowable absolute error, x, of predicted value and reference value at a time_refIs a reference value for a variable x, x ∈ { i } when sampling and controlling the first set of windings_d1,i_q1,V_n}; when sampling and controlling the second set of windings, x ∈ { i }_d2,i_q2,V_n}. The expression (4) is the cost function and the number of cost functions related to 2.9The candidate vector with the smallest value will be selected as the final load vector.

Fig. 9 and 10 show control timing diagrams of the method for alternately executing sampling and control programs by using the double three-phase windings according to the invention. Fig. 9 is a timing chart of synchronous predictive control, sampling points on a winding 1 time line and a winding 2 time line are sampling times of the winding 1 and the winding 2 in a synchronous control timing, loading points on the winding 1 time line and the winding 2 time line are vector loading times of the winding 1 and the winding 2 in the synchronous control timing, reference value points on the winding 1 time line and the winding 2 time line are target times of the winding 1 and the winding 2 in the synchronous control timing which are expected to follow a current reference value, and a delay time between the sampling and the vector loading in the synchronous control timing is one sampling period. Fig. 10 is a timing chart of the method of alternately executing the sampling and control program, the sampling points on the winding 1 time line and the winding 2 time line are sampling timings of the winding 1 and the winding 2 in the time sequence of alternately executing the sampling and control program, the loading points on the winding 1 time line and the winding 2 time line are vector loading timings of the winding 1 and the winding 2 in the time sequence of alternately executing the sampling and control program, the reference value points on the winding 1 time line and the winding 2 time line are target timings at which the winding 1 and the winding 2 are expected to follow the current reference value in the time sequence of alternately executing the sampling and control program, and the sampling period in which the delay time between the sampling and the vector loading is half is alternately executed in the time sequence of alternately executing the sampling and control. As shown in fig. 9, the control strategy in which the sampling timing of the duplex winding, the loading vector timing, and the reference value tracking timing coincide is called synchronous control, and the control strategy in which the sampling timing, the loading vector timing, and the reference value tracking timing alternate with each other by half a sampling period is called a control method in which sampling and control programs are alternately executed. In synchronous control, both sets of windings are at kT_sSamples are taken at times, all at (k +1) T_sLoad new switching vectors at all times and all at (k +2) T_sAnd finishing tracking the reference value at the moment. However, in the alternate sampling and control program method, the first set of windings is at (k +1/2) T_sSampling at time (k +1) T_sLoad new vector at time and T at (k +3/2)_sTracking the reference value as a target after the reference value is completed at any moment; second windingGroup at kT_sSampling at time, T at (k +1/2)_sLoad new vectors at time and with (k +1) T_sAnd finishing tracking the reference value at the moment. Thus, the sampling, loading and tracking of the reference values of the two windings are staggered by half the sampling period.

The double-layer model predictive control strategy is shown as a formula (5), a formula (6), a formula (7) and a formula (8). The first layer model predictive control is embedded in the process of selecting candidate vectors based on dead beat control. With d₁q₁Shaft and d₂q₂The current reference value of the shaft current tracking average is taken as a target, and the calculation of the voltage reference vector is shown as the formula (5):

wherein i_d1 ^*＝i_d2 ^*＝(1/2)i_d ^*，i_q1 ^*＝i_q2 ^*＝(1/2)i_q ^*，L_dAnd L_qAre d, q-axis inductances, u, respectively_d1 ^*,u_q1 ^*,u_d2 ^*And u_q2 ^*Are the respective two-winding dq-axis reference voltages, R_sIs the stator resistance, i_d1,i_q1,i_d2And i_q2Are the two-winding dq-axis currents, L, respectively_dmAnd L_qmD and q axis mutual inductance, omega electrical angular velocity, psi_fIs a permanent magnetic linkage. The goal of the second layer model predictive control is to make the dual winding d-axis total current and the q-axis total current follow the dual winding d-axis total current reference and the q-axis total current reference obtained by the maximum torque to current ratio module. Taking the sampling time of winding 1 shown in fig. 10 as an example, the switching state of the second inverter is kept constant from the sampling time of winding 1 to the tracking time of the reference value of winding 1, and thus the dq-axis voltage u of winding 2_d2And u_q2And remain unchanged therebetween. Therefore, according to the discrete motor model, the following constraint conditions exist for the motor state variables between the winding 1 sampling moment and the winding 1 reference value tracking moment:

on the other hand, the expression of the second layer model predictive control target is:

from the equations (6) and (7), d of the winding 1 shown in the equation (8) can be obtained₁q₁Shaft current reference value:

the current reference value shown in equation (8) is the current reference value in step 2.8 and step 2.9. Similarly, it can be calculated that d of winding 2 when sampling and controlling the second set of windings₂q₂The shaft current reference value is:

in the method for alternately executing the sampling and control program, the switching state of one inverter is kept unchanged, the system dimension is reduced from four dimensions to two dimensions, and therefore, the dimension of the state variable related to the motor is also reduced from four dimensions to two dimensions.

The experimental verification results of the present invention are shown in fig. 11 to 21. The experiment is based on a laboratory prototype of a midpoint clamping type three-level inverter feed double three-phase permanent magnet synchronous motor driving system. The control algorithm is realized by calculation of DSP TI-TMS320F28346, a switch tube control signal is generated by an FPGA Xilinx-Spartan6 XC6SLX25 through a driving circuit, and a coaxial permanent magnet synchronous motor externally connected with a rheostatic box is used as a load of the double three-phase motor. The parameters of the double three-phase motor used are as follows: the number of pole pairs is 4, the q-axis inductance is 13mH, the d-axis inductance is 10mH, the leakage inductance is 5mH, the permanent magnet flux linkage is 0.08Wb, the stator resistance is 0.225 omega, the bus capacitance is 4000 muF, and the sampling frequency of a single set of winding is 7.5 kHz.

The effectiveness verification results of the double three-phase permanent magnet synchronous motor control method for alternately executing the sampling and control program under the steady-state working condition are shown in fig. 11 to 16, and the method comprises the stator current waveform, the capacitor voltage and midpoint potential fluctuation waveform, the rotating speed and the torque waveform of the traditional synchronous control and alternating execution sampling and control program method. Compared with the synchronous control, the sampling and control program method is alternately executed to reduce the current THD from 19.36% to 9.20% and reduce the switching frequency from 724Hz to 489Hz, as shown in fig. 11 and 12; the midpoint potential fluctuation decreased from 1.6V to 1.3V as shown in FIGS. 13 and 14, and u_C1、u_C2The voltage waveform of the capacitance of the upper bridge arm of the bus and the voltage waveform of the capacitance of the lower bridge arm of the bus are obtained; so that the torque ripple is reduced from 1.21Nm to 0.76Nm as shown in FIGS. 15 and 16, n_refIs the reference value waveform of the rotation speed, n is the rotation speed waveform, T_eIs the torque ripple waveform. As can be seen from the comparative experiments in fig. 11 to 16, the multi-step model predictive control strategy for the dual three-phase permanent magnet synchronous motor, which alternately executes the sampling and control procedures, effectively improves the steady-state control performance of the motor.

The validity verification results of the dual three-phase permanent magnet synchronous motor control method for alternately executing the sampling and control program under the dynamic working condition are shown in fig. 17 to 20, and comprise the rotating speed and torque waveforms, n, of the traditional synchronous control and alternate sampling and control program execution method_refIs the reference value waveform of the rotation speed, n is the rotation speed waveform, T_{e_ref}Is the torque reference value waveform, T_eIs a torque waveform. In contrast to synchronous control, alternating the sampling and control program method causes the torque up delay to fall from 133 μ s to 67 μ s, causing the torque response time to fall from 2200 μ s to 1800 μ s, as shown in fig. 18 and 20. As can be seen from the comparative experiments of fig. 17 to 20, the dual three-phase permanent magnet synchronous system alternately performs the sampling and control proceduresThe motor multi-step model predictive control strategy effectively improves the dynamic control performance of the motor.

The experimental verification result of the double three-phase permanent magnet synchronous motor control method for alternately executing the sampling and control program in the aspect of reducing the algorithm complexity is shown in fig. 21, and the experimental verification result comprises the execution time of analog-to-digital conversion, delay compensation, model prediction control algorithm and vector loading part program in the traditional synchronous control and alternating execution sampling and control program method. In the traditional synchronous control, the execution time of an analog-digital conversion program is 11.24 mu s, the execution time of a delay compensation program is 6.62 mu s, the execution time of a multi-step long model prediction control program is 95.69 mu s, and the execution time of a vector loading program is 3.01 mu s; in the method of alternately executing the sampling and control programs, the execution time of the analog-to-digital conversion program is 11.24 μ s, the execution time of the delay compensation program is 6.62 μ s, the execution time of the multi-step long model prediction control program is 20.28 μ s, the execution time of the vector loader program is 3.01 μ s, and the four programs are executed twice in each sampling period. The execution time of analog-to-digital conversion, delay compensation and vector loading related programs of the traditional synchronous control and alternative execution sampling and control program method is the same, but the model predictive control algorithm execution time is obviously lower than that of the traditional synchronous control because the alternative execution sampling and control program method reduces the dimensionality of the system. Therefore, in a single set of winding sampling period, although the sampling and control program method is alternately executed for analog-to-digital conversion, delay compensation and vector loading twice, the total execution time of the program is still less than that of synchronous control. According to the comparative experiment of fig. 21, the multi-step model predictive control strategy of the dual three-phase permanent magnet synchronous motor, which alternately executes the sampling and control program, effectively reduces the calculation amount of the algorithm.