Intelligent control method, device and equipment for energy storage converter and medium

Abstract

The invention relates to an intelligent control method, device and equipment for an energy storage converter and a medium. According to the method, load prediction values corresponding to N second sampling points are obtained through prediction according to M load sampling values, then N prediction load difference values are obtained, power grid fluctuation probabilities corresponding to the N second sampling points are determined by combining a power grid fluctuation prediction model, and the prediction load difference values are adjusted according to the prediction load difference values, the adjustment load difference values and the power grid fluctuation probabilities. And the energy storage converter is controlled according to the N target load values, so that the energy storage converter can be controlled to discharge in advance when the target power grid is faced with the impact load, the stable operation of the target power grid is ensured, and the stability of the power grid is improved.

Description

Technical Field

[0001] The present invention is applicable to the field of converter control, and in particular relates to an intelligent control method, device, equipment and medium for an energy storage converter. Background Art

[0002] At present, in the power system, the energy storage system can enhance the regulation ability of the power supply, and the two-way conversion between the energy storage system and the power system is usually achieved through the energy storage inverter. The energy storage inverter can control the charging and discharging process of the energy storage system and perform AC / DC conversion. It can invert the DC power of the energy storage system into AC power and transmit it to the power grid or use it for AC loads, and it can also rectify the AC power of the power grid into DC power to charge the energy storage system.

[0003] In the grid-connected mode, usually during the load valley period, the energy storage inverter rectifies the AC power of the grid into DC power to charge the battery pack. During the load peak period, the energy storage inverter inverts the DC power in the battery pack into AC power and sends it back to the grid.

[0004] However, in actual applications, the grid load may have large instantaneous changes, such as a large number of electrical equipment connected to the grid, which causes the voltage and frequency of the grid to fluctuate violently. However, the energy storage converter is often unable to immediately provide enough AC power to maintain the stability of the grid, resulting in poor grid stability when facing impact loads. Therefore, how to improve the stability of the grid has become an urgent problem to be solved. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides an energy storage converter intelligent control method, device, equipment and medium to solve the problem.

[0006] In a first aspect, a method for intelligent control of an energy storage converter is provided, the method comprising: Input the acquired load sampling values ​​corresponding to the M first sampling points of the target power grid into the trained load forecasting model to obtain the load forecasting values ​​corresponding to the N second sampling points, where M and N are both positive integers; According to the N load prediction values ​​and the Mth load sampling value, the predicted load difference values ​​corresponding to the N second sampling points are obtained; Inputting N predicted load differences into the trained power grid fluctuation prediction model, and obtaining the power grid fluctuation probabilities corresponding to the N second sampling points respectively; Inputting N predicted load difference values ​​into a preset difference adjustment model to obtain adjusted load difference values ​​corresponding to N second sampling points respectively; Calculating a first training loss according to the N predicted load differences, the N adjusted load differences, the N power grid fluctuation probabilities and a preset training loss function; According to the first training loss, updating the parameters of the difference adjustment model until the first training loss converges, thereby obtaining an updated difference adjustment model; Inputting the N predicted load differences into the updated difference adjustment model to obtain target load differences corresponding to the N second sampling points respectively; Obtaining N load target values ​​according to the N target load differences and the Mth load sampling value; According to the N load target values, the energy storage converter corresponding to the target power grid is controlled.

[0007] In a second aspect, an intelligent control device for an energy storage converter is provided, the device comprising: A load prediction module is used to input the acquired load sampling values ​​corresponding to the M first sampling points of the target power grid into the trained load prediction model to obtain the load prediction values ​​corresponding to the N second sampling points, where M and N are both positive integers; A difference calculation module, used for obtaining predicted load differences corresponding to N second sampling points respectively according to N load prediction values ​​and the Mth load sampling value; A fluctuation prediction module is used to input N predicted load differences into a trained power grid fluctuation prediction model to obtain power grid fluctuation probabilities corresponding to N second sampling points respectively; A difference adjustment module, used for inputting N predicted load differences into a preset difference adjustment model to obtain adjusted load differences corresponding to N second sampling points respectively; A first loss calculation module, used to calculate a first training loss according to the N predicted load differences, the N adjusted load differences, the N power grid fluctuation probabilities and a preset training loss function; A first model updating module, used for updating the parameters of the difference adjustment model according to the first training loss until the first training loss converges to obtain an updated difference adjustment model; A difference determination module, used for inputting the N predicted load differences into the updated difference adjustment model to obtain target load differences corresponding to the N second sampling points respectively; A load determination module, configured to obtain N load target values ​​according to the N target load differences and the Mth load sampling value; The converter control module is used to control the energy storage converter corresponding to the target power grid according to the N load target values.

[0008] In a third aspect, an embodiment of the present invention provides a computer device, which includes a processor, a memory, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, it implements the intelligent control method of the energy storage converter as in the first aspect.

[0009] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the intelligent control method of the energy storage converter according to the first aspect is implemented.

[0010] Compared with the prior art, the present invention has the following beneficial effects: According to the M load sampling values, the load prediction values ​​corresponding to the N second sampling points are predicted, and then N predicted load differences are obtained. The power grid fluctuation probability corresponding to the N second sampling points is determined in combination with the power grid fluctuation prediction model. According to the predicted load difference, the adjusted load difference and the power grid fluctuation probability, the predicted load difference is adjusted to obtain the target load difference, and then the adjusted N load target values ​​are obtained. The energy storage converter is controlled according to the N target load values, so that when the target power grid faces an impact load, the energy storage converter can be controlled to discharge in advance to ensure that the target power grid can operate stably and improve the stability of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the prior art description are briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative labor. Figure 1 This is a schematic diagram of an application environment of an energy storage converter intelligent control method provided by Embodiment 1 of the present invention; Figure 2 It is a flow chart of an intelligent control method for an energy storage converter provided in Embodiment 1 of the present invention; Figure 3 It is a structural schematic diagram of an intelligent control device for an energy storage converter provided in Embodiment 2 of the present invention; Figure 4 It is a structural diagram of a computer device corresponding to an energy storage converter intelligent control method provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0012] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0013] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0014] It should also be understood that the term "and / or" used in the present description and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0015] As used in the present specification and the appended claims, the term "if" may be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" may be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]", depending on the context.

[0016] In addition, in the description of the present specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0017] References to "one embodiment" or "some embodiments" etc. described in the present specification mean that one or more embodiments of the present invention include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0018] The embodiments of the present invention can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0019] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0020] It should be understood that the order of execution of the steps in the following embodiments does not imply a precedence of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0021] In order to illustrate the technical solution of the present invention, specific embodiments are provided below for illustration.

[0022] The first embodiment of the present invention provides an intelligent control method for an energy storage converter, which can be applied in the following aspects: Figure 1 In an application environment, the client communicates with the energy storage converter. The client includes but is not limited to PDAs, desktop computers, laptops, ultra-mobile personal computers (UMPCs), netbooks, cloud terminal devices, personal digital assistants (PDAs) and other terminal devices. The energy storage converter can be used to rectify the AC power of the target power grid into DC power to charge the battery pack corresponding to the energy storage converter, and can also be used to invert the DC power in the battery pack corresponding to the energy storage converter into AC power and send it back to the target power grid.

[0023] See also Figure 2 , is a flow chart of an intelligent control method for an energy storage converter provided in Embodiment 1 of the present invention. The intelligent control method can be applied to Figure 1 The client obtains the load sampling values ​​corresponding to the M first sampling points from the target power grid, and controls the energy storage converter according to the N load target values. The client has a storage function, and stores a trained load prediction model, a trained power grid fluctuation prediction model and a difference adjustment model. The trained load prediction model can be used to predict the load of the second sampling point according to the load sampling value, and the trained power grid fluctuation prediction model can be used to predict the fluctuation of the power grid according to the predicted load difference. The difference adjustment model is an online learning model, which can be used to adjust the predicted load difference. Figure 2 As shown, the intelligent control method may include the following steps: Step S201: input the acquired load sampling values ​​corresponding to the M first sampling points of the target power grid into the trained load prediction model to obtain the load prediction values ​​corresponding to the N second sampling points.

[0024] Among them, M and N are both positive integers, the target power grid corresponds to the energy storage converter, and the load of the target power grid can come from various types of electricity consumers such as industry, commerce, and residents.

[0025] The first sampling point may refer to a moment when the load of the target power grid is actually sampled, and the load sampling value may refer to a result of actually sampling the load of the target power grid at the corresponding first sampling point.

[0026] The trained load prediction model can be used to predict the load value corresponding to the next sampling point based on the load values ​​corresponding to several input sampling points. The sampling points may include a first sampling point and a second sampling point. Accordingly, the load value corresponding to the sampling point may include the load sampling value corresponding to the first sampling point and the load prediction value corresponding to the second sampling point.

[0027] Specifically, the time interval between the sampling points may be a fixed value, for example, the fixed value may be 30 minutes, and the implementer may adjust the fixed value according to actual conditions.

[0028] The input of the trained load prediction model is a first fixed number of load values, and the output is a second fixed number of load values. In this embodiment, the first fixed number may be M, and the second fixed number may be 1.

[0029] Optionally, the acquired load sampling values ​​corresponding to the M first sampling points of the target power grid are input into a trained load prediction model to obtain load prediction values ​​corresponding to the N second sampling points, including: Initialize n=1, input M load sampling values ​​into the trained load forecasting model, and obtain the load forecast value corresponding to the nth second sampling point; Input the load prediction value from the n+1th load sampling value to the Mth load sampling value, and the load prediction value corresponding to the first second sampling point to the load prediction value corresponding to the nth second sampling point into the trained load prediction model to obtain the load prediction value corresponding to the n+1th second sampling point; Update n=n+1, return to execute the step of converting the n+1th load sampling value to the Mth load sampling value, and the load prediction value corresponding to the first second sampling point to the load prediction value corresponding to the nth second sampling point, inputting them into the trained load prediction model, and obtaining the load prediction value corresponding to the n+1th second sampling point, until n+1=N, and obtaining the load prediction values ​​corresponding to the N second sampling points respectively.

[0030] Among them, this embodiment takes the first fixed number as M as an example. For the first prediction, M load sampling values ​​are input into the trained load prediction model to obtain the load prediction value corresponding to the first second sampling point. For the second prediction, n=1, and the load sampling values ​​from the second to the Mth load sampling values, and the load prediction value corresponding to the first second sampling point, are input into the trained load prediction model to obtain the load prediction value corresponding to the second second sampling point. For the third prediction, n=2, and the load sampling values ​​from the third to the Mth load sampling values, and the load prediction values ​​corresponding to the first second sampling point to the second second sampling point, are input into the trained load prediction model to obtain the load prediction value corresponding to the third second sampling point, and so on, until the load prediction values ​​corresponding to the N second sampling points are obtained.

[0031] Specifically, in this embodiment, it is assumed that the time interval between the first sampling points is longer than the prediction duration of the load prediction model, so there is no need to consider collecting new load sampling values ​​when the load prediction model is used.

[0032] In this embodiment, long-term load prediction can be achieved through iterative prediction, thereby eliminating the need to frequently collect load information, improving load prediction efficiency, and further improving the efficiency of overall converter control.

[0033] Step S202: Obtain predicted load difference values ​​corresponding to N second sampling points respectively according to the N load prediction values ​​and the Mth load sampling value.

[0034] The predicted load difference may be used to characterize the degree of change of the load prediction value corresponding to the second sampling point.

[0035] Optionally, according to the N load prediction values ​​and the Mth load sampling value, the predicted load difference values ​​corresponding to the N second sampling points are obtained, including: For the load prediction value corresponding to any second sampling point, if the second sampling point is the first second sampling point, the load prediction value corresponding to the second sampling point is subtracted from the Mth load sampling value to obtain the predicted load difference corresponding to the second sampling point; If the second sampling point is not the first second sampling point, subtract the load prediction value corresponding to the second sampling point from the load prediction value corresponding to the previous second sampling point to obtain the predicted load difference corresponding to the second sampling point; The load prediction values ​​corresponding to all the second sampling points are traversed to obtain the predicted load difference values ​​corresponding to the N second sampling points respectively.

[0036] The predicted load difference corresponding to any second sampling point is obtained by subtracting the load prediction value corresponding to the previous sampling point.

[0037] Step S203: inputting the N predicted load differences into the trained power grid fluctuation prediction model to obtain power grid fluctuation probabilities corresponding to the N second sampling points.

[0038] Among them, the trained power grid fluctuation prediction model can be used to predict the fluctuation of the target power grid according to the degree of load change.

[0039] Optionally, the training process of the power grid fluctuation prediction model includes: Obtaining a historical load difference and a reference fluctuation type corresponding to the target power grid at the third sampling point, where the reference fluctuation type includes a first fluctuation type and a second fluctuation type; Inputting the historical load difference into the power grid fluctuation prediction model, obtaining reference prediction probabilities corresponding to the first fluctuation type and the second fluctuation type respectively, and taking the fluctuation type corresponding to the larger reference prediction probability as the predicted fluctuation type; According to the predicted fluctuation type and the reference fluctuation type, a second training loss is calculated, and the power grid fluctuation prediction model is trained according to the second training loss until the second training loss converges to obtain a trained power grid fluctuation prediction model.

[0040] Among them, the third sampling point may refer to the time point of the historical operation of the target power grid, the historical load difference may refer to the load change value of the target power grid at the third sampling point, the reference fluctuation type may refer to the fluctuation of the target power grid at the third sampling point, the first fluctuation type may refer to a large fluctuation in the target power grid, the second fluctuation may refer to a relatively stable target power grid, and the reference fluctuation type may be manually marked.

[0041] The reference forecast probability can be used to characterize the possibility that the historical load difference belongs to the corresponding fluctuation type.

[0042] Specifically, according to the predicted fluctuation type and the reference fluctuation type, the second training loss is calculated by the binary cross entropy loss function. According to the second training loss, the model parameters of the power grid fluctuation prediction model are trained using the reverse gradient descent method until the second training loss converges to obtain a trained power grid fluctuation prediction model.

[0043] Optionally, the N predicted load differences are input into a trained power grid fluctuation prediction model to obtain power grid fluctuation probabilities corresponding to the N second sampling points, including: For any predicted load difference, the predicted load difference is input into the trained power grid fluctuation prediction model to obtain a target prediction probability corresponding to the first fluctuation type; The target predicted probability corresponding to the first fluctuation type is used as the power grid fluctuation probability corresponding to the predicted load difference, and all predicted load differences are traversed to obtain N power grid fluctuation probabilities.

[0044] The first fluctuation type may refer to a situation where a large fluctuation occurs in the target power grid, and the target prediction probability may be used to characterize the possibility that a predicted load difference may cause a large fluctuation in the target power grid.

[0045] Step S204: input the N predicted load differences into a preset difference adjustment model to obtain the adjusted load differences corresponding to the N second sampling points.

[0046] Among them, the difference adjustment model can be used to adjust and update the predicted load difference. It should be noted that the difference adjustment model is an online learning model, that is, after each batch of N predicted load differences is input, the difference adjustment model needs to be trained.

[0047] Step S205, calculating a first training loss according to the N predicted load differences, the N adjusted load differences, the N power grid fluctuation probabilities and a preset training loss function.

[0048] The first training loss can be used to update the parameters of the difference adjustment model.

[0049] Optionally, the first training loss is calculated according to the N predicted load differences, the N adjusted load differences, the N power grid fluctuation probabilities and a preset training loss function, including: For any second sampling point, subtract the predicted load difference and the adjusted load difference corresponding to the second sampling point and take the absolute value to obtain the subtraction result, traverse all second sampling points to obtain N subtraction results, add the N subtraction results to obtain the first training sub-loss; Add the N power grid fluctuation probabilities together, and use the sum as the second training sub-loss; Add the first training sub-loss and the second training sub-loss to get the first training loss.

[0050] Among them, when calculating the subtraction result, since it is necessary to take the absolute value operation, in order to ensure the differentiability of the first training loss, the subtraction result is obtained by subtracting the predicted load difference from the adjusted load difference and then taking the square root.

[0051] The first training sub-loss can be used to supervise that for each second sampling point, the difference between the adjusted load difference corresponding to the second sampling point and the predicted load difference is as small as possible, thereby ensuring that when the predicted load value is adjusted to control the energy storage converter and thus stabilize the target power grid, a larger power loss occurs.

[0052] The second training sub-loss can be used to supervise the stability of the target power grid.

[0053] Specifically, when the first training sub-loss and the second training sub-loss are added, a weighted addition method can be adopted. In this embodiment, the weight corresponding to the second training sub-loss is larger than the weight corresponding to the first training sub-loss, so that the training process gives priority to ensuring the stability of the target power grid.

[0054] Optionally, the N subtraction results are added together to obtain a first training sub-loss, including: For any second sampling point, mapping the predicted load difference corresponding to the second sampling point to a reference weight according to a preset mapping function; Multiplying the reference weight corresponding to the second sampling point and the subtraction result to obtain a reference subtraction result corresponding to the second sampling point; Traversing all second sampling points to obtain reference subtraction results corresponding to N second sampling points respectively; Add the N reference subtraction results to get the first training sub-loss.

[0055] Among them, the preset mapping function can adopt a normalization function, that is, for any second sampling point, the predicted load difference corresponding to the second sampling point is compared with the maximum value of the predicted load differences corresponding to all second sampling points and the absolute value is taken to obtain the reference weight corresponding to the second sampling point.

[0056] Specifically, when there is an impact load, the power grid will be unstable. Therefore, during training, the model tends to adjust the load with large changes to be similar to the load of the adjacent sampling point. However, in actual use, the above adjustment method will cause the impact load to still have a significant impact on the stability of the target power grid. Therefore, this embodiment normalizes the predicted load difference corresponding to the second sampling point and maps it to a reference weight. For the second sampling point with a large predicted load difference, the subtraction result corresponding to the second sampling point also accounts for a large proportion of the first training sub-loss. Compared with other second sampling points, the difference between the predicted load difference and the adjusted load difference corresponding to the second sampling point is smaller. That is, through the reference weight, the model tends to adjust the load of the adjacent sampling point to be similar to the load with large changes.

[0057] Step S206, updating the parameters of the difference adjustment model according to the first training loss until the first training loss converges, thereby obtaining an updated difference adjustment model.

[0058] Among them, according to the first training loss, the model parameters of the difference adjustment model are trained using the reverse gradient descent method until the first training loss converges to obtain an updated difference adjustment model.

[0059] Step S207: input the N predicted load differences into the updated difference adjustment model to obtain the target load differences corresponding to the N second sampling points.

[0060] The target load difference can be used to provide load information for controlling the energy storage converter.

[0061] Step S208, obtaining N load target values ​​according to the N target load difference values ​​and the Mth load sampling value.

[0062] Among them, for the target load difference corresponding to any second sampling point, if the second sampling point is the first second sampling point, the target load difference corresponding to the second sampling point and the Mth load sampling value are added to obtain the load target value corresponding to the second sampling point.

[0063] If the second sampling point is not the first second sampling point, the target load difference corresponding to the second sampling point and the load target value corresponding to the previous second sampling point are added to obtain the load target value corresponding to the second sampling point; The target load differences corresponding to all the second sampling points are traversed to obtain the load target values ​​corresponding to the N second sampling points respectively.

[0064] Step S209: controlling the energy storage converter corresponding to the target power grid according to the N load target values.

[0065] Among them, the existing energy storage converter control method can be used to control the energy storage converter according to the load, which will not be described in detail here.

[0066] In this embodiment, load prediction values ​​corresponding to N second sampling points are predicted based on M load sampling values, and then N predicted load differences are obtained. The power grid fluctuation probability corresponding to the N second sampling points is determined in combination with the power grid fluctuation prediction model. The predicted load difference is adjusted according to the predicted load difference, the adjusted load difference and the power grid fluctuation probability to obtain the target load difference, and then the adjusted N load target values ​​are obtained. The energy storage inverter is controlled according to the N target load values, so that when the target power grid faces an impact load, the energy storage inverter can be controlled to discharge in advance to ensure that the target power grid can operate stably and improve the stability of the power grid.

[0067] Corresponding to the method of the above embodiment 1, Figure 3The schematic diagram of the structure of an intelligent control device for an energy storage converter provided by the second embodiment of the present invention is shown. The intelligent control device is applied to a client. The client obtains the load sampling values ​​corresponding to the M first sampling points from the target power grid, and controls the energy storage converter according to the N load target values. The client has a storage function. The client stores a trained load prediction model, a trained power grid fluctuation prediction model and a difference adjustment model. The trained load prediction model can be used to predict the load of the second sampling point according to the load sampling value. The trained power grid fluctuation prediction model can be used to predict the fluctuation of the power grid according to the predicted load difference. The difference adjustment model is an online learning model, which can be used to adjust the predicted load difference. For ease of explanation, only the part related to the embodiment of the present invention is shown.

[0068] See also Figure 3 , the intelligent control device comprises: The load prediction module 31 is used to input the acquired load sampling values ​​corresponding to the M first sampling points of the target power grid into the trained load prediction model to obtain the load prediction values ​​corresponding to the N second sampling points, where M and N are both positive integers; The difference calculation module 32 is used to obtain the predicted load difference values ​​corresponding to the N second sampling points respectively according to the N load prediction values ​​and the Mth load sampling value; The fluctuation prediction module 33 is used to input the N predicted load differences into the trained power grid fluctuation prediction model to obtain the power grid fluctuation probabilities corresponding to the N second sampling points respectively; The difference adjustment module 34 is used to input the N predicted load differences into a preset difference adjustment model to obtain the adjusted load differences corresponding to the N second sampling points; A first loss calculation module 35, configured to calculate a first training loss according to the N predicted load differences, the N adjusted load differences, the N power grid fluctuation probabilities and a preset training loss function; A first model updating module 36, used to update the parameters of the difference adjustment model according to the first training loss until the first training loss converges to obtain an updated difference adjustment model; The difference determination module 37 is used to input the N predicted load differences into the updated difference adjustment model to obtain the target load differences corresponding to the N second sampling points respectively; A load determination module 38, configured to obtain N load target values ​​according to the N target load differences and the Mth load sampling value; The converter control module 39 is used to control the energy storage converter corresponding to the target power grid according to the N load target values.

[0069] Optionally, the load forecasting module 31 includes: The first model prediction unit is used to initialize n=1, input M load sampling values ​​into the trained load prediction model, and obtain the load prediction value corresponding to the first second sampling point; The second model prediction unit is used to input the load prediction value from the n+1th load sampling value to the Mth load sampling value, and the load prediction value corresponding to the first second sampling point to the load prediction value corresponding to the nth second sampling point into the trained load prediction model to obtain the load prediction value corresponding to the n+1th second sampling point; The third model prediction unit is used to update n=n+1, return to execute the step of inputting the n+1th load sampling value to the Mth load sampling value, and the load prediction value corresponding to the first second sampling point to the load prediction value corresponding to the nth second sampling point into the trained load prediction model, and obtain the load prediction value corresponding to the n+1th second sampling point, until n+1=N, and obtain the load prediction values ​​corresponding to the N second sampling points respectively.

[0070] Optionally, the difference calculation module 32 includes: A first difference calculation unit is used for, for a load prediction value corresponding to any second sampling point, if the second sampling point is the first second sampling point, subtracting the load prediction value corresponding to the second sampling point from the Mth load sampling value to obtain a predicted load difference value corresponding to the second sampling point; A second difference calculation unit is used for, if the second sampling point is not the first second sampling point, subtracting the load prediction value corresponding to the second sampling point from the load prediction value corresponding to the previous second sampling point to obtain a predicted load difference corresponding to the second sampling point; The first traversal unit is used to traverse the load prediction values ​​corresponding to all the second sampling points to obtain the predicted load difference values ​​corresponding to the N second sampling points respectively.

[0071] Optionally, the above intelligent control device further includes: A data acquisition module, used to acquire a historical load difference and a reference fluctuation type corresponding to the target power grid at a third sampling point, where the reference fluctuation type includes a first fluctuation type and a second fluctuation type; A probability prediction module is used to input the power grid fluctuation prediction model according to the historical load difference, obtain the reference prediction probabilities corresponding to the first fluctuation type and the second fluctuation type respectively, and take the fluctuation type corresponding to the larger reference prediction probability as the predicted fluctuation type; The second model updating module is used to calculate the second training loss according to the predicted fluctuation type and the reference fluctuation type, and train the power grid fluctuation prediction model according to the second training loss until the second training loss converges to obtain a trained power grid fluctuation prediction model.

[0072] Optionally, the fluctuation prediction module 33 includes: A type prediction unit, for inputting the predicted load difference into a trained power grid fluctuation prediction model for any predicted load difference, to obtain a target prediction probability corresponding to a first fluctuation type; The second traversal unit is used to traverse all predicted load differences using the target predicted probability corresponding to the first fluctuation type as the power grid fluctuation probability corresponding to the predicted load difference to obtain N power grid fluctuation probabilities.

[0073] Optionally, the first loss calculation module 35 includes: A first sub-loss calculation unit is used to, for any second sampling point, subtract the predicted load difference and the adjusted load difference corresponding to the second sampling point and then take the absolute value to obtain a subtraction result, traverse all second sampling points to obtain N subtraction results, and add the N subtraction results to obtain a first training sub-loss; A second sub-loss calculation unit, used for adding N power grid fluctuation probabilities and taking the addition result as a second training sub-loss; The first loss determining unit is configured to add the first training sub-loss and the second training sub-loss to obtain a first training loss.

[0074] Optionally, the first sub-loss calculation unit includes: A weight mapping subunit, configured to map, for any second sampling point, a load prediction value corresponding to the second sampling point to a reference weight according to a preset mapping function; A weighting subunit, used to multiply the reference weight corresponding to the second sampling point and the subtraction result to obtain a reference subtraction result corresponding to the second sampling point; A third traversal subunit is used to traverse all the second sampling points to obtain reference subtraction results corresponding to the N second sampling points respectively; The sub-loss determination sub-unit is used to add the N reference subtraction results to obtain a first training sub-loss.

[0075] It should be noted that the information interaction, execution process, etc. between the above-mentioned modules, units, and sub-units are based on the same concept as the embodiment of the method of the present invention. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0076] Figure 4 This is a schematic diagram of the structure of a computer device corresponding to an energy storage converter intelligent control method provided in Embodiment 3 of the present invention. Figure 4 As shown, the computer device of this embodiment includes: at least one processor ( Figure 4 Only one is shown in the figure), a memory, and a computer program stored in the memory and executable on at least one processor, wherein when the processor executes the computer program, the steps in any of the above-mentioned embodiments of the intelligent control method are implemented.

[0077] The computer device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that Figure 4 This is merely an example of a computer device and does not constitute a limitation on the computer device. The computer device may include more or fewer components than those shown in the figure, or a combination of certain components, or different components. For example, it may also include a network interface, a display screen, and an input device.

[0078] The processor may be a CPU, or other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0079] The memory includes a readable storage medium, an internal memory, etc., wherein the internal memory may be the memory of a computer device, and the internal memory provides an environment for the operation of the operating system and computer-readable instructions in the readable storage medium. The readable storage medium may be a hard disk of a computer device, and in other embodiments, it may also be an external storage device of the computer device, for example, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device. Further, the memory may also include both an internal storage unit of the computer device and an external storage device. The memory is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of a computer program, etc. The memory may also be used to temporarily store data that has been output or is to be output.

[0080] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention. The specific working process of the units and modules in the above-mentioned device can refer to the corresponding process in the above-mentioned method embodiment, which will not be repeated here. If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned method embodiment can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device capable of carrying computer program code, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.

[0081] The present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by a computer program product. When the computer program product runs on a computer device, the computer device can implement the steps in the above-mentioned method embodiment when executing.

[0082] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0083] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0084] In the embodiments provided by the present invention, it should be understood that the disclosed devices / computer equipment and methods can be implemented in other ways. For example, the device / computer equipment embodiments described above are only schematic, for example, the division of modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0085] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0086] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. An intelligent control method for an energy storage converter, characterized in that: The energy storage converter intelligent control method comprises: Input the acquired load sampling values ​​corresponding to the M first sampling points of the target power grid into the trained load forecasting model to obtain the load forecasting values ​​corresponding to the N second sampling points, where M and N are both positive integers; According to the N load prediction values ​​and the Mth load sampling value, the predicted load difference values ​​corresponding to the N second sampling points are obtained; Inputting N predicted load differences into the trained power grid fluctuation prediction model, and obtaining the power grid fluctuation probabilities corresponding to the N second sampling points respectively; Inputting N predicted load difference values ​​into a preset difference adjustment model to obtain adjusted load difference values ​​corresponding to N second sampling points respectively; Calculating a first training loss according to the N predicted load differences, the N adjusted load differences, the N power grid fluctuation probabilities and a preset training loss function; According to the first training loss, updating the parameters of the difference adjustment model until the first training loss converges, thereby obtaining an updated difference adjustment model; Inputting the N predicted load differences into the updated difference adjustment model to obtain target load differences corresponding to the N second sampling points respectively; Obtaining N load target values ​​according to the N target load differences and the Mth load sampling value; According to the N load target values, the energy storage converter corresponding to the target power grid is controlled.

2. The energy storage converter intelligent control method according to claim 1, characterized in that: The step of inputting the acquired load sampling values ​​of the target power grid corresponding to the M first sampling points into the trained load prediction model to obtain the load prediction values ​​corresponding to the N second sampling points respectively includes: Initialize n=1, input M load sampling values ​​into the trained load forecasting model, and obtain the load forecast value corresponding to the nth second sampling point; Input the load prediction value from the n+1th load sampling value to the Mth load sampling value, and the load prediction value corresponding to the first second sampling point to the load prediction value corresponding to the nth second sampling point into the trained load prediction model to obtain the load prediction value corresponding to the n+1th second sampling point; Update n=n+1, return to execute the step of converting the n+1th load sampling value to the Mth load sampling value, and the load prediction value corresponding to the first second sampling point to the load prediction value corresponding to the nth second sampling point, inputting them into the trained load prediction model, and obtaining the load prediction value corresponding to the n+1th second sampling point, until n+1=N, and obtaining the load prediction values ​​corresponding to the N second sampling points respectively.

3. The intelligent control method for energy storage converter according to claim 1, characterized in that: The method of obtaining predicted load differences corresponding to N second sampling points respectively according to the N load prediction values ​​and the Mth load sampling value includes: For the load prediction value corresponding to any second sampling point, if the second sampling point is the first second sampling point, subtract the load prediction value corresponding to the second sampling point from the Mth load sampling value to obtain the predicted load difference corresponding to the second sampling point; If the second sampling point is not the first second sampling point, subtract the load prediction value corresponding to the second sampling point from the load prediction value corresponding to the previous second sampling point to obtain the predicted load difference corresponding to the second sampling point; The load prediction values ​​corresponding to all the second sampling points are traversed to obtain the predicted load difference values ​​corresponding to the N second sampling points respectively.

4. The intelligent control method for energy storage converter according to claim 1, characterized in that: The training process of the power grid fluctuation prediction model includes: Acquire a historical load difference and a reference fluctuation type corresponding to the target power grid at a third sampling point, wherein the reference fluctuation type includes a first fluctuation type and a second fluctuation type; Input the power grid fluctuation prediction model according to the historical load difference, obtain reference prediction probabilities corresponding to the first fluctuation type and the second fluctuation type respectively, and use the fluctuation type corresponding to the larger reference prediction probability as the predicted fluctuation type; A second training loss is calculated according to the predicted fluctuation type and the reference fluctuation type, and the power grid fluctuation prediction model is trained according to the second training loss until the second training loss converges to obtain the trained power grid fluctuation prediction model.

5. The intelligent control method for energy storage converter according to claim 4, characterized in that: The step of inputting the N predicted load differences into the trained power grid fluctuation prediction model to obtain the power grid fluctuation probabilities corresponding to the N second sampling points respectively includes: For any predicted load difference, input the predicted load difference into the trained power grid fluctuation prediction model to obtain a target prediction probability corresponding to the first fluctuation type; The target predicted probability corresponding to the first fluctuation type is used as the power grid fluctuation probability corresponding to the predicted load difference, and all predicted load differences are traversed to obtain N power grid fluctuation probabilities.

6. The intelligent control method for energy storage converter according to any one of claims 1 to 5, characterized in that: The first training loss is calculated based on the N predicted load differences, the N adjusted load differences, the N power grid fluctuation probabilities and a preset training loss function, including: For any second sampling point, subtract the predicted load difference and the adjusted load difference corresponding to the second sampling point and then take the absolute value to obtain a subtraction result, traverse all second sampling points to obtain N subtraction results, add the N subtraction results to obtain the first training sub-loss; Adding the N power grid fluctuation probabilities, and taking the added result as the second training sub-loss; The first training sub-loss and the second training sub-loss are added to obtain the first training loss.

7. The intelligent control method for energy storage converter according to claim 6, characterized in that: The adding the N subtraction results to obtain a first training sub-loss includes: For any second sampling point, mapping the load prediction value corresponding to the second sampling point to a reference weight according to a preset mapping function; Multiplying the reference weight corresponding to the second sampling point and the subtraction result to obtain a reference subtraction result corresponding to the second sampling point; Traversing all second sampling points to obtain reference subtraction results corresponding to N second sampling points respectively; The N reference subtraction results are added together to obtain the first training sub-loss.

8. An intelligent control device for energy storage converter, characterized in that: The energy storage converter intelligent control device comprises: A load prediction module is used to input the acquired load sampling values ​​corresponding to the M first sampling points of the target power grid into the trained load prediction model to obtain the load prediction values ​​corresponding to the N second sampling points, where M and N are both positive integers; A difference calculation module, used for obtaining predicted load differences corresponding to N second sampling points respectively according to N load prediction values ​​and the Mth load sampling value; A fluctuation prediction module is used to input N predicted load differences into a trained power grid fluctuation prediction model to obtain power grid fluctuation probabilities corresponding to N second sampling points respectively; A difference adjustment module, used for inputting N predicted load differences into a preset difference adjustment model to obtain adjusted load differences corresponding to N second sampling points respectively; A first loss calculation module, used to calculate a first training loss according to the N predicted load differences, the N adjusted load differences, the N power grid fluctuation probabilities and a preset training loss function; A first model updating module, used for updating the parameters of the difference adjustment model according to the first training loss until the first training loss converges to obtain an updated difference adjustment model; A difference determination module, used for inputting the N predicted load differences into the updated difference adjustment model to obtain target load differences corresponding to the N second sampling points respectively; A load determination module, configured to obtain N load target values ​​according to the N target load differences and the Mth load sampling value; The converter control module is used to control the energy storage converter corresponding to the target power grid according to the N load target values.

9. A computer device, characterized in that: The computer device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the intelligent control method for the energy storage converter according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the intelligent control method for the energy storage converter according to any one of claims 1 to 7 is implemented.

Images