Power grid load prediction method based on power grid resource service platform

Abstract

The invention provides a power grid load prediction method based on a power grid resource service platform, and belongs to the field of power grid data processing. The method comprises the following steps: acquiring operation state data of equipment of each district station in power grid resource business stations; cleaning the running state data to obtain cleaned running state data; performing wavelet transform processing on the cleaned running state data to obtain transform features; wherein the transformation feature comprises an approximation coefficient and a detail coefficient; inputting the transformation characteristics into a power grid load prediction model to obtain a model prediction load value of the power grid; performing calibration processing on the model prediction load value according to pre-acquired calibration data to obtain a final prediction load value of the power grid; wherein the calibration data comprises air temperature data and holiday and festival information data. The power grid prediction value is calibrated according to the air temperature data and the holiday and festival information data, so that the power grid prediction value is closer to the actual power grid load condition.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid data processing, and in particular to a power grid load forecasting method based on a power grid resource business center. Background Art

[0002] The power grid resource business platform is a core component of power system management and plays an important role in ensuring the economy, reliability and sustainability of the power system. The power grid resource business platform is an integrated platform that provides one-stop rapid service support for cross-professional applications such as dispatching, operation and inspection, marketing, and planning, construction, and operation polymorphic maintenance.

[0003] In order to better balance power generation and power demand and ensure the stability and reliability of power supply, it is necessary to predict the power grid load. For example, the patent with application number CN202211511216.5 constructs a first prediction model and a second prediction model through the historical data of the power grid, and determines the final prediction result based on the output of the first prediction model and the output of the second prediction model to realize the prediction of the power grid load.

[0004] In the prior art, when electricity is actually used, the grid load is easily affected by various external environmental factors such as temperature, resulting in inaccurate grid load prediction. Summary of the invention

[0005] The embodiment of the present invention provides a power grid load forecasting method based on a power grid resource service center to solve the problem of inaccurate power grid load forecasting.

[0006] In a first aspect, an embodiment of the present invention provides a power grid load forecasting method based on a power grid resource service middle station, comprising:

[0007] Obtain the operating status data of the equipment in each district of Taichung in the power grid resource business;

[0008] Cleaning the running status data to obtain cleaned running status data;

[0009] Performing wavelet transform processing on the cleaned running status data to obtain transformation features; wherein the transformation features include approximate coefficients and detail coefficients;

[0010] Inputting the transformed characteristics into a power grid load prediction model to obtain a model-predicted load value of the power grid;

[0011] The model predicted load value is calibrated according to the pre-acquired calibration data to obtain the final predicted load value of the power grid; wherein the calibration data includes temperature data and holiday information data.

[0012] In a possible implementation, the temperature data includes the daily average temperature, and the holiday information data includes the number of holiday days;

[0013] Calibrate the grid forecast load value through calibration data, including:

[0014] Combined with the preset formula, and based on the impact of temperature changes on the grid load value and the impact of holidays on the grid load value, the grid forecast load value is calibrated. The preset formula is:

[0015] M=[1+α×(aa ′ )+β×b]M ′

[0016] Where M is the final predicted load value, α is the temperature weight coefficient, a is the predicted daily average temperature, and a ′ is the historical daily average temperature, β is the holiday weight coefficient, b is the holiday length weight coefficient, M ′ is the model predicted load value of the power grid.

[0017] In a possible implementation, the holiday length weight coefficient includes:

[0018] When the number of holiday days is less than or equal to the first threshold, the holiday length weight coefficient is the first weight coefficient;

[0019] When the number of holiday days is greater than the first threshold and less than or equal to the second threshold, the holiday length weight coefficient is the second weight coefficient;

[0020] When the number of holiday days is greater than the second threshold, the holiday length weight coefficient is the third weight coefficient.

[0021] In a possible implementation, the first threshold is less than the second threshold;

[0022] The first weight coefficient is smaller than the second weight coefficient, and the second weight coefficient is smaller than the third weight coefficient.

[0023] In a possible implementation, before inputting the transformation feature into the power grid load prediction model, the power grid load prediction method further includes:

[0024] Obtain historical data; wherein the historical data includes historical operating status data of equipment in each district of the power grid resource business, temperature data of each preset period of history, time data, holiday information data and historical power grid load value;

[0025] The historical operation status data of the equipment in each area is processed by wavelet transformation to obtain transformation features; wherein the transformation features include approximate coefficients and detail coefficients;

[0026] A data set is constructed based on the transformation characteristics, temperature data of each preset period in history, time data, holiday information data and historical power grid load values, and the initial prediction model is trained based on the data set to obtain a power grid load prediction model.

[0027] In a possible implementation, the power grid load forecasting model is an LSTM model.

[0028] In a possible implementation, before acquiring historical data, the power grid load forecasting method further includes:

[0029] Acquire initial data; wherein the initial data includes initial data of historical operation status of equipment in each district of the power grid resource business, initial data of temperature in each preset period of history, initial data of time, initial data of holiday information, and initial data of historical power grid load value;

[0030] The initial data is cleaned and the historical data is obtained based on the cleaned data.

[0031] In a possible implementation, the cleaning process includes missing value processing and outlier processing.

[0032] In a possible implementation, the operating status data includes a voltage value, a current value, and an on / off state.

[0033] In a possible implementation, the operation status data, time data, temperature data, holiday information data and power grid load data of the equipment of each zone station are updated at set intervals, including:

[0034] The operating status data, time data, temperature data, holiday information data and grid load data of the newly added equipment in each district station are merged with the original operating status data, time data, temperature data, holiday information data and grid load data of the equipment in each district station to obtain updated operating status data, time data, temperature data, holiday information data and grid load data of the equipment in each district station.

[0035] In a second aspect, an embodiment of the present invention provides a power grid load prediction device based on a power grid resource service middle station, comprising:

[0036] The data acquisition module is used to obtain the operating status data of the equipment in each area of ​​the power grid resource business;

[0037] A data cleaning module is used to clean the running status data to obtain cleaned running status data;

[0038] A feature extraction module is used to perform wavelet transform processing on the cleaned running status data to obtain transformation features; wherein the transformation features include approximate coefficients and detail coefficients;

[0039] A prediction module is used to input the transformation characteristics into a power grid load prediction model to obtain a model-predicted load value of the power grid;

[0040] The calibration module is used to calibrate the model predicted load value according to the pre-acquired calibration data to obtain the final predicted load value of the power grid; wherein the calibration data includes temperature data and holiday information data.

[0041] The embodiment of the present invention provides a power grid load prediction method based on a power grid resource business center. The operation status data of the equipment in each district station in the acquired power grid resource business center are cleaned and then processed by wavelet transform to obtain transformation features, which can capture important information such as trends, periodicity and abnormal changes in the data. The transformation features are then input into a power grid load prediction model, and a power grid predicted load value is output. Then, considering the impact of temperature and holiday factors on the power grid load, the power grid prediction value is calibrated according to the temperature data and holiday information data, so that the power grid prediction value is closer to the actual power grid load situation, thereby improving the accuracy of the prediction, which is beneficial for power companies to optimize the allocation and scheduling of power generation resources, reduce resource waste, and improve the economy and efficiency of power grid operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 It is a flow chart of an implementation of a power grid load forecasting method based on a power grid resource service middle station provided in an embodiment of the present invention;

[0044] Figure 2 It is a structural diagram of a power grid load prediction device based on a power grid resource business center provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0046] In order to make the purpose, technical solutions and advantages of the present invention more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0047] Figure 1 The implementation flow chart of the power grid load forecasting method based on the power grid resource service center provided in the embodiment of the present invention is described in detail as follows:

[0048] Step 101, obtaining the operation status data of the equipment in each area of ​​the power grid resource business.

[0049] Operation status data is the basic data of power grid operation and the premise of power grid forecasting. By obtaining the operation status data of each area station equipment, the operation status of the power grid can be monitored in real time, including the health status of the equipment, load conditions, etc. It is a basic link in power grid management and power system engineering, and is crucial to ensure the reliable, efficient and safe operation of the power grid.

[0050] Step 102, cleaning the running status data to obtain cleaned running status data.

[0051] Cleaning can remove noise and errors in the operating status data, thereby improving the overall quality of the operating status data. The cleaned operating status data provides solid data support for decision-making. Ensuring the quality of operating status data is crucial for subsequent data analysis, model building and decision support.

[0052] Step 103, performing wavelet transform processing on the cleaned running status data to obtain transformation features; wherein the transformation features include approximate coefficients and detail coefficients.

[0053] The power grid operation status data is often nonlinear and non-stationary. Wavelet transform can adapt to the characteristics of such data and provide more accurate data analysis results. Wavelet transform analyzes the operation status data in both time and frequency domains, which is very useful for identifying periodic and instantaneous changes in the power grid operation status. It can capture the local details and global trends of power grid load changes at the same time, decompose the operation status data into components of different scales, help compress data and remove redundant information, thereby simplifying the complexity of the model.

[0054] The approximate coefficients obtained by wavelet processing can reflect the long-term trend of the data, while the detailed coefficients can reveal the short-term fluctuations and anomalies of the data, which is very important for power grid load forecasting, power grid fault diagnosis and preventive maintenance. The approximate coefficients and detailed coefficients as inputs to the power grid load forecasting model can help improve the model's prediction accuracy and robustness for power grid load changes. The approximate coefficients and detailed coefficients contain richer information than the original data, which helps to reveal patterns and associations that are not easily observed in the data.

[0055] Step 104, input the transformed characteristics into a power grid load prediction model to obtain a model-predicted load value of the power grid.

[0056] Transformation features are features extracted through wavelet transform and can represent key information about the operation status of the power grid. The power grid load forecasting model is used to analyze the input transformation features and generate model-predicted load values ​​for the power grid load, which helps power grid operators to manage and dispatch the load. Model-predicted load values ​​serve as an important basis for power grid operation and maintenance decisions, helping power grid operators optimize resource allocation and improve power grid operation efficiency. Model-predicted load values ​​can identify potential load peaks or troughs in advance, so that preventive measures can be taken to reduce the risk of power grid failures and power outages.

[0057] Step 105, calibrating the model predicted load value according to the pre-acquired calibration data to obtain the final predicted load value of the power grid; wherein the calibration data includes temperature data and holiday information data.

[0058] External factors such as temperature and holidays have a significant impact on grid load. Calibration can take these factors into account to improve the relevance and practicality of the forecast.

[0059] Temperature has a significant impact on grid load because it directly affects people's electricity consumption behavior and electricity demand.

[0060] For example, in hot weather, the use of air conditioners will increase, especially in the summer, which will cause a significant increase in the load on the power grid. In cold weather, the temperature is lower, and the use of heating equipment such as electric heaters will increase, especially in the winter, which will also increase the load on the power grid.

[0061] Temperature changes may affect energy consumption in industrial production processes. For example, in extremely high temperatures, industrial equipment may require more energy to maintain normal operation. Temperature changes may also affect the energy demand for agricultural irrigation and greenhouses. For example, during dry seasons, the demand for electricity for irrigation may increase.

[0062] By calibrating the temperature data, the forecasting model can more accurately reflect the impact of temperature changes on the grid load, thereby providing more reliable load forecasting results, which is crucial for grid operation and dispatching, power resource allocation and power market operation.

[0063] The calibration of holiday information data can help the model consider the special impact of holidays on grid load, such as reduced industrial electricity consumption and increased residential electricity consumption.

[0064] During holidays, especially statutory holidays, many industrial enterprises will stop or reduce production, resulting in a decrease in industrial electricity demand. Residential electricity consumption may change significantly during holidays. For example, more time at home may lead to an increase in the use of household appliances. During holidays, the use of air conditioning and heating equipment may increase as people spend more time at home, especially in extreme weather conditions.

[0065] Considering holiday information in power grid load forecasting can help the forecasting model more accurately reflect the impact of holidays on power grid load, thereby providing more reliable load forecasting results. This is crucial for power grid operation and dispatching, power resource allocation, and power market operations.

[0066] The embodiment of the present invention cleans the operating status data of the equipment in each district station in Taichung obtained from the power grid resource business and performs wavelet transform processing to obtain transformation features, which can capture important information such as trends, periodicity and abnormal changes in the data. The transformation features are then input into the power grid load prediction model, and the power grid predicted load value is output. Then, the influence of temperature and holiday factors on the power grid load is considered, and the power grid prediction value is calibrated according to the temperature data and holiday information data, so that the power grid prediction value is closer to the actual power grid load situation, thereby improving the accuracy of the prediction, which is beneficial for power companies to optimize the allocation and scheduling of power generation resources, reduce resource waste, and improve the economy and efficiency of power grid operation.

[0067] The above is an overall introduction to this solution. Next, the specific methods of this solution are introduced as follows:

[0068] In a possible implementation, the temperature data includes the daily average temperature, and the holiday information data includes the number of holiday days;

[0069] Calibrate the grid forecast load value through calibration data, including:

[0070] Combined with the preset formula, and based on the impact of temperature changes on the grid load value and the impact of holidays on the grid load value, the grid forecast load value is calibrated. The preset formula is:

[0071] M=[1+α×(aa ′ )+β×b]M ′

[0072] Where M is the final predicted load value, α is the temperature weight coefficient, a is the predicted daily average temperature, and a ′ is the historical daily average temperature, β is the holiday weight coefficient, b is the holiday length weight coefficient, M ′ is the model predicted load value of the power grid.

[0073] α represents the impact of temperature changes on load, β represents the impact of holidays on load, and b represents the impact of the length of holidays on the predicted load value of the power grid. It varies with the number of holidays. The more holidays there are, the greater the weight coefficient of the holiday length, indicating that the length of the holiday has a greater impact on the predicted load value of the power grid.

[0074] For example, if the holiday weight coefficient β is 0.1 and the holiday length weight coefficient b is 1.2 (indicating that the number of holiday days exceeds the second threshold), then β×b=0.1×1.2=0.12, and this 0.12 will be added to the model predicted load value to calibrate the final predicted load value.

[0075] In this embodiment, the prediction results are calibrated by considering the influence of external factors such as temperature and holidays. Calibration can help the prediction model better adapt to the actual grid operation conditions, improve the accuracy of the prediction, and thus improve the efficiency and reliability of the grid operation.

[0076] In other possible implementations, the data used to calibrate the predicted load value of the power grid may be other types of data. The present invention is described using temperature and holiday information as an example.

[0077] In a possible implementation, the holiday length weight coefficient includes:

[0078] When the number of holiday days is less than or equal to the first threshold, the holiday length weight coefficient is the first weight coefficient;

[0079] When the number of holiday days is greater than the first threshold and less than or equal to the second threshold, the holiday length weight coefficient is the second weight coefficient;

[0080] When the number of holiday days is greater than the second threshold, the holiday length weight coefficient is the third weight coefficient.

[0081] Accordingly, the first threshold is less than the second threshold;

[0082] The first weight coefficient is smaller than the second weight coefficient, and the second weight coefficient is smaller than the third weight coefficient.

[0083] In this embodiment, the first threshold is a critical point of the number of holiday days. If the number of holiday days is less than or equal to this value, the holiday length weight coefficient will be set to the "first weight coefficient", which means that for shorter holidays, a predefined coefficient is used to adjust the grid load forecast. The second threshold is another critical point of the number of holiday days and is greater than the first threshold. If the number of holiday days is greater than the first threshold and less than or equal to the second threshold, the holiday length weight coefficient will be set to the "second weight coefficient", which is suitable for medium-length holidays. When the number of holiday days exceeds the second threshold, the holiday length weight coefficient will be set to the "third weight coefficient", which is used to adjust the impact of longer holidays on the grid load.

[0084] The impact of different holidays on the grid load can be flexibly adjusted according to their length, because holidays of different lengths may have different effects on people's electricity consumption behavior and grid load. For example, shorter holidays may only lead to slight changes in electricity consumption patterns, while longer holidays may cause more significant changes, such as peak tourism and reduced industrial production.

[0085] In a possible implementation, before inputting the transformation feature into the power grid load prediction model, the power grid load prediction method further includes:

[0086] Obtain historical data; wherein the historical data includes historical operating status data of equipment in each district of the power grid resource business, temperature data of each preset period of history, time data, holiday information data and historical power grid load value;

[0087] The historical operation status data of the equipment in each area is processed by wavelet transformation to obtain transformation features; wherein the transformation features include approximate coefficients and detail coefficients;

[0088] A data set is constructed based on the transformation characteristics, temperature data of each preset period in history, time data, holiday information data and historical power grid load values, and the initial prediction model is trained based on the data set to obtain a power grid load prediction model.

[0089] The collected historical data of the power grid is the basis for building the prediction model. The historical operating status data of the equipment in each area is subjected to wavelet transformation to extract the time-frequency characteristics of the data, including approximate coefficients (reflecting the low-frequency trend information of the data) and detail coefficients (reflecting the high-frequency detail information of the data).

[0090] By using the transformation features (approximate coefficients and detail coefficients) extracted from wavelet transform, combined with the temperature data, time data, holiday information data and historical power grid load values ​​of each preset period of history, a comprehensive data set is constructed for training the prediction model.

[0091] The initial grid load forecasting model is trained using the constructed dataset, a process that involves adjusting model parameters and learning patterns and relationships in the data so that the model can predict grid load based on the input features.

[0092] After training, the initial prediction model becomes an optimized grid load prediction model for predicting future grid load.

[0093] In this embodiment, the grid load prediction model includes a variety of factors that affect the grid load, such as the operating status of the equipment, temperature, time, holidays, etc., and extracts key features through wavelet transform, and uses these features to train a powerful prediction model, which can help grid operators better plan and dispatch power resources and improve the operating efficiency and reliability of the grid.

[0094] In a possible implementation, the power grid load forecasting model is an LSTM model.

[0095] In this embodiment, the LSTM model is used to capture the time series characteristics of the power grid load and make accurate predictions.

[0096] In a possible implementation, before acquiring historical data, the power grid load forecasting method further includes:

[0097] Acquire initial data; wherein the initial data includes initial data of historical operation status of equipment in each district of the power grid resource business, initial data of temperature in each preset period of history, initial data of time, initial data of holiday information, and initial data of historical power grid load value;

[0098] The initial data is cleaned and the historical data is obtained based on the cleaned data.

[0099] The historical operation status initial data includes historical information such as the operating status and performance parameters of the equipment. Recording temperature changes in different time periods is crucial for understanding the impact of temperature on grid load. Factors involving time, such as the day of the week, can help identify the time dependence of power consumption patterns. Holiday information data includes the date and type of holidays, which is very important for predicting grid load changes during holidays. The load record of the grid in the same time period in the past is a key reference data for predicting future load.

[0100] The data are cleaned to ensure their quality and accuracy.

[0101] In this embodiment, the collected initial data is strictly cleaned to ensure that the data used for model training is accurate and reliable. This step is crucial to improving the performance and accuracy of the prediction model.

[0102] In a possible implementation, the cleaning process includes missing value processing and outlier processing.

[0103] For example, statistical methods can be used for preliminary screening, followed by machine learning methods for further validation and identification of outliers. Missing values ​​can be filled and outliers replaced based on linear interpolation.

[0104] Statistical methods such as using statistical methods such as Z-Score or IQR (interquartile range) to identify those data points that deviate too much from the mean or are outside the normal range.

[0105] In this embodiment, the data quality is improved through cleaning processing, ensuring that the results of data analysis and model training are more accurate and reliable.

[0106] In a possible implementation, the operating status data includes a voltage value, a current value, and an on / off state.

[0107] Voltage value is a key indicator to measure the power supply quality of the power grid, which directly affects the user's power consumption experience and the safe operation of electrical equipment. The change in voltage can reflect the distribution of the power grid load. A decrease in voltage may mean that the local load is too high, and continuous voltage abnormality may be a sign of equipment failure or line problems.

[0108] The current value directly reflects the amount of current in the power grid and is a direct indicator for evaluating the load level of the power grid. When the current value exceeds the rated value of the equipment, it indicates that the equipment may be overloaded and needs timely adjustment or maintenance. The current value is directly related to the energy loss of the power grid and helps to evaluate and optimize the operating efficiency of the power grid.

[0109] On / off status data can monitor the on / off status of devices in the power grid, such as circuit breakers, disconnectors, etc. Changes in on / off status can help locate fault points in the power grid, such as line disconnection or equipment failure.

[0110] In this embodiment, the load change of the power grid can be predicted more accurately through the operating status data.

[0111] In a possible implementation, the operation status data, time data, temperature data, holiday information data and power grid load data of the equipment of each zone station are updated at set intervals, including:

[0112] The operating status data, time data, temperature data, holiday information data and grid load data of the newly added equipment in each district station are merged with the original operating status data, time data, temperature data, holiday information data and grid load data of the equipment in each district station to obtain updated operating status data, time data, temperature data, holiday information data and grid load data of the equipment in each district station.

[0113] In this embodiment, by regularly updating data, it is ensured that the power grid load forecasting model uses the latest data, thereby improving the timeliness and accuracy of the forecast.

[0114] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0115] The following is an embodiment of the device of the present invention. For details not described in detail therein, reference may be made to the corresponding method embodiment described above.

[0116] Figure 2The following is a schematic diagram of the structure of a power grid load prediction device based on a power grid resource service middle station provided in an embodiment of the present invention. For ease of explanation, only the part related to the embodiment of the present invention is shown, which is described in detail as follows:

[0117] like Figure 2 As shown, the power grid load prediction device 2 based on the power grid resource business center station includes:

[0118] The data acquisition module 21 is used to obtain the operation status data of the equipment in each area of ​​the power grid resource business;

[0119] The data cleaning module 22 is used to clean the operation status data to obtain the cleaned operation status data;

[0120] The feature extraction module 23 is used to perform wavelet transform processing on the cleaned running status data to obtain transformation features; wherein the transformation features include approximate coefficients and detail coefficients;

[0121] The prediction module 24 is used to input the transformation characteristics into the power grid load prediction model to obtain the model predicted load value of the power grid;

[0122] The calibration module 25 is used to calibrate the model predicted load value according to the pre-acquired calibration data to obtain the final predicted load value of the power grid; wherein the calibration data includes temperature data and holiday information data.

[0123] In a possible implementation, the prediction module 24 is used to:

[0124] Obtain historical data; wherein the historical data includes historical operating status data of equipment in each district of Taichung in the power grid resource business, temperature data of each preset period of history, time data, holiday information data and historical power grid load value;

[0125] The historical operation status data of the equipment in each area is processed by wavelet transformation to obtain transformation features; wherein the transformation features include approximate coefficients and detail coefficients;

[0126] A data set is constructed based on the transformation characteristics, temperature data of each preset period in history, time data, holiday information data and historical power grid load values, and the initial prediction model is trained based on the data set to obtain a power grid load prediction model.

[0127] In a possible implementation, the prediction module 24 is used to:

[0128] Acquire initial data; wherein the initial data includes initial data of historical operation status of equipment in each district of Taichung in the power grid resource business, initial data of temperature in each preset period of history, initial data of time, initial data of holiday information, and initial data of historical power grid load value;

[0129] The initial data is cleaned and the historical data is obtained based on the cleaned data.

[0130] In a possible implementation, the calibration module 25 is used to:

[0131] Combined with the preset formula, and based on the impact of temperature changes on the grid load value and the impact of holidays on the grid load value, the grid forecast load value is calibrated. The preset formula is:

[0132] M=[1+α×(aa ′ )+β×b]M ′

[0133] Where M is the final predicted load value, α is the temperature weight coefficient, a is the predicted daily average temperature, and a ′ is the historical daily average temperature, β is the holiday weight coefficient, b is the holiday length weight coefficient, M ′ is the model predicted load value of the power grid.

[0134] In a possible implementation, the calibration module 25 is used to:

[0135] Holiday length weight factors include:

[0136] When the number of holiday days is less than or equal to the first threshold, the holiday length weight coefficient is the first weight coefficient;

[0137] When the number of holiday days is greater than the first threshold and less than or equal to the second threshold, the holiday length weight coefficient is the second weight coefficient;

[0138] When the number of holiday days is greater than the second threshold, the holiday length weight coefficient is the third weight coefficient.

[0139] In a possible implementation, the calibration module 25 is used to:

[0140] The first threshold is less than the second threshold;

[0141] The first weight coefficient is smaller than the second weight coefficient, and the second weight coefficient is smaller than the third weight coefficient.

[0142] The embodiment of the present invention cleans the operating status data of the equipment in each district station in Taichung obtained from the power grid resource business and performs wavelet transform processing to obtain transformation features, which can capture important information such as trends, periodicity and abnormal changes in the data. The transformation features are then input into the power grid load prediction model, and the power grid predicted load value is output. Then, the influence of temperature and holiday factors on the power grid load is considered, and the power grid prediction value is calibrated according to the temperature data and holiday information data, so that the power grid prediction value is closer to the actual power grid load situation, thereby improving the accuracy of the prediction, which is beneficial for power companies to optimize the allocation and scheduling of power generation resources, reduce resource waste, and improve the economy and efficiency of power grid operation.

[0143] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0144] Those of ordinary skill in the art will appreciate that the templates, units, and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0145] If the module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, it can implement the steps of the above-mentioned embodiments of the power grid load forecasting method based on the power grid resource business middle station. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory, random access memory, electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0146] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A power grid load forecasting method based on a power grid resource service center, characterized in that: include: Obtain the operating status data of the equipment in each district of Taichung in the power grid resource business; Cleaning the running status data to obtain cleaned running status data; Performing wavelet transform processing on the cleaned running status data to obtain transformation features; wherein the transformation features include approximate coefficients and detail coefficients; Inputting the transformation characteristics into a power grid load prediction model to obtain a model-predicted load value of the power grid; The model predicted load value is calibrated according to the pre-acquired calibration data to obtain the final predicted load value of the power grid; wherein the calibration data includes temperature data and holiday information data.

2. The power grid load forecasting method based on the power grid resource service center according to claim 1 is characterized in that: The temperature data includes the daily average temperature, and the holiday information data includes the number of holidays; The step of calibrating the predicted load value of the power grid by using the calibration data comprises: Combined with the preset formula, the grid forecast load value is calibrated according to the impact of temperature changes on the grid load value and the impact of holidays on the grid load value. The preset formula is: M=[1+α×(a-a')+β×b]M' Where M is the final predicted load value, α is the temperature weight coefficient, a is the predicted daily average temperature, a′ is the historical daily average temperature, β is the holiday weight coefficient, b is the holiday length weight coefficient, and M′ is the model predicted load value of the power grid.

3. The power grid load forecasting method based on the power grid resource service center according to claim 2 is characterized in that: The holiday length weight coefficients include: When the number of holiday days is less than or equal to the first threshold, the holiday length weight coefficient is the first weight coefficient; When the number of holiday days is greater than the first threshold and less than or equal to the second threshold, the holiday length weight coefficient is the second weight coefficient; When the number of holiday days is greater than the second threshold, the holiday length weight coefficient is the third weight coefficient.

4. The power grid load forecasting method based on the power grid resource service center according to claim 3 is characterized in that: The first threshold is less than the second threshold; The first weight coefficient is smaller than the second weight coefficient, and the second weight coefficient is smaller than the third weight coefficient.

5. The power grid load forecasting method based on the power grid resource business middle station according to claim 1 is characterized in that: Before inputting the transformation characteristics into the power grid load prediction model, the power grid load prediction method further includes: Acquire historical data; wherein the historical data includes historical operating status data of equipment in each district of the power grid resource business, temperature data of each preset period of history, time data, holiday information data and historical power grid load value; Perform wavelet transform processing on the historical operation status data of the equipment of each zone station to obtain transformation features; wherein the transformation features include approximate coefficients and detail coefficients; A data set is constructed according to the transformation characteristics, temperature data of each preset time period in history, time data, holiday information data and the historical power grid load value, and an initial prediction model is trained according to the data set to obtain the power grid load prediction model.

6. The power grid load forecasting method based on the power grid resource service center according to claim 1 is characterized in that: The power grid load prediction model is an LSTM model.

7. The power grid load forecasting method based on the power grid resource service center according to claim 5 is characterized in that: Before acquiring the historical data, the power grid load forecasting method further includes: Acquire initial data; wherein the initial data includes initial data of historical operating status of equipment in each district of the power grid resource business, initial data of temperature in each preset period of history, initial data of time, initial data of holiday information, and initial data of historical power grid load value; The initial data is cleaned and the historical data is obtained according to the cleaned data.

8. The power grid load forecasting method based on the power grid resource service center according to claim 1 or 7, characterized in that: The cleaning process includes missing value processing and outlier processing.

9. The power grid load forecasting method based on the power grid resource service center according to claim 1, characterized in that: The operating status data includes voltage value, current value and on / off status.

10. The power grid load forecasting method based on the power grid resource service center according to claim 1, characterized in that: Update the operating status data, time data, temperature data, holiday information data and power grid load data of the equipment in each zone at set intervals, including: The operating status data, time data, temperature data, holiday information data and grid load data of the newly added equipment in each district station are merged with the original operating status data, time data, temperature data, holiday information data and grid load data of the equipment in each district station to obtain updated operating status data, time data, temperature data, holiday information data and grid load data of the equipment in each district station.

Images