CN117578899B

CN117578899B - Dynamic optimization and virtual voltage vector sliding mode prediction control method and system

Info

Publication number: CN117578899B
Application number: CN202410053018.1A
Authority: CN
Inventors: 杨勇; 王铀程
Original assignee: Suzhou University
Current assignee: Suzhou University
Priority date: 2024-01-15
Filing date: 2024-01-15
Publication date: 2024-03-29
Anticipated expiration: 2044-01-15
Also published as: CN117578899A

Abstract

The application provides a dynamic optimization sliding mode prediction control method and system based on virtual voltage vectors. The method comprises the following steps: in the output voltage vector diagram, at every twoαThe midpoint of the component is added with one component, at every twoβThe midpoint of the component is added with a component, and then at eachαAndβthe intersection point of the components supplements the voltage vector to obtain a virtual voltage vector, and the virtual voltage vector and the actual voltage vector form a candidate voltage vector; dynamic establishment of control domain based on sliding mode theory, and parallel optimization to obtain output voltage in the control domainαAndβvoltage vector with minimum component cost functionαAndβthe components are synthesized to obtain an optimal voltage vector, monotonicity of a cost function in a control domain is analyzed in a parallel optimization process, and an optimization process is optimized; performing constraint analysis on the obtained optimal voltage vector, and judging whether constraint is needed; and different synthesis modes of the optimal voltage vector are realized by using the redundant small voltage vector, so that neutral point voltage balance is realized. The method suppresses output ripple waves, eliminates weight coefficients in NP voltage balance, and realizes constant switching frequency; the calculated amount is obviously reduced, and the dynamic characteristics of the output voltage are considered by the proposed ESM-PC; the method has better steady-state performance, similar dynamic performance, better model robustness and better neutral point voltage balance performance.

Description

Dynamic optimization and virtual voltage vector sliding mode prediction control method and system

Technical Field

The application relates to the technical field of inverter control, in particular to a dynamic optimization and virtual voltage vector sliding mode prediction control method and system.

Background

Three-phase three-level voltage source inverters with LC filters are the dominant topology in ac microgrids, distributed generation systems, uninterruptible power supplies, and other applications. The main advantages are higher voltage levels, higher efficiency, lower harmonic content and less electromagnetic interference. Nowadays, conventional linear control methods, including Proportional Integral (PI) control, proportional Resonance (PR) control, and the like, have been widely used in three-phase three-level inverter systems. However, as modern power applications increasingly require fast dynamic response and accommodate diverse nonlinear loads, traditional linear control approaches exhibit significant limitations in meeting these changing requirements.

On the other hand, a nonlinear control strategy is an effective solution. Sliding Mode Control (SMC) is a widely used nonlinear control method. In the SMC, switching conditions and control laws are designed according to control targets, so that the system is forced to slide along a predefined sliding surface, and the robustness to load disturbance and modeling errors is enhanced. Many studies have explored the use of SMCs in power converters. A single-stage slip-mode voltage controller with state feedback has been proposed for pulsed load micro-grids with good dynamic response. In addition, a hysteretic SMC technology for grid-connected inverters has been proposed to enhance the robustness to disturbances. However, in a practical discrete system, the control and sampling signals are constant over one period, so that an ideal Sliding motion cannot be achieved, and only a Quasi-Sliding Mode (QSM) near the ideal Sliding surface can be achieved.

However, the above SMC lacks consideration for minimizing tracking error, resulting in large output ripple and difficulty in handling multi-objective optimization problems. Finite control set model predictive control (FCS-MPC) is an effective solution with good dynamic performance and intuitive control architecture. The FCS-MPC first builds a mathematical model of the system, and determines the optimal voltage vector using online rolling optimization to minimize tracking error for the next switching cycle. This control method reduces control delay. The conventional FCS-MPC applies only one voltage vector in one switching period, resulting in a large output ripple and variable switching frequency. To solve this problem, it is possible to apply a plurality of voltage vectors in one switching cycle. Three voltage vectors have been applied in a switching cycle, with the time of action of each voltage vector being inversely proportional to its cost value. FCS-MPC (OSS-MPC) based on an optimal switching sequence has been proposed, and an explicit solution to minimize the cost function is solved, with a better control effect. Both of the above methods rely on modulation. In addition, a virtual voltage vector is designed, and a constant switching frequency can be realized.

The modulation and virtual voltage vector based methods essentially enlarge the control domain, but complicate the optimization process. Therefore, preselection of the voltage vectors is essential, in particular for three-phase three-level inverters. There have been people to limit the search space in combination with Dead Beat (DB) control. However, the accuracy of DB depends on system parameters, with poor robustness. Preselection schemes based on medium voltage vectors and tolerant sequences have been proposed respectively, so that the computational complexity is effectively reduced. However, the above methods only consider that the voltage vector is limited, but that the real and imaginary components of the voltage vector are ignored.

Meanwhile, the realization of Neutral Point (NP) voltage balance is also crucial for a three-phase three-level inverter, and a common strategy is to add a weight term in a cost function. However, the weighting factors increase the complexity of the algorithm. Another way to avoid the use of weight factors is to use different redundant small voltage vectors in combination, automatically balancing the NP voltage.

In conclusion, a novel efficient optimization method is explored, and the advantages of the SMC and the FCS-MPC are fully combined, so that the method has important significance. In recent years, some scholars have conducted preliminary research on this. The space vector diagram is divided into 12 sectors and preselected according to the position of the power grid voltage and the sliding mode theory so as to reduce the calculated amount. However, this approach ignores the dynamics of the grid voltage, limiting the degree of optimization. A method of adding the arrival condition of the SMC to the cost function has been proposed to find the maximum negative value of the cost function. In addition, the sliding mode function is explicitly integrated into the cost function for the first time, so that the robustness of the algorithm is enhanced. However, the cost function does not take into account the minimum tracking error and introduces parameter adjustment problems.

The conventional FCS-MPC only acts on one voltage vector in one switching period, not only resulting in an indefinite switching frequency; meanwhile, the output voltage has large ripple, poor steady state and dynamic performance, is easily influenced by system parameters, and has the problem of complex weight coefficient adjustment. Although the traditional SMC can improve the robustness to modeling errors and interference, the traditional SMC lacks consideration to tracking errors and has large output ripple.

Disclosure of Invention

In view of this, the present application aims to provide a dynamic optimization and virtual voltage vector sliding mode prediction control method and system, which can solve the existing problems in a targeted manner. The application proposes an extended sliding mode predictive control (ESM-PC) for a three-phase three-level Voltage Source Inverter (VSIs) operating in island mode. The control method provided by the invention is simple to realize, can be suitable for different working conditions, and ensures the suppression of the resonance of the filter while relieving the buffeting of the system.

Based on the above objects, the present application proposes a dynamic optimization and virtual voltage vector sliding mode prediction control method, including:

dynamically establishing a control domain based on sliding mode theory, and respectively obtaining output voltages in the control domain through parallel optimizationαAndβvoltage vector with minimum component cost functionαAndβthe components are synthesized to obtain an optimal voltage vector;

analyzing monotonicity of the cost function in the control domain, and optimizing the optimizing process;

by setting virtual voltage vectors, the voltage vectors are enlargedαAndβnumber of candidates for the component;

based on the candidate voltage vectors, different synthesis modes of the optimal voltage vector are realized by using the redundant small voltage vectors, and neutral point voltage balance is realized.

Based on the above objects, the present application further provides a dynamic optimization and virtual voltage vector sliding mode prediction control system, which includes:

αandβthe component module is used for receiving the input voltage of the limited control set and comprises a real voltage vector and a virtual voltage vector;

the control domain module dynamically establishes a control domain based on a sliding mode theory and according to the output voltageαAndβmonotonicity of the component cost function to obtain an optimal voltage vectorαAndβthe components are synthesized to obtain an optimal voltage vector;

and the neutral point voltage balancing module realizes different synthesis modes of the optimal voltage vector by utilizing the redundant small voltage vector according to the NP voltage based on the optimal voltage vector, and realizes neutral point voltage balancing.

Overall, the advantages of the present application and the experience brought to the user are: 1. the application provides a method based onαAndβa virtual voltage vector design method of components. The method suppresses output ripple waves, eliminates weight coefficients in NP voltage balance, and realizes constant switching frequency; 2. the application provides a new optimization method for carrying out parallel optimization on real-axis and imaginary-axis components. Monotonicity of the cost function and constraint conditions in the search space are analyzed, so that the calculated amount is remarkably reduced by about 56% compared with the traditional method; 3. the ESM-PC provided by the application considers the dynamic characteristic of the output voltage, and the optimization method is dynamic; 4. compared with the traditional method, the improved ESM-PC algorithm provided by the application has better steady-state performance, similar dynamic performance, better model robustness and better neutral point voltage balance performance.

Drawings

In the drawings, the same reference numerals refer to the same or similar parts or elements throughout the several views unless otherwise specified. The figures are not necessarily drawn to scale. It is appreciated that these drawings depict only some embodiments according to the disclosure and are not therefore to be considered limiting of its scope.

Fig. 1 shows a three-phase three-level inverter circuit topology with LC filters of the present application.

Fig. 2 shows a voltage vector distribution diagram according to an embodiment of the present application.

FIG. 3 shows a control domain schematic diagram for different operating conditions.，/>；/>，；/>，/>；/>，/>。

Fig. 4 shows a monotonicity diagram of a cost function according to an embodiment of the present application. (a)；(b)/>。

Fig. 5 shows a schematic diagram of constraints according to an embodiment of the present application. (a) an optimal voltage vector that does not exist; (b) illustrative case.

FIG. 6 shows an algorithm flow diagram of a parallel optimization strategy according to an embodiment of the present application.

Fig. 7 shows a voltage vector distribution diagram according to an embodiment of the present application. (a) a real voltage vector and a virtual voltage vector; (b) voltage vector in large sector I.

Fig. 8 shows an overall system control block diagram according to an embodiment of the present application.

Fig. 9 shows a steady-state experimental waveform schematic of a linear nonlinear load according to an embodiment of the present application. (a) a legacy MPC; (b) DB-MPC; (c) ESM-PC.

Fig. 10 shows a schematic of FFT analysis of steady state experiments according to embodiments of the present application. (a) a legacy MPC; (b) DB-MPC; (c) ESM-PC.

Fig. 11 shows a dynamic experimental waveform diagram of a linear resistive load stepped from 30Ω to 50Ω according to an embodiment of the present application. (a) a legacy MPC; (b) DB-MPC; (c) ESM-PC.

Fig. 12 shows a schematic diagram of a dynamic experimental waveform of a three-phase load from a linear load to a parallel connection of a linear and nonlinear load according to an embodiment of the present application. (a) a legacy MPC; (b) DB-MPC; (c) ESM-PC.

Fig. 13 shows a schematic diagram of neutral point voltage balance experiment waveforms according to an embodiment of the present application. (a) a legacy MPC; (b) DB-MPC; (c) ESM-PC.

Fig. 14 shows a configuration diagram of a dynamic optimization and virtual voltage vector sliding mode predictive control system according to an embodiment of the present application.

Fig. 15 shows a schematic structural diagram of an electronic device according to an embodiment of the present application.

Fig. 16 shows a schematic diagram of a storage medium according to an embodiment of the present application.

Detailed Description

The present application is described in further detail below with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings.

It should be noted that, in the case of no conflict, the embodiments and features in the embodiments may be combined with each other. The present application will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.

Multilevel inverter: the multi-level inverter is an inverter capable of outputting multiple levels per bridge arm, and compared with the traditional two-level inverter, the multi-level inverter can not only output more voltage levels, thereby reducing the voltage stress of a switching device, generating output voltage which is closer to sine waves, and reducing harmonic content and electromagnetic interference. Meanwhile, the multilevel inverter can realize higher voltage and current regulation precision, and reduces the loss in the energy conversion process. Common multi-level topologies include diode clamping, capacitive clamping, or cascaded H-bridges, among others.

Finite control set model predictive control (Finite Control Set Model Predictive Control, FCS-MPC): the FCS-MPC first builds a discrete mathematical model of the system, then calculates the control inputs that minimize the cost function by rolling optimization, and acts on the next control cycle. FCS-MPC can simply handle the multi-objective optimization problem, and for power electronic inverters the control inputs are limited, so that the cost value of each control input can be calculated, however, complex calculation procedures and cumbersome weight coefficient adjustments limit the development of FCS-MPC.

Slip-form control (Sliding Mode Control, SMC): SMC is a design method of a control system, and the main idea is to make the dynamic behavior of the system on a sliding mode surface simpler and easier to control by introducing the sliding mode surface. In SMC, the state of the system slides through a sliding mode surface, so that stable control of the system is realized. The control method has stronger robustness to system parameter variation and external interference, so the control method is widely applied to some industrial applications and control system designs, and in a discrete control system, a control signal and a sampling signal are unchanged in one control period, so an ideal sliding mode state cannot be realized, and only a discrete quasi-sliding mode state can be realized.

The application provides a new extended sliding mode predictive control (ESM-PC), which combines SMC and FCS-MPC, and firstly, dynamically establishes a control domain by combining a sliding mode theory; then, a new parallel optimization strategy is proposed to further improve the computational efficiency. In order to suppress the output ripple, a base onαAndβa virtual voltage vector design method of components. And two output modes are designed for candidate voltage vectors, and the redundant small voltage vectors are utilized to realize NP voltage automatic balance without adjusting weight coefficients.

A. System model

Fig. 1 is a circuit diagram of a T-type three-phase three-level inverter with LC filter as used in the present application, wherein,is a direct current side voltage; />And->Upper and lower capacitors, respectivelyA voltage of the device; />，/>And->The current flowing through the upper capacitor and the lower capacitor and the neutral point current respectively; />And->An LC filter constituting an output; />The inductor current is filtered; />And->The voltage and the output current of the filter capacitor are respectively; />，And->Linear resistance loads of A, B and C phases respectively; />And->The output filter capacitor and the resistor load of the diode rectifier bridge load are respectively adopted.

For each leg of a T-type three-phase three-level inverter, there are three possible output states, defined as "P", "O", and respectively"N", wherein "P" represents a switching deviceAnd->Conducting; "O" means switching device->And->Simultaneously turned on, "N" represents a switching device->And->While conducting. Thus, as shown in FIG. 2, there are a total of 3 ³ (27) A possible output state. The magnitude of the voltage vector can be further divided into four voltage vectors, namely zero, small, medium and large.

Switching functionThe definition is as follows:

(1)

assuming that the NP voltage has reached an effective balance, each phase is "relative to the load neutral point" N"output voltage，/>And->Can be expressed as:

(2)

will filter the capacitor voltageAnd filter inductor current->Defined as state variables, can be obtained by Clark transformation from the dynamic relationship of the LC filter in FIG. 1αβContinuous time dynamic model under coordinate system:

(3)

wherein,representing the output voltage of the inverter, if the sampling time is small enough, discretizing the continuous model by a forward Euler method, and obtaining the recursive expression of the state variables at the (k+1) th and (k) th moments, wherein the discrete state space model of the formula (3) is as follows:

(4)

wherein:

(5)

notably, the modeling described above is based on balanced NP voltages. In practical applications, the dc side capacitor voltage needs to be analyzed. As shown in fig. 1, the current and voltage relationship of the dc side upper and lower capacitors can be expressed as:

(6)

dynamic and neutral point current of NP voltageRelated to (I)>Can be expressed as:

(7)

similarly, the forward Euler method may be used for discretization. Thus, the voltage difference of the upper and lower capacitances at the (k+1) th time can be obtained:

(8)

wherein the method comprises the steps of。

B. Legacy FCS-MPC

The legacy FCS-MPC first establishes a predictive equation for the state variables and then determines the optimal control inputs by rolling optimization. Thus, the nominal prediction model is derived from equation (4) as follows:

(9)

(10)

Wherein the superscript "p" indicates a predicted value.

In view of the inherent one-step delay compensation of the FCS-MPC, the predicted value at time (k+1) is replaced with the predicted value at time (k+2). The NP voltage balance is then incorporated into the cost function as a weight term. Cost functionCan be designed as follows:

(11)

wherein,representing tracking costs->Represents NP cost,/->Representing the weight coefficient, ++>Representing the output voltage reference, may be calculated using lagrangian interpolation. For sinusoidal references, the order "n" should be greater than 2. Thus, the output voltage reference value may be expressed as:

(12)

the over-current protection item can be used for considering the over-current protection in the actual operation processInto the cost function. Finally, let(s)>The cost function can be expressed as:

(13)

wherein,indicating the upper limit of the inductor current.

Each voltage vector corresponds to a particular cost value. Thus, by rolling optimization, a voltage vector is determined that minimizes the cost functionAnd acts on the next switching cycle, which can be expressed as:

(14)

C. error convergence condition

The objective of applying the sliding mode theory is to narrow the selection range of candidate voltage vectors and ensure convergence of the output voltage errors. For this purpose, two switching functions are first defined And->Expressed as the difference between the output reference voltage and the actual sampled output voltage:

(15)

in order for the output voltage to track the reference voltage and ensure error convergence, the above-mentioned switching function must meet the lyapunov stability criterion:

(16)

in the discrete model, since the reference voltage is constant over one control period, the derivative of the switching function can be expressed according to the equation (9) and the equation (10):

(17)

(18)

wherein,，/>，/>，/>is the normal number of the two groups,representing inductor current +.>And output current->According to equation (10), this difference directly influences the output voltage +.>Is a variation of (c). Thus (S)>The relationship to the control domain is further analyzed.

D. Control domain

From formulas (15), (16), (17), (18), the control domain of the optimal voltage vector can be determined. For better explanation, without loss of generalityAnd->An example is described.

Then, the candidate voltage vectorαAndβthe components need to meet the following conditions:

(19)

wherein,，/>。

therefore, in the space vector diagram, the coordinate system is reconstructed with the output voltage sampling value at the (k) th time as the origin, as indicated by four arrows centering on the dot on the lower right of the center point in fig. 3.

According toThere are a total of 4 possible control domains. For example, when- >，/>The control field is shown as the lower right dark gray region in fig. 3 (a), while the light gray region does not satisfy the stability condition (16), and the remaining cases are shown as (b), (c), and (d), respectively, in fig. 3.

In view of the parameter mismatch and the external interference,and->Not strictly defined, may be further expressed as:

(20)

wherein,is bounded, < >>Is a positive constant and represents the upper bound of error and disturbance.

E. Parallel optimization strategy

To determine the optimal voltage vector in the control domain, the conventional approach is to substitute each voltage vector in the control domain into a cost function (13), and find the voltage vector that minimizes the cost function by rolling optimization. However, this method is computationally intensive, especially in the case of a wide control domain. Aiming at the problem, the application provides a parallel optimization strategy.

1)αAndβparallel optimization of components

As shown in (11), the cost function consists of two parts, namely "tracking cost" and "NP cost". The former is composed of two parts of real axis error and imaginary axis error, and the weight coefficients of the two parts are 1. Thus, the minimum cost functions for both the real and imaginary axes can be determined, and their intersection points form the voltage vector that minimizes the "tracking cost". The new parallel cost function and optimization procedure can be expressed as:

(21)

(22)

Wherein,representation ofαOptimal value of axis +.>Representation ofβOptimal value of axis +.>Representing the control domain, it should be noted that "NP cost" is not considered here, since in subsequent analysis the weight term will be eliminated by the redundant small voltage vector.

2) Monotonicity analysis

In fact, not all in the control domainαAndβthe components all need to participate in the optimization because the cost function sums all exhibit monotonicity.

For example, as shown in (a) of FIG. 4, whenWhen the voltage vector in the control domainαThe components always causeIncrease and voltage vectorαThe larger the component is>The larger. Therefore, when optimizing from the leftmost end of the control domain to the right, the cost function +.>Always decreasing and then increasing.

Similarly, as shown in FIG. 4 (b), whenWhen optimizing down from the top of the control domain, the cost function +.>Also has the same features.

3) Constraint of

Since the space vector diagram is not rectangular, the intersection point to which the voltage vector corresponds does not always exist. As shown in figure 5 (a),outside the space vector diagram, therefore, further consideration of constraints is required.

According to the monotonicity of the analysis described above, the optimal voltage vector is located at the edge of the space vector diagram. Therefore, as shown in (b) of fig. 5, it is necessary to further evaluate three voltage vectors ，/>And->Is a cost function of (1). It should be noted that at this time, it is necessary toαAndβoverall cost function of the component->The evaluation can be expressed as:

(23)

in summary, the implementation process of the parallel optimization strategy based on the sliding mode theory provided in the present application is shown in fig. 6, and includes:

according to the output voltageαAndβmonotonicity of the component cost function to obtain an optimal voltage vectorαAndβa component;

synthesizing an optimal voltage vectorαAndβthe components are used for obtaining an optimal voltage vector;

and carrying out constraint analysis on the optimal voltage vector, and judging whether constraint is needed or not to obtain a constraint result. If not, directly outputIf so, minimize +.>Output->。

In view of the delay compensation of the MPC, the output voltage and the inductor current at time (k) in the above analysis can be replaced by the predicted value at time (k+1), and the output current can be approximated by the value at time (k) since it is a slowly varying process with respect to the inductor current.

F. Virtual voltage vector and NP voltage balancing

From the above analysis, unlike the conventional method, the optimization strategy proposed in the present application does not depend on the number of voltage vectors, but on the number of voltage vectorsαAndβnumber of components. As shown in fig. 2, in the real space vector diagram αAndβthere are only 5 components, which can lead to a large output ripple.

1) Virtual voltage vector

By setting the virtual voltage vector, it is possible to expandαAndβnumber of candidates for the component. The specific practice is that in the voltage vector diagram, two adjacent voltage vector diagrams are arrangedαOr (b)βComponent(s)A component is added to the midpoint of (c). At the same time, a virtual voltage vector is set at eachαAndβintersection of components, thus, candidatesαOr (b)βThe components will extend from 5 to 9. As shown in fig. 7 (a), the dots marked with V represent real voltage vectors, and the remaining dots represent generated virtual voltage vectors.

It is worth mentioning that the further increaseαAndβthe number of components may further optimize control performance, however, the computational burden may also increase.

2) NP voltage balancing

NP voltage balancing is also important for three-phase three-level inverters. Since the proposed optimization method only considers the "tracking cost", if the "NP cost" can be eliminated from the cost function, thenCan be used as an optimal voltage vector.

The present application has analytically modeled the NP voltage in the system modeling section, with its changes related to the neutral point current. Therefore, the effect on the NP voltage is limited to redundant small voltage vectors, which are classified into P-type and N-type as shown in table i, and it should be noted that the effects of the two small voltage vectors on the NP voltage at each node are opposite, but their effects on the output voltage are the same.

TABLE I Classification of redundant Small Voltage vectors

In practice, virtual voltage vectorsMay be defined as a combination of actual voltage vectors, which may be expressed as:

(24)

wherein the method comprises the steps ofRepresenting virtual voltage vectors, ">Representing the control period->Representing an actual voltage vector; />Representing the time of action of the actual voltage vector;Nrepresenting the number of actual voltage vectors.

When (when)Comprises at least one small voltage vector which may have two output modes, e.g. as shown in fig. 7 (b)>Can be expressed as:

(25)

when (when)By->，/>And->The synthesis belongs to the P-type. Conversely, when by->，/>And->When synthesized, it belongs to the N-type. The principle can be generalized to all candidate voltage vectors containing small voltage vectors.

In summary, the introduction of virtual voltage vectors achieves the dual objective: reducing output ripple and eliminating weighting factors in the cost function.

G. System control block diagram

Based on the above analysis, the ESM-PC complete control block diagram proposed in the present application is shown in FIG. 8, comprising:

Specific examples:

in order to verify the effectiveness of ESM-PC provided by the application, a T-type three-phase three-level inverter experimental platform based on DSP-TMS320F28374S is built. Inductor current [ ]) And output current (+)>) Sampled by VAC current sensor, load voltage (+)>) And NP voltage (+)>) Sampled by the LEM voltage sensor. DSP-TMS320F28374S negativeImplementation of the responsibility sampling and control algorithm. The experimental parameters are detailed in Table II.

TABLE II System parameters

Meanwhile, the ESM-PC provided by the application is compared with the following two algorithms, and the following description is given:

traditional MPC: the traditional MPC designs a cost function as a formula (11), realizes NP voltage balance by using a weighting factor, and obtains an optimal voltage vector by evaluating 27 voltage real quantities.

DB-MPC: the DB-MPC applies 3 voltage vectors in a single switching cycle, with the time of action of each voltage vector inversely proportional to the corresponding cost value. The NP voltage is adjusted by the redundant small voltage vector to eliminate the weighting factor and optimize the computational efficiency by the integrated DB control.

The experimental conditions were set as follows:

experimental conditions A1: setting the amplitude of output voltage as 100V, the frequency as 50Hz and the load as linear resistance load)。

Experimental conditions A2: setting the amplitude of output voltage as 100V, the frequency as 50Hz and the load as nonlinear diode rectifier bridge load)。

Experimental conditions B1: under experimental condition A1, the linear load jumps from 30Ω to 50Ω.

Experimental conditions B2: the amplitude of the output voltage is set to be 100V, the frequency is 50Hz, and the output voltage is connected with the nonlinear load in parallel from the linear load to the linear load.

Experimental condition C: under experimental condition A1, a 50V direct current bias is firstly injected into the upper capacitor and the lower capacitor on the direct current side, then the bias is removed, and the time for the balance of the three algorithms is compared.

Experimental condition D: the time complexity of the different algorithms is compared.

Experimental condition E: under the experimental condition A1, the inductance value and the capacitance value are respectively changed by +/-50%, and the robustness of the algorithm is compared.

Experimental effect

A. Steady state performance experiment

Fig. 9 shows steady-state experimental waveforms for three algorithms under linear and nonlinear loads. Waveform includes three-phase output voltage) And a phase A output current (">). Notably, these results indicate that these three methods are all effective in tracking the reference voltage, indicating that they are able to stably provide a sinusoidal voltage output to the load.

Further, fig. 10 shows an FFT analysis of the output voltage, generated from MATLAB/Simulink and experimental data obtained from an oscilloscope. Experimental results show that the Total Harmonic Distortion (THD) of the proposed algorithm is the lowest, regardless of linear or non-linear load. The proposed method achieves optimal steady state performance compared to conventional MPCs and DB-MPCs. Notably, conventional MPCs have only one switching action per control period, exhibiting a dispersed spectral distribution. In contrast, both ESM-PC and DB-MPC proposed in this application can maintain a constant switching frequency, which is advantageous in filter design and electromagnetic interference (EMI) suppression.

B. Dynamic performance experiment

To evaluate the dynamic performance of ESM-PC as proposed in the present application, FIGS. 11 and 12 show the A-phase output voltages of the three methods, respectively) Phase a output current (+)>) Inverter output line voltage (+)>) And upper capacitance voltage (">) Is a waveform of (a).

In fig. 11, the dynamic waveform shows a linear load jump from 30Ω to 50Ω. These three methods exhibit similar dynamics with the output voltage remaining almost unchanged. At the same time, at the moment of the step, the NP voltage fluctuation of the three methods is about 2.5V. This result shows that these three methods exhibit similar dynamic responses to changes in linear load.

Further, in fig. 12, the waveform describes a jump procedure from a linear load to a linear and nonlinear load in parallel. Similar to the results of linear load jumps, both methods maintain a constant voltage output. In terms of NP voltage balancing, the upper capacitance voltage ripple of the conventional MPC is about 10V, and the upper capacitance voltage ripple of the DB-MPC and the ESP-PC as proposed in the present application is about 8V. Therefore, the ESP-PC and DB-MPC provided by the application have better inhibition effect on NP voltage fluctuation during load disturbance.

C. NP balance experiment

To further evaluate the NP voltage balancing capability of the 3 methods, a 50V dc bias was first added to the sampled signal to simulate NP voltage imbalance. The offset was then removed and the time to restore equilibrium for the 3 algorithms was compared. The experimental results are shown in FIG. 13.

Experimental results indicate that the equilibrium recovery time for conventional MPC and DB-MPC is about 110 ms, whereas the equilibrium recovery time for the method of the present application is about 100 ms. This shows a slight improvement in performance compared to the first two methods. Compared with the traditional method, ESM-PC has better NP voltage balance capability.

D. Algorithm execution time experiment

The operation time of the three methods was compared on a DSP-TMS320F28374S with a crystal oscillator frequency of 200 MHz and a sampling period of 62.5 μs.

Wherein the execution time of tasks such as A/D conversion, system protection and the like is about 4.22 mu s, and the rest is the execution time of the MPC algorithm. Notably, in the method proposed in the present application, the voltage vector to be evaluated in each control periodThe number is variable. The most complex cases are considered when comparing computation times, including for allαAndβevaluation of the cost value of the component, and the case of constraints.

The experimental results are shown in Table III. Experimental results show that the calculation time of the conventional MPC algorithm is 33.74 μs, the calculation time of the DB-MPC algorithm is 27.79 μs, and the calculation time of the ESM-PC algorithm is only 14.77 μs. The calculation time was reduced by about 56% compared to the conventional MPC; the calculation time was reduced by about 46.8% compared to DB-MPC.

Experimental results show that the parallel optimization strategy provided by the application remarkably reduces the calculation load.

Table III algorithm run time comparison

E. Model parameter mismatch experiment

In practical applications, model parameter mismatch is unavoidable due to various disturbances. Under experimental condition E, the parameter sensitivities of the three MPC methods were evaluated by varying the true values of inductance and capacitance by ±50%, respectively, where the rate of change sum is defined as:

(26)

Wherein,and->The true values of inductance and capacitor, respectively, < >>And->The values of the inductance and the capacitor used in practice are the same as those of the experimental condition A1.

The experimental results are shown in Table IV, where the traditional MPC exhibited the weakest ability to counter model mismatch compared to DB-MPC and ESM-PC. However, the ESM-PC proposed in the present application has the lowest THD, showing excellent robustness even in the presence of parameter mismatch.

TABLE IV parameter mismatch analysis

An embodiment of the present application provides a dynamic optimization and virtual voltage vector sliding mode prediction control system, which is configured to execute the dynamic optimization and virtual voltage vector sliding mode prediction control method described in the foregoing embodiment, as shown in fig. 14, and the system includes:αandβa component module 501 for receiving a finite control set input voltage, including a real voltage vector and a virtual voltage vector;

a control domain module 502 for dynamically establishing a control domain based on sliding mode theory according to the output voltageαAndβmonotonicity of the component cost function to obtain an optimal voltage vectorαAndβthe components are synthesized to obtain an optimal voltage vector;

the neutral point voltage balancing module 503 realizes neutral point voltage balancing by using the redundancy small voltage vector to realize different synthesis modes of the optimal voltage vector according to the NP voltage based on the optimal voltage vector.

The dynamic optimization and virtual voltage vector sliding mode prediction control system provided by the embodiment of the application and the dynamic optimization and virtual voltage vector sliding mode prediction control method provided by the embodiment of the application are the same in the same invention conception, and have the same beneficial effects as the method adopted, operated or realized by the stored application program.

The embodiment of the application also provides an electronic device corresponding to the dynamic optimization and virtual voltage vector sliding mode prediction control method provided by the previous embodiment, so as to execute the dynamic optimization and virtual voltage vector sliding mode prediction control method. The embodiments of the present application are not limited.

Referring to fig. 15, a schematic diagram of an electronic device according to some embodiments of the present application is shown. As shown in fig. 15, the electronic device 20 includes: a processor 200, a memory 201, a bus 202 and a communication interface 203, the processor 200, the communication interface 203 and the memory 201 being connected by the bus 202; the memory 201 stores a computer program that can be executed on the processor 200, where the processor 200 executes the dynamic optimization and virtual voltage vector sliding mode prediction control method provided in any of the foregoing embodiments of the present application.

The memory 201 may include a high-speed random access memory (RAM: random Access Memory), and may further include a non-volatile memory (non-volatile memory), such as at least one disk memory. The communication connection between the system network element and at least one other network element is implemented via at least one communication interface 203 (which may be wired or wireless), the internet, a wide area network, a local area network, etc. may be used.

Bus 202 may be an ISA bus, a PCI bus, an EISA bus, or the like. The buses may be classified as address buses, data buses, control buses, etc. The memory 201 is configured to store a program, and the processor 200 executes the program after receiving an execution instruction, and the dynamic optimization and virtual voltage vector sliding mode prediction control method disclosed in any embodiment of the present application may be applied to the processor 200 or implemented by the processor 200.

The processor 200 may be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware in the processor 200 or by instructions in the form of software. The processor 200 may be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU for short), a network processor (Network Processor, NP for short), etc.; but may also be a Digital Signal Processor (DSP), application Specific Integrated Circuit (ASIC), an off-the-shelf programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components. The disclosed methods, steps, and logic blocks in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with the embodiments of the present application may be embodied directly in hardware, in a decoded processor, or in a combination of hardware and software modules in a decoded processor. The software modules may be located in a random access memory, flash memory, read only memory, programmable read only memory, or electrically erasable programmable memory, registers, etc. as well known in the art. The storage medium is located in the memory 201, and the processor 200 reads the information in the memory 201, and in combination with its hardware, performs the steps of the above method.

The electronic equipment provided by the embodiment of the application and the dynamic optimization and virtual voltage vector sliding mode prediction control method provided by the embodiment of the application have the same beneficial effects as the method adopted, operated or realized by the electronic equipment and the virtual voltage vector sliding mode prediction control method provided by the embodiment of the application are based on the same invention conception.

The present embodiment further provides a computer readable storage medium corresponding to the dynamic optimization and virtual voltage vector sliding mode prediction control method provided in the foregoing embodiment, referring to fig. 16, the computer readable storage medium is shown as an optical disc 30, on which a computer program (i.e. a program product) is stored, where the computer program, when executed by a processor, performs the dynamic optimization and virtual voltage vector sliding mode prediction control method provided in any of the foregoing embodiments.

It should be noted that examples of the computer readable storage medium may also include, but are not limited to, a phase change memory (PRAM), a Static Random Access Memory (SRAM), a Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a flash memory, or other optical or magnetic storage medium, which will not be described in detail herein.

The computer readable storage medium provided by the above embodiment of the present application and the dynamic optimization and virtual voltage vector sliding mode prediction control method provided by the embodiment of the present application are the same inventive concept, and have the same beneficial effects as the method adopted, operated or implemented by the application program stored therein.

It should be noted that:

the algorithms and displays presented herein are not inherently related to any particular computer, virtual system, or other apparatus. Various general-purpose systems may also be used with the teachings herein. The required structure for a construction of such a system is apparent from the description above. In addition, the present application is not directed to any particular programming language. It will be appreciated that a variety of programming languages may be used to implement the teachings of the present application as described herein, and the above description of specific languages is provided for disclosure of preferred embodiments of the present application.

In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the present application may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.

Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the application, various features of the application are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the application and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be construed as reflecting the intention that: i.e., the claimed application requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this application.

Those skilled in the art will appreciate that the modules in the apparatus of the embodiments may be adaptively changed and disposed in one or more apparatuses different from the embodiments. The modules or units or components of the embodiments may be combined into one module or unit or component and, furthermore, they may be divided into a plurality of sub-modules or sub-units or sub-components. Any combination of all features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or units of any method or apparatus so disclosed, may be used in combination, except insofar as at least some of such features and/or processes or units are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings), may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.

Furthermore, those skilled in the art will appreciate that while some embodiments described herein include some features but not others included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the present application and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination.

Various component embodiments of the present application may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art will appreciate that some or all of the functions of some or all of the components in a virtual machine creation system according to embodiments of the present application may be implemented in practice using a microprocessor or Digital Signal Processor (DSP). The present application may also be embodied as a device or system program (e.g., a computer program and a computer program product) for performing a portion or all of the methods described herein. Such a program embodying the present application may be stored on a computer readable medium, or may have the form of one or more signals. Such signals may be downloaded from an internet website, provided on a carrier signal, or provided in any other form.

It should be noted that the above-mentioned embodiments illustrate rather than limit the application, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The application may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The use of the words first, second, third, etc. do not denote any order. These words may be interpreted as names.

The foregoing is merely specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think of various changes or substitutions within the technical scope of the present application, and these should be covered in the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A dynamic optimization sliding mode prediction control method based on virtual voltage vectors is used for a T-shaped three-phase three-level inverter with an LC filter, and is characterized by comprising the following steps:

in the output voltage vector diagram, at every twoαThe midpoint of the component is added with one component, at every twoβThe midpoint of the component is added with a component, and then at eachαAndβthe intersection point of the components supplements the voltage vector to obtain a virtual voltage vector, and the actual voltage vector and the virtual voltage vector jointly form a candidate voltage vector;

dynamically establishing a control domain based on a sliding mode theory, and narrowing the selection range of candidate voltage vectors, wherein the method comprises the following steps: determining a control domain based on a switching function and a lyapunov stability criterion, the switching function being the difference between the output voltage reference value and the actual sampled output voltage;

obtaining the output voltage in the control domain through parallel optimizationαAndβvoltage with minimum component cost functionVector ofαAndβthe components are synthesized to obtain an optimal voltage vector, monotonicity of a cost function in a control domain is analyzed in a parallel optimization process, and an optimization process is optimized;

after synthesizing to obtain an optimal voltage vector, carrying out constraint analysis on the optimal voltage vector, judging whether constraint is needed, directly outputting the optimal voltage vector if constraint is not needed, and if so, minimizing αAndβoutputting the optimal voltage vector after the total cost function of the components is thatαCost function of componentβThe sum of the cost functions of the components;

and different synthesis modes of the optimal voltage vector are realized by using the redundant small voltage vector, so that neutral point voltage balance is realized.

2. The method of claim 1, wherein the virtual voltage vectorDefined as a combination of actual voltage vectors, expressed as:

；

3. A virtual voltage vector-based dynamic optimization sliding mode predictive control system employing the sliding mode predictive control method of claim 1, comprising:

αandβthe component module is used for receiving the actual voltage vector and the virtual voltage vector;

and the neutral point voltage balance module realizes different synthesis modes of the optimal voltage vector by using the redundant small voltage vector, and realizes neutral point voltage balance.

4. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor runs the computer program to implement the method of any one of claims 1-2.

5. A computer readable storage medium having stored thereon a computer program, wherein the program is executed by a processor to implement the method of any of claims 1-2.