Festival and holiday load correction method and system in medium and long term load prediction

Abstract

The invention discloses a holiday and festival load correction method and correction system in medium and long term load prediction, and the method comprises the steps: determining a holiday and festival influence interval in which holiday is released according to a lunar calendar; calculating a reference value of the festival and holiday influence interval over the years; calculating a daily electric quantity per unit value of the corresponding festival and holiday influence interval, and fitting a corresponding festival and holiday load change rule curve of the to-be-predicted year by using the daily electric quantity per unit value; supplementing festival and holiday influence interval data in the load data over the years based on a festival and holiday load change rule curve; carrying out load prediction by utilizing the supplemented data, and obtaining a reference value of a prediction year by utilizing the predicted data; and obtaining the load capacity of the holiday influence interval of the prediction year by using the reference value of the prediction year and the holiday load change rule curve of the prediction year, and performing load correction by using the load capacity. According to the correction method and the correction system provided by the invention, the monthly electric quantity can be finely corrected, the prediction precision is improved, and safe and economical operation of the electric power system is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of medium- and long-term load forecasting, and in particular to a holiday load correction method and correction system in medium- and long-term load forecasting. Background Art

[0002] The medium- and long-term load forecasting of power systems is an important part of power planning, dispatching and operation decision-making. It helps power companies to reasonably arrange power generation plans, optimize resource allocation, and ensure the safe and stable operation of power systems. Medium- and long-term load forecasting usually covers a forecast period ranging from weeks to years, and has extremely high requirements for the accuracy and reliability of the forecasting model. However, in practical applications, power load is affected by a variety of internal and external factors, and holidays (especially holidays according to the lunar calendar) are one of the significant sources of disturbance. Its impact on power load shows obvious nonlinearity and uncertainty, which often leads to a significant increase in forecast errors.

[0003] Current medium- and long-term load forecasting methods, such as time series analysis, regression analysis, neural network models, support vector machines, etc., can handle seasonal changes and trends in load to a certain extent, but they often perform poorly when dealing with sudden load changes caused by holiday factors. These models usually lack targeted adjustment mechanisms for specific events and it is difficult to effectively incorporate holiday factors into the forecasting framework, resulting in large deviations in forecast results during these special periods.

[0004] With the development of big data, artificial intelligence and machine learning technologies, the power industry has accumulated a huge amount of operating data, which makes it possible to build a more accurate load forecasting model. However, how to effectively use this data and develop a forecasting method that can automatically identify and adapt to holidays is still a key issue that needs to be solved in power system research and engineering practice.

[0005] Therefore, it is an urgent problem to study a holiday load correction method and correction system based on the lunar calendar in medium- and long-term load forecasting to achieve the correction of monthly electricity forecast, improve the operating efficiency of the power system, reduce operating costs, and ensure the security of power supply. Summary of the invention

[0006] In view of this, the present invention provides a holiday load correction method and correction system in medium- and long-term load forecasting to solve the problem that traditional load forecasting methods cannot reasonably correct holiday interval loads.

[0007] The present invention provides a method for correcting holiday loads in medium- and long-term load forecasting, comprising:

[0008] Determine the holiday impact interval according to the lunar calendar based on historical load data;

[0009] Calculate the benchmark value of the holiday impact interval in previous years, where the benchmark value is used to reflect the normal working week load level;

[0010] The actual load and the benchmark value of the holiday impact interval in previous years are used to calculate the daily electricity per unit value of the corresponding holiday impact interval, and the daily electricity per unit value is used to fit the load change law curve corresponding to the holiday in the predicted year;

[0011] The holiday impact interval data in the load data of previous years are supplemented to obtain the load data not affected by the holiday impact interval, wherein the holiday impact interval load data supplemented in any year is obtained by using the benchmark value of the holiday impact interval of the current year and the corresponding holiday load change law curve fitted based on the historical load data;

[0012] The load data of previous years supplemented with the holiday impact interval data are used to predict the full-year load of the year to be predicted, and the reference value of the holiday impact interval of the year to be predicted is obtained by using the full-year load;

[0013] The load of the holiday impact interval of the to-be-predicted year is obtained by using the reference value of the holiday impact interval of the to-be-predicted year and the corresponding holiday load change law curve of the to-be-predicted year, and the load of the to-be-predicted year is corrected by using the load of the holiday impact interval.

[0014] Preferably, the holiday impact interval and the normal day interval are determined based on the weekly load similarity and the weekly load mean.

[0015] Further preferably, the reference value of the holiday impact interval is an average value of a normal working week load not affected by holidays.

[0016] Further preferably, the load variation curve of the holiday to be predicted is fitted using the data of the two years before the year to be predicted.

[0017] Further preferably, the corresponding daily electricity consumption per unit values ​​of the holiday impact intervals in the two years before the year to be predicted are averaged and fitted to obtain a load change law curve corresponding to the holiday in the year to be predicted.

[0018] The present invention also provides a holiday load correction system in medium- and long-term load forecasting, comprising:

[0019] A holiday impact interval determination unit is used to determine the holiday impact interval according to the lunar calendar based on historical load data;

[0020] A unit for calculating a benchmark value of a holiday impact interval over the years, used to calculate a benchmark value of the holiday impact interval over the years, wherein the benchmark value is used to reflect a normal working week load level;

[0021] A load variation law curve fitting unit is used to calculate the daily electricity per unit value of the corresponding holiday impact interval using the actual load and the reference value of the holiday impact interval in previous years, and to fit the load variation law curve of the corresponding holiday in the predicted year using the daily electricity per unit value;

[0022] The holiday impact interval data supplementation unit is used to supplement the holiday impact interval data in the load data of previous years to obtain load data not affected by the holiday impact interval, wherein the holiday impact interval load data supplemented in any year is obtained by using the reference value of the holiday impact interval of the current year and the corresponding holiday load change law curve fitted based on the historical load data;

[0023] The whole-year load forecasting and reference value obtaining unit is used to use the load data of previous years supplemented with the holiday impact interval data to forecast the whole-year load of the year to be forecasted, and use the whole-year load to obtain the reference value of the holiday impact interval of the year to be forecasted;

[0024] The load correction unit is used to obtain the load of the holiday impact interval of the to-be-predicted year by using the reference value of the holiday impact interval of the to-be-predicted year and the corresponding holiday load change law curve of the to-be-predicted year, and to correct the load of the to-be-predicted year by using the load of the holiday impact interval.

[0025] Preferably, the holiday impact interval determination unit determines the holiday impact interval and the normal day interval according to the weekly load similarity and the weekly load mean.

[0026] Further preferably, the reference value of the holiday impact interval calculated by the historical holiday impact interval reference value calculation unit is an average value of the normal working week load not affected by holidays.

[0027] Further preferably, the load variation law curve fitting unit uses data from two years before the year to be predicted to fit the load variation law curve for the holidays to be predicted.

[0028] Further preferably, the load variation law curve fitting unit averages the corresponding daily electricity consumption per unit values ​​of the holiday impact intervals in the two years before the year to be predicted, and then performs fitting to obtain the load variation law curve corresponding to the holidays in the year to be predicted.

[0029] The holiday load correction method and correction system in the medium- and long-term load forecast provided by the present invention can make a fine correction to the monthly electricity for holidays such as the Spring Festival according to the lunar calendar time, improve the prediction accuracy, and ensure the safe and economical operation of the power system. The holiday load correction method and correction system can first obtain the reference value and per-unit value of the holiday impact interval of the previous years through the load data of the previous years, and use the reference value and per-unit value to obtain the load change law curve reflecting the load fluctuation law of the holiday impact interval. After that, based on the load change law, the historical data is supplemented and the supplemented data is used to perform load forecasting, the predicted data is used to determine the reference value of the forecast year, and the reference value is combined with the load change law curve to calculate the load value of the holiday impact interval. Finally, the calculated load value of the holiday impact interval is used to correct the forecast result. The holiday load correction method and correction system can significantly improve the accuracy and reliability of the load forecast of the forecast model in the holiday interval by finely analyzing the special impact of holidays on the load, and correct the holiday interval load, so as to provide a more accurate decision-making basis for the planning, scheduling and operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flow chart of the holiday load correction method in the medium- and long-term load forecasting provided by the present invention. DETAILED DESCRIPTION

[0031] The present invention will be further described below in conjunction with specific embodiments.

[0032] In order to solve the shortcomings of traditional load forecasting methods in dealing with holiday load forecasting based on the lunar calendar, Figure 1 As shown, the present invention provides a holiday load correction method in medium- and long-term load forecasting, comprising the following steps:

[0033] S1: Determine the holiday impact interval according to the lunar calendar based on historical load data;

[0034] Among them, determining the holiday impact interval requires considering the diffusion effect before and after the holidays to determine a reasonable impact expansion interval (holiday impact interval) to fully cover the scope of holiday impact on load. Among them, the difference in load sequences within the holiday impact interval is mainly reflected in the regularity and mean of the sequence. The correlation coefficient of sequences with similar change rules is high, and the mean difference of sequences with the same load level is not large. Therefore, the holiday impact interval and the normal day interval can be distinguished based on the weekly load similarity and weekly load mean. Taking the Spring Festival holiday as an example, after comprehensive consideration of the holiday impact through calculation and analysis, this embodiment selects 28 days as the length of the Spring Festival holiday impact interval;

[0035] S2: Calculate the benchmark value of the holiday impact interval in previous years, wherein the benchmark value is used to reflect the normal working week load level and is not affected by the week type;

[0036] Among them, the reference value of the holiday impact interval is the average value of the normal working week load that is basically not affected by the holidays. Preferably, the selected normal working week is one or more weeks before and after the holiday impact interval. In order to improve the stability of the data, in this embodiment, the reference value of the Spring Festival holiday impact interval in previous years is the average value of the load of the two weeks before and after the Spring Festival holiday impact interval;

[0037] S3: Calculate the per-unit value of daily electricity consumption in the corresponding holiday impact interval using the actual load and benchmark value of the holiday impact interval in previous years, and use the per-unit value of daily electricity consumption to fit the load change law curve (per-unit value curve) corresponding to the holiday in the predicted year to quantify the inhibitory effect of holidays on load;

[0038] Wherein, the per unit value of daily electricity consumption in the holiday impact interval = the actual value of daily load in the holiday impact interval / the reference value of the holiday impact interval;

[0039] For the same holiday, the correlation coefficient of the electricity series in different years is very high, that is, the fluctuation law of electricity is highly similar. Therefore, the load change law curve of the holiday corresponding to the year to be predicted can be fitted by using the daily electricity per unit value of the holiday impact interval of previous years. Preferably, the load change law curve of the holiday corresponding to the year to be predicted is fitted by using the data of the two years before the year to be predicted.

[0040] S4: Supplement the holiday impact interval data in the load data of previous years to obtain the load data not affected by the holiday impact interval, wherein the holiday impact interval load data supplemented in any year is obtained by using the benchmark value of the holiday impact interval of the current year and the corresponding holiday load change law curve fitted based on the historical load data;

[0041] Since the interval that needs to be supplemented is the (lunar) holiday impact interval in the historical data, the electricity before and after the holiday impact interval in the historical years is known, so the weekly average of the intermediate load sequence can be "filled in", that is, the average daily electricity volume of each week in the holiday impact interval can be interpolated according to the daily electricity volume average of the weeks before and after the holiday impact interval (benchmark value). After obtaining the average Sunday electricity volume to be supplemented by the above method, it is also necessary to obtain the daily electricity volume of the week to be supplemented. Although the load level is different between a certain week of this year and the weeks before and after it and the similar period of previous years, the correlation coefficient of the electricity volume series is very high (a high correlation coefficient can be This indicates that the fluctuation patterns between the two are highly similar, that is, the fluctuation patterns of a certain week this year and its preceding and following weeks and similar weeks in previous years are very similar, which is also related to the periodicity of load power change). Therefore, the average of the daily power of the two weeks before and after the similar weeks in previous years is taken as the base value, and the Sunday power of the similar period is divided by the base value of the week to obtain the per-unit value of the daily power and the per-unit value curve of the daily power of the similar weeks. This curve can be regarded as the load change law curve of the corresponding holidays in the year to be supplemented. Afterwards, the benchmark value of the holiday impact interval of this year and the per-unit value curve of the daily power of the similar weeks in previous years can be used to fit and generate the Sunday power to be supplemented.

[0042] S5: forecasting the full-year load of the year to be forecasted by using the load data of previous years supplemented with the holiday impact interval data, and obtaining the reference value of the holiday impact interval of the year to be forecasted by using the full-year load;

[0043] S6: Using the reference value of the holiday impact interval of the year to be predicted obtained in S5 and the corresponding holiday load change curve obtained in S3, the load of the holiday impact interval of the year to be predicted is obtained, and the load of the holiday impact interval is used to correct the load of the year to be predicted obtained in S5.

[0044] The present invention also provides a holiday load correction system in medium- and long-term load forecasting, comprising:

[0045] A holiday impact interval determination unit is used to determine the holiday impact interval according to the lunar calendar based on historical load data;

[0046] Among them, the determination of the holiday impact interval needs to consider the diffusion effect before and after the holidays to determine a reasonable impact expansion interval (holiday impact interval) to fully cover the scope of holiday impact on load. Among them, the difference in load sequences within the holiday impact interval is mainly reflected in the regularity and mean of the sequence. The correlation coefficient of sequences with similar change rules is high, and the mean difference of sequences with the same load level is not large. Therefore, the holiday impact interval and the normal day interval can be distinguished based on the weekly load similarity and weekly load mean;

[0047] A unit for calculating a benchmark value of a holiday impact interval over the years, used to calculate a benchmark value of the holiday impact interval over the years, wherein the benchmark value is used to reflect a normal working week load level;

[0048] The reference value of the holiday impact interval is the average value of the normal working week load that is basically not affected by the holidays. Preferably, the selected normal working week is one or more weeks before and after the holiday impact interval in order to improve the stability of the data;

[0049] A load variation law curve fitting unit is used to calculate the daily electricity per unit value of the corresponding holiday impact interval using the actual load and the reference value of the holiday impact interval in previous years, and to fit the load variation law curve of the corresponding holiday in the predicted year using the daily electricity per unit value;

[0050] Wherein, the per unit value of daily electricity consumption in the holiday impact interval = the actual value of daily load in the holiday impact interval / the reference value of the holiday impact interval;

[0051] For the same holiday, the correlation coefficient of the electricity series in different years is very high, that is, the fluctuation law of electricity is highly similar. Therefore, the load change law curve of the holiday corresponding to the year to be predicted can be fitted by using the daily electricity per unit value of the holiday impact interval of previous years. Preferably, the load change law curve of the holiday corresponding to the year to be predicted is fitted by using the data of the two years before the year to be predicted.

[0052] The holiday impact interval data supplementation unit is used to supplement the holiday impact interval data in the load data of previous years to obtain load data not affected by the holiday impact interval, wherein the holiday impact interval load data supplemented in any year is obtained by using the reference value of the holiday impact interval of the current year and the corresponding holiday load change law curve fitted based on the historical load data;

[0053] Since the interval that needs to be supplemented is the (lunar) holiday impact interval in the historical data, the electricity before and after the holiday impact interval in the historical years is known, so the weekly average of the intermediate load sequence can be "filled in", that is, the average daily electricity volume of each week in the holiday impact interval can be interpolated according to the daily electricity volume average of the weeks before and after the holiday impact interval (benchmark value). After obtaining the average Sunday electricity volume to be supplemented by the above method, it is also necessary to obtain the daily electricity volume of the week to be supplemented. Although the load level is different between a certain week of this year and the weeks before and after it and the similar period of previous years, the correlation coefficient of the electricity volume series is very high (a high correlation coefficient can be This indicates that the fluctuation patterns between the two are highly similar, that is, the fluctuation patterns of a certain week this year and its preceding and following weeks and similar weeks in previous years are very similar, which is also related to the periodicity of load power change). Therefore, the average of the daily power of the two weeks before and after the similar weeks in previous years is taken as the base value, and the Sunday power of the similar period is divided by the base value of the week to obtain the per-unit value of the daily power and the per-unit value curve of the daily power of the similar weeks. This curve can be regarded as the load change law curve of the corresponding holidays in the year to be supplemented. Afterwards, the benchmark value of the holiday impact interval of this year and the per-unit value curve of the daily power of the similar weeks in previous years can be used to fit and generate the Sunday power to be supplemented.

[0054] The whole-year load forecasting and reference value obtaining unit is used to use the load data of previous years supplemented with the holiday impact interval data to forecast the whole-year load of the year to be forecasted, and use the whole-year load to obtain the reference value of the holiday impact interval of the year to be forecasted;

[0055] The load correction unit is used to obtain the load of the holiday impact interval of the to-be-predicted year by using the reference value of the holiday impact interval of the to-be-predicted year and the corresponding holiday load change law curve of the to-be-predicted year, and to correct the load of the to-be-predicted year by using the load of the holiday impact interval.

[0056] The holiday load correction method and correction system in the medium- and long-term load forecast provided by the present invention can make a fine correction to the monthly electricity for holidays such as the Spring Festival according to the lunar calendar time, improve the prediction accuracy, and ensure the safe and economical operation of the power system. The holiday load correction method and correction system can first obtain the reference value and per-unit value of the holiday impact interval of the previous years through the load data of the previous years, and use the reference value and per-unit value to obtain the load change law curve reflecting the load fluctuation law of the holiday impact interval. After that, based on the load change law, the historical data is supplemented and the supplemented data is used to perform load forecasting, the predicted data is used to determine the reference value of the forecast year, and the reference value is combined with the load change law curve to calculate the load value of the holiday impact interval. Finally, the calculated load value of the holiday impact interval is used to correct the forecast result. The holiday load correction method and correction system can significantly improve the accuracy and reliability of the load forecast of the forecast model in the holiday interval by finely analyzing the special impact of holidays on the load, and correct the holiday interval load, so as to provide a more accurate decision-making basis for the planning, scheduling and operation of the power system.

[0057] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems or computer program products. Therefore, the application can adopt the form of complete hardware embodiment, complete software embodiment, or the embodiment in combination with software and hardware. Moreover, the application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiment of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0058] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0059] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0060] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0061] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0062] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A holiday load correction method in medium- and long-term load forecasting, characterized in that: include: Determine the interval affected by the lunar calendar holidays based on historical load data; Calculate the benchmark value of the holiday impact interval in previous years, where the benchmark value is used to reflect the normal working week load level; The actual load and the benchmark value of the holiday impact interval in previous years are used to calculate the daily electricity per unit value of the corresponding holiday impact interval, and the daily electricity per unit value is used to fit the load change law curve corresponding to the holiday in the predicted year; The holiday impact interval data in the load data of previous years are supplemented to obtain the load data not affected by the holiday impact interval, wherein the holiday impact interval load data supplemented in any year is obtained by using the benchmark value of the holiday impact interval of the current year and the corresponding holiday load change law curve fitted based on the historical load data; The load data of previous years supplemented with the holiday impact interval data are used to predict the full-year load of the year to be predicted, and the reference value of the holiday impact interval of the year to be predicted is obtained by using the full-year load; The load of the holiday impact interval of the to-be-predicted year is obtained by using the reference value of the holiday impact interval of the to-be-predicted year and the corresponding holiday load change law curve of the to-be-predicted year, and the load of the to-be-predicted year is corrected by using the load of the holiday impact interval.

2. The holiday load correction method in the medium- and long-term load forecasting according to claim 1 is characterized by: The holiday impact interval and normal day interval are determined based on weekly load similarity and weekly load mean.

3. The holiday load correction method in the medium- and long-term load forecasting according to claim 1 is characterized by: The reference value of the holiday impact interval is the average value of the normal working week load that is not affected by holidays.

4. The holiday load correction method in the medium- and long-term load forecasting according to claim 1 is characterized by: The data of the two years before the year to be predicted are used to fit the load change curve of the holiday to be predicted.

5. The holiday load correction method in the medium- and long-term load forecasting according to claim 4 is characterized by: The corresponding daily electricity consumption per unit value of the holiday impact interval in the two years before the forecast year is averaged and fitted to obtain the load change law curve corresponding to the holiday in the forecast year.

6. A holiday load correction system in medium- and long-term load forecasting, characterized in that: include: A holiday impact interval determination unit is used to determine the holiday impact interval according to the lunar calendar based on historical load data; A calculation unit for the benchmark value of the holiday impact interval in previous years, used to calculate the benchmark value of the holiday impact interval in previous years, wherein the benchmark value is used to reflect the load level of a normal working week; A load variation law curve fitting unit is used to calculate the daily electricity per unit value of the corresponding holiday impact interval using the actual load and the reference value of the holiday impact interval in previous years, and to fit the load variation law curve of the corresponding holiday in the predicted year using the daily electricity per unit value; The holiday impact interval data supplementation unit is used to supplement the holiday impact interval data in the load data of previous years to obtain load data not affected by the holiday impact interval, wherein the holiday impact interval load data supplemented in any year is obtained by using the reference value of the holiday impact interval of the current year and the corresponding holiday load change law curve fitted based on the historical load data; The whole-year load forecasting and reference value obtaining unit is used to use the load data of previous years supplemented with the holiday impact interval data to forecast the whole-year load of the year to be forecasted, and use the whole-year load to obtain the reference value of the holiday impact interval of the year to be forecasted; The load correction unit is used to obtain the load of the holiday impact interval of the to-be-predicted year by using the reference value of the holiday impact interval of the to-be-predicted year and the corresponding holiday load change law curve of the to-be-predicted year, and to correct the load of the to-be-predicted year by using the load of the holiday impact interval.

7. The holiday load correction system in the medium- and long-term load forecasting according to claim 6 is characterized by: The holiday impact interval determination unit determines the holiday impact interval and the normal day interval according to the weekly load similarity and the weekly load mean.

8. The holiday load correction system in the medium- and long-term load forecasting according to claim 6 is characterized by: The reference value of the holiday impact interval calculated by the historical holiday impact interval reference value calculation unit is the average value of the normal working week load not affected by the holidays.

9. The holiday load correction system in the medium- and long-term load forecasting according to claim 6 is characterized by: The load variation law curve fitting unit fits the load variation law curve of the holiday to be predicted by using the data of the two years before the year to be predicted.

10. The holiday load correction system in the medium- and long-term load forecasting according to claim 9 is characterized by: The load variation law curve fitting unit averages the corresponding daily electricity consumption per unit values ​​of the holiday impact intervals in the two years before the year to be predicted, and then performs fitting to obtain the load variation law curve corresponding to the holiday in the year to be predicted.

Images