Method and device for acquiring photovoltaic output data in photovoltaic power station

Abstract

The invention discloses a method and a device for acquiring photovoltaic output data in a photovoltaic power station. The method comprises the steps of obtaining photovoltaic output data of a plurality of photovoltaic power stations in a transformer area where a target power system is located; preprocessing the photovoltaic output data to filter out abnormal state data and fault state data in the photovoltaic output data to obtain preprocessed photovoltaic output data; inputting the pre-processed photovoltaic output data into a photovoltaic output ultra-short-term prediction model, and processing the pre-processed photovoltaic output data by using the photovoltaic output ultra-short-term prediction model to obtain a processing result; and obtaining photovoltaic output estimated values of the plurality of photovoltaic power stations in the future predetermined duration according to the processing result. According to the method and the device, the technical problem that the estimation value has a relatively large error due to the fact that the evaluation of the output data of the photovoltaic power station in the future duration is easily influenced by measurement errors and bad data in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of distribution network operation perception technology, and in particular to a method and device for acquiring photovoltaic output data in a photovoltaic power station. Background Art

[0002] Accurately assessing the potential risks in the operation of distribution networks and providing reference for operation and dispatching personnel so that they can take preventive measures and indirectly improve power supply safety has become an urgent problem to be solved.

[0003] Traditional distribution network risk assessment focuses on establishing a risk assessment and early warning system from a macro perspective or conducting distribution network risk assessment based on probability statistics. Although it can reflect the overall risk level of the distribution network, the early warning results obtained cannot provide a basis for operators to discover and identify the causes and sources of risks, and have certain limitations on the formulation of risk prevention and control measures.

[0004] Moreover, due to the single information source used in the current distribution network operation status assessment, the loose connection between data, the loose data collection, the lack of systematicity and structure, the established operation status assessment indicator system is incomplete and its accuracy still needs further discussion. In order to realize the perception of the distribution network operation status, real-time measurement is required as data support. Compared with the long-term distribution network operation status perception, the distribution network operation status perception is more susceptible to measurement errors and bad data, and is prone to large evaluation errors due to data quality issues.

[0005] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0006] The embodiments of the present invention provide a method and device for obtaining photovoltaic output data in a photovoltaic power station, so as to at least solve the technical problem in the related art that when evaluating the output data of the photovoltaic power station in the future, it is easily affected by measurement errors and bad data, resulting in large errors in the estimated value.

[0007] According to one aspect of an embodiment of the present invention, a method for acquiring photovoltaic output data in a photovoltaic power station is provided, comprising: acquiring photovoltaic output data of multiple photovoltaic power stations in a substation where a target power system is located; preprocessing the photovoltaic output data to filter out abnormal state data and fault state data in the photovoltaic output data to obtain the preprocessed photovoltaic output data, wherein the abnormal state data is data in the photovoltaic output data that exceeds the maximum photovoltaic output power, and the fault state data is data in the photovoltaic output data that the photovoltaic output power is 0 when the light intensity exceeds a light intensity threshold; filtering out the preprocessed photovoltaic output data to filter out abnormal state data and fault state data The photovoltaic output data is input into an ultra-short-term photovoltaic output prediction model, so as to use the ultra-short-term photovoltaic output prediction model to process the pre-processed photovoltaic output data to obtain a processing result, wherein the ultra-short-term photovoltaic output prediction model is a model obtained by machine learning training using multiple groups of sample data, each of the multiple groups of sample data includes: sample photovoltaic output data and sample processing results corresponding to the sample photovoltaic output data, and the multiple groups of sample data include sample photovoltaic output data collected under different weather conditions; according to the processing results, the estimated photovoltaic output values ​​of the multiple photovoltaic power stations within a predetermined period of time in the future are obtained.

[0008] Optionally, the photovoltaic output data is preprocessed to filter out abnormal state data and fault state data in the photovoltaic output data to obtain the preprocessed photovoltaic output data, including: determining the abnormal state data in the photovoltaic output data according to the photovoltaic maximum output power, and determining the fault state data in the photovoltaic output data; performing data cleaning on the photovoltaic output data to clean out the abnormal state data and the fault state data in the photovoltaic output data; and normalizing the photovoltaic output data after filtering out the abnormal state data and the fault state data to obtain the preprocessed photovoltaic output data.

[0009] Optionally, normalizing the photovoltaic output data after filtering out the abnormal state data and the fault state data to obtain the preprocessed photovoltaic output data, comprising: normalizing the photovoltaic output data after filtering out the abnormal state data and the fault state data by a first formula to obtain the preprocessed photovoltaic output data, wherein the first formula is: , represents the photovoltaic output data after preprocessing, represents the photovoltaic output data after filtering out the abnormal state data and the fault state data, represents the minimum value of the photovoltaic output data after filtering out the abnormal state data and the fault state data, It represents the maximum value of the photovoltaic output data after filtering out the abnormal state data and the fault state data.

[0010] Optionally, the preprocessed photovoltaic output data is input into a photovoltaic output ultra-short-term prediction model, so as to use the photovoltaic output ultra-short-term prediction model to process the preprocessed photovoltaic output data to obtain a processing result, including: using a dynamic exponential smoothing method to determine the weight coefficient of each output data in the photovoltaic output data in the photovoltaic output ultra-short-term prediction model; weighted processing is performed on the photovoltaic output data according to the weight coefficient to obtain a weighted processing result; using an information fusion multi-order convolution channel in the photovoltaic output ultra-short-term prediction model to extract the spatiotemporal coupling relationship of the multiple photovoltaic power stations, wherein the information fusion multi-order convolution channel is a convolution channel obtained by splicing a first-order convolution channel, and the first-order convolution channel is a channel created for each factor affecting the output of the multiple photovoltaic power stations; obtaining a fusion characteristic value of the photovoltaic output data according to the spatiotemporal coupling relationship; and obtaining the processing result according to the fusion characteristic value.

[0011] Optionally, obtaining the fused characteristic value of the photovoltaic output data according to the time-space coupling relationship includes: obtaining the fused characteristic value according to the time-space coupling relationship by a second formula, wherein the second formula is: , represents the fused feature value, represents the weight coefficient of the first-order convolution channel in the process of fusion feature splicing, represents the photovoltaic output data, represents the length of the feature vector, Indicates the channel number of the photovoltaic output data, The number of channels representing the output feature, the spatiotemporal coupling relationship is represented by the length of the feature vector and the number of channels.

[0012] Optionally, before inputting the preprocessed photovoltaic output data into the photovoltaic output ultra-short-term prediction model to process the preprocessed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain the processing result, the photovoltaic output data acquisition method in the photovoltaic power station also includes: training the multiple groups of sample data to obtain the photovoltaic output ultra-short-term prediction model, wherein the training of the multiple groups of sample data to obtain the photovoltaic output ultra-short-term prediction model includes: determining the initial parameters of the basic neural network model corresponding to the photovoltaic output ultra-short-term prediction model through a particle swarm optimization algorithm; assigning the initial parameters to the basic neural network model to obtain an initial photovoltaic output ultra-short-term prediction model; verifying the initial photovoltaic output ultra-short-term prediction model using the multiple groups of sample data to obtain the prediction accuracy of the initial photovoltaic output ultra-short-term prediction model; and obtaining the photovoltaic output ultra-short-term prediction model when the prediction accuracy is greater than a prediction accuracy threshold.

[0013] According to another aspect of an embodiment of the present invention, a photovoltaic output data acquisition device in a photovoltaic power station is also provided, including: a first acquisition unit, used to acquire photovoltaic output data of multiple photovoltaic power stations in the substation where a target power system is located; a preprocessing unit, used to preprocess the photovoltaic output data to filter out abnormal state data and fault state data in the photovoltaic output data, and obtain the preprocessed photovoltaic output data, wherein the abnormal state data is data in the photovoltaic output data that exceeds the maximum photovoltaic output power, and the fault state data is data in the photovoltaic output data that the photovoltaic output power is 0 when the light intensity exceeds the light intensity threshold; a processing unit, used to The preprocessed photovoltaic output data is input into a photovoltaic output ultra-short-term prediction model, so as to use the photovoltaic output ultra-short-term prediction model to process the preprocessed photovoltaic output data to obtain a processing result, wherein the photovoltaic output ultra-short-term prediction model is a model obtained by machine learning training using multiple groups of sample data, each of the multiple groups of sample data includes: sample photovoltaic output data and sample processing results corresponding to the sample photovoltaic output data, and the multiple groups of sample data include sample photovoltaic output data collected under different weather conditions; a second acquisition unit is used to obtain the estimated photovoltaic output values ​​of the multiple photovoltaic power stations within a predetermined time period in the future according to the processing results.

[0014] Optionally, the preprocessing unit includes: a first determination subunit, used to determine the abnormal state data in the photovoltaic output data according to the photovoltaic maximum output power, and determine the fault state data in the photovoltaic output data; a cleaning subunit, used to perform data cleaning on the photovoltaic output data to remove the abnormal state data and the fault state data in the photovoltaic output data; a normalization processing subunit, used to normalize the photovoltaic output data after filtering out the abnormal state data and the fault state data to obtain the preprocessed photovoltaic output data.

[0015] Optionally, the normalization processing subunit includes: a normalization processing module, which is used to normalize the photovoltaic output data after filtering out the abnormal state data and the fault state data through a first formula to obtain the pre-processed photovoltaic output data, wherein the first formula is: , represents the photovoltaic output data after preprocessing, represents the photovoltaic output data after filtering out the abnormal state data and the fault state data, represents the minimum value of the photovoltaic output data after filtering out the abnormal state data and the fault state data, It represents the maximum value of the photovoltaic output data after filtering out the abnormal state data and the fault state data.

[0016] Optionally, the processing unit includes: a second determination subunit, used to determine the weight coefficient of each output data in the photovoltaic output data in the photovoltaic output ultra-short-term prediction model by using a dynamic exponential smoothing method; a weighted processing subunit, used to perform weighted processing on the photovoltaic output data according to the weight coefficient to obtain a weighted processing result; an extraction subunit, used to extract the spatiotemporal coupling relationship of the multiple photovoltaic power stations using the information fusion multi-order convolution channel in the photovoltaic output ultra-short-term prediction model, wherein the information fusion multi-order convolution channel is a convolution channel obtained by splicing a first-order convolution channel, and the first-order convolution channel is a channel created for each factor affecting the output of the multiple photovoltaic power stations; a first acquisition subunit, used to obtain the fusion characteristic value of the photovoltaic output data according to the spatiotemporal coupling relationship; and a second acquisition subunit, used to obtain the processing result according to the fusion characteristic value.

[0017] Optionally, the first acquisition subunit includes: an acquisition module, configured to obtain the fusion feature value by a second formula according to the spatiotemporal coupling relationship, wherein the second formula is: , represents the fused feature value, represents the weight coefficient of the first-order convolution channel in the process of fusion feature splicing, represents the photovoltaic output data, represents the length of the feature vector, Indicates the channel number of the photovoltaic output data, The number of channels representing the output feature, the spatiotemporal coupling relationship is represented by the length of the feature vector and the number of channels.

[0018] Optionally, the photovoltaic output data acquisition device in the photovoltaic power station also includes: a training unit, which is used to input the pre-processed photovoltaic output data into the photovoltaic output ultra-short-term prediction model, so as to use the photovoltaic output ultra-short-term prediction model to process the pre-processed photovoltaic output data, and before obtaining the processing results, train the multiple groups of sample data to obtain the photovoltaic output ultra-short-term prediction model, wherein the training unit includes: a third determination subunit, which is used to determine the initial parameters of the basic neural network model corresponding to the photovoltaic output ultra-short-term prediction model through a particle swarm optimization algorithm; a third acquisition subunit, which is used to assign the initial parameters to the basic neural network model to obtain an initial photovoltaic output ultra-short-term prediction model; a verification subunit, which is used to verify the initial photovoltaic output ultra-short-term prediction model using the multiple groups of sample data to obtain the prediction accuracy of the initial photovoltaic output ultra-short-term prediction model; and a fourth acquisition subunit, which is used to obtain the photovoltaic output ultra-short-term prediction model when the prediction accuracy is greater than a prediction accuracy threshold.

[0019] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, the computer-readable storage medium comprising a stored program, wherein the program executes any one of the above-mentioned methods for acquiring photovoltaic output data in a photovoltaic power station.

[0020] According to another aspect of an embodiment of the present invention, a processor is further provided, the processor being used to run a program, wherein when the program is run, the method for acquiring photovoltaic output data in a photovoltaic power station described in any one of the above is executed.

[0021] According to another aspect of an embodiment of the present invention, a computer program product is provided, comprising computer instructions, which, when executed by a processor, execute any one of the above methods for acquiring photovoltaic output data in a photovoltaic power station.

[0022] In an embodiment of the present invention, photovoltaic output data of multiple photovoltaic power stations in the substation where the target power system is located are obtained; the photovoltaic output data are preprocessed to filter out abnormal state data and fault state data in the photovoltaic output data to obtain preprocessed photovoltaic output data, wherein the abnormal state data is data in the photovoltaic output data that exceeds the maximum photovoltaic output power, and the fault state data is data in the photovoltaic output data where the photovoltaic output power is 0 when the light intensity exceeds the light intensity threshold; the preprocessed photovoltaic output data is input into a photovoltaic output ultra-short-term prediction model to process the preprocessed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain a processing result, wherein the photovoltaic output ultra-short-term prediction model is a model obtained by machine learning training using multiple groups of sample data, each of the multiple groups of sample data includes: sample photovoltaic output data and sample processing results corresponding to the sample photovoltaic output data, and the multiple groups of sample data include sample photovoltaic output data collected under different weather conditions; and the photovoltaic output estimated values ​​of multiple photovoltaic power stations within a predetermined time period in the future are obtained according to the processing results. Through the above-mentioned technical scheme provided by the present invention, the purpose of predicting the photovoltaic output value in a future period of time based on the historical photovoltaic output data through a pre-trained photovoltaic output ultra-short-term prediction model is achieved, and the technical effect of improving the estimation accuracy of the output value of the photovoltaic power station in the future period of time is achieved, thereby solving the technical problem in the related technology that when evaluating the output data of the photovoltaic power station in the future period of time, it is easy to be affected by measurement errors and bad data, resulting in large errors in the estimated value. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0024] Figure 1 It is a hardware structure block diagram of a mobile terminal of a method for acquiring photovoltaic output data in a photovoltaic power station according to an embodiment of the present invention;

[0025] Figure 2 is a flow chart of a method for acquiring photovoltaic output data in a photovoltaic power station according to an embodiment of the present invention;

[0026] Figure 3 is a comparison diagram of prediction results of a method for acquiring photovoltaic output data in a photovoltaic power station according to an embodiment of the present invention under sunny conditions;

[0027] Figure 4 is a comparison diagram of prediction results under cloudy conditions according to a method for acquiring photovoltaic output data in a photovoltaic power station according to an embodiment of the present invention;

[0028] Figure 5is a comparison diagram of prediction results under cloudy conditions according to a method for acquiring photovoltaic output data in a photovoltaic power station according to an embodiment of the present invention;

[0029] Figure 6 is a schematic diagram of a photovoltaic output data acquisition device in a photovoltaic power station according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] As introduced in the background technology, in the related art, when evaluating the output data of a photovoltaic power station in the future, it is easily affected by measurement errors and bad data, resulting in large errors in the estimated value. In view of the above defects, in the embodiments of the present invention, a method and device for obtaining photovoltaic output data in a photovoltaic power station, a computer-readable storage medium, a processor and a computer program product are provided.

[0033] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0034] The method embodiments provided in the embodiments of the present invention can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 1 is a hardware structure block diagram of a mobile terminal for obtaining photovoltaic output data in a photovoltaic power station according to an embodiment of the present invention. Figure 1 As shown, the mobile terminal may include one or more ( Figure 1Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the mobile terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0035] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the photovoltaic output data acquisition method in the photovoltaic power station in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, the above method is implemented. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. The transmission device 106 is used to receive or send data via a network. The above-mentioned specific examples of the network may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0036] According to an embodiment of the present invention, a method embodiment of a method for obtaining photovoltaic output data in a photovoltaic power station is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0037] Figure 2 is a flow chart of a method for acquiring photovoltaic output data in a photovoltaic power station according to an embodiment of the present invention. Figure 2 As shown, the method for obtaining photovoltaic output data in the photovoltaic power station includes the following steps:

[0038] Step S202, obtaining photovoltaic output data of multiple photovoltaic power stations in the area where the target power system is located.

[0039] Optionally, the photovoltaic output data may include but are not limited to: time, light, photovoltaic power generation, meteorological environment, sampling information and geographical location.

[0040] Step S204, preprocessing the photovoltaic output data to filter out abnormal state data and fault state data in the photovoltaic output data to obtain preprocessed photovoltaic output data, wherein the abnormal state data is data in the photovoltaic output data that exceeds the maximum photovoltaic output power, and the fault state data is data in the photovoltaic output data that the photovoltaic output power is 0 when the light intensity exceeds the light intensity threshold.

[0041] Here, the maximum photovoltaic output is an important parameter of photovoltaic power generation. When the photovoltaic output reaches the maximum output power, that is, the light condition reaches the threshold, the continued increase in light intensity will not contribute to the photovoltaic output. Exceeding the maximum photovoltaic output power (maximum photovoltaic output power) is defined as an abnormal state, and when the light is sufficient (the light intensity exceeds the light intensity threshold) the photovoltaic output power is 0 as a fault state. The data of abnormal and fault states are cleaned and filtered out.

[0042] That is, the above scheme proposed in this embodiment can be used to clean abnormal data and normalize photovoltaic output data, including filtering out abnormal state and fault state data and performing normalization processing.

[0043] Step S206, input the preprocessed photovoltaic output data into the photovoltaic output ultra-short-term prediction model, so as to process the preprocessed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain a processing result, wherein the photovoltaic output ultra-short-term prediction model is a model obtained by machine learning training using multiple groups of sample data, each of the multiple groups of sample data includes: sample photovoltaic output data and sample processing results corresponding to the sample photovoltaic output data, and the multiple groups of sample data include sample photovoltaic output data collected under different weather conditions.

[0044] In this embodiment, the photovoltaic output data may be processed using a pre-trained photovoltaic output ultra-short-term prediction model to obtain photovoltaic output estimates of the above-mentioned multiple photovoltaic power stations within a predetermined future time period.

[0045] Step S208, obtaining estimated photovoltaic output values ​​of the plurality of photovoltaic power stations within a predetermined future time period according to the processing results.

[0046] As can be seen from the above, in the embodiment of the present invention, the photovoltaic output data of multiple photovoltaic power stations in the substation where the target power system is located can be obtained; the photovoltaic output data is preprocessed to filter out abnormal state data and fault state data in the photovoltaic output data to obtain preprocessed photovoltaic output data, wherein the abnormal state data is data in the photovoltaic output data that exceeds the maximum photovoltaic output power, and the fault state data is data in the photovoltaic output data where the photovoltaic output power is 0 when the light intensity exceeds the light intensity threshold; the preprocessed photovoltaic output data is input into the photovoltaic output ultra-short-term prediction model to process the preprocessed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain To the processing results, wherein the photovoltaic output ultra-short-term prediction model is a model obtained by machine learning training using multiple groups of sample data, each of the multiple groups of sample data includes: sample photovoltaic output data and sample processing results corresponding to the sample photovoltaic output data, and the multiple groups of sample data include sample photovoltaic output data collected under different weather conditions; according to the processing results, the estimated photovoltaic output values ​​of multiple photovoltaic power stations within a predetermined period of time in the future are obtained, thereby achieving the purpose of predicting the photovoltaic output value within a future period of time based on the historical photovoltaic output data through the pre-trained photovoltaic output ultra-short-term prediction model, and achieving the technical effect of improving the accuracy of the estimated output value of the photovoltaic power station within a future period of time.

[0047] Therefore, the above technical solution provided by the embodiment of the present invention solves the technical problem in the related art that when evaluating the output data of a photovoltaic power station in the future, it is easily affected by measurement errors and bad data, resulting in large errors in the estimated value.

[0048] According to the above-mentioned embodiment of the present invention, photovoltaic output data is preprocessed to filter out abnormal state data and fault state data in the photovoltaic output data to obtain preprocessed photovoltaic output data, including: determining abnormal state data in the photovoltaic output data according to the maximum photovoltaic output power, and determining fault state data in the photovoltaic output data; performing data cleaning on the photovoltaic output data to clean out abnormal state data and fault state data in the photovoltaic output data; and normalizing the photovoltaic output data after filtering out the abnormal state data and fault state data to obtain preprocessed photovoltaic output data.

[0049] In this embodiment, the abnormal state data and fault state data in the photovoltaic output data can be determined first, and then the abnormal state data and fault state data are cleaned to remove the abnormal state data and fault state data in the photovoltaic output data, and then the remaining photovoltaic output data is normalized.

[0050] According to the above embodiment of the present invention, the photovoltaic output data after filtering out the abnormal state data and the fault state data is normalized to obtain the pre-processed photovoltaic output data, including: normalizing the photovoltaic output data after filtering out the abnormal state data and the fault state data by a first formula to obtain the pre-processed photovoltaic output data, wherein the first formula is: , Represents the pre-processed photovoltaic output data, Indicates the photovoltaic output data after filtering out abnormal status data and fault status data. It indicates the minimum value of the photovoltaic output data after filtering out abnormal status data and fault status data. Indicates the maximum value of the PV output data after filtering out abnormal status data and fault status data.

[0051] In this embodiment, the above first formula can be used to normalize the photovoltaic output data remaining after filtering out the abnormal state data and the fault state data.

[0052] According to the above embodiment of the present invention, the preprocessed photovoltaic output data is input into the photovoltaic output ultra-short-term prediction model, so as to process the preprocessed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain a processing result, including: using a dynamic exponential smoothing method to determine the weight coefficient of each output data in the photovoltaic output data in the photovoltaic output ultra-short-term prediction model; weighted processing is performed on the photovoltaic output data according to the weight coefficient to obtain a weighted processing result; the information fusion multi-order convolution channel in the photovoltaic output ultra-short-term prediction model is used to extract the spatiotemporal coupling relationship of multiple photovoltaic power stations, wherein the information fusion multi-order convolution channel is a convolution channel obtained by splicing a first-order convolution channel, and the first-order convolution channel is a channel created for each factor affecting the output of multiple photovoltaic power stations; according to the spatiotemporal coupling relationship, a fusion characteristic value of the photovoltaic output data is obtained; and according to the fusion characteristic value, a processing result is obtained.

[0053] In this embodiment, the weight parameters can be adjusted dynamically, the photovoltaic output is highly coupled with temperature, sunshine and other data, and the dynamic exponential smoothing method is used to dynamically assign the weight coefficient ω in the optimization training.

[0054] Specifically, the dynamic exponential smoothing method here can strengthen the effect of recent observations on the predicted values. During the training process, by dynamically assigning weights to observations at different times, the iteration factor is dynamically adjusted during the search, and the dynamic nature of the weight assignment is strengthened, which is conducive to improving the algorithm's early global and later local search performance and ensuring the objectivity of the prediction results. Its calculation method is shown in the following formula: ,in, is the smoothed value at time step i, is the actual data at this time step, t is the time, It is a dynamic quantity of the initial value, and its value varies with the predicted target Dynamic changes.

[0055] That is, in an embodiment of the present invention, an initial weight coefficient ω value can be set, which can be set according to experience or actual conditions; then, according to the formula of the above-mentioned dynamic exponential smoothing method, the smoothing value at the current time step i is calculated; and then the error between the predicted value and the actual value at the current time step is calculated. The mean square error (MSE) can be used. The weight coefficient ω is adjusted according to the error situation; if the error is large, it means that the current weight coefficient is not appropriate enough, and the weight of the latest data needs to be increased; ω can be increased to make the smoothing value closer to the actual observation value. If the error is small, the current weight coefficient can be maintained, or slightly adjusted for fine-tuning. Finally, the above steps can be repeated until the stop condition is reached (for example, the maximum number of iterations is reached or the error is small enough).

[0056] In addition, in the embodiment of the present invention, a multi-order convolution channel can be constructed through feature splicing and merging operations to extract fusion feature values ​​of multiple photovoltaic power stations in a region, which is described in detail below.

[0057] According to the above embodiment of the present invention, obtaining the fusion characteristic value of photovoltaic output data according to the time-space coupling relationship includes: obtaining the fusion characteristic value according to the time-space coupling relationship by a second formula, wherein the second formula is: , represents the fusion feature value, Represents the weighted coefficient of the first-order convolution channel in the process of fusion feature splicing, Indicates photovoltaic output data, represents the length of the feature vector, Indicates the number of channels for photovoltaic output data. It represents the number of channels of the output features. The spatiotemporal coupling relationship is represented by the length of the feature vector and the number of channels.

[0058] In this embodiment, a first-order convolution channel is independently constructed for the data of each factor affecting the photovoltaic output and used as the basic input channel. At this time, the first-order channel actually characterizes the degree of influence of a certain data on the photovoltaic output, that is, the weighted coefficient of a certain data in the fusion feature splicing process; then the first-order convolution channel that has been constructed is selected, and feature splicing and merging operations are performed to establish a cross-channel information fusion multi-order convolution channel, and the spatiotemporal coupling relationship of multiple photovoltaic power stations in the region is extracted to obtain the fusion characteristic value of the photovoltaic output. The specific calculation formula is the above-mentioned second formula.

[0059] According to the above embodiment of the present invention, before the pre-processed photovoltaic output data is input into the photovoltaic output ultra-short-term prediction model, and the pre-processed photovoltaic output data is processed by the photovoltaic output ultra-short-term prediction model to obtain the processing result, the photovoltaic output data acquisition method in the photovoltaic power station also includes: training multiple groups of sample data to obtain the photovoltaic output ultra-short-term prediction model, wherein the training of multiple groups of sample data to obtain the photovoltaic output ultra-short-term prediction model includes: determining the initial parameters of the basic neural network model corresponding to the photovoltaic output ultra-short-term prediction model through a particle swarm optimization algorithm; assigning the initial parameters to the basic neural network model to obtain the initial photovoltaic output ultra-short-term prediction model; verifying the initial photovoltaic output ultra-short-term prediction model using multiple groups of sample data to obtain the prediction accuracy of the initial photovoltaic output ultra-short-term prediction model; when the prediction accuracy is greater than the prediction accuracy threshold, the photovoltaic output ultra-short-term prediction model is obtained.

[0060] In this embodiment, the particle swarm optimization BP neural network algorithm can be used to solve the optimal particle, and the specific process is as follows: 1) Initialize the particle swarm parameters and generate the initial particle swarm, each particle consists of a group of multi-dimensional real vectors; 2) Input training samples for training; 3) Calculate the fitness value of each particle to obtain Pi and Pg; 4) Update each particle according to the calculation formula of the particle swarm; 5) Determine whether the stop condition of the algorithm is met. If it is met, proceed to the next step, otherwise go to the above step 2); 6) After calculating the optimal particle, output it as the initial parameter of the neural network.

[0061] Specifically, vi(t+1): velocity vector of the ith particle at time t+1; W: inertia weight, used to control the speed of the particle's advance and affect the particle's global search ability; vi(t): velocity vector of the ith particle at time t; c1: individual learning factor, representing the acceleration constant of the particle learning its own experience; r1: a random number between 0 and 1, which increases the randomness of particle search; pi: the historical optimal position of the ith particle, representing the best position found by the particle itself; xi(t): position vector of the ith particle at time t; c2: social learning factor, representing the acceleration constant of the particle learning group experience; r2: a random number between 0 and 1, which increases the randomness of particle search; pg: the global optimal position of the particle group, that is, the best position found by all particles; xi(t+1): position vector of the ith particle at time t+1.

[0062] In addition, it should be noted that the above stop condition can be to determine whether the maximum number of iterations is reached, or whether the error between two iterations is within the range of ±5%.

[0063] That is, in order to objectively and accurately evaluate the prediction ability of the proposed ultra-short-term prediction model, the simulation results are evaluated using the calculation speed and the symmetric mean absolute percentage error (SMAPE) in the embodiment of the present invention. The calculation speed is the time required for the proposed model training to converge, represented by t. The smaller its value, the faster the calculation speed and the higher the training efficiency. The symmetric mean absolute percentage error SMAPE represents the prediction error. The smaller its value, the better the prediction effect and the higher the accuracy.

[0064] The PSO optimized BP neural network model is used for ultra-short-term photovoltaic output forecasting. First, data preprocessing is performed: preprocessing of power, irradiance and other data, eliminating null value data, and normalizing the data. The historical data of photovoltaic output includes statistical information such as time, light, photovoltaic power generation, meteorological environment, sampling information and geographical location. Then the model is trained, in which the model training uses the photovoltaic output fusion feature value to dynamically assign weight parameters based on the observed value using the dynamic exponential smoothing method, effectively avoiding the prediction result deviation caused by random assignment, and because of the dynamic change of the time step, the training speed can be improved. In addition, convolution channels of different scales are set to obtain more reasonable field of view features in the prediction.

[0065] The photovoltaic and total meter data input to the model are intercepted according to a certain time window. When intercepting data at different times, the length of the obtained data sequence is the same, but the time section represented by each specific position in the sequence is different, which increases the complexity of the model and the difficulty of model convergence. Therefore, the input data is sliced ​​and permuted, and the order of each sequence is reordered according to the time point of collection. For the x sequence with a time window width of w, assuming that w represents the time width of one day, the sequence x={x1,x2,....,x t ,…x w},Assume For data with sampling time 0:00, the rearranged sequence x is {x t0 , x t0+1 ,…xw,x1,x2,…, x t0-1}.

[0066] Among them, the case analysis shows that the PSO-BP neural network method based on data optimization has good performance when processing nonlinear sample data of high dimensions and multiple data sources, and is suitable for ultra-short-term photovoltaic prediction under complex weather conditions. On the one hand, the training speed is improved through data optimization, and on the other hand, the prediction accuracy is enhanced through convolution channel merging operations. The coupling relationship analysis of the output data of three actual photovoltaic power stations is carried out, and the prediction results of a photovoltaic power station are verified under different weather conditions. The simulation results show that the proposed model has good prediction performance on sunny days compared with other algorithms, but under complex weather conditions such as cloudy and overcast days, the prediction accuracy of the proposed model has obvious advantages, which is more in line with the actual photovoltaic output situation, and the training efficiency is greatly improved.

[0067] Specific implementation example of algorithm analysis: The neural network structure parameters of the PSO-BP model in the algorithm are: 3 input layer nodes, 5 hidden layer nodes, and 1 output layer node. The particle swarm optimization parameters are population size 30, evolution times 5, and learning factors C1 and C2 are 1.5. Using the Alicespring data set as training data, the optimized model is used to perform ultra-short-term predictions of photovoltaic power under different weather types. The prediction data is selected from the three photovoltaic power generation systems from April 2023 to June 2023. The daily photovoltaic output stability time, that is, the actual photovoltaic power generation data from 06:00 to 18:00, is used for example simulation analysis, and the data sampling interval is 15 minutes. The prediction results under different weather types are shown in Table 1 below:

[0068] Table 1 PV power prediction results for various weather conditions

[0069]

[0070] Analyzing the data in the above table, we can see that: (1) Figure 3 This is a comparison chart of the prediction results of the photovoltaic output data acquisition method in the photovoltaic power station according to the embodiment of the present invention under sunny conditions. On sunny days, the prediction results of the four algorithms are basically consistent with the actual situation, and the SMAPE values ​​are relatively close and the errors are relatively small, indicating that the four algorithms have strong prediction capabilities when the photovoltaic output is stable and can meet the prediction requirements. However, in terms of convergence speed, it can be seen that the convergence time of the proposed model and the traditional deep learning algorithm are relatively short, which are 0.021s and 0.031s respectively, indicating that the deep learning algorithm has advantages over other algorithms in training and optimization, especially after data preprocessing and dynamic parameter assignment, the convergence speed of the proposed model has been significantly improved. (2) Figure 4This is a comparison chart of the prediction results of the photovoltaic output data acquisition method in the photovoltaic power station according to the embodiment of the present invention under cloudy conditions. On cloudy days, the SMAPE value of the proposed model is reduced by 24.4%, 35.4% and 35.8% respectively compared with the deep learning algorithm, multi-layer feedback neural network and artificial neural network, and the convergence speed is increased by 30.8% and 18.8% respectively. This shows that the proposed prediction error is smaller, the accuracy is higher, and the convergence speed is faster. Under cloudy weather conditions, the prediction accuracy and training efficiency have great advantages. (3) Figure 5 This is a comparison chart of the prediction results of the photovoltaic output data acquisition method in the photovoltaic power station according to the embodiment of the present invention under cloudy conditions. When it is cloudy, the prediction results of the three algorithms cannot fully fit the actual photovoltaic output curve changes, and there is a large deviation, which is caused by the rapid changes in the meteorological environment in cloudy weather. However, the SMAPE value of the proposed model is still 5.15%, 12.7% and 15.8% lower than that of the deep learning algorithm, the multi-layer feedback neural network and the artificial neural network, respectively, and the convergence time of the training is also the shortest, indicating that under extreme cloudy weather conditions, the proposed model can still maintain good prediction accuracy and training efficiency, and is more adaptable to complex weather than other algorithms.

[0071] In summary, the simulation results show that the proposed ultra-short-term photovoltaic output prediction model has certain improvements in computational efficiency and prediction accuracy compared to traditional deep learning and multi-layer feedback neural networks. Because the data coupling relationship between different photovoltaic power stations is comprehensively considered, the fusion eigenvalues ​​of multi-order convolution channel parallel operations are constructed to effectively enhance the accuracy of model prediction; at the same time, after data preprocessing and dynamic parameter adjustment, the algorithm has a smaller computational burden, greatly shortens the training convergence time, improves training efficiency, and is more adaptable to complex and changeable climate conditions than other algorithms.

[0072] Through the technical solution provided by the above embodiment of the present invention, the PSO-BP neural network method based on data optimization has good performance when processing nonlinear sample data of high dimensions and multiple data sources, and is suitable for ultra-short-term photovoltaic prediction under complex weather conditions. On the one hand, the training speed is improved through data optimization, and on the other hand, the prediction accuracy is enhanced through convolution channel merging operations. The coupling relationship analysis is performed on the actual output data of three photovoltaic power stations, and the prediction results of a photovoltaic power station are verified under different weather conditions. The simulation results show that the proposed model has good prediction performance on sunny days compared with other algorithms, but under complex weather conditions such as cloudy and overcast days, the prediction accuracy of the proposed model has obvious advantages, which is more in line with the actual photovoltaic output situation, and the training efficiency is greatly improved.

[0073] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0074] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0075] According to an embodiment of the present invention, there is also provided a photovoltaic output data acquisition device in a photovoltaic power station for implementing the above-mentioned photovoltaic output data acquisition method in a photovoltaic power station. Figure 6 is a schematic diagram of a photovoltaic output data acquisition device in a photovoltaic power station according to an embodiment of the present invention. Figure 6 As shown, the photovoltaic output data acquisition device in the photovoltaic power station includes: a first acquisition unit 601, a pre-processing unit 603, a processing unit 605 and a second acquisition unit 607. The photovoltaic output data acquisition device in the photovoltaic power station is described below.

[0076] The first acquisition unit 601 is used to acquire photovoltaic output data of multiple photovoltaic power stations in the area where the target power system is located.

[0077] The preprocessing unit 603 is used to preprocess the photovoltaic output data to filter out abnormal state data and fault state data in the photovoltaic output data to obtain preprocessed photovoltaic output data, wherein the abnormal state data is data in the photovoltaic output data that exceeds the maximum photovoltaic output power, and the fault state data is data in the photovoltaic output data that the photovoltaic output power is 0 when the light intensity exceeds the light intensity threshold.

[0078] The processing unit 605 is used to input the pre-processed photovoltaic output data into the photovoltaic output ultra-short-term prediction model, so as to process the pre-processed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain a processing result, wherein the photovoltaic output ultra-short-term prediction model is a model obtained by machine learning training using multiple groups of sample data, each of the multiple groups of sample data includes: sample photovoltaic output data and sample processing results corresponding to the sample photovoltaic output data, and the multiple groups of sample data include sample photovoltaic output data collected under different weather conditions.

[0079] The second acquisition unit 607 is used to obtain the estimated photovoltaic output values ​​of the multiple photovoltaic power stations within a future predetermined time period according to the processing results.

[0080] It should be noted here that the above-mentioned first acquisition unit 601, preprocessing unit 603, processing unit 605 and second acquisition unit 607 correspond to steps S202 to S208 in the above-mentioned embodiments, and the four units are the same as the instances and application scenarios implemented by the corresponding steps, but are not limited to the contents disclosed in the above-mentioned embodiments.

[0081] As can be seen from the above, in the scheme recorded in the above-mentioned embodiment of the present invention, the first acquisition unit can be used to acquire the photovoltaic output data of multiple photovoltaic power stations in the substation where the target power system is located; the preprocessing unit is used to preprocess the photovoltaic output data to filter out abnormal state data and fault state data in the photovoltaic output data to obtain preprocessed photovoltaic output data, wherein the abnormal state data is data in the photovoltaic output data that exceeds the maximum photovoltaic output power, and the fault state data is data in the photovoltaic output data where the photovoltaic output power is 0 when the light intensity exceeds the light intensity threshold; the processing unit is used to input the preprocessed photovoltaic output data into the photovoltaic output ultra-short-term prediction model to use the photovoltaic output ultra-short-term prediction model to perform preprocessing on the preprocessed photovoltaic output data. The data is processed to obtain a processing result, wherein the photovoltaic output ultra-short-term prediction model is a model obtained by machine learning training using multiple groups of sample data, each of the multiple groups of sample data includes: sample photovoltaic output data and sample processing results corresponding to the sample photovoltaic output data, and the multiple groups of sample data include sample photovoltaic output data collected under different weather conditions; a second acquisition unit is used to obtain photovoltaic output estimates of multiple photovoltaic power stations within a predetermined period of time in the future according to the processing results, thereby achieving the purpose of predicting the photovoltaic output value within a future period of time based on historical photovoltaic output data through a pre-trained photovoltaic output ultra-short-term prediction model, and achieving the technical effect of improving the accuracy of estimating the output value of the photovoltaic power station within a future period of time.

[0082] Therefore, the above technical solution provided by the embodiment of the present invention solves the technical problem in the related art that when evaluating the output data of a photovoltaic power station in the future, it is easily affected by measurement errors and bad data, resulting in large errors in the estimated value.

[0083] Optionally, the preprocessing unit includes: a first determination subunit, used to determine abnormal state data in the photovoltaic output data according to the maximum photovoltaic output power, and determine fault state data in the photovoltaic output data; a cleaning subunit, used to perform data cleaning on the photovoltaic output data to remove the abnormal state data and fault state data in the photovoltaic output data; a normalization processing subunit, used to normalize the photovoltaic output data after filtering out the abnormal state data and fault state data to obtain preprocessed photovoltaic output data.

[0084] Optionally, the normalization processing subunit includes: a normalization processing module, which is used to normalize the photovoltaic output data after filtering out abnormal state data and fault state data through a first formula to obtain pre-processed photovoltaic output data, wherein the first formula is: , Represents the pre-processed photovoltaic output data, Indicates the photovoltaic output data after filtering out abnormal status data and fault status data. It indicates the minimum value of the photovoltaic output data after filtering out abnormal status data and fault status data. Indicates the maximum value of the PV output data after filtering out abnormal status data and fault status data.

[0085] Optionally, the processing unit includes: a second determination subunit, used to determine the weight coefficient of each output data in the photovoltaic output data in the ultra-short-term photovoltaic output prediction model by using a dynamic exponential smoothing method; a weighted processing subunit, used to perform weighted processing on the photovoltaic output data according to the weight coefficient to obtain a weighted processing result; an extraction subunit, used to extract the spatiotemporal coupling relationship of multiple photovoltaic power stations using the information fusion multi-order convolution channel in the ultra-short-term photovoltaic output prediction model, wherein the information fusion multi-order convolution channel is a convolution channel obtained by splicing first-order convolution channels, and the first-order convolution channel is a channel created for each factor affecting the output of multiple photovoltaic power stations; a first acquisition subunit, used to obtain a fused characteristic value of the photovoltaic output data according to the spatiotemporal coupling relationship; and a second acquisition subunit, used to obtain a processing result based on the fused characteristic value.

[0086] Optionally, the first acquisition subunit includes: an acquisition module, configured to obtain a fusion feature value by a second formula according to the spatiotemporal coupling relationship, wherein the second formula is: , represents the fusion feature value, Represents the weighted coefficient of the first-order convolution channel in the process of fusion feature splicing, Indicates photovoltaic output data, represents the length of the feature vector, Indicates the number of channels for photovoltaic output data. It represents the number of channels of the output features. The spatiotemporal coupling relationship is represented by the length of the feature vector and the number of channels.

[0087] Optionally, the photovoltaic output data acquisition device in the photovoltaic power station also includes: a training unit, which is used to input the preprocessed photovoltaic output data into the photovoltaic output ultra-short-term prediction model, so as to process the preprocessed photovoltaic output data using the photovoltaic output ultra-short-term prediction model, and before obtaining the processing results, train multiple groups of sample data to obtain the photovoltaic output ultra-short-term prediction model, wherein the training unit includes: a third determination subunit, which is used to determine the initial parameters of the basic neural network model corresponding to the photovoltaic output ultra-short-term prediction model through a particle swarm optimization algorithm; a third acquisition subunit, which is used to assign the initial parameters to the basic neural network model to obtain the initial photovoltaic output ultra-short-term prediction model; a verification subunit, which is used to verify the initial photovoltaic output ultra-short-term prediction model using multiple groups of sample data to obtain the prediction accuracy of the initial photovoltaic output ultra-short-term prediction model; and a fourth acquisition subunit, which is used to obtain the photovoltaic output ultra-short-term prediction model when the prediction accuracy is greater than the prediction accuracy threshold.

[0088] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium comprising a stored program, wherein the program executes any one of the above-mentioned methods for acquiring photovoltaic output data in a photovoltaic power station.

[0089] Optionally, in this embodiment, the computer-readable storage medium may be located in any one of the computer terminals in a computer terminal group in a computer network, or in any one of the communication devices in a communication device group.

[0090] Optionally, in this embodiment, the computer-readable storage medium is configured to store program codes for executing the following steps: obtaining photovoltaic output data of multiple photovoltaic power stations in the substation where the target power system is located; preprocessing the photovoltaic output data to filter out abnormal state data and fault state data in the photovoltaic output data to obtain preprocessed photovoltaic output data, wherein the abnormal state data is data in the photovoltaic output data that exceeds the maximum photovoltaic output power, and the fault state data is data in the photovoltaic output data where the photovoltaic output power is 0 when the light intensity exceeds the light intensity threshold; inputting the preprocessed photovoltaic output data into a photovoltaic output ultra-short-term prediction model to process the preprocessed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain a processing result, wherein the photovoltaic output ultra-short-term prediction model is a model obtained by machine learning training using multiple groups of sample data, each of the multiple groups of sample data includes: sample photovoltaic output data and sample processing results corresponding to the sample photovoltaic output data, and the multiple groups of sample data include sample photovoltaic output data collected under different weather conditions; and obtaining photovoltaic output estimates of multiple photovoltaic power stations within a predetermined time period in the future according to the processing results.

[0091] Optionally, in this embodiment, the computer-readable storage medium is configured to store program codes for executing the following steps: determining abnormal state data in photovoltaic output data based on the maximum photovoltaic output power, and determining fault state data in the photovoltaic output data; performing data cleaning on the photovoltaic output data to remove abnormal state data and fault state data in the photovoltaic output data; and normalizing the photovoltaic output data after filtering out the abnormal state data and fault state data to obtain pre-processed photovoltaic output data.

[0092] Optionally, in this embodiment, the computer-readable storage medium is configured to store a program code for executing the following steps: normalizing the photovoltaic output data after filtering out the abnormal state data and the fault state data by a first formula to obtain preprocessed photovoltaic output data, wherein the first formula is: , Represents the pre-processed photovoltaic output data, Indicates the photovoltaic output data after filtering out abnormal status data and fault status data. It indicates the minimum value of the photovoltaic output data after filtering out abnormal status data and fault status data. Indicates the maximum value of the PV output data after filtering out abnormal status data and fault status data.

[0093] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for executing the following steps: using a dynamic exponential smoothing method to determine the weight coefficient of each output data in the photovoltaic output data in the ultra-short-term photovoltaic output prediction model; weighted processing is performed on the photovoltaic output data according to the weight coefficient to obtain a weighted processing result; using the information fusion multi-order convolution channel in the ultra-short-term photovoltaic output prediction model to extract the spatiotemporal coupling relationship of multiple photovoltaic power stations, wherein the information fusion multi-order convolution channel is a convolution channel obtained by splicing first-order convolution channels, and the first-order convolution channel is a channel created for each factor affecting the output of multiple photovoltaic power stations; according to the spatiotemporal coupling relationship, a fusion characteristic value of the photovoltaic output data is obtained; and a processing result is obtained according to the fusion characteristic value.

[0094] Optionally, in this embodiment, the computer-readable storage medium is configured to store a program code for executing the following steps: obtaining a fusion feature value by a second formula according to the spatiotemporal coupling relationship, wherein the second formula is: , represents the fusion feature value, Represents the weighted coefficient of the first-order convolution channel in the process of fusion feature splicing, Indicates photovoltaic output data, represents the length of the feature vector, Indicates the number of channels for photovoltaic output data. It represents the number of channels of the output features. The spatiotemporal coupling relationship is represented by the length of the feature vector and the number of channels.

[0095] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for executing the following steps: before inputting the preprocessed photovoltaic output data into the photovoltaic output ultra-short-term prediction model, and processing the preprocessed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain the processing result, multiple groups of sample data are trained to obtain the photovoltaic output ultra-short-term prediction model, wherein the training of multiple groups of sample data to obtain the photovoltaic output ultra-short-term prediction model includes: determining the initial parameters of the basic neural network model corresponding to the photovoltaic output ultra-short-term prediction model through a particle swarm optimization algorithm; assigning the initial parameters to the basic neural network model to obtain the initial photovoltaic output ultra-short-term prediction model; verifying the initial photovoltaic output ultra-short-term prediction model using multiple groups of sample data to obtain the prediction accuracy of the initial photovoltaic output ultra-short-term prediction model; when the prediction accuracy is greater than the prediction accuracy threshold, obtaining the photovoltaic output ultra-short-term prediction model.

[0096] According to another aspect of an embodiment of the present invention, a processor is further provided, and the processor is used to run a program, wherein when the program is run, any one of the above-mentioned methods for acquiring photovoltaic output data in a photovoltaic power station is executed.

[0097] According to another aspect of an embodiment of the present invention, a computer program product is provided, including computer instructions. When the computer instructions are executed by a processor, any one of the above methods for acquiring photovoltaic output data in a photovoltaic power station is executed.

[0098] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0099] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0100] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, 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 units or modules, which can be electrical or other forms.

[0101] 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 units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0102] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0103] 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 this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc. Various media that can store program codes.

[0104] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for acquiring photovoltaic output data in a photovoltaic power station, characterized in that: include: Obtain photovoltaic output data of multiple photovoltaic power stations in the area where the target power system is located; Preprocessing the photovoltaic output data to filter out abnormal state data and fault state data in the photovoltaic output data to obtain the preprocessed photovoltaic output data, wherein the abnormal state data is data in the photovoltaic output data that exceeds the maximum photovoltaic output power, and the fault state data is data in the photovoltaic output data that the photovoltaic output power is 0 when the light intensity exceeds the light intensity threshold; Input the preprocessed photovoltaic output data into a photovoltaic output ultra-short-term prediction model, so as to process the preprocessed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain a processing result, wherein the photovoltaic output ultra-short-term prediction model is a model obtained by machine learning training using multiple groups of sample data, each of the multiple groups of sample data includes: sample photovoltaic output data and sample processing results corresponding to the sample photovoltaic output data, and the multiple groups of sample data include sample photovoltaic output data collected under different weather conditions; The estimated photovoltaic output values ​​of the multiple photovoltaic power stations within a predetermined future time period are obtained according to the processing results.

2. The method for obtaining photovoltaic output data in a photovoltaic power station according to claim 1, characterized in that: Preprocessing the photovoltaic output data to filter out abnormal state data and fault state data in the photovoltaic output data to obtain the preprocessed photovoltaic output data includes: Determine the abnormal state data in the photovoltaic output data according to the photovoltaic maximum output power, and determine the fault state data in the photovoltaic output data; Performing data cleaning on the photovoltaic output data to remove the abnormal state data and the fault state data in the photovoltaic output data; The photovoltaic output data after filtering out the abnormal state data and the fault state data is normalized to obtain the pre-processed photovoltaic output data.

3. The method for obtaining photovoltaic output data in a photovoltaic power station according to claim 2, characterized in that: Normalizing the photovoltaic output data after filtering out the abnormal state data and the fault state data to obtain the pre-processed photovoltaic output data, including: The photovoltaic output data after filtering out the abnormal state data and the fault state data is normalized by a first formula to obtain the pre-processed photovoltaic output data, wherein the first formula is: , represents the photovoltaic output data after preprocessing, represents the photovoltaic output data after filtering out the abnormal state data and the fault state data, represents the minimum value of the photovoltaic output data after filtering out the abnormal state data and the fault state data, It represents the maximum value of the photovoltaic output data after filtering out the abnormal state data and the fault state data.

4. The method for obtaining photovoltaic output data in a photovoltaic power station according to claim 1, characterized in that: The pre-processed photovoltaic output data is input into the photovoltaic output ultra-short-term prediction model, so as to process the pre-processed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain a processing result, including: Determine the weight coefficient of each output data in the photovoltaic output data in the photovoltaic output ultra-short-term prediction model by using a dynamic exponential smoothing method; Performing weighted processing on the photovoltaic output data according to the weight coefficient to obtain a weighted processing result; The spatiotemporal coupling relationship of the plurality of photovoltaic power stations is extracted by using the information fusion multi-order convolution channel in the photovoltaic output ultra-short-term prediction model, wherein the information fusion multi-order convolution channel is a convolution channel obtained by splicing the first-order convolution channel, and the first-order convolution channel is a channel created for each factor affecting the output of the plurality of photovoltaic power stations; Obtaining a fusion characteristic value of the photovoltaic output data according to the spatiotemporal coupling relationship; The processing result is obtained according to the fused feature value.

5. The method for obtaining photovoltaic output data in a photovoltaic power station according to claim 4, characterized in that: Obtaining the fusion characteristic value of the photovoltaic output data according to the spatiotemporal coupling relationship includes: According to the spatiotemporal coupling relationship, the fusion characteristic value is obtained by a second formula, wherein the second formula is: , represents the fused feature value, represents the weight coefficient of the first-order convolution channel in the process of fusion feature splicing, represents the photovoltaic output data, represents the length of the feature vector, Indicates the channel number of the photovoltaic output data, The number of channels representing the output feature, the spatiotemporal coupling relationship is represented by the length of the feature vector and the number of channels.

6. The method for acquiring photovoltaic output data in a photovoltaic power station according to claim 1, characterized in that: Before inputting the preprocessed photovoltaic output data into the photovoltaic output ultra-short-term prediction model to process the preprocessed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain the processing result, the method further includes: training the multiple groups of sample data to obtain the photovoltaic output ultra-short-term prediction model, The multiple groups of sample data are trained to obtain the ultra-short-term photovoltaic output prediction model, including: Determine the initial parameters of the basic neural network model corresponding to the photovoltaic output ultra-short-term prediction model by using a particle swarm optimization algorithm; Assigning the initial parameters to the basic neural network model to obtain an initial photovoltaic output ultra-short-term prediction model; Using the multiple groups of sample data to verify the initial photovoltaic output ultra-short-term prediction model to obtain the prediction accuracy of the initial photovoltaic output ultra-short-term prediction model; When the prediction accuracy is greater than a prediction accuracy threshold, the photovoltaic output ultra-short-term prediction model is obtained.

7. A photovoltaic output data acquisition device in a photovoltaic power station, characterized in that: include: A first acquisition unit is used to acquire photovoltaic output data of multiple photovoltaic power stations in the area where the target power system is located; a preprocessing unit, configured to preprocess the photovoltaic output data to filter out abnormal state data and fault state data in the photovoltaic output data, and obtain the preprocessed photovoltaic output data, wherein the abnormal state data is data in the photovoltaic output data that exceeds the maximum photovoltaic output power, and the fault state data is data in the photovoltaic output data that the photovoltaic output power is 0 when the light intensity exceeds a light intensity threshold; A processing unit, used for inputting the pre-processed photovoltaic output data into a photovoltaic output ultra-short-term prediction model, so as to process the pre-processed photovoltaic output data using the photovoltaic output ultra-short-term prediction model to obtain a processing result, wherein the photovoltaic output ultra-short-term prediction model is a model obtained by machine learning training using multiple groups of sample data, each of the multiple groups of sample data includes: sample photovoltaic output data and sample processing results corresponding to the sample photovoltaic output data, and the multiple groups of sample data include sample photovoltaic output data collected under different weather conditions; The second acquisition unit is used to obtain the estimated photovoltaic output values ​​of the multiple photovoltaic power stations within a future predetermined period of time according to the processing results.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method for acquiring photovoltaic output data in a photovoltaic power station according to any one of claims 1 to 6.

9. A processor, characterized in that: The processor is used to run a program, wherein the program, when running, executes the method for acquiring photovoltaic output data in a photovoltaic power station according to any one of claims 1 to 6.

10. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by the processor, the method for acquiring photovoltaic output data in a photovoltaic power station according to any one of claims 1 to 6 is performed.

Images