Electronic device and method for predicting solar power generation

Abstract

An electronic device is disclosed, comprising a memory and a processor operatively connected to the memory, wherein the processor receives a plurality of weather forecast data, acquires a plurality of characteristic data based on the plurality of weather forecast data, acquires integrated characteristic data by integrating the plurality of characteristic data based on the similarity between the plurality of characteristic data, and generates a predicted value of solar power generation based on the integrated characteristic data.

Description

Technology Field

[0001] The present disclosure relates to an electronic device and method for predicting solar power generation. Background Technology

[0002] As the demand for renewable energy increases, interest in solar power generation is also growing. Solar power generation, a form of renewable energy, is expected to play a crucial role in achieving greenhouse gas reduction targets alongside climate change in the Fourth Industrial Revolution. Consequently, thanks to continuous support policies for the solar power industry and technological advancements, the lifespan and efficiency of solar panels are improving, and solar power generation facilities are becoming more sophisticated.

[0003] Meanwhile, as the demand for renewable energy sources such as solar power increases, the importance of energy management is being highlighted. Energy management centers on matching supply and demand; however, unlike thermal energy, where supply can be freely adjusted, renewable energy requires supply forecasting because it converts the power of nature into energy.

[0004] Although many recent studies have proposed methods to predict solar power generation using artificial neural networks, the accuracy and consistency of performance remain lacking due to the uncertainty and variability of weather forecasts. For example, existing methods for predicting solar power generation train prediction models using weather observation data and rely on only one type of weather forecast data to forecast output. However, in this case, the weather forecast data used as input for the prediction model is limited; in particular, using only a single dataset restricts available information, which can lead to limitations in predicting solar power generation. Furthermore, since the performance of the prediction model heavily depends on the accuracy of the weather forecast data, there is a high probability that the model will output power generation figures significantly different from the actual output if the forecast data differs greatly from the actual weather observation data. Accordingly, there is a need for the development of technologies that can improve the accuracy and stability of solar power generation prediction. The problem to be solved

[0005] The present disclosure provides an electronic device and method for predicting solar power generation to solve the above-mentioned problems. means of solving the problem

[0006] The present disclosure may be implemented in various ways, including methods, devices (systems), and / or computer programs stored on computer-readable storage media.

[0007] According to one embodiment of the present disclosure, an electronic device includes a memory and a processor operatively connected to the memory, and the processor may be configured to receive a plurality of weather forecast data, acquire a plurality of characteristic data based on the plurality of weather forecast data, acquire integrated characteristic data by integrating the plurality of characteristic data based on the similarity between the plurality of characteristic data, and generate a predicted value of solar power generation based on the integrated characteristic data.

[0008] According to one embodiment, a plurality of weather forecast data may include a first weather forecast data including a weather forecast value of a first time interval within a specified time range and a second weather forecast data including a weather forecast value of a second time interval longer than the first time interval within a specified time range.

[0009] According to one embodiment, the processor is configured to acquire an embedding vector corresponding to each of a plurality of weather forecast data as a plurality of characteristic data, wherein the embedding vector may include time series information of the weather forecast data corresponding to the embedding vector among the plurality of weather forecast data and may have a specified length.

[0010] According to one embodiment, the processor may be configured to generate an embedding vector using an attention-based LSTM.

[0011] According to one embodiment, the processor may be configured to generate a query matrix, a key matrix, and a value matrix corresponding to each of a plurality of feature data that are used in an attention model, based on an embedding vector corresponding to each of a plurality of weather forecast data.

[0012] According to one embodiment, the processor may be configured to calculate a first attention score representing similarity between a query and a key based on a key matrix associated with a first characteristic data among a plurality of characteristic data and a query matrix associated with a second characteristic data among a plurality of characteristic data, calculate a first attention value based on the first attention score and a value matrix associated with the first characteristic data, calculate a second attention score representing similarity between a query and a key based on a key matrix associated with the second characteristic data and a query matrix of the first characteristic data, calculate a second attention value based on the second attention score and a value matrix associated with the second characteristic data, and generate integrated characteristic data by concatenating the first attention value and the second attention value.

[0013] According to one embodiment, the processor may be configured to apply integrated characteristic data to a transformer model.

[0014] According to one embodiment, the processor may be configured to generate a predicted value of solar power generation through a multilayer perceptron model using integrated characteristic data.

[0015] According to one embodiment, the activation function of a multilayer perceptron may include a leaky ReLU function.

[0016] According to one embodiment of the present disclosure, a method for predicting solar power generation, performed by at least one processor, may include the steps of receiving a plurality of weather forecast data, obtaining a plurality of characteristic data based on the plurality of weather forecast data, obtaining integrated characteristic data by integrating the plurality of characteristic data based on similarity between the plurality of characteristic data, and generating a predicted value of solar power generation based on the integrated characteristic data. Effects of the invention

[0017] According to some embodiments of the present disclosure, accurate and stable prediction of solar power generation may be possible by generating a predicted value of solar power generation based on a plurality of weather forecast data. In addition, stable grid operation is possible through accurate prediction of solar power generation, and energy waste can be minimized by optimizing the scheduling of energy storage devices. Furthermore, the supply of solar energy can be adjusted, thereby improving the utilization of solar energy, and performance monitoring of the solar power plant may be possible by comparing the predicted generation with the actual generation.

[0018] The effects of the present disclosure are not limited to those mentioned above, and other unmentioned effects will be clearly understood by a person skilled in the art to which the present disclosure pertains (referred to as "person skilled in the art") from the description in the claims. Brief explanation of the drawing

[0019] Embodiments of the present disclosure will be described with reference to the accompanying drawings described below, wherein similar reference numerals indicate similar elements, but are not limited thereto. FIG. 1 is a drawing for explaining the configuration of an electronic device for predicting solar power generation according to one embodiment of the present disclosure. FIG. 2 is a schematic diagram showing a configuration in which an information processing system is connected to communicate with a plurality of user terminals in relation to data processing according to one embodiment of the present disclosure. FIG. 3 is a block diagram showing the internal configuration of a user terminal and an information processing system according to one embodiment of the present disclosure. FIG. 4 is a diagram illustrating a method for calculating a predicted value of solar power generation according to one embodiment of the present disclosure. FIG. 5 is a diagram illustrating a method for calculating an attention value according to one embodiment of the present disclosure. FIG. 6 is a drawing for explaining an artificial neural network model according to one embodiment of the present disclosure. FIG. 7 is a drawing for explaining a method for predicting solar power generation according to one embodiment of the present disclosure. Specific details for implementing the invention

[0020] Hereinafter, specific details for implementing the present disclosure will be described in detail with reference to the attached drawings. However, in the following description, specific descriptions regarding widely known functions or configurations will be omitted if there is a risk that the gist of the present disclosure may be unnecessarily obscured.

[0021] In the attached drawings, identical or corresponding components are assigned the same reference numerals. Additionally, in the description of the following embodiments, the description of identical or corresponding components may be omitted. However, even if a description of a component is omitted, it is not intended that such component is not included in any embodiment.

[0022] The advantages and features of the disclosed embodiments and the methods for achieving them will become clear by referring to the embodiments described below in conjunction with the accompanying drawings. However, the present disclosure is not limited to the embodiments disclosed below but may be implemented in various different forms, and the embodiments provided are merely to make the present disclosure complete and to fully inform those skilled in the art of the scope of the invention.

[0023] The terms used in this specification will be briefly explained, and the disclosed embodiments will be described in detail. The terms used in this specification have been selected to be as generally used as possible, taking into account their functions in this disclosure; however, these terms may vary depending on the intent of those skilled in the art, case law, the emergence of new technologies, etc. Additionally, in specific cases, terms may be arbitrarily selected by the applicant, and in such cases, their meanings will be described in detail in the relevant description of the invention. Therefore, the terms used in this disclosure should be defined not merely by their names, but based on their meanings and the content throughout this disclosure.

[0024] In this specification, singular expressions include plural expressions unless the context clearly specifies them as singular. Additionally, plural expressions include singular expressions unless the context clearly specifies them as plural. Throughout the specification, when a part is described as including a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0025] Additionally, the terms 'module' or 'part' as used in the specification refer to software or hardware components, and the 'module' or 'part' performs certain roles. However, the meaning of 'module' or 'part' is not limited to software or hardware. The 'module' or 'part' may be configured to reside in an addressable storage medium or configured to run on one or more processors. Thus, as an example, the 'module' or 'part' may include components such as software components, object-oriented software components, class components, and task components, and at least one of processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, or variables. The components and the functions provided within the 'module' or 'part' may be combined into a smaller number of components and 'modules' or 'parts', or further separated into additional components and 'modules' or 'parts'.

[0026] According to one embodiment of the present disclosure, a ‘module’ or ‘part’ may be implemented as a processor and memory. The term ‘processor’ should be broadly interpreted to include a general-purpose processor, a central processing unit (CPU), a microprocessor, a digital signal processor (DSP), a controller, a microcontroller, a state machine, etc. In some environments, the term ‘processor’ may refer to an application-specific integrated circuit (ASIC), a programmable logic device (PLD), a field programmable gate array (FPGA), etc. The term ‘processor’ may also refer to a combination of processing devices, such as, for example, a combination of a DSP and a microprocessor, a combination of multiple microprocessors, a combination of one or more microprocessors combined with a DSP core, or any other combination of such configurations. Additionally, the term ‘memory’ should be broadly interpreted to include any electronic component capable of storing electronic information. 'Memory' may refer to various types of processor-readable media, such as Random Access Memory (RAM), Read-Only Memory (ROM), Non-Volatile Random Access Memory (NVRAM), Programmable Read-Only Memory (PROM), Erasable-Programmable Read-Only Memory (EPROM), Electrically Erasable PROM (EEPROM), Flash Memory, Magnetic or Marked Data Storage Devices, Registers, etc. If a processor can read information from memory and / or write information to memory, the memory is said to be in an electronic communication state with the processor. Memory integrated into a processor is in an electronic communication state with the processor.

[0027] In addition, terms such as first, second, A, B, (a), (b), etc. used in the following embodiments are used merely to distinguish one component from another, and the essence, order, or sequence of the said component is not limited by such terms.

[0028] Additionally, in the following embodiments, where it is stated that one component is 'connected', 'coupled', or 'joined' to another component, it should be understood that the component may be directly connected or joined to the other component, but that another component may also be 'connected', 'coupled', or 'joined' between each component.

[0029] Additionally, the terms 'comprises' and / or 'comprising' as used in the following embodiments do not exclude the presence or addition of one or more other components, steps, actions, and / or elements to the mentioned components, steps, actions, and / or elements.

[0030] Hereinafter, various embodiments of the present disclosure will be described in detail with reference to the attached drawings.

[0031] FIG. 1 is a diagram illustrating the configuration of an electronic device (100) for predicting solar power generation according to one embodiment of the present disclosure. Referring to FIG. 1, the electronic device (100) for predicting solar power generation may include a memory (110) and a processor (130). However, the configuration of the electronic device (100) is not limited thereto. According to various embodiments, the electronic device (100) may include at least one additional component in addition to the components described above. According to one embodiment, the electronic device (100) may further include a communication circuit or a display. The communication circuit may support the establishment of a direct (e.g., wired) communication channel or a wireless communication channel between the electronic device (100) and an external electronic device, and the performance of communication through the established communication channel. For example, the electronic device (100) may receive weather forecast data from an external electronic device connected via the communication circuit. The display may visually provide information to the outside of the electronic device (100) (e.g., a user). For example, the electronic device (100) can display a predicted value of solar power generation through a display. In another example, if the electronic device (100) is a server (e.g., the information processing system of FIG. 2 and 3), the electronic device (100) can transmit the predicted value of solar power generation to a user terminal.

[0032] The memory (110) can store various data used by at least one component (e.g., processor (130)) of the electronic device (100). The data may include, for example, input data or output data for software (or programs) and related instructions. The memory (110) may include volatile memory or non-volatile memory. According to one embodiment, the memory (110) may store weather forecast data.

[0033] The processor (130) can execute software (or a program) to control at least one other component (e.g., a hardware or software component) of the electronic device (100) connected to the processor (130) and can perform various data processing or operations. According to one embodiment, as at least part of the data processing or operations, the processor (130) can load instructions or data received from other components into volatile memory, process the instructions or data stored in volatile memory, and store the resulting data in non-volatile memory.

[0034] According to one embodiment, the processor (130) can acquire a plurality of characteristic data based on a plurality of weather forecast data. Then, based on the similarity between the plurality of characteristic data, it can acquire integrated characteristic data by integrating the plurality of characteristic data, and generate a predicted value of solar power generation based on the integrated characteristic data. To this end, the processor (130) may include a characteristic data acquisition module (132), a characteristic data integration module (134), and a solar power generation prediction module (136). However, the types of modules included in the processor (130) are classified according to functions related to the prediction of solar power generation, and the types and number thereof are not limited thereto. Additionally, at least one of the modules included in the processor (130) may be implemented in the form of an instruction stored in memory (110).

[0035] In the present disclosure, the number of weather forecast data used to predict solar power generation may be at least two (e.g., two, three, etc.), and the types thereof may also vary. For example, if the weather forecast data used has an overlapping time range, the weather forecast data that can be used in the present disclosure may include at least two of ultra-short-term forecast data, short-term forecast data, medium-term forecast data, or long-term forecast data. However, for convenience of explanation, the embodiments described below will explain a method for calculating a predicted value of solar power generation using two weather forecast data. For example, a plurality of weather forecast data may include a first weather forecast data containing a weather forecast value of a first time interval within a specified time range, and a second weather forecast data containing a weather forecast value of a second time interval longer than the first time interval within a specified time range.

[0036] According to one embodiment, the first weather forecast data may include forecast data obtained through a Local Data Assimilation and Prediction System (LDAPS), and the second weather forecast data may include forecast data obtained through a Global Data Assimilation and Prediction System (GDAPS). In one embodiment, the first weather forecast data and the second weather forecast data may be forecast data with different time ranges, time intervals, and / or regional ranges.

[0037] The characteristic data acquisition module (132) can acquire multiple characteristic data based on multiple weather forecast data. For example, characteristic data can be acquired / generated for each weather forecast data. The process of the characteristic data acquisition module (132) acquiring multiple characteristic data can be referred to as a temporal fusion process. The temporal fusion process may include a process of matching the time lengths of each weather forecast data. First, if the time ranges (e.g., forecast period) of the multiple weather forecast data differ, the characteristic data acquisition module (132) can extract data corresponding to the overlapping time ranges from each of the multiple weather forecast data. Additionally, if the time intervals (e.g., forecast intervals) of the multiple weather forecast data differ, the characteristic data acquisition module (132) can perform a function of matching the time lengths / intervals to integrate (or mix) the multiple weather forecast data for use. To this end, the characteristic data acquisition module (132) can acquire characteristic data containing weather forecast information while preserving time series information from each of the multiple weather forecast data.

[0038] According to one embodiment, the feature data acquisition module (132) can acquire embedding vectors corresponding to each of the plurality of weather forecast data as the plurality of feature data. Here, the embedding vector may be a method of representing data as a fixed-dimensional real vector or a result value obtained through this. When there is high semantic similarity between the data, the embedding vectors corresponding to the data may be located adjacent to each other in the space in which they are represented. The embedding vector corresponding to each of the plurality of weather forecast data may contain time series information possessed by the corresponding weather forecast data and may have a specified length (e.g., dimension). For example, when multiple weather forecast data include a first weather forecast data having first time series information and a second weather forecast data having second time series information, a first embedding vector corresponding to the first weather forecast data and a second embedding vector corresponding to the second weather forecast data each include the first time series information and the second time series information, but the length (e.g., dimension) of the first embedding vector and the length (e.g., dimension) of the second embedding vector may be the same.

[0039] According to one embodiment, the feature data acquisition module (132) can generate an embedding vector using an attention-based LSTM (Long Short-Term Memory). Here, the attention-based LSTM may be referred to as an Attentive LSTM. The LSTM exhibits high accuracy in the field of time series forecasting and can represent a model that serves as the foundation for many time series forecasting models. For example, feature vector x t When receiving input, the LSTM uses the two state vectors h from the previous time step. t-1 and c t-1 Update and the new state vector h t and c t It outputs the hidden state vector h generated at the very last time point. Tg can be the final output value of the LSTM. However, since LSTMs have a limitation in that they cannot use values ​​prior to the hidden state vector at the final time step, an Attentive LSTM can be used to compensate for this. The Attentive LSTM does not use the hidden state vector at the last time step as the output value, but instead calculates attention values ​​by utilizing the hidden state vectors at each time step. Then, the Attentive LSTM calculates a weighted sum of the hidden state vectors using each attention value as a weight and outputs a single context vector C. Through this process, weather forecast data of different lengths can be represented as embedding vectors of the same length while preserving time-series information. Through the following Equation 1, the attention value a for each time step of the Attentive LSTM i and context vector C can be produced.

[0040]

[0041]

[0042] Here, h i represents the hidden state vector at time i (the final state when i=T), and a i can represent the attention value at time i.

[0043] The feature data integration module (134) can obtain integrated feature data by integrating multiple feature data based on the similarity between multiple feature data. The process by which the feature data integration module (134) obtains integrated feature data may be referred to as a feature fusion process. The feature fusion process may include a process of integrating multiple feature data by identifying highly relevant core information (or important information) from multiple feature data.

[0044] According to one embodiment, the feature data integration module (134) can obtain integrated feature data by integrating multiple feature data using a cross attention mechanism. First, the feature data integration module (134) can generate a query matrix, a key matrix, and a value matrix corresponding to each of the multiple feature data that are used in an attention model, based on an embedding vector corresponding to each of the multiple weather forecast data. For example, the feature data integration module (134) can generate a query matrix, a key matrix, and a value matrix associated with a first feature data (e.g., a first embedding vector) corresponding to the first weather forecast data, and generate a query matrix, a key matrix, and a value matrix associated with a second feature data (e.g., a second embedding vector) corresponding to the second weather forecast data. According to one embodiment, the feature data integration module (134) can adjust the dimensions of the query matrix, the key matrix, and the value matrix through a linear projection on the query matrix, the key matrix, and the value matrix.

[0045] Then, the feature data integration module (134) can calculate a first attention score representing the similarity between a query and a key based on a key matrix associated with the first feature data and a query matrix associated with the second feature data, and can calculate a second attention score representing the similarity between a query and a key based on a key matrix associated with the second feature data and a query matrix associated with the first feature data. According to one embodiment, the feature data integration module (134) can calculate an attention score (e.g., a first attention score or a second attention score) using a matrix multiplication operation (or a dot product between matrices).

[0046] Then, the feature data integration module (134) can calculate a first attention value based on a first attention score and a matrix of values ​​associated with the first feature data, and can calculate a second attention value based on a second attention score and a matrix of values ​​associated with the second feature data. In this process, the feature data integration module (134) can calculate an attention weight by passing the attention score through a softmax function. For example, the feature data integration module (134) can calculate a first attention weight by applying a softmax function to the first attention score and calculate a second attention weight by applying a softmax function to the second attention score. Here, applying a softmax function is intended to obtain a probability distribution in which the sum of all values ​​is 1, which is called an attention distribution, and each value obtained by applying the softmax function, i.e., the attention weight, can represent the importance of each key corresponding to the query. According to one embodiment, the feature data integration module (134) can process the attention score by scaling it before applying the softmax function. Here, the scaling operation may be performed to increase numerical stability and computational efficiency. For example, due to the nature of the inner product calculation between matrices, the input value may have a larger number as it becomes longer. Additionally, the softmax function is structured to obtain a probability distribution by applying an exponential function to each element of the input vector and normalizing the result; since an exponential function is used, the result value increases rapidly as the input value increases, which can cause an overflow. Since this phenomenon causes the gradient vanishing problem of the softmax function, the feature data integration module (134) can increase numerical stability and computational efficiency by reducing the size of the attention score through a scaling operation before applying the softmax function.Additionally, the feature data integration module (134) can obtain an attention value by calculating a weighted sum of the value matrix through a matrix multiplication operation (or inner product) of the attention weight and the value matrix. For example, the feature data integration module (134) can calculate a first attention value through a weighted sum of the first attention weight and the value matrix associated with the first feature data, and can calculate a second attention value through a weighted sum of the second attention weight and the value matrix associated with the second feature data. According to one embodiment, the feature data integration module (134) can adjust the dimension of the attention value through a linear projection of the calculated attention value (e.g., the first attention value or the second attention value).

[0047] Then, the feature data integration module (134) can generate integrated feature data by concatenating the calculated attention values. For example, the feature data integration module (134) can generate integrated feature data by concatenating the first attention value and the second attention value.

[0048] According to one embodiment, the feature data integration module (134) can apply the integrated feature data to the Transformer model. For example, the feature data integration module (134) can extract key information (or important information) once more from the integrated feature data, i.e., the connected attention value, through additional operation of the Transformer. Here, the Transformer model may be a model implemented using an attention mechanism without using a recurrent neural network (RNN), although it has a structure of an encoder that compresses an input sequence into a single vector representation and a decoder that generates an output sequence through the vector representation. Since the Transformer does not receive input values ​​sequentially according to the position of the input sequence, it can perform positional encoding to add positional information of the input values ​​to the feature data (e.g., embedding vector). Additionally, the attention mechanism used in the Transformer may include a self-attention mechanism. The self-attention mechanism is similar in structure to the cross-attention mechanism described above, but differs in that the input query matrix, key matrix, and value matrix come from the same source. For example, when the feature data integration module (134) applies integrated feature data to the transformer model, the operation of the transformer may include a self-attention operation, and the query matrix, key matrix, and value matrix used in the self-attention operation may be generated from the integrated feature data.

[0049] The solar power generation prediction module (136) can generate a predicted value of solar power generation based on integrated characteristic data. The process by which the solar power generation prediction module (136) generates a predicted value of solar power generation may be referred to as a multi-layer perceptron process. The multi-layer perceptron process may include a process of generating at least one predicted value of solar power generation having a specified time interval (e.g., prediction interval) within a specified time range (e.g., prediction period or prediction day) to be predicted using a multi-layer perceptron model utilizing integrated characteristic data. For example, the solar power generation prediction module (136) can generate predicted values ​​of solar power generation corresponding to one-hour intervals during the 24 hours of day t, which is the prediction target day (e.g., a total of 24 predicted values). Here, the multi-layer perceptron model may be an artificial neural network model having two or more hidden layers between the input layer and the output layer, and may be a model capable of identifying linear and non-linear relationships of data. According to one embodiment, the solar power generation prediction module (136) may use a leaky ReLU (Rectified Linear Unit) function as the activation function of a multilayer perceptron.

[0050] As described above, an electronic device (100) for predicting solar power generation according to one embodiment of the present disclosure can acquire a plurality of characteristic data based on a plurality of weather forecast data, acquire integrated characteristic data by integrating the plurality of characteristic data based on similarity between the plurality of characteristic data, and generate a predicted value of solar power generation based on the integrated characteristic data. For example, the electronic device (100) can use various information, namely multiple weather forecast data, as input values ​​to extract key information (or important information) and identify their relationships to integrate the weather forecast data into an embedding value (or embedding vector). Additionally, the electronic device (100) can utilize the weather forecast data for both the training and testing of the model to reduce dependence on the model forecast. For example, the electronic device (100) can train the model using the weather forecast data throughout the entire process to identify patterns between the weather forecast data and the solar power generation amount.

[0051] In the present disclosure, weather forecast data used as source data (data used as input values ​​for a prediction model) to predict solar power generation may include various information / data related to weather. According to various embodiments, the weather forecast data may include not only forecast data obtained through a forecast model, but also various forms of prediction / observation data or information related to weather. For example, the weather forecast data may include satellite imagery, Social Network Service (SNS) data generated by electronic devices located in areas adjacent to solar power plants, or web portal search data.

[0052] FIG. 2 is a schematic diagram showing a configuration in which an information processing system (230) is connected to communicate with a plurality of user terminals (210_1, 210_2, 210_3) in relation to data processing according to one embodiment of the present disclosure. The information processing system (230) may include system(s) capable of providing a data processing service (e.g., a solar power generation prediction-based service). In one embodiment, the information processing system (230) may include one or more server devices and / or databases capable of storing, providing, and executing computer-executable programs (e.g., downloadable applications) and data related to the data processing service, or one or more distributed computing devices and / or distributed databases based on cloud computing services. For example, the information processing system (230) may include separate systems (e.g., servers) for the data processing service.

[0053] Data processing services, etc. provided by the information processing system (230) can be provided to the user through a data processing application, a web browser application, etc. installed on each of the multiple user terminals (210_1, 210_2, 210_3).

[0054] Multiple user terminals (210_1, 210_2, 210_3) can communicate with an information processing system (230) through a network (220). The network (220) can be configured to enable communication between the multiple user terminals (210_1, 210_2, 210_3) and the information processing system (230). Depending on the installation environment, the network (220) may be configured as a wired network such as Ethernet, Power Line Communication, telephone line communication devices and RS-serial communication, a mobile communication network, a Wireless LAN (WLAN), Wi-Fi, Bluetooth and ZigBee, or a combination thereof. The communication method is not limited and may include not only communication methods utilizing communication networks that the network (220) may include (e.g., mobile communication network, wired internet, wireless internet, broadcasting network, satellite network, etc.) but also short-range wireless communication between user terminals (210_1, 210_2, 210_3).

[0055] For example, multiple user terminals (210_1, 210_2, 210_3) can transmit a data processing request and a command associated with a user request for data processing to an information processing system (230) through a network (220), and the information processing system (230) can receive this.

[0056] In FIG. 2, a mobile phone terminal (210_1), a tablet terminal (210_2), and a PC terminal (210_3) are shown as examples of user terminals, but are not limited thereto, and the user terminals (210_1, 210_2, 210_3) may be any computing device capable of wired and / or wireless communication and capable of installing and running data processing applications, etc. For example, user terminals may include smartphones, mobile phones, navigation systems, computers, laptops, digital broadcasting terminals, PDAs (Personal Digital Assistants), PMPs (Portable Multimedia Players), tablet PCs, game consoles, wearable devices, IoT (Internet of Things) devices, VR (Virtual Reality) devices, AR (Augmented Reality) devices, etc. Additionally, FIG. 2 illustrates three user terminals (210_1, 210_2, 210_3) communicating with an information processing system (230) through a network (220), but is not limited thereto, and may be configured so that a different number of user terminals communicate with an information processing system (230) through a network (220).

[0057] When the information processing system (230) provides a solar power generation prediction-based service, it may receive weather forecast data from user terminals (210_1, 210_2, 210_3). In this case, the information processing system (230) may generate a predicted value of solar power generation based on multiple weather forecast data. Afterward, the information processing system (230) may transmit the generated predicted value of solar power generation to the user terminals (210_1, 210_2, 210_3).

[0058] FIG. 3 is a block diagram showing the internal configuration of a user terminal (210) and an information processing system (230) according to one embodiment of the present disclosure. The user terminal (210) may refer to any computing device capable of executing data processing applications, etc., and capable of wired / wireless communication, and may include, for example, the mobile phone terminal (210_1), tablet terminal (210_2), PC terminal (210_3) of FIG. 2. As illustrated, the user terminal (210) may include a memory (312), a processor (314), a communication module (316), and an input / output interface (318). Similarly, the information processing system (230) may include a memory (332), a processor (334), a communication module (336), and an input / output interface (338). As illustrated in FIG. 3, the user terminal (210) and the information processing system (230) may be configured to communicate information and / or data through the network (220) using their respective communication modules (316, 336). Additionally, the input / output device (320) may be configured to input information and / or data to the user terminal (210) or output information and / or data generated from the user terminal (210) through the input / output interface (318).

[0059] The memory (312, 332) may include any non-transient computer-readable recording medium. According to one embodiment, the memory (312, 332) may include a non-perishable permanent mass storage device such as ROM (read-only memory), a disk drive, an SSD (solid-state drive), or a flash memory. As another example, a non-perishable permanent mass storage device such as ROM, an SSD, a flash memory, or a disk drive may be included in the user terminal (210) or the information processing system (230) as a separate permanent storage device distinct from the memory. Additionally, the memory (312, 332) may store an operating system and at least one program code (e.g., code for an application associated with a data processing service).

[0060] These software components may be loaded from a computer-readable recording medium separate from memory (312, 332). This separate computer-readable recording medium may include a recording medium that can be directly connected to the user terminal (210) and the information processing system (230), for example, a computer-readable recording medium such as a floppy drive, disk, tape, DVD / CD-ROM drive, or memory card. As another example, the software components may be loaded into memory (312, 332) via a communication module (316, 336) rather than a computer-readable recording medium. For example, at least one program may be loaded into memory (312, 332) based on a computer program (e.g., an application associated with a data processing service, etc.) that is installed by files provided through a network (220) by developers or a file distribution system that distributes installation files for the application.

[0061] The processor (314, 334) may be configured to process instructions of a computer program by performing basic arithmetic, logic, and input / output operations. Instructions may be provided to the processor (314, 334) by memory (312, 332) or a communication module (316, 336). For example, the processor (314, 334) may be configured to execute instructions received according to program code stored in a recording device such as memory (312, 332).

[0062] The communication module (316, 336) may provide a configuration or function for the user terminal (210) and the information processing system (230) to communicate with each other via the network (220), and may provide a configuration or function for the user terminal (210) and / or the information processing system (230) to communicate with another user terminal or another system (e.g., a separate cloud system). For example, a request or data (e.g., a data processing request or data, etc.) generated by the processor (314) of the user terminal (210) according to program code stored in a recording device such as memory (312) may be transmitted to the information processing system (230) via the network (220) under the control of the communication module (316). Conversely, a control signal or command provided under the control of the processor (334) of the information processing system (230) can be received by the user terminal (210) through the communication module (336) and the network (220) via the communication module (316) of the user terminal (210).

[0063] The input / output interface (318) may be a means for interfacing with an input / output device (320). As an example, the input device may include a device such as a camera including an audio sensor and / or an image sensor, a keyboard, a microphone, or a mouse, and the output device may include a device such as a display, a speaker, or a haptic feedback device. As another example, the input / output interface (318) may be a means for interfacing with a device in which the configuration or function for performing input and output is integrated into one, such as a touchscreen. Although the input / output device (320) is depicted in FIG. 3 as not being included in the user terminal (210), it is not limited thereto and may be configured as a single device with the user terminal (210). Additionally, the input / output interface (338) of the information processing system (230) may be a means for interfacing with a device (not shown) for input or output that is connected to the information processing system (230) or that the information processing system (230) may include. In FIG. 3, the input / output interface (318, 338) is shown as an element configured separately from the processor (314, 334), but is not limited thereto, and the input / output interface (318, 338) may be configured to be included in the processor (314, 334).

[0064] The user terminal (210) and the information processing system (230) may include more components than those of FIG. 3. However, it is not necessary to clearly illustrate most of the conventional technical components. In one embodiment, the user terminal (210) may be implemented to include at least some of the input / output devices (320) described above. Additionally, the user terminal (210) may further include other components such as a transceiver, a GPS (Global Positioning System) module, a camera, various sensors, a database, etc. For example, if the user terminal (210) is a smartphone, it may include components that are generally included in a smartphone, and may be implemented to include various components such as an accelerometer, a gyroscope, a microphone module, a camera module, various physical buttons, buttons using a touch panel, input / output ports, and a vibrator for vibration.

[0065] According to one embodiment, the processor (314) of the user terminal (210) may be configured to operate a data processing application or a web browser application that provides data processing services. At this time, program code associated with the application may be loaded into the memory (312) of the user terminal (210). While the application is operating, the processor (314) of the user terminal (210) may receive information and / or data provided from an input / output device (320) through an input / output interface (318) or receive information and / or data from an information processing system (230) through a communication module (316), and may process the received information and / or data and store it in the memory (312). Additionally, such information and / or data may be provided to the information processing system (230) through the communication module (316).

[0066] While the data processing application is in operation, the processor (314) may receive voice data, text, images, videos, etc., that are input or selected through an input device such as a touch screen, keyboard, audio sensor and / or image sensor, camera, microphone, etc., connected to the input / output interface (318), and may store the received voice data, text, images and / or videos, etc. in memory (312) or provide them to an information processing system (230) through a communication module (316) and a network (220). In one embodiment, the processor (314) may receive user input input through an input device and provide data / requests corresponding to the received user input to an information processing system (230) through a network (220) and a communication module (316).

[0067] The processor (314) of the user terminal (210) can transmit information and / or data to an input / output device (320) through an input / output interface (318) and output it. For example, the processor (314) of the user terminal (210) can output the processed information and / or data through an output device (320), such as a display output device (e.g., touch screen, display, etc.) or a voice output device (e.g., speaker).

[0068] The processor (334) of the information processing system (230) may be configured to manage, process, and / or store information and / or data received from a plurality of user terminals (210) and / or a plurality of external systems. The information and / or data processed by the processor (334) may be provided to the user terminals (210) through a communication module (336) and a network (220).

[0069] FIG. 4 is a diagram illustrating a method for calculating a predicted value of solar power generation according to an embodiment of the present disclosure. Referring to FIG. 4, an electronic device for predicting solar power generation (e.g., the electronic device (100) of FIG. 1) can generate a predicted value (490) of solar power generation by using a plurality of weather forecast data (e.g., a first weather forecast data (412) and a second weather forecast data (414)). For convenience of explanation, the following description describes the case where the plurality of weather forecast data consists of two weather forecast data, but the number of weather forecast data is not limited thereto. For example, the number of weather forecast data used may be three or more.

[0070] Looking at the process of generating a predicted value (490) of solar power generation, first, an electronic device (or at least one processor of the electronic device) can obtain first characteristic data (432) and second characteristic data (434), respectively, based on first weather forecast data (412) and second weather forecast data (414). In this process, if the time range (e.g., forecast period) of the first weather forecast data (412) and the second weather forecast data (414) is different, the electronic device can extract data corresponding to the overlapping time range in the first weather forecast data (412) and the second weather forecast data (414). Additionally, the electronic device may obtain first characteristic data (432) from first weather forecast data (412) (or extracted first weather forecast data) using an attention-based LSTM (422), and obtain second characteristic data (434) from second weather forecast data (414) (or extracted second weather forecast data) using an attention-based LSTM (424). According to one embodiment, the first characteristic data (432) may include a first embedding vector, and the second characteristic data (434) may include a second embedding vector. For example, an electronic device can use an attention-based LSTM (422, 424) to generate embedding vectors (e.g., a first embedding vector and a second embedding vector) having the same length (e.g., dimension) from each weather forecast data (412, 414) while preserving the time series information of each weather forecast data (412, 414).

[0071] Then, the electronic device can obtain integrated characteristic data (460) by integrating the first characteristic data (432) and the second characteristic data (434) based on the similarity between the first characteristic data (432) and the second characteristic data (434). In this process, the electronic device can generate a query matrix, a key matrix, and a value matrix corresponding to each of the first characteristic data (432) and the second characteristic data (434). For example, the electronic device can generate a query matrix, a key matrix, and a value matrix associated with the first characteristic data (432) (e.g., a first embedding vector), and generate a query matrix, a key matrix, and a value matrix associated with the second characteristic data (434) (e.g., a second embedding vector).

[0072] Subsequently, the electronic device can obtain a first attention value (452) and a second attention value (454) based on the first characteristic data (432) and the second characteristic data (434) through a cross attention (442, 444) mechanism. For example, the electronic device may use a query associated with the second characteristic data (434) in the cross attention (442) to calculate the first attention value (452), and a query associated with the first characteristic data (432) in the cross attention (444) to calculate the second attention value (454). Additionally, the electronic device may generate integrated characteristic data (460) by concatenating the first attention value (452) and the second attention value (454). The cross attention (442, 444) mechanism will be described in detail with reference to FIG. 5.

[0073] Then, the electronic device can apply the integrated characteristic data (460) to the transformer (470) model. For example, the electronic device can extract key information (or important information) once more from the integrated characteristic data (460) through the operation of an additional transformer (470). Here, the attention mechanism used in the transformer (470) may include a self-attention mechanism. The self-attention mechanism is similar in structure to the cross-attention (442, 444) mechanism, but differs in that the input query matrix, key matrix, and value matrix come from the same source (e.g., integrated characteristic data (460)).

[0074] Then, the electronic device can generate a predicted value (490) of solar power generation based on the integrated characteristic data (460). For example, the electronic device can generate a predicted value (490) of solar power generation based on the integrated characteristic data (460) additionally applied to the transformer (470) model. According to one embodiment, the electronic device can obtain a predicted value (490) of solar power generation as an output value by inputting the integrated characteristic data (460) as an input value to a multilayer perceptron (480) model. Here, the electronic device may use a leaky ReLU function as the activation function of the multilayer perceptron. The multilayer perceptron (480) model will be described in detail with reference to FIG. 6.

[0075] FIG. 5 is a diagram illustrating a method for calculating an attention value according to an embodiment of the present disclosure. Referring to FIG. 5, an electronic device for predicting solar power generation (e.g., the electronic device (100) of FIG. 1) can calculate an attention value (580) through an attention mechanism. Here, the attention mechanism can support the model in focusing on important information by dynamically adjusting the weight of the output for each element of the input. The attention mechanism may include a self-attention mechanism, a cross-attention mechanism, etc.

[0076] Looking at the process of calculating an attention value (580) through an attention mechanism, first, the electronic device can obtain a query matrix (512), a key matrix (514), and a value matrix (516) based on characteristic data (e.g., an embedding vector). For example, the electronic device can generate a query matrix, a key matrix, and a value matrix associated with first characteristic data (e.g., a first embedding vector) corresponding to first weather forecast data, and generate a query matrix, a key matrix, and a value matrix associated with second characteristic data (e.g., a second embedding vector) corresponding to second weather forecast data.

[0077] Then, the electronic device can perform linear projection (522, 524, 526) operations on the query matrix (512), key matrix (514), and value matrix (516). For example, the electronic device can adjust the dimensions of the query matrix (512), key matrix (514), and value matrix (516) through the linear projection (522, 524, 526).

[0078] Then, the electronic device can calculate an attention score representing the similarity between the query matrix (512) and the key matrix (514) based on the query matrix (512) and the key matrix (514). For example, the electronic device can calculate the attention score through a matrix multiplication operation (530) (or inner product) on the query matrix (512) and the key matrix (514). The query matrix (512) used in this process may differ between the self-attention mechanism and the cross-attention mechanism. For example, in the self-attention mechanism, the query matrix (512) and the key matrix (514) are obtained from the same source (e.g., the same embedding vector), whereas in the cross-attention mechanism, the query matrix (512) and the key matrix (514) may be obtained from different sources (e.g., different embedding vectors).

[0079] Then, the electronic device can process the attention score by scaling (540). For example, the electronic device can increase numerical stability and increase computational efficiency by reducing the size of the attention score through the scaling (540) operation.

[0080] Then, the electronic device can apply a softmax function (550) to the scaled attention score. For example, the electronic device can input the attention score into a softmax function to calculate the attention weight. Here, the attention weight may represent the importance of each key matrix (514) corresponding to the query matrix (512).

[0081] Then, the electronic device can calculate an attention value (580) through a matrix multiplication operation (560) (or inner product) of the attention weights and the value matrix (516). For example, the electronic device can calculate an attention value (580) through a weighted sum of the attention weights and the value matrix (516). Subsequently, the electronic device can adjust the dimension of the attention value (580) through a linear projection on the calculated attention value (580).

[0082] FIG. 6 is a diagram illustrating an artificial neural network model (600) according to one embodiment of the present disclosure. Referring to FIG. 6, the artificial neural network model (600) may represent a statistical learning algorithm implemented based on a biological neural network structure or a structure for executing said algorithm, as an example of a machine learning model in machine learning technology and cognitive science.

[0083] According to one embodiment, the artificial neural network model (600) may represent a machine learning model having problem-solving ability by learning that nodes, which are artificial neurons forming a network through the combination of synapses like a biological neural network, repeatedly adjust the weights of the synapses so that the error between the correct output corresponding to a specific input and the inferred output is reduced.

[0084] According to one embodiment, the above-described solar power generation prediction model can be generated in the form of an artificial neural network model (600). For example, the artificial neural network model (600) can receive weather forecast data for a region associated with a solar power plant and estimate the expected power generation based thereon.

[0085] The artificial neural network model (600) can be implemented as a multilayer perceptron composed of multiple layers of nodes and connections between them. The artificial neural network model (600) according to the present embodiment can be implemented using one of various artificial neural network model structures including a multilayer perceptron. The artificial neural network model (600) can be composed of an input layer (620) that receives input data (610) (or input signal) from the outside, an output layer (640) that outputs output data (650) (or output signal) corresponding to the input data (610), and n hidden layers (630_1 to 630_n) located between the input layer (620) and the output layer (640), which receive a signal from the input layer (620), extract a feature, and transmit it to the output layer (640) (where n is a positive integer). Here, the output layer (640) can receive a signal from the hidden layers (630_1 to 630_n) and output it to the outside.

[0086] The learning method of the artificial neural network model (600) may include a supervised learning method that learns to be optimized for solving a problem by inputting a correct answer teacher signal (or label) and an unsupervised learning method that does not require a teacher signal. According to one embodiment, an electronic device according to one embodiment of the present disclosure (e.g., the electronic device (100) of FIG. 1) can train the artificial neural network model (600) using a plurality of weather forecast data.

[0087] According to one embodiment, the electronic device may generate training data for training an artificial neural network model (600). For example, the electronic device may generate a training data set including a plurality of weather forecast data. Then, the electronic device may train an artificial neural network model (600) to calculate a predicted value of solar power generation based on the generated training data set.

[0088] According to one embodiment, the input variables of the artificial neural network model (600) may include a plurality of weather forecast data. When the input variables described above are input through the input layer (620), the output variable output from the output layer (640) of the artificial neural network model (600) may be a predicted value of solar power generation.

[0089] In this way, a plurality of input variables and corresponding plurality of output variables are matched to the input layer (620) and output layer (640) of the artificial neural network model (600), and the synapse values ​​between the nodes included in the input layer (620), hidden layers (630_1 to 630_n), and output layer (640) are adjusted so that the correct output corresponding to a specific input can be extracted. Through this learning process, characteristics hidden in the input variables of the artificial neural network model (600) can be identified, and the synapse values ​​(or weights) between the nodes of the artificial neural network model (600) can be adjusted so that the error between the output variable calculated based on the input variables and the target output is reduced. Additionally, the electronic device learns an algorithm that receives a plurality of weather forecast data as input, and can learn in a manner that minimizes the loss with respect to the predicted value of solar power generation (i.e., annotation information).

[0090] By using the artificial neural network model (600) trained in this way, the predicted value of solar power generation can be estimated.

[0091] FIG. 7 is a diagram illustrating a method for predicting solar power generation according to an embodiment of the present disclosure. Referring to FIG. 7, a processor (e.g., processor (130) of FIG. 1) of an electronic device (e.g., electronic device (100) of FIG. 1) that predicts solar power generation can receive a plurality of weather forecast data at step 710 (S710). For example, the processor can receive a plurality of weather forecast data from at least one external electronic device connected via a communication circuit. For another example, the processor can receive a plurality of weather forecast data from a memory (e.g., memory (110) of FIG. 1). The number of weather forecast data used to predict solar power generation in the present disclosure can be at least two (e.g., two, three, etc.), and the types thereof can also be diverse. For example, a plurality of weather forecast data may include first weather forecast data containing weather forecast values ​​for a first time interval within a specified time range and second weather forecast data containing weather forecast values ​​for a second time interval longer than the first time interval within a specified time range. According to one embodiment, the first weather forecast data may include forecast data obtained through a Local Forecast Model (LDAPS), and the second weather forecast data may include forecast data obtained through a Global Forecast Model (GDAPS).

[0092] In step 720 (S720), the processor may acquire multiple feature data. For example, the processor may acquire multiple feature data based on multiple weather forecast data. According to one embodiment, the processor may acquire embedding vectors corresponding to each of the multiple weather forecast data as multiple feature data. The embedding vectors corresponding to each of the multiple weather forecast data may include time series information possessed by the corresponding weather forecast data and may have a specified length (e.g., dimension). According to one embodiment, the processor may generate embedding vectors using an attention-based LSTM.

[0093] In step 730 (S730), the processor can obtain integrated characteristic data that integrates multiple characteristic data. For example, the processor can obtain integrated characteristic data that integrates multiple characteristic data based on similarity between multiple characteristic data. According to one embodiment, the processor can obtain integrated characteristic data that integrates multiple characteristic data using a cross-attention mechanism. For example, the processor can first generate a query matrix, a key matrix, and a value matrix associated with the first characteristic data based on the first characteristic data (e.g., a first embedding vector) corresponding to the first weather forecast data among the multiple weather forecast data, and generate a query matrix, a key matrix, and a value matrix associated with the second characteristic data based on the second characteristic data (e.g., a second embedding vector) corresponding to the second weather forecast data among the multiple weather forecast data. Then, the processor can calculate a first attention score based on a key matrix associated with the first feature data and a query matrix associated with the second feature data, and calculate a second attention score based on a key matrix associated with the second feature data and a query matrix associated with the first feature data. Then, the processor can calculate a first attention value based on the first attention score and a value matrix associated with the first feature data, and calculate a second attention value based on the second attention score and a value matrix associated with the second feature data. Then, the processor can generate integrated feature data by concatenating the first attention value and the second attention value.

[0094] According to one embodiment, the processor can apply integrated feature data to the transformer model. For example, the processor can extract key information (or important information) once more from the integrated feature data, i.e., the connected attention value, through additional transformer operations.

[0095] In step 740 (S740), the processor can generate a predicted value of solar power generation. For example, the processor can generate a predicted value of solar power generation based on integrated characteristic data. According to one embodiment, the processor can generate a predicted value of solar power generation through a multilayer perceptron model using integrated characteristic data. According to one embodiment, the processor can use a leaky ReLU function as the activation function of the multilayer perceptron.

[0096] The above-described flowchart and description are merely examples and may be implemented differently in some embodiments. For instance, in some embodiments, the order of each step may be changed, some steps may be repeated, some steps may be omitted, or some steps may be added.

[0097] The method described above may be provided as a computer program stored on a computer-readable recording medium for execution on a computer. The medium may continuously store a program executable by a computer, or temporarily store it for execution or download. Additionally, the medium may be various recording or storage means in the form of a single or multiple hardware components combined, and may not be limited to a medium directly connected to a computer system but may exist distributed over a network. Examples of media may include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and media configured to store program instructions, including ROM, RAM, and flash memory. Furthermore, other examples of media may include recording or storage media managed by app stores that distribute applications or sites and servers that supply or distribute various other software.

[0098] The methods, operations, or techniques of the present disclosure may be implemented by various means. For example, these techniques may be implemented in hardware, firmware, software, or a combination thereof. Those skilled in the art will understand that the various exemplary logical blocks, modules, circuits, and algorithmic steps described in connection with the disclosure herein may be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate such interchangeability between hardware and software, various exemplary components, blocks, modules, circuits, and steps have been generally described above in terms of their functional aspects. Whether such functions are implemented in hardware or in software depends on the design requirements imposed on the specific application and the overall system. Those skilled in the art may implement the functions described in various ways for each specific application, but such implementations should not be construed as departing from the scope of the present disclosure.

[0099] In a hardware implementation, the processing units used to perform the techniques may be implemented in one or more ASICs, DSPs, digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, electronic devices, other electronic units designed to perform the functions described in this disclosure, computers, or a combination thereof.

[0100] Accordingly, the various exemplary logic blocks, modules, and circuits described in connection with the present disclosure may be implemented or performed by any combination of general-purpose processors, DSPs, ASICs, FPGAs or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or those designed to perform the functions described herein. A general-purpose processor may be a microprocessor, but alternatively, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, for example, a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors coupled with a DSP core, or any other combination of configurations.

[0101] In firmware and / or software implementations, techniques may be implemented as instructions stored on a computer-readable medium such as random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable PROM (EEPROM), flash memory, compact disc (CD), magnetic or marked data storage device, etc. The instructions may be executable by one or more processors, and may cause the processor(s) to perform specific aspects of the functions described in this disclosure.

[0102] When implemented in software, the techniques described above may be stored on a computer-readable medium as one or more instructions or code, or transmitted through a computer-readable medium. Computer-readable media include both computer storage media and communication media, including any medium that facilitates the transmission of a computer program from one place to another. Storage media may be any available media accessible by a computer. As a non-limiting example, such computer-readable media may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium accessible by a computer that can be used to transfer or store desired program code in the form of instructions or data structures. Additionally, any connection is appropriately referred to as a computer-readable medium.

[0103] For example, if software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair cable, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, coaxial cable, fiber optic cable, twisted pair cable, digital subscriber line, or wireless technologies such as infrared, radio, and microwave are included within the definition of a medium. As used herein, disks and discs include CDs, laser discs, optical discs, DVDs (digital versatile discs), floppy disks, and Blu-ray discs, wherein disks usually play data magnetically, whereas discs play data optically using a laser. The above combinations should also be included within the scope of computer-readable media.

[0104] The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, a hard disk, a removable disk, a CD-ROM, or any other known form of storage medium. An exemplary storage medium may be connected to a processor so that the processor can read information from the storage medium or write information to the storage medium. Alternatively, the storage medium may be integrated into the processor. The processor and the storage medium may exist within an ASIC. The ASIC may exist within a user terminal. Alternatively, the processor and the storage medium may exist as separate components within the user terminal.

[0105] Although the embodiments described above have been described as utilizing aspects of the subject matter disclosed herein in one or more standalone computer systems, the present disclosure is not limited thereto and may be implemented in conjunction with any computing environment, such as a network or a distributed computing environment. Furthermore, aspects of the subject matter in the present disclosure may be implemented in a plurality of processing chips or devices, and storage may be similarly affected across a plurality of devices. Such devices may include PCs, network servers, and portable devices.

[0106] Although the present disclosure has been described in relation to some embodiments, various modifications and changes may be made without departing from the scope of the present disclosure as understood by a person skilled in the art to which the invention of the present disclosure pertains. Furthermore, such modifications and changes should be considered to fall within the scope of the claims appended to this specification. Explanation of the symbols

[0107] 100: Electronic device 110: Memory 130: Processor 132: Characteristic data acquisition module 134: Characteristic Data Integration Module 136: Solar Power Generation Prediction Module

Claims

Claim 1 In electronic devices, memory; and includes a processor operatively connected to the memory, wherein the processor receives a plurality of weather forecast data including first weather forecast data obtained through a local forecast model and second weather forecast data obtained through a global forecast model, and obtains an embedding vector corresponding to each of the plurality of weather forecast data as a plurality of feature data, wherein the embedding vector includes time series information possessed by the weather forecast data corresponding to the embedding vector among the plurality of weather forecast data and has a specified length, and the plurality of feature data includes a first feature data corresponding to the first weather forecast data and a second feature data corresponding to the second weather forecast data, and based on the embedding vector corresponding to each of the plurality of weather forecast data, generates a query matrix, a key matrix, and a value matrix corresponding to each of the plurality of feature data that are used in an attention model, and calculates a first attention score representing the similarity between a query and a key based on the key matrix associated with the first feature data and the query matrix associated with the second feature data, and the first attention score and the first An electronic device configured to calculate a first attention value based on a value matrix associated with characteristic data, calculate a second attention score representing similarity between a query and a key based on a key matrix associated with the second characteristic data and a query matrix of the first characteristic data, calculate a second attention value based on the second attention score and a value matrix associated with the second characteristic data, concatenate the first attention value and the second attention value to generate integrated characteristic data, and generate a predicted value of solar power generation based on the integrated characteristic data. Claim 2 An electronic device according to claim 1, wherein the first weather forecast data includes a weather forecast value of a first time interval within a specified time range, and the second weather forecast data includes a weather forecast value of a second time interval longer than the first time interval within the specified time range. Claim 3 delete Claim 4 An electronic device according to claim 1, wherein the processor is configured to generate the embedding vector using an attention-based LSTM (Long Short-Term Memory). Claim 5 delete Claim 6 delete Claim 7 An electronic device according to claim 1, wherein the processor is configured to apply the integrated characteristic data to a Transformer model. Claim 8 An electronic device according to claim 1, wherein the processor is configured to generate a predicted value of the solar power generation amount through a Multi-Layer Perceptron model using the integrated characteristic data. Claim 9 An electronic device according to claim 8, wherein the activation function of the multilayer perceptron includes a leaky ReLU (Rectified Linear Unit) function. Claim 10 A method for predicting solar power generation, performed by at least one processor, comprising: receiving a plurality of weather forecast data including first weather forecast data obtained through a local forecast model and second weather forecast data obtained through a global forecast model; acquiring an embedding vector corresponding to each of the plurality of weather forecast data as a plurality of characteristic data, wherein the embedding vector includes time series information possessed by the weather forecast data corresponding to the embedding vector among the plurality of weather forecast data and has a specified length, and the plurality of characteristic data includes first characteristic data corresponding to the first weather forecast data and second characteristic data corresponding to the second weather forecast data; generating a query matrix, a key matrix, and a value matrix corresponding to each of the plurality of characteristic data, which are used in an attention model, based on the embedding vector corresponding to each of the plurality of weather forecast data; calculating a first attention score representing the similarity between a query and a key based on the key matrix associated with the first characteristic data and the query matrix associated with the second characteristic data, and the first A method comprising: a step of calculating a first attention value based on an attention score and a value matrix associated with the first characteristic data; a step of calculating a second attention score representing the similarity between a query and a key based on a key matrix associated with the second characteristic data and a query matrix of the first characteristic data, and a step of calculating a second attention value based on the second attention score and a value matrix associated with the second characteristic data; a step of concatenating the first attention value and the second attention value to generate integrated characteristic data; and a step of generating a predicted value of solar power generation based on the integrated characteristic data.

Images