Photovoltaic ultra-short-term power prediction method and system

Abstract

According to the photovoltaic ultra-short-term power prediction method and system provided by the invention, it is determined that the time range specified by the target ultra-short-term power event transmission data is met and the time range specified by the target ultra-short-term power event transmission data is similar to the target energy consumption loss data; and inputting the target ultra-short-term power event transmission data into a target power transmission feature chain corresponding to the target energy consumption loss data, and inputting the target ultra-short-term power event transmission data into the target energy consumption loss data, the target energy consumption loss data is most similar to the target energy consumption loss data of the target ultra-short-term power event transmission data, so that the optimization workload of the target ultra-short-term power event transmission data is reduced, and the optimization data correction network is executed based on a plurality of different target power transmission feature chains through a plurality of power prediction networks, so that the optimization efficiency of the target ultra-short-term power event transmission data is improved. Therefore, the accuracy and confidence of predicting the ultra-short-term power data can be improved while the optimization of the data correction network can be ensured at the same time.

Description

Technical Field

[0001] The present application relates to the technical field of data prediction, and in particular to a photovoltaic ultra-short-term power prediction method and system. Background Art

[0002] Ultra-short-term power forecasting refers to the prediction of power output results every 15 minutes within the next 0-4 hours. This forecasting method is mainly used in new energy sites, such as wind power and photovoltaic power stations, to cope with their randomness, intermittency and volatility.

[0003] At present, the ultra-short-term power prediction method in traditional technology has the problem of low prediction accuracy, which leads to a large deviation between the actual ultra-short-term power and the prediction. Therefore, a technical solution is urgently needed to improve the above technical problems. Summary of the invention

[0004] In order to improve the technical problems existing in the related technologies, the present application provides a photovoltaic ultra-short-term power prediction method and system.

[0005] In a first aspect, a photovoltaic ultra-short-term power prediction method is provided, the method comprising: obtaining a plurality of target ultra-short-term power event transmission data and a plurality of target energy consumption loss data, wherein the plurality of target energy consumption loss data are different; forming a target power transmission feature chain with the target ultra-short-term power event transmission data corresponding to the same target energy consumption loss data, wherein the target energy consumption loss data corresponding to the target ultra-short-term power event transmission data is: covering the time resolution feature specified by the target ultra-short-term power event transmission data, and having a time range and a target energy consumption loss data similar to the time range specified by the target ultra-short-term power event transmission data in the plurality of target energy consumption loss data; inputting the target ultra-short-term power event transmission data in each target power transmission feature chain that is smaller than the target energy consumption loss data corresponding to the target power transmission feature chain into the target energy consumption loss data corresponding to the target power transmission feature chain; and simultaneously executing an optimized data correction network based on the predicted target power transmission feature chain through each of the plurality of power prediction networks, wherein the data correction network is used to perform prediction processing on random ultra-short-term power data.

[0006] In the present application, the data correction network includes a partial data correction network corresponding to each of the power prediction networks, and the optimization of the data correction network is simultaneously performed based on the predicted target power transmission feature chain through each of the multiple power prediction networks, including: for each of the power prediction networks, the multiple target ultra-short-term power event transmission data in the predicted target power transmission feature chain are spliced ​​into an ultra-short-term power influencing factor directory, the first feature quantity of the ultra-short-term power influencing factor directory is the same as the number of the multiple target ultra-short-term power event transmission data, the second feature quantity of the ultra-short-term power influencing factor directory is equal to the target energy consumption loss data corresponding to the target power transmission feature chain, or the first feature quantity is equal to the target energy consumption loss data corresponding to the target power transmission feature chain, and the second feature quantity is equal to the number of the multiple target ultra-short-term power event transmission data; combined with the ultra-short-term power influencing factor directory, the partial data correction network corresponding to the power prediction network is optimized.

[0007] In the present application, after optimizing the partial data correction network corresponding to the power prediction network in combination with the ultra-short term power influencing factor catalog, the method further includes: correcting the important distribution of multiple target ultra-short term power event transmission data in the ultra-short term power influencing factor catalog to obtain a corrected ultra-short term power influencing factor catalog; and continuing to optimize the partial data correction network corresponding to the power prediction network in combination with the corrected ultra-short term power influencing factor catalog.

[0008] In the present application, the data correction network includes a partial data correction network corresponding to each of the power prediction networks, and the optimization of the data correction network is simultaneously performed based on the predicted target power transmission characteristic chain through each of the multiple power prediction networks, including: for each of the power prediction networks, the original important power data text on at least one time resolution characteristic distribution of the multiple target ultra-short term power event transmission data in the predicted target power transmission characteristic chain is optimized to the target important power data text; through the partial data correction network corresponding to the power prediction network, based on the optimized multiple target ultra-short term power event transmission data, the parsed important power data text on the at least one time resolution characteristic distribution is obtained; and the partial data correction network corresponding to the power prediction network is optimized by combining the original important power data text and the parsed important power data text on the at least one time resolution characteristic distribution.

[0009] In the present application, the partial data correction network corresponding to the power prediction network is used to parse the optimized multiple target ultra-short term power event transmission data to obtain the parsed important power data text on the at least one time resolution characteristic distribution, including: classifying and processing the optimized multiple target ultra-short term power event transmission data through the partial data correction network corresponding to the power prediction network to obtain the important power data text description corresponding to each important power data text on the time resolution characteristic distribution in the optimized multiple target ultra-short term power event transmission data; obtaining the important power data text description corresponding to the target important power data text on the at least one time resolution characteristic distribution from the obtained multiple important power data text descriptions; and parsing in combination with the important power data text description corresponding to the target important power data text on the at least one time resolution characteristic distribution to obtain the parsed important power data text on the at least one time resolution characteristic distribution.

[0010] In the present application, the data correction network includes a partial data correction network corresponding to each of the power prediction networks, and the optimization of the data correction network is simultaneously performed based on the predicted target power transmission characteristic chain through each of the multiple power prediction networks, including: optimizing the corresponding partial data correction network based on the predicted target power transmission characteristic chain through each of the power prediction networks; determining all optimization data based on the first optimization data obtained by each of the power prediction networks, on the premise that each of the power prediction networks obtains the first optimization data of the corresponding partial data correction network; optimizing the corresponding partial data correction network through each of the power prediction networks in combination with the all optimization data.

[0011] In the present application, the partial data correction network includes multiple units, and the determination of all optimization data based on the first optimization data obtained by each power prediction network, on the premise that each power prediction network obtains the first optimization data of the respective corresponding partial data correction network, includes: obtaining the first optimization data of the first unit in the respective corresponding partial data correction network one by one through each power prediction network; determining all optimization data of the first unit based on the first optimization data that each power prediction network has obtained and has not processed, on the premise that the vectors of the first optimization data that each power prediction network has obtained and has not processed meet the target value; continuing to obtain the first optimization data of the second unit in the respective corresponding partial data correction network one by one through each power prediction network; determining all optimization data of the second unit based on the first optimization data that each power prediction network has obtained and has not processed, on the premise that the vectors of the first optimization data that each power prediction network has obtained and has not processed meet the target value, or on the premise that each power prediction network has obtained the first optimization data of each unit in the respective corresponding partial data correction network.

[0012] In the present application, the method also includes: on the premise that the vector of the first optimization data obtained by the power prediction network does not belong to the weight coefficient of the target vector, optimizing the first optimization data through the power prediction network so that the vector of the first optimization data belongs to the weight coefficient of the target vector.

[0013] In a second aspect, a photovoltaic ultra-short-term power prediction system is provided, comprising a processor and a memory communicating with each other, wherein the processor is used to read a computer program from the memory and execute the computer program to implement the above method.

[0014] A photovoltaic ultra-short-term power prediction method and system provided in an embodiment of the present application determines, among multiple target energy consumption loss data, a target ultra-short-term power event transmission data that meets the time range specified by the target ultra-short-term power event transmission data and is similar to the target energy consumption loss data within the time range specified by the target ultra-short-term power event transmission data, and inputs the target ultra-short-term power event transmission data into a target power transmission feature chain corresponding to the target energy consumption loss data, and inputs the target ultra-short-term power event transmission data into the target energy consumption loss data. Since the target energy consumption loss data is most similar to the target energy consumption loss data of the target ultra-short-term power event transmission data, the optimization workload of the target ultra-short-term power event transmission data is reduced, and then, through multiple power prediction networks based on multiple different target power transmission feature chains, the optimized data correction network is simultaneously executed, thereby ensuring that the accuracy and confidence of the prediction of the ultra-short-term power data are improved while the optimized data correction network is simultaneously executed. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the time range. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 A flowchart of a photovoltaic ultra-short-term power prediction method provided in an embodiment of the present application.

[0017] Figure 2 A block diagram of a photovoltaic ultra-short-term power prediction device provided in an embodiment of the present application.

[0018] Figure 3 An architecture diagram of a photovoltaic ultra-short-term power prediction system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] In order to better understand the above technical scheme, the technical scheme of the present application is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical scheme of the present application, rather than limitations on the technical scheme of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0020] See also Figure 1 , shows a photovoltaic ultra-short-term power prediction method, which may include the technical solutions described in the following steps 201-204.

[0021] 201. Acquire multiple target ultra-short term power event transmission data and multiple target energy consumption loss data.

[0022] The multiple target ultra-short term power event transmission data are used to optimize the data correction network. The multiple target ultra-short term power event transmission data can be any type of target ultra-short term power event transmission data. For different types of optimization instructions, corresponding types of target ultra-short term power event transmission data can be obtained. The target energy consumption loss data indication is used to optimize the description content of the target ultra-short term power event transmission data of the data correction network. Further, the multiple target energy consumption loss data are previously specified description contents.

[0023] Among them, the description contents of the multiple target ultra-short-term power event transmission data are different, and the multiple target energy consumption loss data are different.

[0024] 202. Target ultra-short term power event transmission data corresponding to the same target energy consumption loss data constitute a target power transmission feature chain.

[0025] Among the multiple target energy consumption loss data, the target energy consumption loss data corresponding to each target ultra-short term power event transmission data is determined. The target energy consumption loss data corresponding to the target ultra-short term power event transmission data is: greater than or equal to the description content of the target ultra-short term power event transmission data, and the time range specified by the target ultra-short term power event transmission data is similar to the target energy consumption loss data among the multiple target energy consumption loss data.

[0026] After determining the target energy consumption loss data corresponding to each target ultra-short term power event transmission data, the target ultra-short term power event transmission data corresponding to the same target energy consumption loss data constitute a target power transmission characteristic chain, thereby obtaining multiple target power transmission characteristic chains, and one target energy consumption loss data corresponds to one target power transmission characteristic chain.

[0027] 203. Input the target ultra-short term power event transmission data in each target power transmission characteristic chain that is smaller than the target energy consumption loss data corresponding to the target power transmission characteristic chain into the target energy consumption loss data corresponding to the target power transmission characteristic chain.

[0028] Since the description contents of multiple target ultra-short term power event transmission data used in the optimized data correction network need to be the same, after obtaining multiple target power transmission characteristic chains, for each target power transmission characteristic chain, the target ultra-short term power event transmission data in the target power transmission characteristic chain that is smaller than the target energy consumption loss data corresponding to the target power transmission characteristic chain is determined, and these target ultra-short term power event transmission data are input into the target energy consumption loss data corresponding to the target power transmission characteristic chain, thereby ensuring that the description contents of multiple target ultra-short term power event transmission data in the target power transmission characteristic chain are all the target energy consumption loss data corresponding to the target power transmission characteristic chain.

[0029] 204. Through each of the multiple power prediction networks, based on the predicted target power transmission characteristic chain, an optimized data correction network is simultaneously executed.

[0030] In this embodiment, multiple power prediction networks are included, and the multiple power prediction networks can be processed simultaneously. Therefore, in order to improve the optimization efficiency of the data correction network, the multiple target power transmission characteristic chains obtained are divided into the multiple power prediction networks for processing, and each power prediction network is divided into different target power transmission characteristic chains. Through each of the multiple power prediction networks, based on the predicted target power transmission characteristic chain, the optimized data correction network is simultaneously executed.

[0031] In this embodiment, multiple target ultra-short term power event transmission data are input into the data correction network for processing at the same time. This requires ensuring that the description contents of the multiple target ultra-short term power event transmission data are the same. Therefore, after obtaining the multiple target ultra-short term power event transmission data, the description contents of all target ultra-short term power event transmission data are input into the best description content among the multiple target ultra-short term power event transmission data, resulting in a large workload for optimizing the target ultra-short term power event transmission data. In the present embodiment, a plurality of target energy consumption loss data are determined, and the target ultra-short term power event transmission data is input into a description content that is greater than the target ultra-short term power event transmission data and is similar to the time range specified by the target ultra-short term power event transmission data and the target energy consumption loss data, thereby effectively reducing the optimization workload of the target ultra-short term power event transmission data. Then, the target power transmission characteristic chains corresponding to different target energy consumption loss data are divided into a plurality of power prediction networks, so that the plurality of power prediction networks simultaneously execute the optimization data correction network based on the predicted target power transmission characteristic chains. Therefore, although the plurality of target ultra-short term power event transmission data are divided into target power transmission characteristic chains corresponding to different target energy consumption loss data, the optimization data correction network can still be simultaneously executed based on the plurality of target ultra-short term power event transmission data, thereby ensuring that the optimization efficiency is accelerated while also weakening the interference in the optimization process.

[0032] In this embodiment, among multiple target energy consumption loss data, the target ultra-short term power event transmission data that meets the time range specified by the target ultra-short term power event transmission data and is similar to the target energy consumption loss data within the time range specified by the target ultra-short term power event transmission data is determined, and the target ultra-short term power event transmission data is input into the target power transmission feature chain corresponding to the target energy consumption loss data. If the description content of the target ultra-short term power event transmission data is less than the target energy consumption loss data, the target ultra-short term power event transmission data is input into the target energy consumption loss data. Since the target energy consumption loss data is the most similar to the target energy consumption loss data of the target ultra-short term power event transmission data, the optimization workload of the target ultra-short term power event transmission data is reduced. Then, through multiple power prediction networks based on multiple different target power transmission feature chains, the optimized data correction network is simultaneously executed, thereby ensuring that the accuracy and confidence of the prediction of ultra-short term power data are improved while the optimized data correction network is simultaneously executed.

[0033] Another photovoltaic ultra-short-term power prediction method provided in an embodiment of the present application is a process of optimizing a data correction network based on multiple target power transmission characteristic chains. The embodiment of the present application specifically describes the steps of optimizing a data correction network based on an ultra-short-term power influencing factor catalog, which may include the following contents.

[0034] 301. For each power prediction network, multiple target ultra-short-term power event transmission data in the predicted target power transmission feature chain are spliced ​​into an ultra-short-term power influencing factor catalog.

[0035] In this embodiment, the data correction network includes a portion of the data correction network corresponding to each power prediction network.

[0036] A power prediction network is taken as an example for further explanation. Through the power prediction network, multiple target ultra-short term power event transmission data in the target power transmission characteristic chain predicted by the power prediction network are obtained. Since the description content of the multiple target ultra-short term power event transmission data has been unified into the target energy consumption loss data corresponding to the target power transmission characteristic chain, the multiple target ultra-short term power event transmission data can be spliced ​​into an ultra-short term power influencing factor directory, so that the ultra-short term power influencing factor directory can be directly processed through the partial data correction network corresponding to the power prediction network, thereby realizing simultaneous processing of multiple target ultra-short term power event transmission data.

[0037] Among them, the first characteristic quantity of the ultra-short term power influencing factor catalog is the same as the number of multiple target ultra-short term power event transmission data, and the second characteristic quantity of the ultra-short term power influencing factor catalog is equal to the target energy consumption loss data corresponding to the target power transmission characteristic chain. That is, the multiple important power data texts in each target ultra-short term power event transmission data are distributed according to the importance, and then the multiple target ultra-short term power event transmission data are spliced ​​one by one according to the distribution of the importance, so as to obtain the ultra-short term power influencing factor catalog. Alternatively, the first characteristic quantity of the ultra-short term power influencing factor catalog is equal to the target energy consumption loss data corresponding to the target power transmission characteristic chain, and the second characteristic quantity of the ultra-short term power influencing factor catalog is the same as the number of multiple target ultra-short term power event transmission data.

[0038] 302. Based on the catalog of ultra-short-term power influencing factors, optimize the partial data correction network corresponding to the power prediction network.

[0039] Based on the ultra-short-term power influencing factor catalog, a partial data correction network is optimized. Further, a target indication corresponding to the ultra-short-term power influencing factor catalog is obtained, the ultra-short-term power influencing factor catalog is processed through the partial data correction network to obtain a processing result, and based on the processing result and the target indication corresponding to the ultra-short-term power influencing factor catalog, the partial data correction network is optimized.

[0040] 303. Correct the important distribution of multiple target ultra-short term power event transmission data in the ultra-short term power influencing factor catalog to obtain a corrected ultra-short term power influencing factor catalog.

[0041] The multiple target ultra-short-term power event transmission data in the ultra-short-term power impact factor catalog are distributed in a certain order. Based on the ultra-short-term power impact factor catalog, after optimizing the partial data correction network at least once, the important distribution of the multiple target ultra-short-term power event transmission data in the ultra-short-term power impact factor catalog is corrected again to obtain a corrected ultra-short-term power impact factor catalog. Further, the important distribution of the multiple target ultra-short-term power event transmission data in the ultra-short-term power impact factor catalog is corrected arbitrarily, or a specified correction method is used to correct the important distribution of the multiple target ultra-short-term power event transmission data in the ultra-short-term power impact factor catalog.

[0042] Therefore, the multiple target ultra-short term power event transmission data in the revised ultra-short term power influencing factor catalog are the same as the multiple target ultra-short term power event transmission data in the ultra-short term power influencing factor catalog before correction, but the important distribution of the multiple target ultra-short term power event transmission data is different.

[0043] 304. Based on the revised catalog of ultra-short-term power influencing factors, continue to optimize the partial data correction network corresponding to the power prediction network.

[0044] After obtaining the corrected ultra-short term power influencing factor list, the partial data correction network is continuously optimized based on the corrected ultra-short term power influencing factor list.

[0045] The method provided in the embodiment of the present application splices multiple target ultra-short term power event transmission data in the target power transmission feature chain into an ultra-short term power influencing factor directory, and optimizes a partial data correction network based on the ultra-short term power influencing factor directory, so that during the optimization process of the partial data correction network, multiple target ultra-short term power event transmission data in the ultra-short term power influencing factor directory can be processed in parallel, thereby improving the optimization speed of the partial data correction network.

[0046] Another photovoltaic ultra-short-term power prediction method is provided in an embodiment of the present application. The executor of the embodiment of the present application is a process of optimizing a data correction network based on multiple target power transmission characteristic chains. The process of optimizing a data correction network based on original important power data text and parsed important power data text is specifically described. The method may include the following contents.

[0047] 601. For each power prediction network, optimize the original important power data text on at least one time resolution feature distribution in multiple target ultra-short-term power event transmission data in the predicted target power transmission feature chain into a target important power data text.

[0048] In this embodiment, the data correction network includes a partial data correction network corresponding to each power prediction network. The partial data correction network is a sub-thread of the preliminary optimized parsing class, which is configured to block at least one original important power data text in the target ultra-short-term power event transmission data, and to allow the partial data correction network to parse the blocked original important power data text, thereby continuously learning the features in the text. Therefore, taking a power prediction network as an example, through the power prediction network, the original important power data text on at least one time resolution feature distribution in each target ultra-short-term power event transmission data in the target power transmission feature chain predicted by the power prediction network is optimized to the target important power data text.

[0049] Further, the at least one time resolution feature distribution is a previously specified time resolution feature distribution. Further, the target important power data text is a previously specified important power data text.

[0050] 602. Performing analysis based on the optimized transmission data of multiple target ultra-short-term power events through a partial data correction network corresponding to the power prediction network to obtain an analyzed important power data text on the at least one time resolution characteristic distribution.

[0051] After obtaining the optimized multiple target ultra-short-term power event transmission data, the partial data correction network is used to perform parsing based on the optimized multiple target ultra-short-term power event transmission data to obtain the parsed important power data text on the at least one time resolution characteristic distribution. Among them, the parsed important power data text is which important power data text should be on the at least one time resolution characteristic distribution parsed by the partial data correction network. The optimization purpose of the partial data correction network is to parse out the original important power data text on the at least one time resolution characteristic distribution, that is, if the parsed important power data text obtained by the partial data correction network is the same as the original important power data text, it means that the calculation method of the partial data correction network is accurate.

[0052] For some possible implementation embodiments, the optimized multiple target ultra-short term power event transmission data are classified and processed through a partial data correction network corresponding to the power prediction network, and important power data text descriptions corresponding to each important power data text on a time resolution characteristic distribution in the optimized multiple target ultra-short term power event transmission data are obtained. Among the multiple important power data text descriptions obtained, the important power data text description corresponding to the target important power data text on at least one time resolution characteristic distribution is obtained, and the important power data text description corresponding to the target important power data text on the at least one time resolution characteristic distribution is parsed to obtain the parsed important power data text on the at least one time resolution characteristic distribution.

[0053] Furthermore, based on the important power data text description corresponding to the target important power data text on the at least one time resolution characteristic distribution in the target ultra-short term power event transmission data, the text features corresponding to the target ultra-short term power event transmission data are determined, and parsing is performed based on the text features through a partial data correction network to obtain the parsed important power data text on the at least one time resolution characteristic distribution.

[0054] 603. Based on the original important power data text and the parsed important power data text on the at least one time resolution feature distribution, optimize a partial data correction network corresponding to the power prediction network.

[0055] Based on the difference between the original important power data text and the parsed important power data text on the at least one time resolution feature distribution, the partial data correction network is optimized. Since the original important power data text is the real important power data text, and the parsed important power data text is the important power data text parsed by the partial data correction network, there is a correlation between the parsed important power data text and the original important power data text, and the partial data correction network calculates the accuracy. Therefore, according to the difference between the original important power data text and the parsed important power data text, the partial data correction network is optimized, so that the difference between the parsed important power data text parsed by the partial data correction network and the original important power data text becomes smaller and smaller, so as to improve the calculation performance of the partial data correction network, thereby improving the calculation reliability of the partial data correction network.

[0056] Another photovoltaic ultra-short-term power prediction method provided in an embodiment of the present application is executed by a subject, and the embodiment of the present application specifically describes the process of optimizing a data correction network through multiple power prediction networks. The method includes the following steps.

[0057] 1001. Through each power prediction network, based on the predicted target power transmission characteristic chain, optimize the corresponding data correction network.

[0058] In this embodiment, the data correction network includes a partial data correction network corresponding to each power prediction network. Taking a power prediction network as an example, after the target power transmission characteristic chain is divided into the power prediction network, the partial data correction network corresponding to the power prediction network is optimized based on the predicted target power transmission characteristic chain through the power prediction network.

[0059] 1002. On the premise that each power prediction network obtains the first optimization data of its corresponding partial data correction network, all optimization data are determined based on the first optimization data obtained by each power prediction network.

[0060] Taking a power prediction network as an example, by optimizing a partial data correction network corresponding to the power prediction network through the power prediction network, first optimization data of the partial data correction network can be obtained, and the first optimization data is used to optimize the partial data correction network.

[0061] For some possible implementation embodiments, for each of the multiple power prediction networks, when the first optimization data is obtained through the power prediction network, the first optimization data is sent to other power prediction networks through the power prediction network, so that each of the multiple power prediction networks obtains the first optimization data. Therefore, under the premise that each power prediction network obtains the first optimization data of the partial data correction network corresponding to each power prediction network, each power prediction network will also obtain the first optimization data of the partial data correction network corresponding to other power prediction networks, that is, each power prediction network can obtain the first optimization data corresponding to the multiple power prediction networks. Then for each of the multiple power prediction networks, the power prediction network determines all the optimization data based on the multiple first optimization data obtained, so each of the power prediction networks can obtain all the optimization data.

[0062] 1003. Through each power prediction network, based on all optimization data, optimize the corresponding partial data correction network.

[0063] The partial data correction network includes multiple units. On the premise that each power prediction network obtains the first optimization data of its corresponding partial data correction network, all optimization data are determined based on the first optimization data obtained by each power prediction network, including the following steps 601-604.

[0064] 601. Obtain first optimization data of a first unit in each corresponding partial data correction network through each power prediction network.

[0065] 602. On the premise that the vectors of the first optimization data acquired and not processed by each power prediction network meet the target value, determine all the optimization data of the first unit based on the first optimization data acquired and not processed by each power prediction network.

[0066] 603. Continue to obtain the first optimization data of the second unit in the corresponding partial data correction network one by one through each power prediction network.

[0067] 604. On the premise that the vectors of the first optimization data that have been acquired and not processed by each power prediction network meet the target value, or on the premise that each power prediction network has acquired the first optimization data of each unit in its corresponding partial data correction network, determine all the optimization data of the second unit based on the first optimization data that have been acquired and not processed by each power prediction network.

[0068] In this embodiment, the partial data correction network includes multiple units. During the optimization process, the power prediction network obtains the first optimization data of the multiple units one by one. For example, the partial data correction network includes a first unit, a second unit, a third unit, a fourth unit and a fifth unit connected in sequence. The process of obtaining the first optimization data is a reverse calculation process. Therefore, during an iterative optimization process, the power prediction network will obtain the first optimization data of the fifth unit, the fourth unit, the third unit, the second unit and the first unit one by one. When the first optimization data of a certain unit is obtained, if the vector of the first optimization data that has been obtained and not processed does not reach the target value, the first optimization data of the next unit will continue to be obtained. If the vector of the first optimization data that has been obtained and not processed reaches the target value, the first optimization data that has been obtained and not processed is sent to other power prediction networks. Therefore, under the premise that the vectors of the first optimization data that each power prediction network has acquired and not processed meet the target value, each power prediction network can obtain the first optimization data that the multiple power prediction networks have acquired and not processed, then through each power prediction network, based on the first optimization data that each power prediction network has acquired and not processed, all the optimization data of the first unit are determined, wherein the first unit refers to at least one unit corresponding to the first optimization data that has been acquired and not processed. For example, the first optimization data that has been acquired and not processed is the first optimization data of the fifth unit and the fourth unit, then the first unit refers to the fifth unit and the fourth unit.

[0069] Then, through each power prediction network, continue to obtain the first optimization data of the next unit of the corresponding partial data correction network one by one. Similar to the above steps, on the premise that the vectors of the first optimization data that have been obtained and not processed by each power prediction network meet the target value, each power prediction network sends the first optimization data that have been obtained and not processed to other power prediction networks, so that each power prediction network determines all the optimization data of the second unit based on the first optimization data that have been obtained and not processed by each power prediction network. Or, on the premise that each power prediction network has obtained the first optimization data of each unit in the corresponding partial data correction network, each power prediction network sends the first optimization data that have been obtained and not processed to other power prediction networks, so that each power prediction network determines all the optimization data of the second unit based on the first optimization data that have been obtained and not processed by each power prediction network. Wherein, the second unit refers to at least one unit corresponding to the first optimization data that has been obtained and not processed.

[0070] Under the above premise, please refer to Figure 2 , provides a photovoltaic ultra-short-term power prediction device 200, which is applied to a photovoltaic ultra-short-term power prediction system, and the device comprises: The information acquisition module 210 is used to acquire a plurality of target ultra-short term power event transmission data and a plurality of target energy consumption loss data, wherein the plurality of target energy consumption loss data are different; the target ultra-short term power event transmission data corresponding to the same target energy consumption loss data constitute a target power transmission feature chain, wherein the target energy consumption loss data corresponding to the target ultra-short term power event transmission data is: covering the time resolution feature specified by the target ultra-short term power event transmission data, and having a time range specified by the target ultra-short term power event transmission data and the target energy consumption loss data in the plurality of target energy consumption loss data; An information loading module 220 is used to input the target ultra-short-term power event transmission data in each target power transmission characteristic chain that is smaller than the target energy consumption loss data corresponding to the target power transmission characteristic chain into the target energy consumption loss data corresponding to the target power transmission characteristic chain; The data correction network 230 is used to simultaneously execute an optimized data correction network based on a predicted target power transmission characteristic chain through each of the multiple power prediction networks, and the data correction network is used to perform prediction processing on random ultra-short-term power data.

[0071] Under the above premise, please refer to Figure 3, shows a photovoltaic ultra-short-term power prediction system 300, including a processor 310 and a memory 320 that communicate with each other, and the processor 310 is used to read and execute a computer program from the memory 320 to implement the above method.

[0072] Under the above premise, a computer-readable storage medium is also provided, on which a computer program stored implements the above method when running.

[0073] In summary, based on the above scheme, among multiple target energy consumption loss data, determine the target ultra-short term power event transmission data that meets the time range specified by the target ultra-short term power event transmission data and is similar to the target energy consumption loss data within the time range specified by the target ultra-short term power event transmission data, and input the target ultra-short term power event transmission data into the target energy consumption loss data. Since the target energy consumption loss data is the most similar to the target energy consumption loss data of the target ultra-short term power event transmission data, the optimization workload of the target ultra-short term power event transmission data is reduced. Then, through multiple power prediction networks, based on multiple different target power transmission feature chains, the optimized data correction network is simultaneously executed, thereby ensuring that the accuracy and confidence of the prediction of ultra-short term power data are improved while the optimized data correction network is executed simultaneously.

[0074] It should be understood that the system and its modules shown above can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or a dedicated design hardware. Those skilled in the art will understand that the above methods and systems can be implemented using computer executable instructions and / or included in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The system and its modules of the present application can not only be implemented by hardware circuits such as ultra-large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., but can also be implemented by software such as executed by various types of processors, and can also be implemented by a combination of the above hardware circuits and software (e.g., firmware).

[0075] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects that may be produced may be any one or a combination of the above, or may be any other beneficial effects that may be obtained.

[0076] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only for example and does not constitute a limitation of the present application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements and amendments to the present application. Such modifications, improvements and amendments are suggested in the present application, so such modifications, improvements and amendments still belong to the spirit and time scope of the exemplary embodiments of the present application.

[0077] At the same time, the present application uses specific words to describe the embodiments of the present application. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of the present application. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. In addition, some features, structures or characteristics in one or more embodiments of the present application can be appropriately combined.

[0078] In addition, it will be appreciated by those skilled in the art that various aspects of the present application may be illustrated and described by a number of patentable categories or situations, including any new and useful process, machine, product or combination of substances, or any new and useful improvements thereto. Accordingly, each aspect of the present application may be performed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above hardware or software may all be referred to as "data blocks", "modules", "engines", "units", "components" or "systems". In addition, various aspects of the present application may be represented as a computer product located in one or more computer-readable media, which includes computer-readable program code.

[0079] A computer storage medium may include a propagated data signal containing computer program code, for example, in baseband or as part of a carrier wave. The propagated signal may be in a variety of forms, including electromagnetic, optical, etc., or a suitable combination. A computer storage medium may be any computer-readable medium other than a computer-readable storage medium, which can be connected to an instruction execution system, device or apparatus to communicate, propagate or transmit the program for use. The program code on the computer storage medium may be transmitted via any suitable medium, including radio, cable, fiber optic cable, RF, or similar media, or any combination of the above media.

[0080] The computer program codes required for the operation of each part of the present application can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, etc., conventional procedural programming languages ​​such as C language, Visual Basic, Fortran 2003, Perl, COBOL 2002, PHP, ABAP, dynamic programming languages ​​such as Python, Ruby and Groovy, or other programming languages, etc. The program code can be run entirely on the user's computer, or run on the user's computer as an independent software package, or run partially on the user's computer and partially on a remote computer, or run entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any network form, such as a local area network (LAN) or a wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as software as a service (SaaS).

[0081] In addition, unless explicitly stated in the claims, the order of the processing elements and sequences described in this application, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of this application. Although the above disclosure discusses some invention embodiments that are currently considered useful through various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the essence and time scope of the embodiments of this application. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing server or mobile device.

[0082] Similarly, it should be noted that in order to simplify the description of the disclosure of this application and thus help understand one or more embodiments of the invention, in the above description of the embodiments of this application, multiple features are sometimes combined into one embodiment, figure or description thereof. However, this disclosure method does not mean that the features required by the object of this application are more than the features mentioned in the claims. In fact, the features of the embodiments are less than all the features of the single embodiment disclosed above.

[0083] In some embodiments, numbers describing the number of components and attributes are used. It should be understood that such numbers used for the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise specified, "about", "approximately" or "substantially" indicate that the numbers allow adaptive changes. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which can be changed according to the required characteristics of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining the digits. Although the numerical domains and parameters used to confirm the breadth of the time range in some embodiments of the present application are approximate values, in specific embodiments, the setting of such numerical values ​​is as accurate as possible within the feasible time range.

[0084] Each patent, patent application, patent application disclosure, and other materials, such as articles, books, instructions, publications, documents, etc., cited in this application are hereby incorporated by reference in their entirety. Except for application history documents that are inconsistent with or conflicting with the content of this application, documents that limit the broadest time range of the claims of this application (currently or later attached to this application) are also excluded. It should be noted that if the descriptions, definitions, and / or use of terms in the attached materials of this application are inconsistent with or conflicting with the content described in this application, the descriptions, definitions, and / or use of terms in this application shall prevail.

[0085] Finally, it should be understood that the embodiments described in this application are only used to illustrate the principles of the embodiments of the present application. Other variations may also fall within the time range of this application. Therefore, as an example and not a limitation, the alternative configurations of the embodiments of the present application may be considered to be the same as the teachings of this application. Accordingly, the embodiments of the present application are not limited to the embodiments explicitly introduced and described in this application.

[0086] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A photovoltaic ultra-short-term power prediction method, characterized in that: The method comprises: Acquire multiple target ultra-short term power event transmission data and multiple target energy consumption loss data, wherein the multiple target energy consumption loss data are different; form a target power transmission feature chain with the target ultra-short term power event transmission data corresponding to the same target energy consumption loss data, wherein the target energy consumption loss data corresponding to the target ultra-short term power event transmission data is: covers the time resolution feature specified by the target ultra-short term power event transmission data, and is similar to the target energy consumption loss data in the time range specified by the target ultra-short term power event transmission data among the multiple target energy consumption loss data; Inputting the target ultra-short-term power event transmission data in each target power transmission characteristic chain that is smaller than the target energy consumption loss data corresponding to the target power transmission characteristic chain into the target energy consumption loss data corresponding to the target power transmission characteristic chain; Through each of the multiple power prediction networks, an optimized data correction network is simultaneously executed based on the predicted target power transmission characteristic chain, and the data correction network is used to predict and process random ultra-short-term power data.

2. The method according to claim 1, characterized in that The data correction network includes a partial data correction network corresponding to each of the power prediction networks, and the optimization of the data correction network is simultaneously performed based on the predicted target power transmission characteristic chain through each of the multiple power prediction networks, including: For each of the power prediction networks, multiple target ultra-short-term power event transmission data in the predicted target power transmission feature chain are spliced ​​into an ultra-short-term power influencing factor catalog, the number of first features of the ultra-short-term power influencing factor catalog is the same as the number of the multiple target ultra-short-term power event transmission data, the number of second features of the ultra-short-term power influencing factor catalog is equal to the target energy consumption loss data corresponding to the target power transmission feature chain, or the first feature number is equal to the target energy consumption loss data corresponding to the target power transmission feature chain, and the second feature number is equal to the number of the multiple target ultra-short-term power event transmission data; In combination with the ultra-short-term power influencing factor catalog, a partial data correction network corresponding to the power prediction network is optimized.

3. The method according to claim 2, characterized in that After optimizing the partial data correction network corresponding to the power prediction network in combination with the ultra-short-term power influencing factor catalog, the method further includes: Correcting the important distribution of multiple target ultra-short term power event transmission data in the ultra-short term power influencing factor catalog to obtain a corrected ultra-short term power influencing factor catalog; Combined with the revised ultra-short-term power influencing factor catalog, continue to optimize the partial data correction network corresponding to the power prediction network.

4. The method according to claim 1, characterized in that: The data correction network includes a partial data correction network corresponding to each of the power prediction networks, and the optimization of the data correction network is simultaneously performed based on the predicted target power transmission characteristic chain through each of the multiple power prediction networks, including: For each of the power prediction networks, optimizing the original important power data text on at least one time resolution feature distribution of multiple target ultra-short-term power event transmission data in the predicted target power transmission feature chain into target important power data text; Through a partial data correction network corresponding to the power prediction network, parsing is performed based on the optimized transmission data of multiple target ultra-short-term power events to obtain parsed important power data text on the at least one time resolution characteristic distribution; The original important power data text and the parsed important power data text on the at least one time resolution feature distribution are combined to optimize a partial data correction network corresponding to the power prediction network.

5. The method according to claim 4, characterized in that The method of performing parsing based on the optimized transmission data of multiple target ultra-short-term power events through a partial data correction network corresponding to the power prediction network to obtain the parsed important power data text on the at least one time resolution characteristic distribution includes: The optimized multiple target ultra-short-term power event transmission data are classified and processed through a partial data correction network corresponding to the power prediction network to obtain an important power data text description corresponding to each important power data text on the time resolution feature distribution in the optimized multiple target ultra-short-term power event transmission data; Obtaining, from the obtained multiple important power data text descriptions, an important power data text description corresponding to the target important power data text on the at least one time resolution feature distribution; The important power data text description corresponding to the target important power data text on the at least one time resolution characteristic distribution is parsed to obtain the parsed important power data text on the at least one time resolution characteristic distribution.

6. The method according to claim 5, characterized in that The data correction network includes a partial data correction network corresponding to each of the power prediction networks, and the optimization of the data correction network is simultaneously performed based on the predicted target power transmission characteristic chain through each of the multiple power prediction networks, including: By means of each of the power prediction networks, based on the predicted target power transmission characteristic chain, the corresponding partial data correction network is optimized; On the premise that each of the power prediction networks obtains the first optimization data of the respective corresponding partial data correction network, all optimization data are determined based on the first optimization data obtained by each of the power prediction networks; Through each of the power prediction networks, in combination with all the optimization data, the corresponding partial data correction networks are optimized.

7. The method according to claim 6, characterized in that The partial data correction network includes a plurality of units, and the determination of all optimization data based on the first optimization data obtained by each power prediction network, on the premise that each power prediction network obtains the first optimization data of the partial data correction network corresponding to the power prediction network, includes: Through each of the power prediction networks, first optimization data of the first unit in the respective corresponding partial data correction network is obtained one by one; On the premise that each vector of the first optimization data acquired and not processed by the power prediction network meets the target value, determining all the optimization data of the first unit based on each first optimization data acquired and not processed by the power prediction network; Through each of the power prediction networks, continue to obtain the first optimization data of the second unit in the respective corresponding partial data correction network one by one; On the premise that each vector of the first optimization data acquired and not processed by the power prediction network meets the target value, or on the premise that each power prediction network has acquired the first optimization data of each unit in its corresponding partial data correction network, all the optimization data of the second unit are determined based on the first optimization data acquired and not processed by each power prediction network.

8. The method according to claim 6, characterized in that The method also includes: on the premise that the vector of the first optimization data obtained by the power prediction network does not belong to the weight coefficient of the target vector, optimizing the first optimization data through the power prediction network so that the vector of the first optimization data belongs to the weight coefficient of the target vector.

9. A photovoltaic ultra-short-term power prediction system, characterized in that: The invention comprises a processor and a memory communicating with each other, wherein the processor is used to read a computer program from the memory and execute the computer program to implement the method according to any one of claims 1 to 8.

Images