Power supply and demand prediction method and device, storage medium, and electronic device

Abstract

The application discloses a power supply and demand prediction method, wherein the method comprises the following steps: acquiring power dispatching data, static characteristic data and dynamic characteristic data; performing feature extraction based on the power dispatching data, determining a first vector representation corresponding to the power dispatching data, performing feature extraction based on the static characteristic data, determining a second vector representation corresponding to the static characteristic data, and performing feature extraction based on the dynamic characteristic data, and determining a third vector representation corresponding to the dynamic characteristic data; fusing the first vector representation, the second vector representation and the third vector representation to obtain a target vector representation; and based on the target vector representation, the power supply and demand of different provinces are predicted to obtain a power supply and demand prediction result. The power supply and demand can be accurately predicted.

Description

Technical Field

[0001] This application relates to the field of power system technology, and more specifically, to a power supply and demand forecasting method and apparatus, storage medium, and electronic device. Background Technology

[0002] Electricity supply and demand forecasting is the foundation of power system planning, dispatching, operation, and management. Accurate forecasting of electricity supply and demand is of great significance for ensuring the safe and economical operation of the power system. Electricity supply and demand forecasting is generally divided into two main categories: electricity consumption forecasting and electricity supply and demand balance forecasting. Electricity consumption forecasting mainly refers to the prediction of the total electricity consumption in a future period, while electricity supply and demand balance forecasting mainly refers to the prediction of the balance between electricity supply and demand in a certain period.

[0003] Electricity supply and demand are influenced by a variety of factors, including weather conditions, economic activities, and user behavior. These factors generate data from different sources and in different formats, such as weather data from meteorological stations, economic indicators from economic statistics agencies, and electricity consumption data from households and industrial users. This multi-source heterogeneity makes data integration and feature extraction a complex task.

[0004] Currently, electricity supply and demand forecasting is mainly based on a single forecasting model, such as the support vector machine model or the genetic algorithm model. However, since electricity supply and demand are affected by a variety of factors, a single forecasting model cannot provide accurate forecast results.

[0005] Therefore, how to accurately predict electricity supply and demand has become an urgent technical problem to be solved. Summary of the Invention

[0006] This application provides a method and apparatus for predicting power supply and demand, a storage medium, and an electronic device to at least solve the problem of low accuracy in power supply and demand prediction in related technologies.

[0007] According to one embodiment of this application, a method for predicting power supply and demand is provided, comprising: acquiring power dispatch data, static feature data, and dynamic feature data; wherein, the power dispatch data is data used to characterize the power allocation situation among different provinces; the static feature data is data used to characterize the energy structure of target cities associated with the different provinces; the dynamic feature data is data used to characterize the regional electricity consumption of different parks in the target city; performing feature extraction based on the power dispatch data to determine a first vector representation corresponding to the power dispatch data; performing feature extraction based on the static feature data to determine a second vector representation corresponding to the static feature data; and performing feature extraction based on the dynamic feature data to determine a third vector representation corresponding to the dynamic feature data; fusing the first vector representation, the second vector representation, and the third vector representation to obtain a target vector representation; and predicting the power supply and demand of the different provinces based on the target vector representation to obtain a power supply and demand prediction result.

[0008] In an exemplary embodiment, the step of extracting features based on the power dispatch data to determine a first vector representation corresponding to the power dispatch data includes: cleaning the power dispatch data to obtain cleaned power dispatch data; performing time series analysis on the cleaned power dispatch data to determine the temporal characteristics of the power dispatch data changing over time; performing spatial analysis on the cleaned power dispatch data to obtain the geographical distribution characteristics corresponding to the power dispatch data; fusing the temporal characteristics, the geographical distribution characteristics, and the external environment characteristics to obtain a first target feature; and performing vector representation on the first target feature to obtain a first vector representation.

[0009] In an exemplary embodiment, the step of extracting features based on the static feature data and determining a second vector representation corresponding to the static feature data includes: cleaning the static feature data to obtain cleaned static feature data; extracting energy structure information from the cleaned static feature data and calculating an energy diversity index based on the energy structure information; wherein the energy diversity index is a parameter used to characterize the degree of energy diversity; and performing feature encoding on the energy diversity index to determine a second vector representation corresponding to the static feature data.

[0010] In an exemplary embodiment, the step of extracting features based on the dynamic feature data and determining the third vector representation corresponding to the dynamic feature data includes: decomposing the dynamic feature data to obtain trend terms, seasonality terms, and noise terms; and encoding the trend terms, seasonality terms, and noise terms respectively to obtain the third vector representation corresponding to the dynamic feature data.

[0011] In an exemplary embodiment, fusing the first vector representation, the second vector representation, and the third vector representation to obtain a target vector representation includes: concatenating the first vector representation, the second vector representation, and the third vector representation in a preset concatenation order to obtain a long vector containing all feature information; and obtaining the target vector representation based on the long vector.

[0012] In an exemplary embodiment, the step of predicting the power supply and demand of different provinces based on the target vector representation to obtain power supply and demand prediction results includes: inputting the target vector representation into a pre-trained power supply and demand prediction model, predicting the power supply and demand of different provinces based on the power supply and demand prediction model to obtain power supply and demand prediction results, wherein the power supply and demand prediction model is trained by machine learning using multiple sets of data, and each set of data includes: historical power dispatch data with added weights, historical static feature data, and historical dynamic feature data.

[0013] According to another embodiment of this application, a power supply and demand forecasting device is also provided, comprising: an acquisition module, configured to acquire power dispatch data, static feature data, and dynamic feature data; wherein the power dispatch data is data used to characterize the power allocation situation among different provinces; the static feature data is data used to characterize the energy structure of target cities associated with the different provinces; and the dynamic feature data is data used to characterize the regional electricity consumption of different parks in the target city; a determination module, configured to perform feature extraction based on the power dispatch data to determine a first vector representation corresponding to the power dispatch data, perform feature extraction based on the static feature data to determine a second vector representation corresponding to the static feature data, and perform feature extraction based on the dynamic feature data to determine a third vector representation corresponding to the dynamic feature data; a fusion module, configured to fuse the first vector representation, the second vector representation, and the third vector representation to obtain a target vector representation; and a prediction module, configured to predict the power supply and demand of the different provinces based on the target vector representation to obtain a power supply and demand forecasting result.

[0014] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, wherein a computer program is stored in the computer program, and the computer program is configured to execute the above-described method at runtime.

[0015] According to another aspect of the embodiments of this application, an electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the above-described method through the computer program.

[0016] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0017] In this embodiment, power dispatch data, static feature data, and dynamic feature data are acquired. A first vector representation of the power dispatch data is determined, a second vector representation of the static feature data is determined, and a third vector representation corresponding to the dynamic feature data is determined. These three vector representations are then merged into a single target vector representation, integrating the macroscopic characteristics of power dispatch, the energy structure background of the target city, and the electricity consumption situation at the park level, thus achieving multi-level power supply and demand forecasting from macro to micro perspectives. Using the target vector representation for power supply and demand forecasting effectively improves forecast accuracy. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0019] Figure 1 This is a hardware structure block diagram of a computer terminal for a power supply and demand forecasting method according to an embodiment of this application;

[0020] Figure 2 This is a flowchart of the generation method of the power supply and demand forecasting method according to an embodiment of this application;

[0021] Figure 3 This is a structural block diagram of a power supply and demand forecasting device according to an embodiment of this application. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus; "a plurality" means two or more.

[0024] The methods and embodiments provided in this application can be executed on a computer terminal or similar computing device, cloud platform, independent physical server, or software platform, wherein the aforementioned software platform runs through one or more servers. Taking running on a computer terminal as an example, Figure 1 This is a hardware structure block diagram of a power supply and demand forecasting method according to an embodiment of this application. For example... Figure 1 As shown, a computer terminal may include one or more ( Figure 1 Only one processor 102 and a memory 103 for storing data are shown in the diagram. In one exemplary embodiment, the computer terminal may further include a transmission device 106 for communication functions and an input / output device 108. The processor 102 may include, but is not limited to, a processing device such as a microprocessor (MCU) or a programmable logic device (FPGA). Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the computer terminal described above. For example, the computer terminal may also include components that are more complex than those described above. Figure 1 The more or fewer components shown, or having the same Figure 1 Equivalent functions or ratios shown Figure 1 The functions shown have more different configurations.

[0025] The memory 103 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the method in this embodiment. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 103, thus implementing the above-described method. The memory 103 may include high-speed random access memory and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 103 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to a computer terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0026] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider for the computer terminal. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0027] This embodiment provides a method for predicting power supply and demand, which is applied to the aforementioned computer device. Figure 2 This is a flowchart of a power supply and demand forecasting method according to an embodiment of this application, which includes the following steps:

[0028] Step S202: Obtain power dispatch data, static feature data, and dynamic feature data; wherein, the power dispatch data is used to characterize the power allocation situation between different provinces; the static feature data is used to characterize the energy structure of target cities associated with the different provinces; and the dynamic feature data is used to characterize the regional electricity consumption of different parks in the target cities.

[0029] Step S204: Based on the power dispatch data, perform feature extraction to determine a first vector representation corresponding to the power dispatch data; based on the static feature data, perform feature extraction to determine a second vector representation corresponding to the static feature data; and based on the dynamic feature data, perform feature extraction to determine a third vector representation corresponding to the dynamic feature data.

[0030] Step S206: Merge the first vector representation, the second vector representation, and the third vector representation to obtain the target vector representation;

[0031] Step S208: Based on the target vector representation, predict the power supply and demand of the different provinces to obtain the power supply and demand prediction results.

[0032] In this embodiment, power dispatch data, static feature data, and dynamic feature data are acquired. A first vector representation of the power dispatch data is determined, a second vector representation of the static feature data is determined, and a third vector representation corresponding to the dynamic feature data is determined. These three vector representations are then merged into a single target vector representation, integrating the macroscopic characteristics of power dispatch, the energy structure background of the target city, and the electricity consumption situation at the park level, thus achieving multi-level power supply and demand forecasting from macro to micro perspectives. Using the target vector representation for power supply and demand forecasting effectively improves forecast accuracy.

[0033] In an exemplary embodiment, the step of extracting features based on the power dispatch data to determine a first vector representation corresponding to the power dispatch data includes: cleaning the power dispatch data to obtain cleaned power dispatch data; performing time series analysis on the cleaned power dispatch data to determine the temporal characteristics of the power dispatch data changing over time; performing spatial analysis on the cleaned power dispatch data to obtain the geographical distribution characteristics corresponding to the power dispatch data; fusing the temporal characteristics, the geographical distribution characteristics, and the external environment characteristics to obtain a first target feature; and performing vector representation on the first target feature to obtain a first vector representation.

[0034] It should be noted that power dispatch data typically contains diverse information such as inter-provincial power transmission, power generation plans, and dispatch instructions, and may contain issues such as missing values, outliers, and duplicate records. Data cleaning is the foundation for ensuring the accuracy of subsequent analysis.

[0035] Data cleaning can include: missing value handling: filling in missing values ​​through interpolation (such as time series interpolation) or prediction methods based on historical data; outlier detection: identifying and handling outliers using machine learning techniques (such as Isolation Forest); and data consistency checks: ensuring that the timestamps, units, and other information of the data are consistent, eliminating duplicate records, and maintaining the integrity of the dataset.

[0036] Time series analysis refers to identifying the patterns and trends in power dispatch data over time, providing time-dimensional feature support for prediction models.

[0037] Spatial analysis refers to the distribution characteristics of power dispatch data in geographic space, providing information at the regional level. When conducting spatial analysis, geographical location information (such as provinces) in the dispatch data can be converted into geographic coordinates to facilitate spatial analysis.

[0038] External environmental feature integration can include external data such as weather forecasts, holiday information, and market activities, which directly impact power dispatch. Correlation analysis and principal component analysis (PCA) are used to select the most relevant features, reducing computational complexity and avoiding overfitting. The first target feature set encompasses the temporal patterns, geographical distribution characteristics, and influence of external environmental factors in power dispatch, forming a multi-level feature description.

[0039] In an exemplary embodiment, the step of extracting features based on the static feature data and determining the second vector representation corresponding to the static feature data includes: cleaning the static feature data to obtain cleaned static feature data; extracting energy structure information from the cleaned static feature data and calculating an energy diversity index based on the energy structure information; wherein the energy diversity index is a parameter used to characterize the degree of energy diversity; and performing feature encoding on the energy diversity index to determine the second vector representation corresponding to the static feature data.

[0040] Among them, static characteristic data, such as energy structure, geographical location, population density, and industrial structure, reflect the inherent attributes of a city and have a long-term and stable impact on electricity demand.

[0041] Data processing may include: Removing duplicate data: Checking for duplicate records in static feature data and removing duplicates based on unique identifiers (such as city ID, year, etc.). Imputing missing values: For missing static feature data, using appropriate imputation strategies, such as imputing based on the average of data from adjacent years. Standardizing data format: Converting all static feature data to a consistent format and units for easier model understanding and processing.

[0042] Energy structure information is used to analyze the energy structure of a target city, identifying the proportion and diversity of different energy types (such as coal, natural gas, hydropower, wind power, solar energy, etc.) in the energy supply. Information about energy supply is extracted from static feature datasets, which may include the annual production, power generation, or consumption of various energy sources. Energy classification: Data is categorized according to energy source and type, such as renewable energy, fossil fuels, etc.

[0043] The energy diversity index is used to quantify the diversity of the energy structure and assess the stability and sustainability of energy supply. A higher energy diversity index indicates a more abundant energy structure and lower dependence risk. The calculated energy diversity index is standardized to obtain a vector representation of the energy diversity index.

[0044] In the above embodiments, the second vector representation based on static feature data extraction, through data cleaning, energy structure information extraction, energy diversity index calculation, and feature encoding, can comprehensively reflect the inherent characteristics of the target city in terms of energy supply. This vector representation not only covers quantitative indicators of energy diversity but may also include other static features related to energy structure, providing a stable and rich long-term background information for electricity supply and demand models. This helps improve the accuracy and reliability of predictions, especially in assessing long-term energy planning and electricity demand elasticity.

[0045] In an exemplary embodiment, the step of extracting features based on the dynamic feature data and determining the third vector representation corresponding to the dynamic feature data includes: decomposing the dynamic feature data to obtain trend terms, seasonality terms, and noise terms; and encoding the trend terms, seasonality terms, and noise terms respectively to obtain the third vector representation corresponding to the dynamic feature data.

[0046] Understandably, data decomposition breaks down time series data with dynamic characteristics into trend, seasonality, and noise components. The trend component reflects the long-term trend of the time series, such as the year-on-year increase or decrease in electricity demand. The seasonality component captures the periodic changes in the data, such as the periodic fluctuations in electricity demand throughout the year or the seasonal patterns of power generation with weather variations. The noise component isolates the random fluctuations in the time series, i.e., the portion that cannot be explained by trend and seasonality, typically representing short-term or sudden fluctuations.

[0047] The decomposed trend, seasonality, and noise terms are converted into numerical vector representations, resulting in a third vector representation. Trend term encoding: Methods such as linear regression and multinomial fitting can be used to encode the trend term, converting trend changes into numerical representations. Seasonal term encoding: Techniques such as Fourier transform are used to convert seasonal patterns into frequency domain representations, or seasonal factors (such as months in a year or days of the week) are directly used for encoding. Noise term encoding: The noise term is statistically described, such as by calculating standard deviation and variance, to quantify the degree of random fluctuation.

[0048] In the above embodiments, features reflecting long-term trends, seasonal patterns, and random fluctuations in time series data can be extracted from dynamic feature data and converted into numerical vector representations. This process not only helps the model model the complexity of time series data but also reduces the impact of noise on prediction results to some extent, thereby improving the accuracy and stability of power supply and demand forecasting. The generation of the third vector representation enables a refined description of the power system's operating state and is an important component of the power supply and demand forecasting model input.

[0049] In an exemplary embodiment, fusing the first vector representation, the second vector representation, and the third vector representation to obtain the target vector representation includes: concatenating the first vector representation, the second vector representation, and the third vector representation in a preset concatenation order to obtain a long vector containing all feature information; and obtaining the target vector representation based on the long vector.

[0050] Specifically, the first vector representation, the second vector representation, and the third vector representation extracted from power dispatch data, static feature data, and dynamic feature data are integrated to form a long vector containing all key information.

[0051] The three vector representations are concatenated in a pre-defined order (e.g., power dispatch data first, then static feature data, and finally dynamic feature data). This order reflects the logical flow of feature extraction, from macro to micro, and from long-term to short-term. Dimensional alignment: Before concatenation, ensure that the dimensions of all vector representations are compatible. If inconsistencies exist, add zero vectors or perform feature dimensionality reduction to enable concatenation. Data type conversion: If the data types in the vector representations are inconsistent (e.g., one is an integer, the other a floating-point number), data type conversion is required to ensure consistency of data type within the concatenated long vector. Finally, a long vector containing features from power dispatch data, static feature data, and dynamic feature data is obtained, integrating multi-level information from macro to micro perspectives.

[0052] The concatenated long vector is further processed to generate a target vector representation, thereby optimizing the model input and improving prediction efficiency and accuracy.

[0053] In an exemplary embodiment, the step of predicting the power supply and demand of different provinces based on the target vector representation to obtain power supply and demand prediction results includes: inputting the target vector representation into a pre-trained power supply and demand prediction model, predicting the power supply and demand of different provinces based on the power supply and demand prediction model to obtain power supply and demand prediction results, wherein the power supply and demand prediction model is trained by machine learning using multiple sets of data, and each set of data includes: historical power dispatch data with added weights, historical static feature data, and historical dynamic feature data.

[0054] It should be noted that when training the power supply and demand forecasting model, a suitable machine learning or deep learning model can be selected. For example, Long Short-Term Memory (LSTM), Spatiotemporal Fusion Transformer (TFT), or other sequence forecasting models can be used. Data weighting and integration: Historical power dispatch data, historical static feature data, and historical dynamic feature data are weighted to ensure their relative importance is reflected in model training. The weighting strategy can be based on domain knowledge or automatically determined through model learning. Training data preparation: Multiple sets of historical data are used for model training. Each set of data includes at least weighted historical power dispatch data, historical static feature data, and historical dynamic feature data to ensure the model can learn power supply and demand patterns from different dimensions. Model training: The prepared dataset is trained using machine learning algorithms to optimize model parameters, minimizing prediction error and improving prediction accuracy. This results in a forecasting model that can accurately predict power supply and demand based on the target vector representation.

[0055] A pre-trained power supply and demand forecasting model is used to predict the power supply and demand situation in different provinces based on a target vector representation. The target vector representation is used as the model input. This target vector representation is an optimized vector representation that integrates power dispatch data, static feature data, and dynamic feature data. Model Prediction: The target vector representation is input into the power supply and demand forecasting model, which then predicts the power supply and demand in different provinces based on the patterns it has learned. Result Interpretation: The model output may be one or more numerical values, representing the predicted power demand, supply, or supply-demand ratio, etc. Depending on the specific task requirements, these prediction results may need further interpretation and transformation to facilitate understanding and decision-making by power dispatchers.

[0056] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of this application.

[0057] This application also provides a structural block diagram of a power supply and demand forecasting device. Figure 3 This is a structural block diagram of a power supply and demand forecasting device according to an embodiment of this application; as shown below. Figure 3 As shown, it includes:

[0058] The acquisition module 32 is used to acquire power dispatch data, static feature data, and dynamic feature data; wherein, the power dispatch data is used to characterize the power allocation situation between different provinces; the static feature data is used to characterize the energy structure of target cities associated with the different provinces; and the dynamic feature data is used to characterize the regional electricity consumption of different parks in the target cities.

[0059] The determining module 34 is used to perform feature extraction based on the power dispatch data to determine a first vector representation corresponding to the power dispatch data, perform feature extraction based on the static feature data to determine a second vector representation corresponding to the static feature data, and perform feature extraction based on the dynamic feature data to determine a third vector representation corresponding to the dynamic feature data.

[0060] Fusion module 36 is used to fuse the first vector representation, the second vector representation and the third vector representation to obtain the target vector representation;

[0061] The prediction module 38 is used to predict the power supply and demand of different provinces based on the target vector representation, and obtain the power supply and demand prediction results.

[0062] The aforementioned device acquires power dispatch data, static feature data, and dynamic feature data. A first vector representation of the power dispatch data is determined, a second vector representation of the static feature data is determined, and a third vector representation corresponding to the dynamic feature data is determined. These three vector representations are then merged into a single target vector representation, integrating the macroscopic characteristics of power allocation, the energy structure background of the target city, and the electricity consumption situation at the park level. This achieves multi-level power supply and demand forecasting from macro to micro perspectives. Utilizing the target vector representation for power supply and demand forecasting effectively improves forecast accuracy.

[0063] In an exemplary embodiment, the determining module 34 is further configured to clean the power dispatch data to obtain cleaned power dispatch data; perform time series analysis on the cleaned power dispatch data to determine the temporal characteristics of the power dispatch data changing over time; and perform spatial analysis on the cleaned power dispatch data to obtain the geographical distribution characteristics corresponding to the power dispatch data; fuse the temporal characteristics, the geographical distribution characteristics, and the external environment characteristics to obtain a first target feature; and perform vector representation on the first target feature to obtain a first vector representation.

[0064] In an exemplary embodiment, the determining module 34 is further configured to clean the static feature data to obtain cleaned static feature data; extract energy structure information from the cleaned static feature data, and calculate an energy diversity index based on the energy structure information; wherein the energy diversity index is a parameter used to characterize the degree of energy diversity; and perform feature encoding on the energy diversity index to determine a second vector representation corresponding to the static feature data.

[0065] In an exemplary embodiment, the determining module 34 is further configured to decompose the dynamic feature data to obtain trend items, seasonal items, and noise items; and encode the trend items, seasonal items, and noise items respectively to obtain a third vector representation corresponding to the dynamic feature data.

[0066] In an exemplary embodiment, the fusion module 36 is further configured to concatenate the first vector representation, the second vector representation, and the third vector representation in a preset concatenation order to obtain a long vector containing all feature information; and to obtain a target vector representation based on the long vector.

[0067] In an exemplary embodiment, the prediction module 38 is further configured to input the target vector representation into a pre-trained power supply and demand prediction model, and predict the power supply and demand of different provinces based on the power supply and demand prediction model to obtain power supply and demand prediction results. The power supply and demand prediction model is trained by machine learning using multiple sets of data, and each set of data includes: historical power dispatch data with added weights, historical static feature data, and historical dynamic feature data.

[0068] Embodiments of this application also provide a storage medium including a stored program, wherein the program executes any of the methods described above when it is run.

[0069] Optionally, in this embodiment, the storage medium may be configured to store program code for performing the following steps:

[0070] S1, acquire power dispatch data, static feature data, and dynamic feature data; wherein, the power dispatch data is used to characterize the power allocation situation between different provinces; the static feature data is used to characterize the energy structure of target cities associated with the different provinces; and the dynamic feature data is used to characterize the regional electricity consumption of different parks in the target cities.

[0071] S2, based on the power dispatch data, perform feature extraction to determine a first vector representation corresponding to the power dispatch data; based on the static feature data, perform feature extraction to determine a second vector representation corresponding to the static feature data; and based on the dynamic feature data, perform feature extraction to determine a third vector representation corresponding to the dynamic feature data.

[0072] S3, fuse the first vector representation, the second vector representation, and the third vector representation to obtain the target vector representation;

[0073] S4. Based on the target vector representation, predict the power supply and demand of different provinces to obtain the power supply and demand prediction results.

[0074] Embodiments of this application also provide an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0075] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0076] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0077] S1, acquire power dispatch data, static feature data, and dynamic feature data; wherein, the power dispatch data is used to characterize the power allocation situation between different provinces; the static feature data is used to characterize the energy structure of target cities associated with the different provinces; and the dynamic feature data is used to characterize the regional electricity consumption of different parks in the target cities.

[0078] S2, based on the power dispatch data, perform feature extraction to determine a first vector representation corresponding to the power dispatch data; based on the static feature data, perform feature extraction to determine a second vector representation corresponding to the static feature data; and based on the dynamic feature data, perform feature extraction to determine a third vector representation corresponding to the dynamic feature data.

[0079] S3, fuse the first vector representation, the second vector representation, and the third vector representation to obtain the target vector representation;

[0080] S4. Based on the target vector representation, predict the power supply and demand of different provinces to obtain the power supply and demand prediction results.

[0081] Optionally, in this embodiment, the storage medium may include, but is not limited to, various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0082] Embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0083] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0084] Embodiments of this application also provide a computer program that includes computer instructions stored in a computer-readable storage medium; a processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps in any of the above method embodiments.

[0085] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0086] Obviously, those skilled in the art should understand that the modules or steps of this application described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented here, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, this application is not limited to any particular combination of hardware and software.

[0087] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A method for forecasting electricity supply and demand, characterized in that, include: The system acquires power dispatch data, static feature data, and dynamic feature data. The power dispatch data characterizes the power allocation between different provinces. The static feature data characterizes the energy structure of target cities associated with the different provinces. The dynamic feature data characterizes the regional electricity consumption of different parks within the target cities. The power dispatch data includes: inter-provincial power transmission, power generation plans, and dispatch instructions. The power dispatch data is cleaned to obtain cleaned power dispatch data; time series analysis is performed on the cleaned power dispatch data to determine the temporal characteristics of the power dispatch data over time; spatial analysis is performed on the cleaned power dispatch data to obtain the geographical distribution characteristics corresponding to the power dispatch data; the temporal characteristics, geographical distribution characteristics, and external environmental characteristics are fused to obtain a first target feature; the first target feature is represented by a vector to obtain a first vector representation; feature extraction is performed based on the static feature data to determine a second vector representation corresponding to the static feature data; and the dynamic feature data is decomposed to obtain a trend term, a seasonal term, and a noise term; the trend term, seasonal term, and noise term are encoded respectively to obtain a third vector representation corresponding to the dynamic feature data; the trend term reflects the long-term trend of the time series; the seasonal term captures the periodic changes in the data; the noise term separates the random fluctuations in the time series; the external environmental characteristics include: weather forecasts, holiday information, and market activities; By fusing the first vector representation, the second vector representation, and the third vector representation, a target vector representation is obtained; The target vector is input into a pre-trained power supply and demand prediction model. Based on the power supply and demand prediction model, the power supply and demand of different provinces are predicted to obtain the power supply and demand prediction results. The power supply and demand prediction model is trained by machine learning using multiple sets of data. Each set of data includes: historical power dispatch data with added weights, historical static feature data, and historical dynamic feature data.

2. The method according to claim 1, characterized in that, The step of extracting features based on the static feature data and determining the second vector representation corresponding to the static feature data includes: The static feature data is cleaned to obtain cleaned static feature data; Energy structure information is extracted from the cleaned static feature data, and an energy diversity index is calculated based on the energy structure information; wherein, the energy diversity index is a parameter used to characterize the degree of energy diversity. The energy diversity index is feature-encoded to determine a second vector representation corresponding to the static feature data.

3. The method according to claim 1, characterized in that, The process of fusing the first vector representation, the second vector representation, and the third vector representation to obtain the target vector representation includes: The first vector representation, the second vector representation, and the third vector representation are concatenated in a preset concatenation order to obtain a long vector containing all feature information; Based on the long vector, the target vector representation is obtained.

4. A power supply and demand forecasting device, characterized in that, include: The acquisition module is used to acquire power dispatch data, static feature data, and dynamic feature data; wherein, the power dispatch data is used to characterize the power allocation between different provinces; the static feature data is used to characterize the energy structure of target cities associated with the different provinces; the dynamic feature data is used to characterize the regional electricity consumption of different parks in the target cities; the power dispatch data includes: inter-provincial power transmission, power generation plans, and dispatch instructions; A determination module is used to clean the power dispatch data to obtain cleaned power dispatch data; perform time series analysis on the cleaned power dispatch data to determine the temporal characteristics of the power dispatch data over time; and perform spatial analysis on the cleaned power dispatch data to obtain the geographical distribution characteristics corresponding to the power dispatch data; fuse the temporal characteristics, the geographical distribution characteristics, and external environmental characteristics to obtain a first target feature; represent the first target feature as a vector to obtain a first vector representation; perform feature extraction based on the static feature data to determine a second vector representation corresponding to the static feature data; and decompose the dynamic feature data to obtain a trend item, a seasonal item, and a noise item; encode the trend item, seasonal item, and noise item respectively to obtain a third vector representation corresponding to the dynamic feature data; the trend item reflects the long-term trend of the time series; the seasonal item captures the periodic changes of the data; the noise item separates the random fluctuations in the time series; the external environmental characteristics include: weather forecasts, holiday information, and market activities; A fusion module is used to fuse the first vector representation, the second vector representation, and the third vector representation to obtain a target vector representation; The prediction module is used to input the target vector representation into a pre-trained power supply and demand prediction model, and to predict the power supply and demand of different provinces based on the power supply and demand prediction model to obtain the power supply and demand prediction results. The power supply and demand prediction model is trained by machine learning using multiple sets of data. Each set of data includes: historical power dispatch data with added weights, historical static feature data, and historical dynamic feature data.

5. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein the program, when executed, performs the method described in any one of claims 1 to 3.

6. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the method described in any one of claims 1 to 3 through the computer program.

7. A computer program product, characterized in that, The computer program product includes a computer program, wherein the computer program, when executed by a processor, performs the method described in any one of claims 1 to 3.

Images