阅读说明：本技术 一种抑制五相逆变器共模电压的有限集模型预测控制方法 (Finite set model prediction control method for restraining common-mode voltage of five-phase inverter ) 是由 冯晓云 余彬 宋文胜 李建新 郭永琪 葛兴来 苟斌 杨顺风 于 2020-05-14 设计创作，主要内容包括：本发明公开了一种抑制五相逆变器共模电压的有限集模型预测控制方法,具体为：利用五相逆变器输出的十个大矢量来合成虚拟电压矢量,选择相邻的四个大矢量合成α-β子空间下的虚拟有效电压矢量,共合成十个虚拟有效电压矢量,选择两方向相反的大矢量合成虚拟零矢量；以虚拟电压矢量作为矢量控制集,通过简化的预测模型预测电流变化趋势；利用简化的评价函数,选择出最优的虚拟电压矢量；最后,设计对称的开关序列输出最优矢量。本发明有效抑制了共模电压和低次谐波电流,提高了逆变器输出电流稳态性能；此外,得益于虚拟电压矢量抑制共模电压和低次谐波的特点,预测模型和评价函数得到简化,算法计算复杂度低,易于数字化实现。(The invention discloses a finite set model prediction control method for inhibiting common-mode voltage of a five-phase inverter, which is characterized by synthesizing a virtual voltage vector by utilizing ten large vectors output by the five-phase inverter, selecting four adjacent large vectors to synthesize a virtual effective voltage vector under α - β subspace, synthesizing ten virtual effective voltage vectors in total, selecting large vectors with opposite directions to synthesize a virtual zero vector, using the virtual voltage vector as a vector control set, predicting the current change trend through a simplified prediction model, selecting an optimal virtual voltage vector by utilizing a simplified evaluation function, and finally designing a symmetrical switching sequence to output an optimal vector.)

1. A finite set model predictive control method for suppressing common-mode voltage of a five-phase inverter is characterized by comprising the following steps:

synthesizing virtual voltage vectors, namely synthesizing the virtual voltage vectors by utilizing ten large vectors output by a five-phase inverter according to a volt-second balance principle, selecting four adjacent large vectors to synthesize a virtual effective voltage vector under α - β subspaces, enabling the amplitude of the synthesized voltage vector under the x-y subspaces to be zero, and synthesizing ten virtual effective voltage vectors according to the synthesizing principle, wherein V is sequentially_v1、V_v2……V_v10The amplitude of the obtained virtual effective voltage vector is 0.5257V under α - β subspaces_dcThe amplitude is 0 under the mapping to the x-y subspace, where V_dcIn addition, two large vectors in opposite directions under α - β subspaces are selected to be combined according to the same duty ratio to obtain a virtual zero vector;

step 2: and (3) selecting an optimal virtual voltage vector: the construction of the virtual voltage vector inhibits common-mode voltage, eliminates third harmonic current, does not need to predict and calculate the third harmonic current and the common-mode voltage, and simplifies evaluation functions; and selecting an optimal virtual voltage vector by using the principle of minimum G through a simplified evaluation function:

wherein the content of the first and second substances,α and β axis components for a given reference current,α, β axis components of the predicted current, respectively;

and step 3: designing a symmetrical switching sequence: in order to reduce the harmonic content and improve the current steady-state performance, a symmetrical switching sequence is designed; if the optimal vector is an effective virtual vector, the action time of four adjacent large vectors is equally divided into two sections, the two sections act in a counterclockwise sequence in the first half sampling period, and the two sections act in a clockwise sequence in the second half sampling period, so that the average action times of the switches of all the phases are ensured; and if the optimal vector is a virtual zero vector, selecting a corresponding switching sequence according to the switching state of the previous period in order to reduce the switching frequency.

Technical Field

The invention belongs to the field of design and manufacture of five-phase motor alternating current control systems in the field of power electronics and power transmission, and particularly relates to a finite set model prediction control method for inhibiting common-mode voltage of a five-phase inverter.

Background

Due to the limitation of the three-phase power supply system of the power system, research work in the aspect of alternating current electric transmission mainly focuses on theoretical research and practical application of a three-phase motor transmission system. With the development of society and the progress of industry, more and more application occasions put forward higher requirements on the output power of a transmission system, and the variable-frequency transmission system of the multi-phase motor can provide higher power on the occasions with limited voltage, and has small torque pulsation and high operation efficiency; the fault tolerance performance is good, the reliability is high, and the method is widely concerned and researched increasingly.

The multi-phase motor variable frequency transmission system mainly comprises a multi-phase power electronic frequency converter and a multi-phase motor, and the application of the frequency converter ensures that the variable frequency transmission system has the advantages of wide speed regulation range, large starting torque, good motor dynamic performance, high efficiency and the like; however, the use of frequency converters also brings about some negative effects: the high-frequency common-mode voltage (CMV) generated by the frequency converter induces a high-amplitude shaft voltage on the rotating shaft of the motor and forms a bearing current, so that the aging of the motor bearing is accelerated, the bearing of the motor is damaged in a short time, and the service life of the motor is shortened.

In order to reduce the negative effect of the common-mode voltage on the motor, domestic and foreign scholars propose a plurality of methods for suppressing the common-mode voltage, which are mainly divided into active suppression and passive suppression. Passive rejection includes using common mode rejection transformer, common mode filter, common mode choke coil, etc., which all require increased hardware cost; the active suppression is mainly based on the control algorithm of the inverter to suppress the common-mode voltage, emphasizes that the purpose of suppressing the common-mode voltage is achieved by modifying the modulation algorithm, does not need additional hardware, and is more flexible in application and low in cost.

Currently, model predictive control has attracted wide attention in the field of motor control, and, among them, finite set model predictive current control (FCS-MPC) has become a research hotspot of model predictive control in the field of motor control due to its advantages of simple structure, fast dynamic response, easy handling of system constraint problems, and the like. However, the conventional model predictive control also brings higher common-mode voltage, for this reason, some researchers have proposed some improved model predictive control algorithms to realize suppression of the common-mode voltage by increasing the constraint condition for suppressing the common-mode voltage, but still have the disadvantages of complex algorithm, heavy calculation burden, high low-order harmonic content, and the like. Aiming at the defects of the existing model predictive control method, the invention provides an improved finite set model predictive control method for suppressing the common-mode voltage of a five-phase inverter, which can effectively realize the balance among suppressing the common-mode voltage, eliminating harmonic current and simplifying a model predictive control algorithm.

Disclosure of Invention

The invention aims to solve the technical problem of inhibiting common-mode voltage in a five-phase permanent magnet synchronous motor control system, and provides a finite set model prediction control method for inhibiting common-mode voltage of a five-phase inverter.

The invention discloses a finite set model prediction control method for inhibiting common-mode voltage of a five-phase inverter, which comprises the following specific steps:

step 1, synthesizing virtual voltage vectors, synthesizing the virtual voltage vectors by utilizing ten large vectors output by a five-phase inverter according to a volt-second balance principle, selecting four adjacent large vectors to synthesize the virtual effective voltage vectors under α - β subspaces, and enabling the amplitude of the synthesized voltage vector mapped to the x-y subspaces to be zero according to the aboveSynthesizing ten virtual effective voltage vectors in total according to a synthesis principle, wherein the synthesis principle comprises the following steps: v_v1、V_v2……V_v10The amplitude of the obtained virtual effective voltage vector is 0.5257V under α - β subspaces_dcThe amplitude is 0 under the mapping to the x-y subspace, where V_dcThe vector is direct current bus voltage, and in addition, in order to avoid the zero vector from generating larger common-mode voltage, two large vectors in opposite directions under α - β subspaces are selected to be combined according to the same duty ratio to obtain a virtual zero vector.

Step 2: and selecting an optimal virtual voltage vector. The construction of the virtual voltage vector inhibits common-mode voltage, eliminates third harmonic current, does not need to predict and calculate the third harmonic current and the common-mode voltage, and simplifies evaluation functions; and selecting an optimal virtual voltage vector by using the principle of minimum G through a simplified evaluation function:

wherein the content of the first and second substances,α and β axis components for a given reference current,α, β axis components of the predicted current, respectively.

And step 3: and (4) designing a symmetrical switching sequence. In order to reduce the harmonic content and improve the current steady-state performance, a symmetrical switching sequence is designed; if the optimal vector is an effective virtual vector, the action time of four adjacent large vectors is equally divided into two sections, the two sections act in a counterclockwise sequence in the first half sampling period, and the two sections act in a clockwise sequence in the second half sampling period, so that the average action times of the switches of all the phases are ensured; and if the optimal vector is a virtual zero vector, selecting a corresponding switching sequence according to the switching state of the previous period in order to reduce the switching frequency.

Compared with the prior art, the invention has the beneficial effects that:

(1) under the condition that a hardware structure is not increased, the synthesized virtual voltage vector is used as a model prediction control set, and the common-mode voltage is effectively restrained.

(2) The current sine degree is high, the third harmonic current is effectively eliminated by the virtual voltage vector, and the dynamic performance and the steady-state performance are superior.

(3) The method has the advantages that the virtual voltage vector can restrain common-mode voltage and low-order harmonic current, the prediction model and the evaluation function are simplified, and meanwhile, a control set formed by the virtual voltage vector is simpler, so that the calculation complexity of the model prediction control method is reduced, and the method is easy to realize digitally.

(4) The method is based on a five-phase inverter for analysis, and can be expanded and applied to other multi-phase variable frequency driving systems.

Drawings

Fig. 1 is a five-phase inverter with inductive load.

FIG. 2 is a space voltage vector distribution diagram of a five-phase inverter at α - β subspace.

FIG. 3 is a space voltage vector distribution diagram for a five-phase inverter in x-y subspace.

Fig. 4 is an overall structural diagram of a prediction control method for suppressing a five-phase inverter common mode voltage model according to the present invention.

FIG. 5 is a schematic diagram of virtual voltage vector synthesis (in V) under the subspace α - β_v1For example).

FIG. 6 is a schematic diagram of virtual voltage vector synthesis (in V) in x-y subspace_v1For example).

FIG. 7 is a diagram of a virtual voltage vector distribution under α - β and x-y subspace.

FIG. 8 is a diagram of a symmetrical switching pattern (in V) over one sampling period_v1For example).

Fig. 9 shows the output phase current and common mode voltage waveform of the inverter (RCMV-impsc) of the present invention, as well as the fundamental and third harmonic traces of the output phase current.

Fig. 10 is a fundamental wave trace and a third harmonic trace of the output phase current and the common mode voltage waveform of the inverter of the DRO-MPCC method.

Fig. 11 is a fundamental wave trace and a third harmonic trace of the output phase current and the common mode voltage waveform of the RCMV-MPCC method inverter.

FIG. 12 shows the program execution time of the present invention (RCMV-IMPCC) and RCMV-MPCC and DRO-MPCC.

Fig. 13 is a dynamic process diagram of inverter output phase current given reference current switching in the present invention.

Detailed Description

The present invention is further described in detail below with reference to the drawings and the detailed description so as to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the detailed description, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the present invention as defined and defined by the appended claims, and all changes that come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

In a five-phase motor system, symmetrical physical quantities (voltage, current, flux linkage, etc.) in a natural coordinate system can be mapped α - β and x-y orthogonal subspaces, respectively, according to a coordinate transformation the fundamental and 10j ± 1(j ═ 1,2,3 …) subharmonics of motor variables are mapped to α - β subspaces via the first two rows of the transformation matrix the 3 rd and 10j ± 3(j ═ 1,2,3 …) subharmonics in motor variables are mapped to x-y subspaces via the 3 rd and 4 th rows of the transformation matrix, the 10j ± 5(j ═ 1,2,3 …) subharmonics in motor variables are projected into the zero sequence subspace formed by the 5 th row, which remains zero for a star-connected five-symmetric load.

FIG. 1 shows a five-phase voltage source inverter with inductive load, S_i(i ═ a, b, c, d, e) shows the switching state of each arm of the inverter in turn, and when the upper arm is on, S is_i1 is ═ 1; when the lower bridge arm is conducted, S_iAnd the upper bridge arm and the lower bridge arm cannot be conducted simultaneously. Through analysis, the five-phase inverter has 32 switch states, which correspond to 32 voltage vectors,the voltage vectors corresponding to 32 switch states can be respectively mapped to α - β and x-y two orthogonal subspaces, as shown in fig. 2 and fig. 3, respectively, in α - β subspaces, 30 effective voltage vectors can be divided into large, medium and small vectors, and the amplitudes are 0.6472V_dc、0.4V_dc、0.2472V_dcIn which V is_dcShowing the inverter dc side bus voltage.

For a five-phase inverter, the common-mode voltage u_CMCan be expressed as:

table 1 gives 32 voltage vectors and their corresponding switch states and corresponding common mode voltages.

TABLE 1 Voltage vectors and their corresponding switch states and corresponding common mode voltages

The distribution rule of the space voltage vector of the five-phase inverter is as follows:

(1) α - β large vectors in the subspace and small vectors in the x-y subspace are mapped one by one;

(2) α - β sub-spaces with large, medium and small voltage vectors (e.g., V) in the same direction₂₅，V₁₆，V₉) In the x-y subspace, the median vector (V)₁₆) Will be associated with large (V)₂₅) Small (V)₉) The vector directions are opposite.

(3) The magnitude ratio of the large vector, the medium vector and the small vector is: 1.618²:1.618:1。

The control block diagram of the prediction control method for suppressing the five-phase inverter common-mode voltage model provided by the invention is shown in fig. 4, and mainly comprises the following steps: a limited control set, current prediction, evaluation functions and delay compensation.

The mathematical model of the resistive-inductive load in α - β and x-y subspace is:

in the formula u_αβAnd u_xy、i_αβAnd i_xyα - β and x-y phase voltage and load current under two subspaces respectively through a forward Euler method and delay compensation, a current prediction model can be obtained:

wherein, T_sFor the sample period, L is the load inductance, R is the load resistance,to compensate for the new prediction starting point for the delay,the α and β axis components of the predicted current at time k +2,the α, β axis components of the virtual voltage vector,are physical quantities in the x-y subspace having the same meaning as the corresponding variables in the α - β subspace.

It can be seen from table 1 that the zero vector and the middle vector generate a larger common mode voltage, and the large vector and the small vector generate a smaller common mode voltage. Therefore, the purpose of suppressing the common mode voltage can be achieved by avoiding using the vector and the zero vector.

In order to restrain low harmonic current and common mode voltage, the invention provides a virtual voltage vector synthesis method. By synthesizing a virtual voltage vector, a prediction model and an evaluation function are simplified, and the aims of eliminating low-order harmonic current and suppressing common-mode voltage are achieved, and the specific implementation steps are as follows:

first, a virtual voltage vector is constructed. In the existing research of a five-phase inverter system of model predictive control, in order to reduce the calculated amount and improve the steady-state performance, a virtual voltage vector is synthesized by a large vector and a medium vector in the same direction as the large vector to be used as a model predictive control set.

In order to restrain common mode voltage, the virtual voltage vector is synthesized by using only ten large vectors output by a five-phase inverter, the virtual effective voltage vector under the subspace of α - β is synthesized by selecting four adjacent large vectors according to the volt-second balance principle, meanwhile, the amplitude of the synthesized voltage vector under the mapping of the three-time harmonic current is enabled to be zero, as shown in figures 5 and 6, and the duty ratios of the four large vectors in a sampling period are sequentially set to be lambda in the anticlockwise sequence₁、λ₂、λ₃、λ₄The duty ratio of four large vectors in one period is calculated as follows, and λ is assumed₁＝λ₄：

The duty ratios of the four large vectors in one sampling period are sequentially obtained by the following formula: lambda [ alpha ]₁＝0.191，λ₂＝0.309，λ₃＝0.309，λ₄0.191. According to the above synthesis principle, ten virtual effective voltage vectors can be synthesized, which are sequentially as follows: v_v1、V_v2……V_v10As shown in FIG. 7, substituting the duty cycle results in a virtual effective voltage vector with a magnitude of 0.5257V in the subspace α - β_dc。

In addition, in order to avoid using a zero vector, two large vectors in opposite directions under α - β subspaces are selected and combined according to the same duty ratio to obtain a virtual zero vector;

and secondly, predicting the current by taking ten virtual effective voltage vectors and a virtual zero vector as a control set. Because the amplitude value of the constructed virtual voltage vector is 0 when the constructed virtual voltage vector is mapped to the x-y subspace, the low-order harmonic current is eliminated fundamentally, so that a prediction model under a third-order harmonic space can not be included any more, and the prediction model is simplified; and (3) predicting the current variation trend by using a simplified prediction model with time delay compensation:

then, an optimal vector is selected through the evaluation function. The structure of the virtual voltage vector fundamentally inhibits common-mode voltage, eliminates third harmonic current, and the evaluation function can not contain a third harmonic current weight term and a common-mode voltage inhibition weight term any more, so that the evaluation function of the invention is simplified compared with the evaluation function of the traditional model predictive control; and selecting an optimal virtual voltage vector by using the principle of minimum G through a simplified evaluation function:

and finally, designing a symmetrical switching sequence. In a sampling period, if the selected optimal virtual voltage vector is an effective vector, in order to ensure that the switching state is a symmetrical mode, the duty ratio of four large vectors in one sampling period is divided into two equal parts, in the first half of the sampling period, the four large vectors act anticlockwise, in the second half of the sampling period, the four large vectors act clockwise, and the symmetrical switching mode is as shown in fig. 8 (with V as the effective vector)_v1For example); if the selected optimal virtual voltage vector is a zero vector, based on the principle of lowest switching frequency, selecting a large vector which is finally acted in the last sampling period and a large vector opposite to the large vector to synthesize a virtual zero vector, and through the same construction method as the virtual effective vector symmetrical switching mode, constructing the switching state corresponding to the virtual zero vector into a symmetrical switching mode.

The invention (RCMV-IMPCC) is verified on a five-phase inverter test platform, and the experimental result is compared with the previous research, wherein the comparison objects are as follows: conventional common mode voltage rejection RCMV-MPCC and DRO-MPCC. The basic idea of the RCMV-MPCC is that all 32 voltage vectors are used as a control set, and two common-mode voltage rejection weight terms are added in an evaluation function of model prediction control, so that a model prediction algorithm is undoubtedly more complex; the basic idea of the DRO-MPCC is to synthesize a virtual voltage vector as a control set by a large vector and a medium vector in the same direction as the large vector in order to reduce the calculation amount and improve the steady-state performance.

FIG. 9 shows the inverter output phase current and common mode voltage waveform of the model predictive control algorithm proposed by the present invention, and it can be seen that the common mode voltage is suppressed to + -0.1V_dcWhile the maximum value of the common-mode voltage of the conventional model predictive control and DRO-MPCC method (FIG. 10) is + -0.5V_dcTherefore, the invention reduces the amplitude of the common mode voltage by 80%. The RCMV-MPCC method (fig. 11) suppresses the common mode voltage to some extent, but its harmonic content is high.

Meanwhile, the inverter has high output current sine degree and good steady-state performance, the harmonic content is 5.17%, and the third harmonic current amplitude is almost 0 as can be seen from the harmonic track of the output current in fig. 8, which shows that the inverter can ensure that the output current has excellent steady-state performance while inhibiting the common-mode voltage.

FIG. 12 shows a comparison of the program execution times of the present invention (RCMV-IMPCC) and RCMV-MPCC and DRO-MPCC. From a comparison of FIG. 12, it can be seen that the RCMV-MPCC has the longest execution time, almost 3 times that of the other two; the program execution time of the method is shortest and is equivalent to DRO-MPCC, but the common mode voltage of the DRO-MPCC method is the largest.

Fig. 13 shows the dynamic course of the inverter output phase current when a given reference current is switched from 3A to 6A, and it can be seen that the dynamic course is only 1ms, so the method of the present invention retains the advantage of fast dynamic response of model predictive control.

The above comparisons are combined to obtain: the method provided by the invention selects four large vectors to synthesize the virtual voltage vector, and adopts the virtual voltage vector as the control set, so that the third harmonic current is eliminated while the common mode voltage is suppressed, the model prediction algorithm is simplified, the model prediction algorithm is easier to realize digitally, the calculation burden of a digital processor is obviously reduced, and the balance among the suppression of the common mode voltage, the elimination of the harmonic current and the simplification of the algorithm is effectively realized.