A model-free control method and system for wind power system

Abstract

The present invention discloses a model-free control method and system for wind power systems. The method includes calculating the autoregressive sampling data vector at time k by combining the historical data of the current and output voltage of the grid-connected inverter, and updating the autoregressive parameter vector at time k in real time based on the least square method; taking the eight switching vectors of the grid-connected inverter as the voltage value at time k+1, updating the autoregressive sampling data vector at time k+1, and multiplying the autoregressive sampling data vector at time k+1 with the autoregressive parameter vector at time k to obtain the current prediction value at time k+1 corresponding to the eight switching vectors; selecting the switching vector that minimizes the preset cost function G as the optimal switching vector to act on the grid-connected inverter. The present invention aims to realize the model-free control of the grid-connected inverter applied to the wind power system, and improve the predictive control performance of the system under parameter mismatch conditions and the parameter robustness of the predictive control of the grid-connected inverter.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind power control, and in particular relates to a model-free control method and system applied to a wind power system. Background Art

[0002] Due to the increase in energy demand, renewable energy solutions have become popular. Among them, wind power generation is an outstanding solution for renewable energy. Grid-connected inverters are an important part of connecting renewable energy to the grid. Grid-connected inverters ensure the stability of distributed power generation systems and safely inject the generated electricity into the grid. Considering the stability of the grid, grid-connected inverters must be effectively controlled. The control method of traditional grid-connected inverters is generally a model-based predictive control method (model predictive control method). Due to its extremely fast response speed, the model predictive control method is a very promising grid-connected inverter control strategy. However, a great disadvantage of the model predictive control method is that it requires high accuracy of the dynamic model on the grid side. In practical applications, as the ambient temperature changes, the resistance, inductance and other parameters of the line will change, resulting in parameter mismatch, resulting in serious prediction errors, and affecting the safe and stable operation of the power grid. Therefore, how to achieve model-free control to improve the parameter robustness of predictive control of wind power grid-connected inverters has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] Technical problem to be solved by the present invention: In view of the above-mentioned problems in the prior art, a model-free control method and system applied to a wind power system are provided. The present invention aims to realize model-free control applied to a wind power system, improve the predictive control performance of the system under parameter mismatch conditions and the parameter robustness of the predictive control of the grid-connected inverter.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0005] A model-free control method for a wind power system, comprising:

[0006] S101, obtaining the grid-side current and output voltage of the grid-connected inverter, calculating the autoregressive sampling data vector at time k based on the historical data of the current and voltage, and updating and calculating the autoregressive parameter vector at time k in real time based on the least squares method;

[0007] S102, taking eight switching vectors of the grid-connected inverter as voltage values ​​at time k+1, updating the autoregressive sampling data vector at time k+1, and multiplying the autoregressive sampling data vector at time k+1 with the autoregressive parameter vector at time k to obtain current prediction values ​​at time k+1 corresponding to the eight switching vectors;

[0008] S103 , among the current prediction values ​​at time k+1 corresponding to the eight switching vectors, select the switching vector that minimizes the preset cost function G as the optimal switching vector to act on the grid-connected inverter.

[0009] Optionally, the calculation function expression of the autoregressive sampling data vector at time k in step S101 is:

[0010] ,

[0011] ,

[0012] In the above formula, and are the autoregressive sampling data vectors at time k. d , q Axis component, ~ They are k -1 moment~k- n A Current at the moment d Axis component, ~ They are k -1 moment~ k - n A Current at the moment q Axis component, ~ They are k -1 moment~ k - n B Voltage at the moment d Axis component, ~ They are k-1 moment to kn B Voltage at the moment q Axis component, n A and n B are the orders of current and voltage respectively.

[0013] Optionally, before step S101, the process further includes combining the historical data of the current and voltage and determining the order of the current and voltage based on the autoregressive external input model fitting. n A and n B , the functional expression of the autoregressive external input model is:

[0014] ,

[0015] In the above formula, and They are k The current prediction value at the moment d , q Axis component, and They are k Voltage at the moment d , q Axis component, , , , , as well as are all parameters in the autoregressive external input model, and:

[0016] ,

[0017] In the above formula, ~ , ~ , ~ , ~ , ~ as well as ~ All are fitting coefficients.

[0018] Optionally, in step S101, the autoregressive parameter vector at time k is updated and calculated in real time based on the least squares method. The function expression is:

[0019] ,

[0020] In the above formula, is the autoregressive parameter vector at time k-1, for k The gain matrix at time instant, for k The estimated error at the moment, subscript express d Axis or q Axis components, and:

[0021] ,

[0022] ,

[0023] ,

[0024] In the above formula, is the grid-side current at time k, is the autoregressive sampling data vector at time k, and are the adaptive gain matrices at time k-1 and time k respectively, For the forgetting factor.

[0025] Optionally, the forgetting factor value range satisfies: .

[0026] Optionally, the calculation function expression of the current prediction value at time k+1 corresponding to the eight switching vectors in step S102 is:

[0027] ,

[0028] ,

[0029] In the above formula, and They are the predicted current values ​​at time k+1. d , q Axis component, and are the autoregressive parameter vectors at time k, and are the autoregressive sampling data vectors at time k+1. d , q Axis component.

[0030] Optionally, the calculation function expression of the cost function G in step S103 is:

[0031] ,

[0032] In the above formula, and The command current is d , q Axis component, and They are the predicted current values ​​at time k+1. d , q Axis component.

[0033] In addition, the present invention also provides a model-free control system applied to a wind power system, comprising a microprocessor and a memory connected to each other, wherein the microprocessor is programmed or configured to execute the model-free control method applied to the wind power system.

[0034] In addition, the present invention also provides a computer-readable storage medium, in which a computer program / instruction is stored. The computer program / instruction is programmed or configured to execute the model-free control method applied to a wind power system through a processor.

[0035] In addition, the present invention also provides a computer program product, comprising a computer program / instruction, wherein the computer program / instruction is programmed or configured to execute the model-free control method applied to a wind power system through a processor.

[0036] Compared with the prior art, the present invention mainly has the following advantages: the present invention includes establishing an autoregressive external input model about the grid-connected inverter based on the real-time sampling data of voltage and current, and updating the model parameters in real time in combination with the least square method. This process does not involve any model parameters of the grid-connected inverter, and does not require prior knowledge of the physical models of the inverter and the grid-side load. Compared with the finite-level predictive control based on the mathematical model of the grid-connected inverter, this new strategy has strong robustness in terms of parameter and model mismatch. At the same time, the method of the present invention opens up a new field of predictive control. Only by establishing an autoregressive model through the obtained sampling data, predictive control can be robustly applied in highly complex power electronic systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 Schematic diagram of the basic flow of the method of the embodiment of the present invention.

[0038] Figure 2 Detailed flowchart of the method of the embodiment of the present invention.

[0039] Figure 3 The figure is a calculation block diagram of the autoregressive parameter vector based on the least squares algorithm in an embodiment of the present invention.

[0040] Figure 4 The figure is an overall block diagram of a grid-connected inverter control strategy applied to a wind power system in an embodiment of the present invention.

[0041] Figure 5 This is the a-phase current diagram obtained by simulation according to the method of the embodiment of the present invention.

[0042] Figure 6 This is a diagram of harmonic analysis results obtained by simulation using the method of an embodiment of the present invention.

[0043] Figure 7 This is the harmonic analysis result diagram of the first set of data simulated by the traditional finite-level predictive current control method.

[0044] Figure 8 This is the harmonic analysis result diagram of the second set of data simulated by the traditional finite-level predictive current control method. DETAILED DESCRIPTION

[0045] The present invention aims at the problem that the parameter mismatch on the grid side affects the control performance under actual conditions, and provides a model-free control method and system for wind power systems. The method first constructs an autoregressive external input model based on the grid-connected inverter, samples the voltage and current, generates an autoregressive sampling data vector, and uses the least squares method to update the autoregressive parameter vector in real time, establishes a real-time updated autoregressive external input model for predictive control of the grid-connected inverter, and finally selects the optimal switch vector for each sampling period based on the constructed cost function. The entire design process avoids directly using the parameters given by the model, and greatly improves the predictive control performance of the system under parameter mismatch conditions.

[0046] like Figure 1 As shown, the model-free control method applied to the wind power system in this embodiment includes:

[0047] S101, obtaining the grid-side current and output voltage of the grid-connected inverter, calculating the autoregressive sampling data vector at time k based on the historical data of the current and voltage, and updating and calculating the autoregressive parameter vector at time k in real time based on the least squares method;

[0048] S102, taking eight switching vectors of the grid-connected inverter as voltage values ​​at time k+1, updating the autoregressive sampling data vector at time k+1, and multiplying the autoregressive sampling data vector at time k+1 with the autoregressive parameter vector at time k to obtain current prediction values ​​at time k+1 corresponding to the eight switching vectors;

[0049] S103 , among the current prediction values ​​at time k+1 corresponding to the eight switching vectors, select the switching vector that minimizes the preset cost function G as the optimal switching vector to act on the grid-connected inverter.

[0050] The calculation function expression of the autoregressive sampling data vector at time k in step S101 of this embodiment is:

[0051] ,

[0052] ,

[0053] In the above formula, and are the autoregressive sampling data vectors at time k. d , q Axis component, ~ They are k -1 moment~k- n A Current at the moment d Axis component, ~ They are k-1 moment~ k - n A Current at the moment q Axis component, ~ They are k -1 moment~ k - n B Voltage at the moment d Axis component, ~ They are k-1 moment to kn B Voltage at the moment q Axis component, n A and n B are the orders of current and voltage respectively.

[0054] See also Figure 2 , before step S101 of this embodiment, the steps further include combining the historical data of current and voltage and determining the order of current and voltage based on the autoregressive external input model fitting. n A and n B , the functional expression of the autoregressive external input model is:

[0055] ,

[0056] In the above formula, and They are k The current prediction value at the moment d , q Axis component, and They are k Voltage at the moment d , q Axis component, , , , , as well as are all parameters in the autoregressive external input model, and:

[0057] ,

[0058] In the above formula, ~ , ~ , ~ , ~ , ~ as well as ~ are all fitting coefficients. Determine the order of the autoregressive model and When the value of and The value of is a trade-off between the computational burden and the accuracy of the model. By increasing the order, an accurate model can be obtained, but the increase in the order increases the amount of calculation. In order to obtain good control performance and lower computational burden, and The selection of should be determined through multiple experiments. The determination of the autoregressive parameters represents the determination of the system model. Therefore, if the autoregressive parameters between the input and output can be accurately identified, the system prediction model can be generated. The expression of the autoregressive external input model can be rewritten as follows:

[0059] ,

[0060] .

[0061] On this basis, it can be further rewritten as:

[0062] ,

[0063] ,

[0064] Thus, the calculation function expression of the autoregressive sampling data vector at time k in step S101 can be obtained.

[0065] The least squares method is used to update the autoregressive parameter vector in real time. First, the grid-connected inverter is sampled in real time to obtain dq Rotating coordinate system d axis, q Grid-side current of the shaft i d ( k ) 、i q ( k ) and inverter output voltage v d ( k ) 、v q ( k ), k Corresponding to the sampling time or sampling point, the historical data of current and voltage are established to generate the autoregressive sampling data vector, and then the autoregressive parameter vector is updated and calculated in real time based on the least squares method. Figure 3 As shown, in step S101 of this embodiment, the autoregressive parameter vector at time k is updated and calculated in real time based on the least squares method. The function expression is:

[0066] ,

[0067] In the above formula, is the autoregressive parameter vector at time k-1, for k The gain matrix at time instant, for k The estimated error at the moment, subscript express d Axis or q Axis components, and:

[0068] ,

[0069] ,

[0070] ,

[0071] In the above formula, is the grid-side current at time k, is the autoregressive sampling data vector at time k, and are the adaptive gain matrices at time k-1 and time k respectively, For the forgetting factor.

[0072] Among them, the range of the forgetting factor value satisfies: . The smaller the value, the faster the tracking speed of fast-changing parameters, but the system is more sensitive to noise. Since the parameter changes are much lower than the sampling frequency in the predictive controller, and in order to increase noise immunity, it is generally selected .

[0073] The calculation function expression of the current prediction value at time k+1 corresponding to the eight switching vectors in step S102 of this embodiment is:

[0074] ,

[0075] ,

[0076] In the above formula, and They are the predicted current values ​​at time k+1. d , q Axis component, and are the autoregressive parameter vectors at time k, and are the autoregressive sampling data vectors at time k+1.d , q Axis component.

[0077] The calculation function expression of the cost function G in step S103 of this embodiment is:

[0078] ,

[0079] In the above formula, and The command current is d , q Axis component, and They are the predicted current values ​​at time k+1. d , q The value of the cost function G corresponding to the current prediction value output by the data-driven model is traversed, and the switching vector that minimizes the cost function G is selected as the optimal switching vector and applied to the inverter.

[0080] like Figure 4 As shown in the figure, the "undetermined coefficient estimation and establishment of time series model" is a simplified description of the steps before step S103 of this embodiment. Then in step S103, when the current prediction value at time k+1 corresponding to the eight switching vectors is selected, the switching vector that minimizes the preset cost function G (minimizing the cost function) is selected as the optimal switching vector (S a ~S c ) acts on the grid-connected inverter. Figure 4 middle, i a ~ i c They represent the collected three-phase currents abc respectively, and the dq axis currents are obtained by coordinate transformation (abc / dq) i dq ( k ) as the input for “Estimation of Undetermined Coefficients and Establishment of Time Series Model”. In addition, Figure 4 Middle U dc is the DC bus voltage, L and R are the inductance and resistance of the grid-side line respectively.

[0081] In order to further verify the technical effect of the method of this embodiment, a simulation comparison analysis was conducted between the method of this embodiment and the traditional finite-level predictive current control method. The switching frequency was 1000KHZ, and the corresponding sampling period was 1e-6 seconds. Figure 5 This is the a-phase current diagram obtained by simulation using the method of this embodiment. Figure 6 The harmonic analysis result diagram obtained by simulating the method of this embodiment shows that the total harmonic distortion rate of the current in this embodiment is 1.78%. The traditional finite-level predictive current control method uses inductance parameters, so inductance mismatch will have an impact on it. Figure 7 This is the harmonic analysis result diagram of the first set of data (inductance is 0.5 times the actual inductance) obtained by the traditional finite-level predictive current control method. It can be seen that the total harmonic distortion (TDH) of the current is 2.12% at this time. Figure 8 The harmonic analysis result diagram of the second set of data (inductance is twice the actual inductance) obtained by the traditional finite-step predictive current control method shows that the total harmonic distortion rate of the current is 2.23%. Finally, the simulation results of the traditional finite-step predictive current control method and the method of this embodiment under two degrees of inductance parameter mismatch are shown in Table 1.

[0082] Table 1: Simulation results comparison table

[0083]

[0084] As shown in Table 1, since the method of this embodiment is not affected by the inductance parameters, the total harmonic distortion rate of the current is 1.78% in both cases of inductance parameters. Compared with the total harmonic distortion rate of the current of the traditional finite-level predictive current control method of 2.12% or 2.23%, the total harmonic distortion rate of the current of the method of this embodiment is significantly lower. It can be seen that compared with the traditional finite-level predictive current control method, the method used in this embodiment can significantly reduce the pulsation of the d and q axis currents, reduce the current harmonics, and improve the robustness of the predictive control to parameter mismatch. By comparing the simulation results under the above two inductance parameter mismatches, it is proved that the method of this embodiment can reduce the current harmonic value in the case of parameter mismatch and improve the robustness of the system.

[0085] In summary, the model-free control method applied to the wind power system in this embodiment includes constructing an autoregressive external input model of the grid-connected inverter, establishing an autoregressive sampling data vector about the current and voltage by collecting the current and voltage on the grid side, and calculating and updating the autoregressive parameter vector in real time based on the least squares algorithm; then, the eight switching vectors output by the inverter are used as the voltage value at the k+1 moment to update the autoregressive sampling data vector, and multiply it with the updated autoregressive parameter vector to obtain the current values ​​at the eight k+1 moments; traverse the cost function values ​​corresponding to the output current prediction values, and select the switching vector that minimizes the cost function as the optimal vector, which is applied to the inverter. The model-free control method applied to the wind power system in this embodiment can avoid the use of grid-connected inverter parameters to realize model-free control applied to the wind power system, improve the predictive control performance of the system under parameter mismatch conditions and the parameter robustness of the predictive control of the grid-connected inverter.

[0086] In addition, this embodiment also provides a model-free control system applied to a wind power system, comprising a microprocessor and a memory connected to each other, wherein the microprocessor is programmed or configured to execute the model-free control method applied to the wind power system.

[0087] In addition, this embodiment further provides a computer-readable storage medium, in which a computer program / instruction is stored. The computer program / instruction is programmed or configured to execute the model-free control method applied to a wind power system through a processor.

[0088] In addition, this embodiment also provides a computer program product, including a computer program / instruction, wherein the computer program / instruction is programmed or configured to execute the model-free control method applied to a wind power system through a processor.

[0089] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application can take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the process Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the functions specified in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0090] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A model-free control method for wind power systems, It is characterized in that include: S101, obtaining the grid-side current and output voltage of the grid-connected inverter, calculating the autoregressive sampling data vector at time k based on the historical data of the current and voltage, and updating and calculating the autoregressive parameter vector at time k in real time based on the least squares method; S102, taking eight switching vectors of the grid-connected inverter as voltage values ​​at time k+1, updating the autoregressive sampling data vector at time k+1, and multiplying the autoregressive sampling data vector at time k+1 with the autoregressive parameter vector at time k to obtain current prediction values ​​at time k+1 corresponding to the eight switching vectors; S103, selecting a switching vector that minimizes a preset cost function G from among the current prediction values ​​at time k+1 corresponding to the eight switching vectors as the optimal switching vector to act on the grid-connected inverter; Before step S101, it also includes combining the historical data of current and voltage, and determining the order of current and voltage based on the autoregressive external input model fitting. n A and n B , the functional expression of the autoregressive external input model is: ， In the above formula, and They are k The current prediction value at the moment d , q Axis component, and They are k Voltage at the moment d , q Axis component, , , , , as well as are all parameters in the autoregressive external input model, and: ， In the above formula, ~ , ~ , ~ , ~ , ~ as well as ~ All are fitting coefficients.

2. The model-free control method for a wind power system according to claim 1, It is characterized in that The calculation function expression of the autoregressive sampling data vector at time k in step S101 is: ， ， In the above formula, and are the autoregressive sampling data vectors at time k. d , q Axis component, ~ They are k -1 moment~k- n A Current at the moment d Axis component, ~ They are k -1 moment~ k - n A Current at the moment q Axis component, ~ They are k -1 moment~ k - n B Voltage at the moment d Axis component, ~ They are k-1 moment to kn B Voltage at the moment q Axis component, n A and n B are the orders of current and voltage respectively.

3. The model-free control method for a wind power system according to claim 1, It is characterized in that In step S101, the autoregressive parameter vector at time k is updated and calculated in real time based on the least squares method. The function expression is: ， In the above formula, is the autoregressive parameter vector at time k-1, for k The gain matrix at time instant, for k The estimated error at the moment, subscript express d Axis or q Axis components, and: ， ， ， In the above formula, is the grid-side current at time k, is the autoregressive sampling data vector at time k, and are the adaptive gain matrices at time k-1 and time k respectively, For the forgetting factor.

4. The model-free control method for a wind power system according to claim 3, It is characterized in that The forgetting factor value range satisfies: .

5. The model-free control method for a wind power system according to claim 1, It is characterized in that The calculation function expression of the current prediction value at time k+1 corresponding to the eight switching vectors in step S102 is: ， ， In the above formula, and They are the predicted current values ​​at time k+1. d , q Axis component, and are the autoregressive parameter vectors at time k, and are the autoregressive sampling data vectors at time k+1. d , q Axis component.

6. The model-free control method for a wind power system according to claim 1, It is characterized in that The calculation function expression of the cost function G in step S103 is: ， In the above formula, and The command current is d , q Axis component, and They are the predicted current values ​​at time k+1. d , q Axis component.

7. A model-free control system for a wind power system, comprising a microprocessor and a memory connected to each other, It is characterized in that The microprocessor is programmed or configured to execute the model-free control method for a wind power system as claimed in any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program / instructions stored therein, It is characterized in that The computer program / instruction is programmed or configured to execute the model-free control method for a wind power system according to any one of claims 1 to 6 through a processor.

9. A computer program product comprising a computer program / instructions, It is characterized in that The computer program / instruction is programmed or configured to execute the model-free control method for a wind power system according to any one of claims 1 to 6 through a processor.

Images