阅读说明：本技术 一种基于占空比优化的五相逆变器模型预测电流控制方法 (Five-phase inverter model prediction current control method based on duty ratio optimization ) 是由 何鸿云 宋文胜 余彬 陈溉泉 王青元 崔恒斌 冯晓云 李婷婷 葛兴来 于 2020-05-26 设计创作，主要内容包括：本发明公开了一种基于占空比优化的五相逆变器模型预测电流控制方法,具体包括：选择α-β子空间4个相邻大矢量按照特定占空比合成虚拟电压矢量,消除x-y子空间的谐波电压；选用虚拟电压矢量构造有限控制集,简化预测模型和评价函数；估算每个控制周期虚拟电压矢量作用时间,进行占空比优化,减小电流纹波；选择方向相反的2个大矢量按照相同的占空比组合,调节每个控制周期伏秒值,从而抑制共模电压；设计优化开关序列,实现各桥臂开关动作次数平均。本发明在有效抑制低次谐波电流的同时,共模电压减小80％；且简化了评价函数,提高了电流跟踪精度,减小了电流纹波。(The invention discloses a five-phase inverter model prediction current control method based on duty ratio optimization, which specifically comprises the following steps: 4 adjacent large vectors in the alpha-beta subspace are selected to synthesize a virtual voltage vector according to a specific duty ratio, and harmonic voltage of the x-y subspace is eliminated; selecting a virtual voltage vector to construct a limited control set, and simplifying a prediction model and an evaluation function; estimating the virtual voltage vector action time of each control period, optimizing the duty ratio and reducing the current ripple; selecting 2 large vectors in opposite directions to combine according to the same duty ratio, and adjusting the volt-second value of each control period, so as to restrain the common-mode voltage; and designing an optimized switching sequence to realize the average switching times of each bridge arm. The common-mode voltage is reduced by 80% while the low-order harmonic current is effectively inhibited; and the evaluation function is simplified, the current tracking precision is improved, and the current ripple is reduced.)

1. A five-phase inverter model prediction current control method based on duty ratio optimization is characterized by comprising the following steps:

s1, selecting 4 adjacent large vectors in the alpha-beta subspace, synthesizing the vectors into virtual voltage vectors according to duty ratios of 0.191, 0.309 and 0.191 in sequence, and obtaining 10 virtual voltage vectors in total;

s2, taking the synthesized virtual voltage vector as a vector control set, and simplifying the evaluation function G to only contain an alpha-beta subspace current term without a weight coefficient as follows:

wherein i_α ^*And i_β ^*Giving reference currents, i, for the k time α and β axes_α ^k+2And i_β ^k+2The estimated values of the current of the shaft at the time α and β of k + 2;

s3, in each control period, estimating the duty ratio of the virtual voltage vector based on the evaluation function minimum principle:

wherein, V_optThe amplitude of the optimal virtual voltage vector is obtained; v_α、V_βIs a V_optProjected values on α, β axes i_α ^k+1And i_β ^k+1The estimated values of the current of the k +1 time α and β axes, R is the resistance of the load side, L is the inductance of the load side, and T is the inductance of the load side_sIs a control period;

s4, designing an optimized switching sequence: according to the optimal virtual voltage vector, two large vectors with opposite directions are selected and respectively inserted into the two ends and the middle of the control period, so that a symmetrical switching sequence and average switching action times of each phase are realized.

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 five-phase inverter model prediction current control method based on duty ratio optimization.

Background

The model predictive control has the advantages of simplicity and flexibility in control, low switching frequency, multi-objective optimization and the like, and has attracted wide attention in the field of motor control, wherein finite set model predictive current control (FCS-MPCC) is one of the commonly used methods in the field of power electronics, and the method directly controls current, is intuitive in concept and simple to implement, and has become a research hotspot in the field of motor control. The classic FCS-MPCC estimates the current change trend at the next moment in each control period, selects the optimal vector according to the evaluation function, and applies the optimal vector to the next control period. However, most of the time, the error between the actual current and the reference value is small, the effective vector acts on the whole period, the current tracking performance is reduced, and the current ripple is increased. To solve this problem, a duty cycle based FCS-MPCC algorithm is proposed. The method estimates the action time of the optimal vector according to the current error, and complements the rest control period by using a zero vector, thereby reducing the current ripple and effectively improving the steady-state performance of the FCS-MPCC method.

However, the duty-cycle-optimized FCS-MPCC method may generate a large common-mode voltage (CMV) due to the use of the zero vector. The common mode voltage can generate shaft current and leakage current to cause the motor to generate heat, which is one of important factors causing the damage of a motor bearing and seriously influences the service life of the motor. In addition, the common mode voltage can also generate electromagnetic interference, which affects the normal operation of other electric devices of the system. In order to suppress the common-mode voltage, the existing FCS-MPCC method generally increases the constraint of the common-mode voltage in the evaluation function, and achieves a balance between the common-mode voltage and the steady-state performance by adjusting the weight coefficient. However, the introduction of common-mode voltage constraint and the adjustment of the weight coefficient make the algorithm implementation complicated, and especially in a five-phase system, the constraint of harmonic current needs to be considered, and the setting of the weight coefficient is more difficult.

Disclosure of Invention

Aiming at the defects of the prior art, the invention provides a five-phase inverter model prediction current control method based on duty ratio optimization.

A five-phase inverter model prediction current control method based on duty ratio optimization comprises the following steps:

and S1, selecting 4 adjacent large vectors in the alpha-beta subspace, synthesizing the vectors into virtual voltage vectors according to the duty ratios of 0.191, 0.309 and 0.191 in sequence, and obtaining 10 virtual voltage vectors in total.

S2, taking the synthesized virtual voltage vector as a vector control set, simplifying the evaluation function G to only contain an alpha-beta subspace current term and no weight coefficient, and avoiding the setting of the weight coefficient; the evaluation function G is as follows:

wherein i_α ^*And i_β ^*Giving reference currents, i, for the k time α and β axes_α ^k+2And i_β ^k+2The k +2 times α and β axis current estimates.

And S3, estimating the duty ratio of the virtual voltage vector in each control period based on the evaluation function minimum principle to realize more accurate current tracking and reduce current ripple.

Wherein V_optThe amplitude of the optimal virtual voltage vector is obtained; v_α、V_βIs a V_optProjected values on α, β axes i_α ^k+1And i_β ^k+1The estimated values of the current of the k +1 time α and β axes, R is the resistance of the load side, L is the inductance of the load side, and T is the inductance of the load side_sIs a control cycle.

And S4, selecting two reverse large vectors according to the selected optimal virtual voltage vector, and designing an optimized switching sequence to output a desired volt-second value.

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

the virtual voltage vector is synthesized by adopting the voltage vector which generates the minimum common-mode voltage, and the common-mode voltage is reduced by 80 percent; the projection amplitude of the virtual voltage vector in the x-y subspace is 0, so that low-order harmonic current is suppressed, and an evaluation function is simplified; and the duty ratio optimization is introduced, so that the current tracking precision is improved, and the current ripple is reduced. In addition, the calculation complexity is simplified, the digital implementation is easy, and the difficulty of program design is reduced.

Drawings

Fig. 1 is a diagram of a five-phase two-level voltage source inverter topology.

Fig. 2 is a voltage vector distribution diagram of a five-phase voltage source inverter in an alpha-beta subspace.

Fig. 3 is a voltage vector distribution diagram of a five-phase voltage source inverter in an x-y subspace.

FIG. 4 shows the α - β subspace virtual voltage vector composition case (in V)_v1For example).

FIG. 5 shows the x-y subspace virtual voltage vector composition (in V)_v1For example).

FIG. 6 is a waveform (in V) of a control period symmetric switching sequence_v1For example).

FIG. 7 shows phase current waveforms, CMV waveforms, and alpha-beta and x-y subspace current trajectories for a conventional common mode voltage rejection method.

FIG. 8 is a graph of phase current waveforms, CMV waveforms, and alpha-beta and x-y subspace current trajectories for a conventional duty cycle optimization method.

FIG. 9 shows the phase current waveform, CMV waveform, and the α - β and x-y subspace current trajectories according to the proposed method of the present invention.

Fig. 10 is a comparison of the execution time of the present invention and two conventional methods.

Detailed Description

The invention is described in further detail below with reference to the accompanying drawings and specific embodiments.

According to one embodiment of the present application, referring to fig. 1, the present invention is directed to a five-phase two-level voltage source inverter. In a five-phase motor system, symmetrical physical quantities under a natural coordinate system can be respectively mapped into alpha-beta and x-y orthogonal subspaces according to an expanded park rotation transformation matrix, and fundamental waves and 10k +/-1 (k is 1,2,3 …) subharmonics are mapped into the alpha-beta space; the 3 rd and 10k ± 3(k ═ 1,2,3 …) subharmonics are mapped to the x-y subspace; the 10k ± 5(k ═ 1,2,3 …) subharmonics are mapped into the zero sequence subspace, which term remains zero at all times for the five symmetric loads of the wye connection.

Defining a switching function as S_i(i＝a,b,c,d,e)，S_iWhen the bridge arm is 1, the upper bridge arm is conducted; s_iThe five-phase inverter has 32 basic voltage vectors including 30 effective voltage vectors and 2 zero vectors, and is divided into large, medium, small and zero vectors according to the amplitude, and the large, medium and small voltage vectors in the subspace of α - β have amplitudes of 0.6472V respectively_dc、0.4V_dc、0.2472V_dcFIGS. 2 and 3 show five-phase inverters at α - β and x-y subspaceVoltage vector distribution diagram.

Common mode voltage u of a five-phase inverter_CMThe calculation formula is as follows:

the common mode voltage values generated by different voltage vectors are shown in table 1, and it is clear that small and large vectors generate the smallest CMV and that zero vectors generate the largest CMV. The largest vector composite virtual voltage vector with the CMV smallest is therefore selected.

TABLE 1 different Voltage vectors CMV

Selecting 4 adjacent large vector composite virtual voltage vectors, α - β in the figures 4 and 5 and x-y subspace virtual voltage vector composite conditions, and enabling the vector amplitude of the x-y subspace composite to be 0 and V to be used_v1For example, 4 large vector combined duty cycle calculations are given:

wherein λ is₁、λ₂、λ₃、λ₄A total of 10 effective virtual voltage vectors are available, with amplitudes at α - β subspace of:

the prediction model needs to consider the inherent one-beat delay of the controller, and the current prediction model after delay compensation is as follows:

wherein i_αβ ^kSampled values of current on axes α and β at time k, i_αβ ^k+2Is the k +2 time α, β axisAn estimated current value of (a); u. of_αβ ^kThe projection values of the optimal voltage vector at the time k on α and β axes, u_αβ ^k+1For controlling the projection values of the concentrated voltage vectors on α and β axes, R is the resistance on the load side, L is the inductance on the load side, and T is the inductance on the load side_sIs a control cycle.

The 10 effective virtual voltage vectors are used as a limited control set, harmonic current and CMV can be effectively suppressed, so that an evaluation function does not need to contain x-y subspace current and CMV evaluation items, the calculation complexity is reduced, and the evaluation function is as follows:

while the conventional FCS-MPCC only applies 1 voltage vector per control cycle, the method proposed by the present invention applies 4 large vectors, and thus requires a proper configuration of the switching sequence. The present invention adopts a symmetrical switching mode, in which 4 large vectors act in sequence in the counterclockwise direction in the first half of the control period and in the clockwise direction in the second half of the control period, as shown in FIG. 6 (with V)_v1For example).

In most cases, it is not necessary to apply the optimal voltage vector to track the reference current throughout the control cycle. Therefore, to achieve more accurate current tracking, the duty cycle of the optimal effective voltage vector is calculated, and the zero vector effect is inserted for the rest of the time. Assuming that the duty ratio of the effective voltage vector is d, and the predicted current value at the moment k +2 is:

substituting the formula into the evaluation function to obtain the corresponding duty ratio when the evaluation function is minimum, namely the optimal duty ratio:

wherein V_optThe amplitude of the optimal virtual voltage vector is obtained; v_α、V_βIs a V_optProjected values on α, β axes i_α ^k+1And i_β ^k+1The k +1 times α and β axis current estimates.

As can be seen from Table 1, the zero vector produces the largest CMV, and the conventional duty cycle optimized FCS-MPCC method produces the larger CMV. Therefore, the method provided by the invention abandons the conventional zero vector and synthesizes a virtual zero vector by 2 large vectors with opposite directions according to the same duty ratio (0.5). To maintain the symmetry of the switching sequence, different virtual zero vector combinations are used for different effective virtual voltage vectors, see table 2, where V_optIs the optimal virtual voltage vector calculated at the current moment.

TABLE 2 different virtual zero vector combination modes

The traditional five-phase inverter FCS-MPCC method for inhibiting common-mode voltage adds a common-mode voltage evaluation item in an evaluation function, reduces the selection frequency of a zero vector and a middle vector, thereby inhibiting CMV, but the method sacrifices partial steady-state performance_dc(+ -10V) and steady state performance is similar to conventional duty cycle optimization methods.

Fig. 10 shows a comparison of the controller execution times for the proposed method and for the two conventional methods.

While the embodiments of the invention have been described in detail in connection with the accompanying drawings, it is not intended to limit the scope of the invention. Various modifications and changes may be made by those skilled in the art without inventive step within the scope of the appended claims.