Regional electrical load prediction system and method
By using machine learning and sample training methods in the regional power load prediction system to build an initial power load prediction model, and improving accuracy through secondary optimization, the problem of low prediction accuracy in the prior art is solved, and higher prediction accuracy and reliability are achieved.
Patent Information
- Application Number
- CN202510202777.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-24
- Publication Date
- 2025-06-10
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing regional power load prediction system has low accuracy and cannot effectively deal with load changes under different power consumption conditions.
A regional power load prediction system was designed, and the central server, prediction model construction module and model secondary optimization module were used to build the initial power load prediction model using machine learning and sample training methods, and the prediction accuracy was improved through secondary optimization. The system divides historical electricity consumption data into industrial, commercial and resident categories, builds load prediction models separately, and generates regional electricity consumption load prediction results through the summary and comprehensive evaluation of the results.
It improves the accuracy and reliability of the forecasting of electricity load, enhances the objective, scientific and accurate prediction, and can more accurately predict regional electricity loads, helping the power system to operate stably.
Smart Images
Figure CN120127635A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power load forecasting, and in particular to a regional power load forecasting system and method. Background Art
[0002] Urban electricity load refers to the sum of active power actually consumed by all electricity users in a city or local area at a certain moment. With the continuous development of science and technology, more diverse electrical equipment is being used more and more widely, and accordingly, the demand for electricity in life is also increasing. In turn, the power system must ensure a more stable and higher-quality power supply, so it is necessary to conduct load forecasting for regional electricity consumption to ensure the stable and reliable operation of the power system.
[0003] The existing regional electricity load forecasting system usually uses a single regional electricity load forecasting model to forecast the regional electricity load for all electricity consumption data in a certain area. However, different electricity consumption situations will result in different electricity loads, resulting in low prediction accuracy. Therefore, we propose a regional electricity load forecasting system and method. Summary of the invention
[0004] In view of the deficiencies in the prior art, the present invention provides a regional electricity load forecasting system and method, which solves the problems raised in the above background technology.
[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a regional power load forecasting system, comprising a central server, a forecasting model construction module and a model secondary optimization module, the output end of the central server is electrically connected to the forecasting model construction module, the output end of the forecasting model construction module is electrically connected to the power load forecasting module, and the output end of the power load forecasting module is electrically connected to the result summary analysis module, the output end of the result summary analysis module is electrically connected to the power comprehensive evaluation module, and the output end of the power comprehensive evaluation module is electrically connected to the prediction result comparison module, the output end of the prediction result comparison module is electrically connected to the accuracy calculation module, and the output end of the accuracy calculation module is electrically connected to the prediction deviation analysis module, the output end of the prediction deviation analysis module is electrically connected to the model secondary optimization module, and the output end of the prediction deviation analysis module is also electrically connected to the power abnormality monitoring module, the power abnormality monitoring module comprises an abnormal data identification module, an abnormal data determination module and an abnormal data feedback module, the output end of the abnormal data identification module is electrically connected to the abnormal data determination module, and the output end of the abnormal data determination module is electrically connected to the abnormal data feedback module.
[0006] Furthermore, the input end of the central server is electrically connected to a regional electricity classification module, and the input end of the regional electricity classification module is electrically connected to a historical data collection module.
[0007] Furthermore, the regional electricity consumption classification module includes industrial electricity consumption, commercial electricity consumption and residential electricity consumption, and industrial electricity consumption, commercial electricity consumption and residential electricity consumption are historical electricity consumption data of industry, commerce and residents respectively, so that each category corresponds to a power load prediction model.
[0008] Furthermore, the prediction model construction module includes a machine learning module, a sample training module and a correction optimization module, the output end of the machine learning module is electrically connected to the sample training module, and the output end of the sample training module is electrically connected to the correction optimization module.
[0009] Furthermore, the machine learning module performs regionalization and constructs an initial power load forecasting model through machine learning. The sample training module is used to perform sample training using historical power consumption data as a data set for the training model, and then extracts the characteristics and patterns of industrial power consumption, commercial power consumption, and residential power consumption through learning and analyzing the training samples. A corresponding power load forecasting model is obtained for each category. The correction and optimization module is used to optimize the power load forecasting model based on machine learning and sample training and obtain the final power load forecasting model.
[0010] Furthermore, the power load forecasting module is used to forecast power loads for three categories of industrial power consumption, commercial power consumption and residential power consumption respectively; the result summary and analysis module is used to summarize all categories of loads to obtain regional power load forecasting results; the comprehensive power consumption evaluation module is used to perform comprehensive power consumption evaluation based on the regional power load forecasting results.
[0011] Furthermore, the prediction result comparison module is used to compare the prediction results after the actual power consumption results are output, and the accuracy calculation module is used to provide the accuracy of the regional power load prediction based on the result of comparing the actual power load with the prediction result.
[0012] Furthermore, the prediction deviation analysis module is used to perform deviation analysis on regional power load conditions where the accuracy of power load prediction is lower than the expected value, and the model secondary optimization module is used to perform secondary optimization on the regional power load prediction model based on the prediction deviation analysis results and according to effective deviation data.
[0013] Furthermore, the abnormal data identification module is used to identify and extract the independent abnormal deviation data based on the prediction deviation analysis module, the abnormal data judgment module is used to judge the extracted abnormal electricity consumption data, and the abnormal data feedback module is used to feed back the abnormal electricity consumption data to the central server according to the abnormal data judgment result.
[0014] Furthermore, the method for regional power load forecasting includes the following specific steps:
[0015] S1. First, the regional electricity consumption classification module divides the data collected by the historical data collection module into industrial, commercial and residential historical electricity consumption data, so that each category corresponds to a power load forecasting model;
[0016] S2, the prediction model construction module constructs the power load prediction model, and then the power load prediction module predicts the power load for three categories of industrial power, commercial power, and residential power respectively. Finally, the result summary analysis module summarizes all categories of loads to obtain the regional power load prediction results, which provides a reference for subsequent power dispatching and monitoring;
[0017] S3, the prediction result comparison module and the accuracy calculation module give the accuracy of the regional power load prediction based on the comparison results between the actual power load and the prediction result, and then the model secondary optimization module performs secondary optimization on the regional power load prediction model based on the prediction deviation analysis results and the effective deviation data, so as to fully improve the accuracy of the regional power load prediction results;
[0018] S4. Based on the prediction deviation analysis module, the power consumption anomaly monitoring module identifies the independent abnormal deviation data, extracts and judges it, and feeds back the abnormal power consumption data to the central server, thereby realizing efficient and accurate identification of abnormal power consumption behavior.
[0019] The present invention provides a regional power load forecasting system and method, which has the following beneficial effects:
[0020] 1. The regional electricity load forecasting system and method are provided with a forecasting model building module. The machine learning module performs regional and constructs an initial electricity load forecasting model through machine learning. The sample training module is used to perform sample training using historical electricity consumption data as a data set for the training model, and then extracts the characteristics and patterns of industrial electricity consumption, commercial electricity consumption, and residential electricity consumption through learning and analysis of the training samples. Each category obtains a corresponding electricity load forecasting model, and matches different load forecasting probability models for different electricity consumption types, which can improve the accuracy of electricity load forecasting, with higher reliability, better accuracy, and more objectivity and science. The correction and optimization module is used to optimize the electricity load forecasting model based on machine learning and sample training and obtain the final electricity load forecasting model.
[0021] 2. The regional electricity load forecasting system and method are provided with a summary analysis module. The regional electricity classification module divides the data collected by the historical data collection module into historical electricity data of industry, commerce and residents, so that each category corresponds to a power load forecasting model. The power load forecasting module is used to forecast the power loads of the three categories of industrial power, commercial power and residential power respectively, and flexibly form targeted forecasts. At the same time, power load forecasts are conducted based on power time series data, including but not limited to weekly power load forecasts and monthly power load forecasts. The result summary analysis module is used to summarize all categories of loads to obtain regional power load forecast results, which provides a reference for subsequent power dispatching and monitoring. The comprehensive power evaluation module is used to conduct a comprehensive power evaluation based on the regional power load forecast results, evaluate the medium- and long-term load forecasting effect, and provide a basis for the medium- and long-term planning of the power system.
[0022] 3. The regional electricity load forecasting system and method are provided with a model secondary optimization module, a prediction result comparison module is used to compare the prediction results after the actual electricity consumption results are output, an accuracy calculation module is used to give the accuracy of the regional electricity load forecast according to the result of the comparison between the actual electricity load and the prediction result, a prediction deviation analysis module is used to perform deviation analysis on the regional electricity load situation where the accuracy of the electricity load forecast is lower than the expected value, and a model secondary optimization module is used to perform secondary optimization on the regional electricity load forecasting model based on the prediction deviation analysis results and according to the effective deviation data, so as to fully improve the accuracy of the regional electricity load forecasting results.
[0023] 4. The regional electricity load forecasting system and method are provided with an electricity consumption anomaly monitoring module. The abnormal data identification module is used to identify and extract the independent abnormal deviation data based on the prediction deviation analysis module. The abnormal data judgment module is used to judge the extracted abnormal electricity consumption data. The abnormal data feedback module is used to feed back the abnormal electricity consumption data to the central server according to the abnormal data judgment result, thereby realizing efficient and accurate identification of abnormal electricity consumption behavior, and being able to quickly locate abnormal electricity consumption data, which is conducive to helping power companies to carry out power dispatching and ensure the stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 A schematic diagram of a system and method for predicting regional power load according to the present invention;
[0025] Figure 2 It is a schematic diagram of a model secondary optimization module of the regional power load forecasting system and method of the present invention;
[0026] Figure 3 A schematic diagram of a regional electricity consumption classification module of the regional electricity consumption load prediction system and method of the present invention;
[0027] Figure 4 A schematic diagram of a prediction model building module of a regional power load prediction system and method of the present invention;
[0028] Figure 5 It is a schematic diagram of the power consumption anomaly monitoring module of the regional power load prediction system and method of the present invention.
[0029] In the figure: 1. Central server; 2. Regional electricity classification module; 201. Industrial electricity; 202. Commercial electricity; 203. Residential electricity; 3. Historical data collection module; 4. Prediction model construction module; 401. Machine learning module; 402. Sample training module; 403. Correction optimization module; 5. Power load prediction module; 6. Result summary and analysis module; 7. Comprehensive electricity evaluation module; 8. Prediction result comparison module; 9. Accuracy calculation module; 10. Prediction deviation analysis module; 11. Model secondary optimization module; 12. Abnormal electricity monitoring module; 1201. Abnormal data identification module; 1202. Abnormal data determination module; 1203. Abnormal data feedback module. DETAILED DESCRIPTION
[0030] See also Figure 1 , Figure 3 and Figure 4 The present invention provides a technical solution: a regional power load forecasting system, comprising a central server 1, a forecasting model building module 4 and a model secondary optimization module 11, the input end of the central server 1 is electrically connected to a regional power classification module 2, and the input end of the regional power classification module 2 is electrically connected to a historical data collection module 3, the regional power classification module 2 includes industrial power 201, commercial power 202 and residential power 203, and the industrial power 201, the commercial power 202 and the residential power 203 are the historical power data of industry, commerce and residents respectively, so that each category corresponds to a power load forecasting model, the output end of the central server 1 is electrically connected to the forecasting model building module 4, the forecasting model building module 4 includes a machine learning module 401, a sample training module 402 and a correction optimization module 403, the output end of the machine learning module 401 is electrically connected to the sample training module 402, and the output end of the sample training module 402 is electrically connected to the correction optimization module 403;
[0031] The specific operations are as follows: the machine learning module 401 performs regionalization and constructs an initial power load prediction model through machine learning; the sample training module 402 is used to perform sample training using historical power consumption data as a data set for the training model, and then extracts the characteristics and patterns of industrial power consumption 201, commercial power consumption 202, and residential power consumption 203 through learning and analyzing the training samples, and obtains a corresponding power load prediction model for each category. Different load prediction probability models are matched to different power consumption types, which can improve the accuracy of power load prediction, make it more reliable, more accurate, and more objective and scientific; the correction and optimization module 403 is used to optimize the power load prediction model based on machine learning and sample training and obtain the final power load prediction model.
[0032] See also Figure 1-Figure 2 The output end of the prediction model construction module 4 is electrically connected to the power load prediction module 5, and the output end of the power load prediction module 5 is electrically connected to the result summary analysis module 6, the output end of the result summary analysis module 6 is electrically connected to the comprehensive power consumption evaluation module 7, and the output end of the comprehensive power consumption evaluation module 7 is electrically connected to the prediction result comparison module 8, the output end of the prediction result comparison module 8 is electrically connected to the accuracy calculation module 9, and the output end of the accuracy calculation module 9 is electrically connected to the prediction deviation analysis module 10;
[0033] The specific operations are as follows: the power load forecasting module 5 is used to forecast the power load for the three categories of industrial power 201, commercial power 202, and residential power 203 respectively, and flexibly form targeted forecasts. At the same time, the power load forecast is conducted based on the power time series data, including but not limited to weekly power load forecasts and monthly power load forecasts. The result summary and analysis module 6 is used to summarize all categories of loads to obtain regional power load forecast results, providing a reference for subsequent power scheduling and monitoring. The comprehensive power evaluation module 7 is used to conduct a comprehensive power evaluation based on the regional power load forecast results, evaluate the medium- and long-term load forecast effects, and provide a basis for the medium- and long-term planning of the power system. The forecast result comparison module 8 is used to compare the forecast results after the actual power consumption results are output. The accuracy calculation module 9 is used to give the accuracy of the regional power load forecast based on the comparison results of the actual power load and the forecast results.
[0034] See also Figure 2 and Figure 5The output end of the prediction deviation analysis module 10 is electrically connected to the model secondary optimization module 11, and the output end of the prediction deviation analysis module 10 is also electrically connected to the power consumption abnormality monitoring module 12, the power consumption abnormality monitoring module 12 includes an abnormal data identification module 1201, an abnormal data determination module 1202 and an abnormal data feedback module 1203, the output end of the abnormal data identification module 1201 is electrically connected to the abnormal data determination module 1202, and the output end of the abnormal data determination module 1202 is electrically connected to the abnormal data feedback module 1203;
[0035] The specific operations are as follows: the prediction deviation analysis module 10 is used to perform deviation analysis on the regional power load situation where the accuracy of power load prediction is lower than the expected value; the model secondary optimization module 11 is used to perform secondary optimization on the regional power load prediction model based on the prediction deviation analysis results and according to the effective deviation data, so as to fully improve the accuracy of the regional power load prediction results; the abnormal data identification module 1201 is used to identify and extract the independent abnormal deviation data on the basis of the prediction deviation analysis module 10; the abnormal data judgment module 1202 is used to judge the extracted abnormal power data; the abnormal data feedback module 1203 is used to feed back the abnormal power data to the central server 1 according to the abnormal data judgment results, thereby realizing efficient and accurate identification of abnormal power consumption behavior, and being able to quickly locate abnormal power consumption data, which is conducive to helping power companies to carry out power dispatch, ensure the stable operation of the power system, and improve power safety.
[0036] In summary, please refer to Figure 1-Figure 5When the regional power load forecasting system and method are used, first, the regional power classification module 2 divides the data collected by the historical data collection module 3 into industrial, commercial and residential historical power consumption data, so that each category corresponds to a power load forecasting model, and then the machine learning module 401 performs regional and constructs an initial power load forecasting model through machine learning, and the sample training module 402 uses the historical power consumption data as a data set for the training model for sample training, and then extracts the characteristics and patterns of industrial power consumption 201, commercial power consumption 202, and residential power consumption 203 through learning and analyzing the training samples, and obtains a corresponding power load forecasting model for each category, and matches different load forecasting probability models for different power consumption types. The model can improve the accuracy of power load prediction, with higher reliability, better accuracy and more objective science. Then, the correction and optimization module 403 optimizes the power load prediction model based on machine learning and sample training and obtains the final power load prediction model. Then, the power load prediction module 5 performs power load prediction for the three categories of industrial power 201, commercial power 202 and residential power 203 respectively, and flexibly forms targeted predictions. At the same time, power load prediction is performed based on power time series data, including but not limited to weekly power load prediction and monthly power load prediction. Finally, the result summary analysis module 6 summarizes all categories of loads to obtain regional power load prediction results for subsequent power dispatch and monitoring. The power comprehensive evaluation module 7 is then used to conduct a comprehensive power evaluation based on the regional power load forecast results to evaluate the medium- and long-term load forecast effect, which is conducive to providing a basis for the medium- and long-term planning of the power system. After the actual power consumption results are output, the forecast results are compared by the forecast result comparison module 8. Then, the accuracy calculation module 9 gives the accuracy of the regional power load forecast based on the comparison result between the actual power load and the forecast result. Then, the forecast deviation analysis module 10 performs a deviation analysis on the regional power load situation where the accuracy of the power load forecast is lower than the expected value. Then, the model secondary optimization module 11 performs a secondary optimization on the regional power load forecast model based on the forecast deviation analysis results and the effective deviation data. , fully improving the accuracy of regional power load forecasting results; in this process, the abnormal data identification module 1201 can identify and extract the independent abnormal deviation data on the basis of the prediction deviation analysis module 10, and then the abnormal data judgment module 1202 judges the extracted abnormal power data, and then the abnormal data feedback module 1203 feeds back the abnormal power data to the central server 1 according to the abnormal data judgment result, thereby realizing efficient and accurate identification of abnormal power consumption behavior, and being able to quickly locate abnormal power consumption data, which is conducive to helping power companies to carry out power dispatching, ensure the stable operation of the power system, and improve power safety, thus completing the use process of the entire regional power load forecasting system and method.
[0037] The embodiments of the present invention are given for the purpose of illustration and description, and are not intended to be exhaustive or to limit the invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses.
Claims
1. A regional power load forecasting system, characterized in that: The invention comprises a central server (1), a prediction model construction module (4) and a model secondary optimization module (11), wherein the output end of the central server (1) is electrically connected to the prediction model construction module (4), the output end of the prediction model construction module (4) is electrically connected to the power load prediction module (5), and the output end of the power load prediction module (5) is electrically connected to the result summary analysis module (6), the output end of the result summary analysis module (6) is electrically connected to the power consumption comprehensive evaluation module (7), and the output end of the power consumption comprehensive evaluation module (7) is electrically connected to the prediction result comparison module (8), and the output end of the prediction result comparison module (8) is electrically connected to the accuracy calculation module (9), and the accuracy calculation module (9) is electrically connected to the power consumption comprehensive evaluation module (7). The output end of the calculation module (9) is electrically connected to a prediction deviation analysis module (10), the output end of the prediction deviation analysis module (10) is electrically connected to a model secondary optimization module (11), and the output end of the prediction deviation analysis module (10) is also electrically connected to an abnormal power usage monitoring module (12), the abnormal power usage monitoring module (12) comprises an abnormal data identification module (1201), an abnormal data determination module (1202) and an abnormal data feedback module (1203), the output end of the abnormal data identification module (1201) is electrically connected to the abnormal data determination module (1202), and the output end of the abnormal data determination module (1202) is electrically connected to the abnormal data feedback module (1203).
2. A regional power load forecasting system according to claim 1, characterized in that: The input end of the central server (1) is electrically connected to a regional electricity classification module (2), and the input end of the regional electricity classification module (2) is electrically connected to a historical data collection module (3).
3. A regional power load forecasting system according to claim 2, characterized in that: The regional electricity consumption classification module (2) includes industrial electricity consumption (201), commercial electricity consumption (202) and residential electricity consumption (203), and the industrial electricity consumption (201), commercial electricity consumption (202) and residential electricity consumption (203) are historical electricity consumption data of industry, commerce and residents respectively, so that each category corresponds to a power load prediction model.
4. A regional power load forecasting system according to claim 1, characterized in that: The prediction model construction module (4) comprises a machine learning module (401), a sample training module (402) and a correction optimization module (403), wherein the output end of the machine learning module (401) is electrically connected to the sample training module (402), and the output end of the sample training module (402) is electrically connected to the correction optimization module (403).
5. A regional power load forecasting system according to claim 4, characterized in that: The machine learning module (401) performs regionalization and constructs an initial power load prediction model through machine learning. The sample training module (402) is used to perform sample training using historical power consumption data as a data set for the training model, and then extracts the characteristics and patterns of industrial power consumption (201), commercial power consumption (202), and residential power consumption (203) through learning and analyzing the training samples, and obtains a corresponding power load prediction model for each category. The correction optimization module (403) is used to optimize the power load prediction model based on machine learning and sample training and obtain a final power load prediction model.
6. A regional power load forecasting system according to claim 1, characterized in that: The power load prediction module (5) is used to respectively predict power loads for three categories of industrial power (201), commercial power (202) and residential power (203); the result summary analysis module (6) is used to summarize all categories of loads to obtain regional power load prediction results; and the power consumption comprehensive evaluation module (7) is used to perform a comprehensive power consumption evaluation based on the regional power load prediction results.
7. A regional power load forecasting system according to claim 1, characterized in that: The prediction result comparison module (8) is used to compare the prediction results after the actual power consumption results are output, and the accuracy calculation module (9) is used to provide the accuracy of regional power load prediction based on the result of comparing the actual power load with the prediction result.
8. A regional power load forecasting system according to claim 1, characterized in that: The prediction deviation analysis module (10) is used to perform deviation analysis on regional power load conditions where the accuracy of power load prediction is lower than the expected value, and the model secondary optimization module (11) is used to perform secondary optimization on the regional power load prediction model based on the prediction deviation analysis result and according to effective deviation data.
9. A regional power load forecasting system according to claim 1, characterized in that: The abnormal data identification module (1201) is used to identify and extract abnormal deviation data based on the prediction deviation analysis module (10); the abnormal data determination module (1202) is used to determine the extracted abnormal electricity consumption data; and the abnormal data feedback module (1203) is used to feed back the abnormal electricity consumption data to the central server (1) based on the abnormal data determination result.
10. A regional power load forecasting system according to any one of claims 1 to 9, characterized in that: The method for regional power load prediction comprises the following specific steps: S1. First, the regional electricity consumption classification module (2) divides the data collected by the historical data collection module (3) into industrial, commercial and residential historical electricity consumption data, so that each category corresponds to a power load prediction model; S2, the prediction model building module (4) builds the power load prediction model, and then the power load prediction module (5) respectively predicts the power load for the three categories of industrial power (201), commercial power (202), and residential power (203). Finally, the result summary analysis module (6) summarizes all categories of loads to obtain regional power load prediction results, which provides a reference for subsequent power dispatching and monitoring; S3, the prediction result comparison module (8) and the accuracy calculation module (9) provide the accuracy of the regional power load prediction based on the comparison result between the actual power load and the prediction result, and then the model secondary optimization module (11) performs secondary optimization on the regional power load prediction model based on the prediction deviation analysis result and the effective deviation data, so as to fully improve the accuracy of the regional power load prediction result; S4. Based on the prediction deviation analysis module (10), the power consumption anomaly monitoring module (12) identifies the abnormal deviation data, extracts and judges it, and feeds back the abnormal power consumption data to the central server (1), thereby realizing efficient and accurate identification of abnormal power consumption behavior.