CN113676065A

CN113676065A - Model prediction control method and system of Vienna rectifier for optimizing voltage vector

Info

Publication number: CN113676065A
Application number: CN202110994917.8A
Authority: CN
Inventors: 张承慧; 刘畅; 邢相洋; 张博学; 张洪亮
Original assignee: Shandong University
Current assignee: Shandong University
Priority date: 2021-08-27
Filing date: 2021-08-27
Publication date: 2021-11-19
Anticipated expiration: 2041-08-27
Also published as: CN113676065B

Abstract

The present disclosure provides a method and system for model prediction of a vienna rectifier for optimizing voltage vectors, comprising: sampling the Vienna rectifier at the current moment to obtain three-phase input current and voltage of the Vienna rectifier at the current moment and power grid voltage; inputting the three-phase input current and voltage of the Vienna rectifier at the current moment and the power grid voltage into the mathematical model of the Vienna rectifier to obtain the reference predicted input voltage at the next moment; analyzing the influence of all different types of candidate space voltage vectors on the balance of the point based on a finite set model, and constructing a cost function without a weight factor to obtain the minimum value of the cost function; and (4) performing midpoint floating judgment, calculating the optimal action time of the candidate vector according to the minimum value of the cost function, and performing duty ratio calculation so as to control the action of a rectifier switching tube.

Description

Model prediction control method and system of Vienna rectifier for optimizing voltage vector

Technical Field

The present disclosure relates to the field of switching converter technologies, and in particular, to a method and a system for model prediction of a vienna rectifier for optimizing a voltage vector.

Background

The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.

In order to solve the problem that the conventional fuel oil automobile consumes fossil energy, an electric automobile using electricity as clean energy enters the field of people, and a vienna rectifier is used as an AC/DC converter and is also applied to an electric automobile charging pile.

In high-voltage and high-power occasions, the diode clamping three-level rectifier is widely applied due to the advantages of high voltage-resistant level, more output voltage levels, less alternating current harmonic waves, low voltage stress of a switching device and the like. However, compared with the conventional two-level rectifier, the three-level rectifier has many switching devices and is complex to regulate and control. Therefore, in order to combine the characteristics of a three-level high-quality output waveform with the advantage of few two-level switching devices, a three-level vienna rectifier is produced. The vienna rectifier has been widely applied to electric vehicle charging systems and industrial applications due to high efficiency and small number of switches, and has a great application prospect.

However, since the vienna rectifier has two diodes per phase, it is necessary to require the input current and voltage to be in phase, otherwise serious zero-crossing distortion will occur, and this distortion is coupled with the control of the dc-side midpoint balance. Improper regulation can increase the input current harmonics, reducing system efficiency. Increasing the inductance can reduce current harmonics, but increases the volume and increases the visual cost. In addition, large switching losses reduce system life, and uncertain switching frequency increases the design difficulty of the filter. These problems all affect the safe, stable, efficient, reliable operation of vienna rectifiers, greatly increasing system losses.

As the technology of microprocessors has been developed,model Predictive Control (MPC) enables multi-objective synchronous optimization control without the need for PWM modules or proportional-integral (PI) controllers. It is a method of predicting and optimizing control strategies by building an online model of the controlled process, calculating the error between the desired and predicted objects in a cost function, and applying constraints in each sampling period. Based on the advantages of the MPC, multi-target control of the Vienna rectifier can be realized. However, as the converter voltage level increases, the number of MPC calculations grows exponentially. For example, for a three-phase rectifier, there are 8 (2) in two levels³) Vector of 27 (3) of three levels³) And (4) a vector. Furthermore, in performing multi-objective MPC, weighting factors are often introduced into the cost function. Not only does the weighting factors increase the computational burden of the MPC, but the process of selecting and tuning is often a tedious task. In order to improve the calculation efficiency, the prior art proposes an MPC method based on simplified simulated annealing, which performs online optimization on the weight factors, rather than removing the weight factors. However, the remaining weight factors still increase the computational burden. Therefore, one of the problems to be overcome by MPC is to reduce the amount of calculation and eliminate the weight factor and increase the response speed of the controller while achieving multi-objective optimization control.

In recent years, there have been a number of researchers performing multi-objective MPC control strategies for vienna rectifiers. The following were used:

(1) an improved MPC is proposed to handle multiple control targets. However, the above method selects only one space vector to track the reference voltage within one sampling time, increasing current harmonics.

(2) A dual vector MPC method is proposed to improve the current quality. Unfortunately, due to the particularity of the vienna rectifier topology, severe current zero crossing distortion occurs when the voltage and current are not in the same direction. Therefore, the conventional dual vector MPC method cannot be directly migrated to the vienna rectifier.

(3) A dual vector MPC method for vienna rectifier topology without weighting factors is presented. However, the reduction of current harmonics is largely influenced by the size of the inductor, and above all, this method cannot fix the switching frequency in each cycle, nor does it consider reducing the switching losses.

In summary, for a three-level vienna rectifier system in an electric vehicle charging system, there are problems of unfixed switching frequency, large switching loss, large current harmonics, distortion of zero-crossing points of alternating-current side current, imbalance of direct-current side midpoint voltage, and large calculation burden and weight factor included in a cost function when a traditional MPC realizes multi-target control.

Disclosure of Invention

An object of the embodiments of the present disclosure is to provide a method and a system for model prediction of a vienna rectifier that optimizes a voltage vector, a method for model prediction that reduces input current harmonics of a three-level vienna rectifier system and outputs a fixed switching frequency, and on the basis of the method, reduces switching losses in adjacent sampling times. The method can effectively eliminate the zero crossing point distortion of the alternating current side, balance the midpoint voltage, reduce the harmonic wave of the input current, fix the switching frequency of each period, optimize the switching sequence of adjacent sampling periods, reduce the switching loss, realize the whole multi-target control by a model prediction method without a weight factor cost function, and solve the control coupling problems of the zero crossing point distortion of the alternating current side and the unbalanced midpoint voltage of the direct current side.

The first aspect of the invention provides a model prediction method of a vienna rectifier for optimizing a voltage vector, which is realized by the following technical scheme:

a method of model prediction for a vienna rectifier that optimizes a voltage vector, comprising:

sampling the Vienna rectifier at the current moment to obtain three-phase input current and voltage of the Vienna rectifier at the current moment and power grid voltage;

inputting the three-phase input current and voltage of the Vienna rectifier at the current moment and the power grid voltage into the mathematical model of the Vienna rectifier to obtain the reference predicted input voltage at the next moment;

analyzing the influence of all different types of candidate space voltage vectors on the balance of the point based on a finite set model, and constructing a cost function without a weight factor to obtain the minimum value of the cost function;

and (4) performing midpoint floating judgment, calculating the optimal action time of the candidate vector according to the minimum value of the cost function, and performing duty ratio calculation so as to control the action of a rectifier switching tube.

In a further technical solution, the building of the mathematical model of the vienna rectifier specifically includes:

carrying out Clark transformation on three-phase input current of the Vienna rectifier and the power grid voltage at the current moment to obtain three-phase input current and power grid voltage under an alpha beta-orthogonal coordinate system;

calculating three-phase reference input current of the Vienna rectifier at the current moment by adopting voltage outer loop control;

obtaining the power grid voltage and the reference input current at the next moment based on the Lagrange extrapolation theorem;

and constructing a dynamic model of the voltage vector based on the discrete time domain to obtain the three-phase reference input voltage at the next moment.

According to the further technical scheme, the candidate space voltage vector is generated based on three-phase reference input current of the Vienna rectifier at the current moment, a space vector diagram of the Vienna rectifier is divided into large sectors according to positive and negative polarities of the three-phase input current, then each large sector is divided into small sectors, and the candidate space voltage vector is obtained according to space vector classification of the Vienna rectifier.

In a further technical scheme, the calculation of the minimum value of the cost function specifically includes: all the vertexes of the polygon formed by the candidate space voltage vectors are in the cost function without the weighting factor, and then the cost function results of each adjacent vertex are added, wherein the smallest of all the calculation results is the area where the three-phase input current of the Vienna rectifier is located at the current moment.

In a further technical scheme, the midpoint floating judgment comprises the following steps: and combining the small sectors into a new large sector, carrying out midpoint floating judgment, selecting a proper N-type or P-type sequence to balance the midpoint according to the fluctuation condition of the midpoint, and realizing the fixation of the switching frequency in each sampling period.

In a further technical scheme, the combination of the small sectors into a new large sector is based on the parity of the small sectors for merging and classification.

In a further technical solution, the fixing of the switching frequency in each sampling period by the appropriate N-type or P-type sequence includes: only one of the two phases in each group of sequences acts once in each period, and the other phase of switch keeps unchanged, so that the switching frequency in each sampling period is fixed.

The second aspect of the present invention provides a model prediction system for a vienna rectifier that optimizes a voltage vector, which is implemented by the following technical solutions:

a model prediction system for a Vienna rectifier that optimizes a voltage vector includes:

a data acquisition module configured to: sampling the Vienna rectifier at the current moment to obtain three-phase input current and voltage of the Vienna rectifier at the current moment and power grid voltage;

a reference input voltage prediction module configured to: inputting the three-phase input current and voltage of the Vienna rectifier at the current moment and the power grid voltage into the mathematical model of the Vienna rectifier to obtain the reference predicted input voltage at the next moment;

a cost function calculation module configured to: analyzing the influence of all different types of candidate space voltage vectors on the balance of the point based on a finite set model, and constructing a cost function without a weight factor to obtain the minimum value of the cost function;

a switch tube control module configured to: and (4) performing midpoint floating judgment, calculating the optimal action time of the candidate vector according to the minimum value of the cost function, and performing duty ratio calculation so as to control the action of a rectifier switching tube.

A third aspect of the invention provides a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor when executing the program implementing the steps of the method for model prediction of a vienna rectifier for arbitrarily optimizing a voltage vector as described above.

A fourth aspect of the invention provides a computer-readable storage medium having stored thereon a computer program which, when being executed by a processor, carries out the steps of the method for model prediction of a vienna rectifier for arbitrary optimization of voltage vectors as described above.

Compared with the prior art, the beneficial effect of this disclosure is:

(1) this application adopts three level vienna rectifier system, compares two level rectifiers of tradition, has the output level number of messenger many, and the characteristics of current quality is high, compares traditional three level rectifier, has the less obvious advantage of switching device, and this complexity that has not only reduced the regulation and control has still practiced thrift the cost of system.

(2) The method is realized by adopting a model predictive control algorithm, the cost function of the method is only used for realizing the tracking of the current, and the method does not contain a weight factor, so that the calculation burden of the controller is greatly reduced. Meanwhile, the model prediction is used as a current inner loop of the whole system, and a PI controller is not contained, so that the adjustment of PI parameters is omitted.

(3) According to the method, a space vector diagram of the three-level Vienna rectifier is divided into small regions in two steps, the first step is divided into six large sectors according to input current, and the second step is used for further dividing the large sectors into 6 small regions according to the prediction of a finite set model. The number of the candidate vectors in each period is reduced from 19 to 8, so that the calculation load is obviously reduced, the problems of zero crossing point distortion of alternating current side current and neutral point voltage balance of direct current side are solved, and decoupling of control over zero crossing point distortion of alternating current side current and neutral point current unbalance of direct current side is realized.

(4) The application provides a method for solving the duty ratio of a vector based on a value function calculation result. The calculation result of the cost function in the method is derived from the prediction of a finite set model for dividing the sector, so that the burden of a controller is not increased by extra calculation, and the optimal duty ratio can be obtained by only simple four-rule operation.

(5) The small sectors divided in the second step are recombined to form different areas with positive and negative polarities to the centering point, and the sequences with the positive and negative polarities in each area can keep the midpoint voltage balance and recover larger voltage fluctuation when the midpoint voltage is unbalanced.

(6) The application provides a method for solving the duty ratio of a vector based on a value function calculation result. In the method, various vectors participate in tracking voltage, so that the input current harmonic can be effectively reduced, and the power quality is improved.

(7) According to the method and the device, sequence combination is optimized during adjacent sampling periods, switching loss can be remarkably reduced, current quality is not affected, and in each sampling period, a five-segment sequence formed by combining a plurality of candidate vectors can fix switching frequency, so that the design of a filter is facilitated.

Drawings

The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure and are not to limit the disclosure.

FIG. 1 is a flow chart of a method for model prediction of a Vienna rectifier for optimizing voltage vectors as disclosed herein;

FIG. 2 is a three-level Vienna rectifier system according to one embodiment of the disclosure;

FIG. 3 is a schematic diagram of a sector division based on input current according to an embodiment of the present invention;

FIG. 4(a) is a graph showing the effect of a three-level Vienna rectifier large space vector on the DC-side midpoint voltage according to an embodiment of the present invention;

FIG. 4(b) is a graph showing the effect of the space vector of the three-level Vienna rectifier on the DC-side midpoint voltage according to an embodiment of the present invention;

FIG. 4(c) is a graph showing the effect of a P-type small vector of a three-level Vienna rectifier on the DC-side midpoint voltage according to an embodiment of the present invention;

FIG. 4(d) is a diagram illustrating the effect of the N-type small vector of the three-level Vienna rectifier on the DC-side midpoint voltage according to an embodiment of the present invention;

FIG. 5(a) is a large sector partition diagram of a space vector diagram of a disclosed Vienna rectifier according to an embodiment of the present invention;

fig. 5(b) is a small sector partition diagram of a space vector diagram of a vienna rectifier according to an embodiment of the present invention;

FIG. 6(a) is a first sector N-type region according to an embodiment of the present invention;

FIG. 6(b) is a first sector P-type region according to an embodiment of the present invention;

FIG. 6(c) is a second sector N-type region according to an embodiment of the present invention;

FIG. 6(d) is a second sector P-type region according to an embodiment of the present invention;

FIG. 7(a) is a first sector P according to a first embodiment of the present invention_ⅠA sequence diagram of fixed switching frequencies;

FIG. 7(b) is a first sector P according to a first embodiment of the present invention_ⅡA sequence diagram of fixed switching frequencies;

FIG. 7(c) is a first sector N according to an embodiment of the present invention_ⅠA sequence diagram of fixed switching frequencies;

FIG. 7(d) is a first sector N according to an embodiment of the present invention_ⅡA sequence diagram of fixed switching frequencies;

FIG. 7(e) is a first sector N according to an embodiment of the present invention_ⅢA sequence diagram of fixed switching frequencies;

FIG. 7(f) is a first sector N according to an embodiment of the present invention_ⅣA sequence diagram of fixed switching frequencies;

FIG. 7(g) is a first sector N according to an embodiment of the present invention_ⅤA sequence diagram of fixed switching frequencies;

FIG. 7(h) is a first sector N according to an embodiment of the present invention_ⅥA sequence diagram of fixed switching frequencies;

FIG. 8(a) shows a first sector P according to an embodiment of the present invention_Ⅲ-N_ⅢA sequence diagram for reducing switching losses in adjacent sampling periods;

FIG. 8(b) is a first sector P according to a first embodiment of the present invention_Ⅲ-N_ⅤA sequence diagram for reducing switching losses in adjacent sampling periods;

FIG. 8(c) shows a first sector P according to a first embodiment of the present invention_Ⅳ-N_ⅣA sequence diagram for reducing switching losses in adjacent sampling periods;

FIG. 8(d) is the first sector P according to the first disclosure of the present invention_Ⅳ-N_ⅥA sequence diagram for reducing switching losses in adjacent sampling periods;

fig. 9(a) is a simulation diagram of a load of 65 Ω with a fixed switching frequency according to an embodiment of the present invention;

fig. 9(b) is a simulation diagram of a load of 65 Ω with a fixed switching frequency and reduced switching loss according to an embodiment of the present invention;

fig. 9(c) is a simulation diagram of a load of 100 Ω with a fixed switching frequency according to an embodiment of the present invention;

fig. 9(d) is a simulation diagram of a load with a fixed switching frequency and a reduced switching loss of 100 Ω according to an embodiment of the present invention.

Detailed Description

It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.

It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present disclosure. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.

Example one

The embodiment discloses a model prediction method for a vienna rectifier for optimizing voltage vectors, and as introduced in the background art, vienna topology faces fixed switching frequency, switching loss is reduced, and problems of alternating current zero crossing distortion and unbalanced direct current side midpoint voltage and control coupling problems of alternating current zero crossing distortion and unbalanced direct current side midpoint voltage exist at the same time.

As shown in fig. 1, the present application provides a model prediction method of a vienna rectifier for optimizing a voltage vector, the method comprising the steps of:

step S1, sampling the Vienna rectifier at the current moment to obtain the three-phase input current i of the Vienna rectifier at the current moment k_x(k) And voltage u_xo(k) And the network voltage e_x(k)；

As shown in fig. 2, the input voltage of the vienna rectifier satisfies:

in the formula, e_xFor the mains voltage, i_xAnd u_xoFor three-phase input current and voltage, x ═ a, b, c, R, and L are filter inductance and equivalent series resistance, respectively.

Step S2, constructing a mathematical model of the Vienna rectifier to obtain the reference predicted input voltage at the next moment;

step S2-1, inputting the three-phase input current i of the Vienna rectifier at the current moment k_x(k) And the network voltage e_x(k) Performing Clark transformation to obtain three-phase input current i under an alpha beta-orthogonal coordinate system_y(k) And the network voltage e_y(k) Wherein y is α, β.

Step S2-2: input of i_y(k)、e_y(k) And sampling the DC side voltage V in real time_dcAnd the required DC side voltage V^* _dcCalculating the three-phase reference input current i of the Vienna rectifier at the kth moment by adopting voltage outer loop control^* _y(k) Wherein y is α, β.

Step S2-3, obtaining the grid voltage e at the (k +1) th moment based on the Lagrange' S extrapolation theorem_y(k +1) and a reference input current i^* _y(k +1) respectively expressed as:

wherein e is_y(k) Representing the three-phase network voltage k at the present time, e_y(k-1) represents the three-phase network voltage at the preceding time k-1, e_y(k-2) represents the three-phase grid voltage at the first two moments k-1; i.e. i^* _y(k) Representing the three-phase reference input current, i, at the present moment k^* _y(k-1) represents the three-phase reference input current, i, at the previous instant k-1^* _y(k-2) represents the three-phase reference input current at the first two time instants k-1.

Step S2-4: assuming NP voltage balance, constructing a dynamic model of a voltage vector based on a discrete time domain under an alpha beta-orthogonal coordinate system to obtain a (k +1) th time three-phase reference input voltage u^* _y(k+1)；

The calculation method comprises the following steps:

in the formula i^* _y(k +1) is the (k +1) th instant three-phase reference input current, e_y(k +1) is the grid voltage at the (k +1) th moment, u^* _y(k +1) (y ═ α, β) is the (k +1) th time three-phase reference input voltage, T_sIs the sampling period, M is a constant value defined as:

M＝RT_s+L (4)；

step S3, analyzing the influence of all different kinds of candidate vectors on the center point balance based on a finite set model, and constructing a cost function without a weight factor lambda, wherein the minimum value is the three-phase input current i of the Vienna rectifier at the current moment k_x(k) The area in which it is located.

Step S3-1: based on the three-phase reference input current of the Vienna rectifier at the current moment, dividing a space vector diagram of the Vienna rectifier into large sectors according to the positive and negative polarities of the three-phase input current, dividing each large sector into small sectors, and classifying according to the space vectors of the Vienna rectifier to obtain candidate space voltage vectors; table 1 shows the classification of the space vector diagram of the vienna rectifier:

table 1 space vector classification of vienna rectifiers

As shown in fig. 3, the space vector diagram of the vienna rectifier can be divided into 6 large sectors according to the positive and negative polarities of the three-phase input current, for example, assuming i_x(k) Located in sector II, where the signs of the three input currents are [ + ] -]. In which two redundant small voltage vectors [ NON ]]And [ OON]Severe current distortion occurs and therefore cannot be used as a candidate vector. That is, only 8 space voltage vectors (V) are selected₀[OOO]、V₁[POO]、 V₃[PPO]、V₄[OON]、V₅[OPO]、V₁₄[PPN]、V₁₉[PON]And V₂₀[OPN]) Synthesis of V_ref. The number of candidate vectors will be reduced from 19 to 7 (including two redundant small vectors), which also reduces the computational burden by more than half.

Step S3-2: constructing a cost function containing a weight factor;

defining the capacitance currents of the upper direct current link and the lower direct current link as follows:

in the formula, V_PAnd V_NVoltages of upper and lower capacitors, i_pAnd i_nIs the current through the upper and lower DC link capacitors, C_PAnd C_NRespectively, the capacitance values of the upper and lower capacitors. Usually, they are the same, and are all set to C.

Discretizing the formula (5) to obtain a discrete time model:

in the formula i_n(k +1) and i_p(k +1) is the predicted current passing through the upper and lower direct current link capacitors at the (k +1) th moment respectively; v_N(k +1) and V_P(k +1) are the predicted upper and lower capacitor voltages, respectively.

Further, a cost function including current tracking and midpoint balancing is obtained and specifically expressed as:

in the formula, λ is a weighting factor.

Step S3-3: analyzing the influence of all different types of candidate vectors on the point balance, and constructing a cost function without a weight factor lambda;

large vector [ PNN ] in fig. 4(a) taking sector 1 as an example]There is no effect on the NP voltage because none of the three phase voltages are connected to the neutral point. And the medium vector [ PON ] in FIG. 4(b)]The three phases are connected to the P, O, N state respectively, so that the charging and discharging of the direct current link capacitor are uncertain, and the NP voltage is unbalanced. Since B, C is connected to neutral point and P is connected to P type small vector POO]Can discharge the upper capacitor CP and the lower capacitor C_NCharging is performed as shown in fig. 4 (c). N type small vector [ ONN]The lower capacitance C in FIG. 4(d) can be made_NDischarging and making the upper capacitor C_PAnd (6) charging. Thus, the midpoint voltage can be balanced by small vectors of the P-type or N-type, which have opposite effects on the neutral point.

Therefore, the elimination of the weight factor lambda in the cost function (7) is realized, the balance of neutral points is realized, and the calculated amount and the complicated weight factor test work are greatly reduced. Thus, the new cost function only includes one term to predict the tracking current, and the new cost function without the weighting factor is:

g₂(k)＝|u^* _α(k+1)-u_α(k+1)|²+|u^* _β(k+1)-u_β(k+1)|² (8)

through the scheme, the control decoupling of the zero crossing point distortion of the alternating current side current and the neutral point voltage unbalance of the direct current side is realized, the influence of all different types of vectors on the neutral point balance is analyzed, and the decoupling of the current distortion is realized. And constructing a cost function, wherein the cost function only comprises the effect of realizing the tracking current, and the weight factor is removed during decoupling.

Step S3-3: calculating the cost function of all the vertexes of the polygon formed by the candidate space voltage vectors without the weight factor lambda, and then adding the cost function results of each adjacent vertex, wherein the minimum of all the calculation results is the three-phase input current i of the Vienna rectifier at the current moment k_x(k) The area in which it is located.

Taking the region I shown in FIG. 5(a) as an example, the polygon is hexagonal, L₁、M₁、M₂、S₁、S₂、Z₁、 O₁Respectively representing the calculation results of the cost functions of a large voltage vector, two medium voltage vectors, two small voltage vectors, a zero vector and a central voltage vector. The formula for determining the six small regions is:

region (i) ═ min { p }_i},i∈{1,2,3,4,5,6}

If S is₁And M₂P of (a)₃The sum being p_i(i is 1,2 … … 6) minimum value, then V_refIn the region shown in fig. 5 (b).

Step S4: and (4) performing midpoint floating judgment, calculating the optimal action time of the candidate vector according to the minimum value of the cost function, and performing duty ratio calculation so as to control the action of a rectifier switching tube.

Step S4-1, combining the small sectors X in the step S3-1 into a new large sector Y as shown in fig. 5(a) and 5(b), performing midpoint floating judgment, and selecting a proper N-type or P-type sequence to balance the midpoint according to the fluctuation condition of the midpoint;

wherein the process of combining the small sectors X into a new large sector Y comprises:

to balance the NP voltage, as long as each region contains both N-type and P-type switching sequences, large sectors need to be sorted by parity due to the different characteristics of odd and even sectors.

Take sector 1 and sector 2 analyses in fig. 6(a) -6 (d) as an example. In sector 1, as shown in FIG. 6(a), due to the P-type vector [ PPO ]]And [ POP ]]Is removed and only one P-type vector [ POO ] remains in the entire sector]. To ensure that sequences in 6 cells tend to be P-type, each sequence must contain a POO]And at the same time cannot contain any N-type small vectors. Thus, N is_ⅢAnd N_ⅤMerge into one region, denoted as P in FIG. 6(b)_Ⅲ. Also, P_ⅣOr from N_ⅢAnd N_ⅤAnd (6) merging. At the same time, P_ⅢAnd P_ⅣThe P-type sequences in (1) can be respectively expressed as [ OOO ]]-[POO]-[PON]-[POO]-[OOO]And [ OOO ]]-[POO]-[PNO]-[POO]-[OOO]. For P_ⅠAnd P_ⅡThe P-type sequence can be respectively designed as [ PNN ]]-[PON]-[POO]-[PON]-[PNN]And [ PNN]-[PNO]-[POO]-[PNO]-[PNN]. However, N-type small vectors are not deleted in sector 1. Thus, there is no need to merge P-type sectors as previously described, and sector 1 is still divided into 6 negative regions (N) in fig. 6(a)_Ⅰ,N_Ⅱ,N_Ⅲ,N_Ⅳ,N_ⅤAnd N_Ⅵ). In contrast to the sector P-type sequences, each N-type sequence only contains at least one small N-type vector, rendering the entire sequence negative.

Thus, sector 1 may also be N-type. According to the above division, sector 1 contains both N-type and P-type switching sequences, which are determined by the magnitude of the fluctuation of the NP voltage. In addition, as can be seen from fig. 7(a) -7 (h), the switching sequence is a five-segment sequence. In each sampling period, one of the three phases is always clamped and kept unchanged, and the switches of the other two phases are only operated once, so that the switching frequency is fixed while the midpoint balance is realized.

For even-numbered large sectors, the situation is opposite to that for odd-numbered large sectors. Since the N-type small vector is not eliminated in the even large sector, taking the second sector as an example, P can be eliminated_ⅢAnd P_ⅤAre combined into N_ⅢA 1 is to P_ⅣAnd P_ⅥAre combined into N_ⅣAs shown in fig. 6(c) and (d). Similarly, consistent with the conclusion of the odd sector in S4-1, sector 2 can contain both N-type and P-type sequences, and in each sampling period, one of the three phases is always clamped and kept unchanged, while the switches of the other two phases are only operated once, so that the switching frequency is fixed while the midpoint balance is realized.

Step S4-2, fixing the switching frequency in each sampling period according to the balanced midpoint N-type or P-type sequence;

any sequence is a five-segment sequence formed by three vectors, as shown in fig. 7(a) -7 (h), only one of two phases in each sequence acts once in each period, and the other phase of switch keeps unchanged, so that the switching frequency in each sampling period is fixed;

taking the switching losses in adjacent sampling periods into account, the novel sequence is shown in fig. 7(a) -7 (h). By applying the group of sequences, the harmonic amplitude at the switching frequency can be obviously reduced, and the switching loss is reduced;

in N_ⅢAnd N_ⅤIn (e), the switching sequences are [ PON ] shown in FIGS. 8(e) and (g), respectively]-[OON]-[ONN]-[OON]-[OOO]And [ OOO ]]-[OON]-[OOO]. In N_ⅢIn, the start and end of the switching sequence are medium vectors [ PON ]]And in N_ⅤMiddle is zero vector [ OOO]. Furthermore, N_ⅢAnd N_ⅤThe sum of the areas of (A) is P_ⅢThe range of (1). Thus, suppose a P-type sequence is required and V_refAt N_ⅢThen the P-type sequence is designed as [ PON ]]-[POO]-[OOO]-[POO]-[PON]As shown in fig. 8 (a); when V is_refAt N_ⅤWhen the P-type sequence of the demand is designed as [ OOO ]]-[POO]-[PON]-[POO]-[OOO]As shown in fig. 8 (b). Thus, if an N-type or P-type sequence is required to balance the midpoint voltage, there is no switching between adjacent sampling periods. Likewise, according to V_refAt N_ⅣOr N_ⅥIn the region, the desired P-type sequence was also designed as [ PNO ] according to the above method]-[POO]-[OOO]-[POO]-[PNO]And [ OOO ]]-[POO]-[PNO]-[POO]-[OOO]As shown in fig. 8(c) and 8 (d). At the same time, in N_ⅣAnd N_ⅥOf (2) aIn, the N-type sequence is designed as [ PNO ]]-[ONO]-[ONN]-[ONO]-[PNO]And [ OOO ]]-[ONO]-[ONN]-[ONO]-[OOO]. It follows that the switching losses between adjacent sampling periods can be reduced by the specific sequence described above.

P for sector 1_ⅠAnd N_ⅠIn addition, the P-and N-type sequences shown in FIGS. 7(a) and (c) are [ PNN ] respectively]-[PON]-[POO]-[PON]-[PNN]And [ PON ]]-[PNN]-[ONN]-[PNN]-[PON]. It can be seen that the above sequence avoids direct switching between vienna rectifier states P and N when the sequence needs to switch in the next sampling period. Likewise, the first sector P_ⅡAnd N_ⅡAlso in line with the above rule, as shown in fig. 7(b) and (d), all switching sequences of the second sector are summarized in the following table.

TABLE II sector 2 candidate switching sequences

Step S4-3, calculating the optimal action time of three vectors by using a value function calculation value obtained by the prediction control of a finite set model, and redistributing the optimal action time to a five-segment PWM wave-forming so as to control the action of a rectifier switching tube;

conventional SVPWM uses the three closest voltage vectors to synthesize V by volt-second balancing_ref. Its constraint formula can be expressed as:

wherein dc₁、dc₂、dc₃Three candidate vectors Vc respectively₁、Vc₂、Vc₃The duty cycle of (c).

As can be seen from equation (9), the cost function used herein calculates V in each large sector_refDistance to 7 fixed vectors. This means that the larger the result of the calculation of the cost function, the larger the fixed vector is from V_refThe further away. Therefore, the corresponding duty ratio is small. Overall, duty cycleInversely proportional to the result of the calculation of the cost function. Suppose that V is determined_refIn the region, the duty ratio is defined as:

wherein K is a scale factor and K is greater than or equal to 0, g_cxIs three candidate vectors Vc₁、Vc₂、Vc₃The result of the calculation of the corresponding cost function (9). Substituting (10) into (11), and calculating to obtain a scaling factor K as follows:

since the results of the cost function are all non-negative, it is ensured that the scale factor K is greater than or equal to 0. Thus, the calculated duty cycle is also non-negative and can be expressed as:

based on the duty cycles of the three candidate vectors, a five-segment sequence can be generated to drive the vienna rectifier.

In the embodiment, the peak value of the voltage at the alternating current side is 150V, the required voltage at the direct current side is 400V, the equivalent load at the direct current side is 65 omega/100 omega, the size of the filter inductor is 5mH, and the sampling period is 100 mus. Fig. 9 is a simulation result diagram of a method of fixing the switching frequency, so that the current frequency is mainly concentrated on the switching frequency, i.e., integer multiples of 10kHz, such as 20kHz and 30kHz, which is advantageous for the design of the filter. Since (c) and (d) in fig. 9 optimize the sequence switching order between adjacent sampling periods, the harmonic amplitude is significantly smaller at a frequency of 10kHz than in (a) and (c) in fig. 9. For example, when the load size is 65 Ω, the amplitude is reduced from 0.7 to 0.6; when the load size is 100 Ω, the amplitude decreases from 1.3 to 0.9.

In addition, the difference in THD between the two methods is small, and the THD for optimizing the sequence switching sequence between adjacent sampling periods is slightly higher than that of the non-optimized method, because the overall switching frequency is reduced, but under the condition of different loads of 65 Ω/100 Ω, the THD is still less than 5%, and the input current waveform quality is high. The method for predicting the model provided by the embodiment is proved to be capable of fixing the switching frequency, reducing the switching loss, improving the electric energy quality and simultaneously improving the safety and reliability of the rectifier.

Example two

The embodiment of the specification provides a transformer oil temperature prediction system based on a variable weight combined model, which is realized by the following technical scheme:

the method comprises the following steps:

a Vienna rectifier building block configured to: constructing a mathematical model of the Vienna rectifier to obtain a reference prediction input voltage at the next moment;

The implementation of the specific modules in this embodiment example can be referred to in the related art in the first embodiment example, and will not be described in detail here.

EXAMPLE III

The present specification provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the method for model prediction of a vienna rectifier for optimizing a voltage vector in the first embodiment when executing the program.

Example four

The present specification provides a computer readable storage medium, on which a computer program is stored, wherein the program is executed by a processor to implement the steps of the method for model prediction of a vienna rectifier for optimizing a voltage vector in the first embodiment.

It is to be understood that throughout the description of the present specification, reference to the term "one embodiment", "another embodiment", "other embodiments", or "first through nth embodiments", etc., is intended to mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, or materials described may be combined in any suitable manner in any one or more embodiments or examples.

The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure, and various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.

Claims

1. A method for model prediction of a Vienna rectifier for optimizing a voltage vector, comprising:

acquiring three-phase input current and voltage of the Vienna rectifier at the current moment and power grid voltage;

2. The method of claim 1, wherein the constructing a mathematical model of the vienna rectifier comprises:

3. The method of model prediction for a vienna rectifier that optimizes a voltage vector as set forth in claim 1, wherein generating the candidate space voltage vector comprises: based on the three-phase reference input current of the Vienna rectifier at the current moment, the space vector diagram of the Vienna rectifier is divided into large sectors according to the positive and negative polarities of the three-phase input current, then each large sector is divided into small sectors, and candidate space voltage vectors are obtained according to the space vector classification of the Vienna rectifier.

4. The method of model prediction for a vienna rectifier for optimizing voltage vectors as defined in claim 1 wherein the minimum value of said cost function is calculated by: all the vertexes of the polygon formed by the candidate space voltage vectors are in the cost function without the weighting factor, and then the cost function results of each adjacent vertex are added, wherein the smallest of all the calculation results is the area where the three-phase input current of the Vienna rectifier is located at the current moment.

5. The method of model prediction for a vienna rectifier for optimizing a voltage vector of claim 1 wherein the midpoint floating determination comprises: and combining the small sectors into a new large sector, carrying out midpoint floating judgment, selecting an N-type or P-type sequence for balancing the midpoint according to the fluctuation condition of the midpoint to balance the midpoint, and fixing the switching frequency in each sampling period.

6. The method of model prediction for a vienna rectifier optimizing voltage vectors as set forth in claim 1, wherein said combining of small sectors into a new large sector is a merge classification based on the parity of small sectors.

7. The method of model prediction for a vienna rectifier for optimizing voltage vectors as claimed in claim 1 wherein said N-type or P-type sequence of equilibrium midpoints performing a fixed switching frequency for each sampling period comprises: only one of the two phases in each group of sequences acts once in each period, and the other phase of switch keeps unchanged, so that the switching frequency in each sampling period is fixed.

8. A model prediction system for a vienna rectifier that optimizes a voltage vector, comprising:

a data acquisition module configured to: acquiring three-phase input current and voltage of the Vienna rectifier at the current moment and power grid voltage;

9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements the steps of the method of model prediction of a vienna rectifier for optimizing a voltage vector of any one of claims 1-7.

10. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method for model prediction of a vienna rectifier for optimizing voltage vectors according to any one of claims 1 to 7.