A winter daily maximum load prediction method and system based on load-sensitive air temperature distribution

Abstract

The application discloses a winter daily maximum load prediction method and system based on load-sensitive air temperature distribution, which comprises the following steps: determining the negative growth change rule of statutory holiday load by adopting the year-by-year statistical method; taking months as units, using the observation analysis method, the variation coefficient and other methods to statistically calculate the dispersion degree of daily maximum load, and determining the analysis period of winter heating load; constructing a trend change model; comprehensively considering the influence of multiple factors, setting the load-sensitive air temperature isotherm distribution range of winter heating load; determining the specific range of load-sensitive air temperature isotherm distribution in the calculation and analysis period; constructing a correlation model of heating load and load-sensitive air temperature isotherm distribution range; and realizing daily maximum load prediction. The method can realize short-term quantitative prediction of daily maximum load in a simple and fast table lookup mode, has strong regional expansibility, and has certain guiding significance for regional short-term winter daily maximum load prediction.

Description

Technical Field

[0001] This invention relates to a method and system for predicting the maximum daily load in winter based on load-sensitive temperature distribution, belonging to the field of load prediction technology. Background Technology

[0002] Electricity load forecasting is fundamental to power operation and dispatching, power system planning, and power safety command, impacting the entire power grid's generation capacity, electricity sales efficiency, and operational stability. Research on electricity load forecasting has a long history, with numerous scholars both domestically and internationally conducting extensive research on forecasting theories and methods, achieving significant progress. Due to the increasing number of factors influencing load changes and their highly nonlinear nature, the emergence of new theories and technologies continuously drives the development of load forecasting. Changes in electricity load are significantly affected by environmental factors; for winter heating loads, ambient temperature, a meteorological factor, has a crucial influence.

[0003] There are many studies on temperature-sensitive load forecasting, but most of them focus on summer cooling load forecasting methods. Patents "A Maximum Load Forecasting Method Based on Load Decomposition" (CN104881722) and "A Classification Forecasting Method for Annual Maximum Load Based on Economic and Meteorological Factors" (CN103413253) emphasize the forecasting of summer cooling loads and lack descriptions of winter load forecasting methods. Furthermore, these methods only select the basic load analysis period based on the season and lack data and quantitative algorithm support. Hu Zhihao et al. (2021) divided the electric heating load increment into stable load increment, meteorologically sensitive load increment, random load increment, and date-type (weekend) load increment, but did not specifically analyze the load of statutory holidays other than weekends, and also lacked specific methods for calculating holiday load increments. Patent "Power Grid Power Load Forecasting Method and System Based on Meteorological Factors" (CN105825294) added a forecasting and analysis method for winter electric heating load, but the temperature used in this method is the daily average temperature of a specific region, and the forecast is only applicable to local areas and does not have wide-ranging universality (nationwide) application. The current methods for predicting large-scale daily maximum power load based on load-sensitive temperature distribution still need improvement. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for predicting the daily maximum load in winter based on load-sensitive temperature distribution. The method and system are accurate in prediction and have strong regional scalability, which has certain guiding significance for regional short-term prediction of the daily maximum load in winter.

[0005] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0006] In a first aspect, the present invention provides a method for predicting the daily maximum load in winter based on load-sensitive temperature distribution, comprising the following steps:

[0007] S1. Acquire information data for the load forecasting area and compile specific elements to obtain data output content;

[0008] S2. Based on the data results, the load correction of the daily maximum load sequence is achieved by using the statutory holiday load removal method and the linear interpolation filling method, and the corrected daily maximum load sequence after removing statutory holidays is obtained.

[0009] S3. Based on the modified daily maximum load sequence, construct a model of the daily maximum load variation pattern on statutory holidays;

[0010] S4. Based on the modified daily maximum load sequence, statistically calculate the dispersion of the daily maximum load and determine the analysis period for winter heating load;

[0011] S5. Calculate the annual base load based on the analysis period of the winter heating load, and construct a base load trend change model;

[0012] S6. Set the temperature isotherm distribution level sensitive to winter heating load;

[0013] S7. Based on the distribution level of the load-sensitive temperature isotherm in winter heating load, calculate the specific level of the load-sensitive temperature isotherm distribution during the analysis period of winter heating load;

[0014] S8. Based on historical daily maximum load data and the specific distribution levels of load-sensitive temperature isotherms, construct a correlation model between heating load and the distribution levels of load-sensitive temperature isotherms;

[0015] S9. Based on the daily maximum load variation model of the statutory holidays, the basic load trend variation model, and the correlation model between heating load and load-sensitive temperature isotherm distribution level, the daily maximum load of the forecast day in the load forecast area is predicted.

[0016] Furthermore, in step S1, the information collection data includes:

[0017] Multi-year historical load data: Multi-year historical daily maximum load data for the load forecasting area;

[0018] Basic data of meteorological observation stations: station name, station number, administrative division, longitude, latitude, and elevation of national standard meteorological observation stations within the load forecast area;

[0019] Historical meteorological data over many years: recent years' observed temperature, rainfall, and wind speed within the load forecast area;

[0020] Meteorological element forecast data: forecast of the minimum daily temperature, the rainfall, and the maximum wind speed in the future within the load forecast area;

[0021] The specific elements are compiled, including:

[0022] By compiling historical real-time observation data of meteorological elements day by day, the daily minimum temperature is obtained;

[0023] By compiling historical real-time observation data of meteorological elements day by day, the maximum daily wind force can be obtained;

[0024] By compiling historical observation data of meteorological elements day by day, it is possible to determine whether it rains on a given day.

[0025] Furthermore, in step S2, based on the data results, the load correction of the daily maximum load sequence is achieved using the statutory holiday load elimination method and the linear interpolation filling method, including:

[0026] The historical daily maximum load data from the past few years are arranged in time series and labeled with the weekdays. Based on the specific arrangements of domestic statutory holidays, the daily maximum load corresponding to weekdays and statutory holidays is separated.

[0027] Remove the daily maximum load corresponding to statutory holidays from the historical daily maximum load data;

[0028] A linear interpolation method was used to fill in the gaps in the daily maximum load for the excluded dates. The actual daily maximum load was marked as follows: Where j = 0...n; the daily maximum load in the sequence numbered j to j+m is removed because of statutory holidays; m is the statutory holiday.

[0029] Solve for the corrected daily maximum load

[0030]

[0031] in, —The maximum load correction value for the (j+1)th day;

[0032] —Measured value of maximum load on the j-th day;

[0033] —Measured value of the maximum load on the (j+m)th day;

[0034] The final result is a revised daily maximum load sequence after removing statutory holidays.

[0035] Furthermore, in step S3, determining the negative growth pattern of statutory holiday load includes the following steps:

[0036] Based on the revised daily maximum load sequence, and in accordance with the arrangements for public holidays and public holidays stipulated by current domestic laws, the daily maximum load corresponding to public holidays and public holidays is separated.

[0037] The changes in the maximum daily load for each type of holiday are statistically analyzed year by year, including: statistical analysis of the negative growth of the load for each type of holiday, statistical analysis of the negative growth trend of the load for each type of holiday year by year, and construction of a model of the daily maximum load change pattern for statutory holidays.

[0038] Furthermore, in step S4, the analysis period for winter heating load is determined, including:

[0039] Based on the revised daily maximum load sequence, and combining the observation of seasonal variation patterns and the coefficient of variation evaluation method, the dispersion of daily maximum load in autumn and winter (September of the current year to March of the following year) is analyzed year by year on a monthly basis:

[0040] Plot the modified daily maximum load variation curve over time, and select months with higher dispersion based on the local climate change patterns in the load forecast area;

[0041] Calculate the coefficient of variation (CV) of the daily maximum load monthly, and select the months with higher CVs. The specific calculation method for CV is as follows:

[0042]

[0043]

[0044] The load during winter (December to February of the following year) is used as the baseline load, and the months with the coefficient of variation exceeding the preset threshold are selected as the analysis period for winter heating load.

[0045] Furthermore, in step S5, constructing the basic load trend change model includes the following steps:

[0046] The months excluding the analysis period for winter heating load are used as the months for calculating the base load;

[0047] The method of solving for the arithmetic mean is used to calculate the daily maximum load of the calculation month of the base load each year, which is used as the base load for the winter heating load of that year, and a model of the annual variation trend of the base load is established.

[0048] Methods for establishing annual variation models of basic loads include:

[0049] Select socioeconomic factors strongly correlated with the base load (such as GDP, CPI, added value of the secondary industry, and the proportion of the secondary industry). Select factors with high Pearson correlation coefficients as the dominant factors affecting the base load.

[0050] A basic load prediction model was established by using the least squares method, linear regression method, and polynomial fitting method, combined with the dominant factors affecting the basic load.

[0051] Furthermore, in step S6, setting the winter heating load sensitive temperature isotherm distribution level includes the following steps:

[0052] Taking into account the distribution of minimum temperatures in the load forecast area, the variation of the frost line, the orientation of rivers and mountains, the distribution of geographical and climatic environment, and the population distribution, the distribution of each level of the load-sensitive temperature isotherm is determined.

[0053] The gear ranges include: northern North China, along the Yellow River, along the Qinling-Huaihe River, along the Yangtze River, and along the Nanling-Wuyi Mountains.

[0054] Furthermore, in step S7, calculating the specific range of the load-sensitive temperature isotherm distribution during the analysis period of the winter heating load includes the following steps:

[0055] Based on the geographical distribution of each meteorological level, and following the principle of proximity, nearby national standard meteorological stations were searched and selected.

[0056] The isotherm range of load-sensitive temperature is selected to represent meteorological stations, respectively, to characterize the current isotherm position. When more than half of the selected stations have daily minimum temperatures that vary within ±1℃ of the load-sensitive temperature, the isotherm range of load-sensitive temperature is considered to have reached the position it represents.

[0057] Furthermore, in step S8, constructing the correlation model between heating load and load-sensitive temperature isotherm distribution levels includes the following steps:

[0058] Based on historical daily maximum load data and load-sensitive temperature isotherm distribution data of minimum temperature, when the load-sensitive temperature isotherm shifts southward, the increment of the daily maximum load compared to the base load is calculated, which is the heating load.

[0059] Plot the correlation curve between heating load and load-sensitive temperature isotherm levels. Integrate daily maximum wind force and whether there is daily rainfall to establish a correlation model between heating load and load-sensitive temperature isotherm levels, wind chill, and water chill.

[0060] Furthermore, in step S9, the process of predicting the daily maximum load for the predicted date in the load prediction area includes the following steps:

[0061] Obtain the year and date of the forecast date for the load forecast area and the forecast values ​​of the dominant factors, which include GDP, CPI, and added value of the secondary industry;

[0062] Obtain the geographical location of the load forecast area, and based on the geographical location, obtain the location of the load-sensitive temperature distribution level in the load forecast area, the rainfall, wind speed, and historical daily maximum load measured values ​​of the stations covered by the level on the forecast day;

[0063] Inputting the predicted values ​​of the dominant factors into the annual variation trend model of the base load yields the base load L for the predicted date. basic ;

[0064] The holiday type for the forecast date is determined based on the year and date of the forecast date. This holiday type is then input into the daily maximum load variation model for statutory holidays to obtain the statutory holiday load L for the forecast date. holiday ;

[0065] The location of the load-sensitive temperature distribution level in the load forecast area, the rainfall, wind speed, and historical daily maximum load measurements of the stations covered by the level are input into the correlation model between heating load and the load-sensitive temperature isotherm level, wind chill, and water chill to obtain the winter heating load L for the forecast day. warm ;

[0066] The predicted value of the daily maximum load in winter is calculated using the formula for calculating the daily maximum load in winter. The formula for calculating the daily maximum load L in winter is as follows:

[0067] L = L basic +L warm +L holiday

[0068] Among them, L basic Based on the load, L warm For winter heating load, L holiday The workload is due to statutory holidays.

[0069] In a second aspect, the present invention provides a winter daily maximum load prediction system based on load-sensitive temperature distribution, including a processor and a storage medium;

[0070] The storage medium is used to store instructions;

[0071] The processor is configured to operate according to the instructions to perform the steps of the method according to the first aspect.

[0072] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0073] 1. This invention establishes a correlation model between load-sensitive temperature isotherm levels and winter heating load increments to predict daily maximum load. This method can achieve simple and rapid short-term quantitative prediction of daily maximum load through table lookup, and also has strong regional scalability, providing guidance for regional short-term winter heating load prediction.

[0074] 2. This invention utilizes observational analysis and coefficient of variation methods to statistically calculate the dispersion of the daily maximum load, determine the analysis period for winter heating load, and construct a basic load trend change model; comprehensively considering the influence of multiple factors, it sets the distribution level of the temperature isotherm sensitive to winter heating load.

[0075] 3. This invention uses multi-year historical load data, basic data from meteorological observation stations, multi-year historical observation data of meteorological elements, and future forecast data of meteorological elements as its data foundation. For the prediction of maximum daily load in winter, the influence of holiday load is eliminated, and a year-by-year statistical method is used to determine the negative growth pattern of statutory holiday load. The data is accurate and reliable. Attached Figure Description

[0076] Figure 1 This is the workflow of a method for predicting the maximum daily load in winter based on load-sensitive temperature distribution. Detailed Implementation

[0077] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0078] Example 1:

[0079] This embodiment provides a method for predicting the daily maximum load in winter based on load-sensitive temperature distribution, such as... Figure 1 As shown, it includes the following steps:

[0080] 1) The necessary basic data measurement results for developing a load forecasting method based on load-sensitive temperature distribution;

[0081] 2) The load correction of the daily maximum load sequence is achieved by using the statutory holiday load elimination method and the linear interpolation filling method;

[0082] 3) Use the method of annual statistics to determine the negative growth pattern of statutory holiday load;

[0083] 4) Using observational analysis and coefficient of variation as monthly units, statistically calculate the dispersion of daily maximum load and determine the analysis period for winter heating load;

[0084] 5) Using the conclusions of step 4), further calculate the annual base load and construct its trend change model;

[0085] 6) Taking into account the influence of multiple factors, set the temperature isotherm distribution level sensitive to winter heating load;

[0086] 7) Based on the data results from step 1), calculate the specific range of the load-sensitive temperature isotherm distribution during the analysis period in step 4);

[0087] 8) Based on the data results from step 1) and the calculation results from step 7), construct a correlation model between heating load and the distribution level of load-sensitive temperature isotherms.

[0088] 9) Based on the data results from step 1) and the correlation model from step 8), the daily maximum load forecast is achieved.

[0089] Specifically, step 1) includes information collection content and data compilation requirements.

[0090] The information collected includes: multi-year historical load data within the load forecast area, basic data from meteorological observation stations, multi-year historical observation data of meteorological elements, and future forecast data of meteorological elements.

[0091] The specific elements to be collected include:

[0092] Multi-year historical load data: Maximum load data for each year and day over the past 3-5 years within the load forecasting area;

[0093] Basic data of meteorological observation stations: station name, station number, administrative division, longitude, latitude, elevation, etc. of national standard meteorological observation stations within the load forecast area;

[0094] Historical observation data of meteorological elements over many years: actual observation data of temperature, rainfall, wind speed, etc. in the load forecast area over the past 3-5 years;

[0095] Meteorological element forecast data: forecast of the minimum daily temperature, rainfall, and maximum wind speed in the future within the load forecast area.

[0096] The specific elements compiled include:

[0097] By compiling historical real-time observation data of meteorological elements day by day, the daily minimum temperature is obtained;

[0098] By compiling historical real-time observation data of meteorological elements day by day, the maximum daily wind force can be obtained;

[0099] By compiling historical observation data of meteorological elements day by day, it is possible to determine whether it rains on a given day.

[0100] Specifically, step 2) is as follows:

[0101] The historical daily maximum load data collected in step 1) are arranged in time series and labeled with the day of the week. Based on the specific arrangements of domestic statutory holidays, the daily maximum load corresponding to weekdays and statutory holidays is separated.

[0102] Remove the daily maximum load corresponding to statutory holidays from the historical load data in step 1);

[0103] Linear interpolation is used to fill in the daily maximum load gaps on the excluded dates:

[0104] Taking a historical daily maximum load sequence as an example, the daily maximum load is denoted as L. Ri Where i = 0...n. The daily maximum loads numbered j to j+m in the sequence are removed because of statutory holidays;

[0105] Solve for the corrected daily maximum load

[0106]

[0107] in, —The maximum load correction value for the (j+1)th day;

[0108] —Measured value of maximum load on the j-th day;

[0109] —Measured value of the maximum load on the (j+m)th day;

[0110] The final result is a revised daily maximum load sequence after removing statutory holidays.

[0111] Specifically, step 3) includes the following steps:

[0112] Based on the revised daily maximum load sequence in step 2), and in accordance with the arrangements for public holidays and public holidays stipulated by current domestic laws, the daily maximum load corresponding to public holidays (weekends) and public holidays (New Year's Day, Spring Festival, Qingming Festival, Labor Day, Dragon Boat Festival, Mid-Autumn Festival, National Day, etc.) is separated.

[0113] The changes in the maximum daily load for each type of holiday are statistically analyzed year by year, including: statistical analysis of the negative growth of the load for each type of holiday (maximum value, minimum value, average value), statistical analysis of the trend of negative growth of the load for each type of holiday year by year (interannual changes, etc.), and construction of a model of the daily maximum load change pattern for statutory holidays.

[0114] This model is a mathematical statistics model. Model input: year, type of statutory holiday (Saturday, Sunday, New Year's Day, Spring Festival, Qingming Festival, Labor Day, Dragon Boat Festival, Mid-Autumn Festival, National Day), and load variation corresponding to each type of statutory holiday in different years; Model output: load variation for each type of statutory holiday, which can be calculated from the average value or linear fitting value of each year.

[0115] In addition, the model can also be established by using big data analysis and machine learning methods (such as ARIMA time series forecasting method, SVR support vector regression method, etc.) to establish a model of the daily maximum load variation pattern on statutory holidays.

[0116] Specifically, step 4) is as follows:

[0117] Based on the corrected daily maximum load sequence from step 2), the dispersion of daily maximum load during the autumn and winter seasons (September of the current year to March of the following year) is analyzed year by year on a monthly basis using methods such as combining observed data changes and climate change patterns, and the standard deviation evaluation method.

[0118] Plot the modified daily maximum load variation curve over time, and select months with higher dispersion based on the local climate (temperature) variation patterns in the load forecast area;

[0119] The coefficient of variation (CV) of the daily maximum load is calculated monthly, and months with higher dispersion are selected. The specific calculation method for CV is as follows:

[0120]

[0121]

[0122] Based on the analysis results of the two methods mentioned above, the months with higher dispersion were selected as the analysis period for winter heating load. Combining the results of the seasonal variation method and the coefficient of variation method, the conventional method uses the winter load (December to February of the following year) as the base load. This method combines the winter period (December to February of the following year) and the months with higher coefficient of variation, taking the intersection of the two to determine the winter heating load analysis period.

[0123] Selecting months with high dispersion means determining a threshold for the coefficient of variation by statistically analyzing the monthly coefficients of variation of a long series. A coefficient of variation higher than the preset threshold is considered high.

[0124] Specifically, step 5) includes the following steps:

[0125] Based on the dispersion results obtained in step 4), the months after excluding the analysis period of winter heating load are used as the months for calculating the base load.

[0126] The arithmetic mean method is used to calculate the daily maximum load for the aforementioned months year by year, which serves as the baseline load for the winter heating load of that year. A model for the annual variation trend of the baseline load is then established.

[0127] Baseline load is not a fixed value; it is influenced by macroeconomic trends, economic growth rate, population growth, and other macroeconomic factors. Therefore, a baseline load trend change model needs to be established based on local GDP, CPI, and population changes. Methods for establishing this model include:

[0128] Step 1: Identify the dominant factors affecting base load

[0129] The Pearson correlation coefficient was used to identify the dominant factors influencing the base load. The Pearson correlation coefficient ρ between two variables X and Y is calculated. X,Y for:

[0130]

[0131] Where X represents the socioeconomic factors affecting the base load; Y represents the base load; cov(X,Y) represents the covariance of the two data columns; and σ X σ Y The standard deviation of socioeconomic factors and load data is given.

[0132] Select socioeconomic factors strongly correlated with the base load (such as GDP, CPI, added value of the secondary industry, and the proportion of the secondary industry). Select factors with high Pearson correlation coefficients as the dominant factors affecting the base load.

[0133] In this method, the prediction of basic load does not distinguish between regions, the dominant factors are consistent, and once selected, it can be used universally.

[0134] Step 2: Establish a model of basic load variation trends

[0135] A basic load prediction model was established by using the least squares method, linear regression method, and polynomial fitting method, combined with the dominant factors affecting the basic load.

[0136] Specifically, step 6) includes the following steps:

[0137] Taking into account the distribution of minimum temperatures in the load forecast area, the variation of the frost line, the orientation of rivers and mountains, the distribution of geographical and climatic environment, and population distribution, the distribution of various temperature isotherms sensitive to load is determined.

[0138] Statistical analysis of long-term daily maximum load and minimum temperature data from the State Grid Corporation of China revealed that during winter, as the 0℃ (load-sensitive temperature) isotherm moves from north to south, there are instances of sharp increases and jumps in daily maximum load. Through data analysis, several fixed locations of the 0℃ isotherm distribution were identified, termed "load-sensitive temperature distribution levels." This patent is also based on these statistically derived levels.

[0139] Currently, this method primarily targets the State Grid Corporation's operating area for load division, which is categorized into five areas: Northern North China, the Yellow River basin, the Qinling-Huaihe River basin, the Yangtze River basin, and the Nanling-Wuyishan basin. If extended to other regions, load division needs to be based on actual frost conditions. In application, the load division for a given area can be directly determined using the address of the load forecasting region.

[0140] Specifically, step 7) includes the following steps:

[0141] We select isotherm levels representing weather stations based on load-sensitive temperatures to characterize the current isotherm positions. When the daily minimum temperature of more than half of the stations in a group varies within ±1℃ of the load-sensitive temperature, the isotherm level representing the load-sensitive temperature is considered to have reached its designated position.

[0142] Based on the determined isotherm range distribution locations, select representative meteorological station groups for each range.

[0143] If more than half of the stations in a group of meteorological stations record a daily minimum temperature that is at the load-sensitive temperature level (0℃), then the load-sensitive temperature level is determined to be within that range, and the isotherm range for each region is obtained. By querying the address of the forecast area, the specific range for that area can be obtained, i.e., the location of the load-sensitive temperature range. A group of stations does not have a fixed number of stations; instead, a representative meteorological station for each range is determined based on the aforementioned ranges. (Determination method: According to the geographical distribution of that range, search for nearby national standard meteorological stations based on the nearest distance principle).

[0144] Specifically, step 8) includes the following steps:

[0145] Based on historical daily maximum load data and load-sensitive temperature isotherm distribution data of minimum temperature, when the load-sensitive temperature isotherm shifts southward, the increment of the daily maximum load compared to the base load is calculated, which is the heating load.

[0146] Plot the correlation curve between heating load and load-sensitive temperature isotherm level. Based on the information compiled in step 1), such as the daily maximum wind force and whether it rains on the day, establish a correlation model between heating load and load-sensitive temperature isotherm level, wind chill (temperature drop caused by strong wind), and water chill (temperature drop caused by rainfall).

[0147] The model is established by employing big data analysis and machine learning methods (such as ARIMA time series forecasting method, SVR support vector regression method, etc.) to build a heating load forecasting model.

[0148] The model's inputs are: the location of the load-sensitive temperature distribution level, the rainfall, wind speed, and historical measured values ​​of the daily maximum load at the stations covered by the level; the model's output is: the predicted value of the daily maximum load.

[0149] Specifically, step 9) includes the following steps:

[0150] Obtain the year and date of the forecast date for the load forecast area and the forecast values ​​of the dominant factors (such as GDP, CPI, and added value of the secondary industry). The forecast values ​​of the dominant factors are obtained from the forecasts and statistical data on GDP, CPI, etc. published by the National Bureau of Statistics, the International Monetary Fund (IMF), etc.

[0151] Obtain the location of the load-sensitive temperature distribution level in the load forecast area, the forecast day's rainfall, wind speed, and historical daily maximum load measured values ​​for the stations covered by the level;

[0152] By inputting the predicted values ​​of key factors (such as GDP, CPI, and added value of the secondary industry) into the annual variation trend model of the base load, the base load L on the predicted date is obtained. basic ;

[0153] The type of holiday for the forecast date is determined based on the year and date of the forecast date. This holiday type is then input into the daily maximum load variation model for statutory holidays to obtain the statutory holiday load L for the forecast date. holiday ;

[0154] The location of the load-sensitive temperature distribution level in the load forecast area, the rainfall, wind speed, and historical daily maximum load measurements of the stations covered by the level are input into the correlation model between heating load and the load-sensitive temperature isotherm level, wind chill, and water chill to obtain the winter heating load L for the forecast day. warm ;

[0155] The predicted value of the daily maximum load in winter is calculated using the formula for calculating the daily maximum load in winter. The formula for calculating the daily maximum load L in winter is as follows:

[0156] L = L basic +L warm +L holiday

[0157] Among them, L basic Based on the load, L warm For winter heating load, L holiday The workload is due to statutory holidays.

[0158] This invention uses multi-year historical load data, basic data from meteorological observation stations, multi-year historical observation data of meteorological elements, and future forecast data of meteorological elements as data foundations. For the prediction of maximum daily load in winter, the influence of holiday load is eliminated, and a year-by-year statistical method is used to determine the negative growth pattern of statutory holiday load.

[0159] By using observational analysis and coefficient of variation, the dispersion of daily maximum load is statistically calculated to determine the analysis period of winter heating load and construct a basic load trend change model. Taking into account the influence of multiple factors, the distribution level of isotherms sensitive to winter heating load is set.

[0160] By establishing a correlation model between load-sensitive temperature isotherm levels and winter heating load increments, daily maximum load prediction can be achieved. This method can easily and quickly achieve short-term quantitative prediction of daily maximum load through table lookup, and it also has strong regional scalability, providing guidance for regional short-term winter heating load prediction.

[0161] Example 2:

[0162] This embodiment provides a winter daily maximum load prediction system based on load-sensitive temperature distribution, including a processor and a storage medium;

[0163] The storage medium is used to store instructions;

[0164] The processor is configured to operate according to the instructions to perform the steps of the method according to Embodiment 1.

[0165] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0166] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0167] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0168] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0169] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for predicting the daily maximum load in winter based on load-sensitive temperature distribution, characterized in that, Includes the following steps: S1. Acquire information data for the load forecasting area and compile specific elements to obtain data output content; S2. Based on the data results, the load correction of the daily maximum load sequence is achieved by using the statutory holiday load removal method and the linear interpolation filling method, and the corrected daily maximum load sequence after removing statutory holidays is obtained. S3. Based on the modified daily maximum load sequence, construct a model of the daily maximum load variation pattern on statutory holidays; Based on the revised daily maximum load sequence, and in accordance with the arrangements for public holidays and public holidays stipulated by current domestic laws, the daily maximum load corresponding to public holidays and public holidays is separated. The maximum load changes for each type of holiday are statistically analyzed year by year, the negative growth of the load for each type of holiday is statistically analyzed, and the negative growth trend of the load for each type of holiday is statistically analyzed year by year, so as to construct a model of the daily maximum load change pattern for statutory holidays. S4. Based on the modified daily maximum load sequence, statistically calculate the dispersion of the daily maximum load and determine the analysis period for winter heating load; Based on the modified daily maximum load sequence, and combining the observation of seasonal variation patterns and the coefficient of variation evaluation method, the dispersion of daily maximum load in autumn and winter is analyzed year by year on a monthly basis: Plot the modified daily maximum load variation curve over time, and select months with higher dispersion based on the local climate change patterns in the load forecast area; Calculate the coefficient of variation (CV) of the daily maximum load each month, and select the months with higher CVs. The specific calculation method for CV is as follows: ； ； Using the winter load as the baseline load, the months with the coefficient of variation exceeding the preset threshold are selected as the analysis period for the winter heating load. S5. Calculate the annual base load based on the analysis period of the winter heating load, and construct a base load trend change model; S6. Set the temperature isotherm distribution level sensitive to winter heating load; S7. Based on the distribution level of the load-sensitive temperature isotherm in winter heating load, calculate the specific level of the load-sensitive temperature isotherm distribution during the analysis period of winter heating load; S8. Based on historical daily maximum load data and the specific distribution levels of load-sensitive temperature isotherms, construct a correlation model between heating load and the distribution levels of load-sensitive temperature isotherms; S9. Based on the daily maximum load variation model of the statutory holidays, the basic load trend variation model, and the correlation model between heating load and load-sensitive temperature isotherm distribution level, the daily maximum load of the forecast day in the load forecast area is predicted. In step S6, setting the winter heating load sensitive temperature isotherm distribution level includes the following steps: Taking into account the distribution of minimum temperatures in the load forecast area, the variation of the frost line, the orientation of rivers and mountains, the distribution of geographical and climatic environment, and the population distribution, the distribution of each level of the load-sensitive temperature isotherm is determined. The gear ranges include: northern North China, along the Yellow River, along the Qinling-Huaihe River, along the Yangtze River, and along the Nanling-Wuyi Mountains; In step S8, constructing the correlation model between heating load and load-sensitive temperature isotherm distribution levels includes the following steps: Based on historical daily maximum load data and load-sensitive temperature isotherm distribution data of minimum temperature, when the load-sensitive temperature isotherm shifts southward, the increment of the daily maximum load compared to the base load is calculated, which is the heating load. Plot the correlation curve between heating load and load-sensitive temperature isotherm levels. Integrate daily maximum wind force and whether there is daily rainfall to establish a correlation model between heating load and load-sensitive temperature isotherm levels, wind chill, and water chill.

2. The method for predicting the daily maximum load in winter based on load-sensitive temperature distribution according to claim 1, characterized in that, In step S1, the information collection data includes: Multi-year historical load data: Multi-year historical daily maximum load data for the load forecasting area; Basic data of meteorological observation stations: station name, station number, administrative division, longitude, latitude, and elevation of national standard meteorological observation stations within the load forecast area; Historical meteorological data over many years: recent years' observed temperature, rainfall, and wind speed within the load forecast area; Meteorological element forecast data: forecast of the minimum daily temperature, the rainfall, and the maximum wind speed in the future within the load forecast area; The specific elements are compiled, including: By compiling historical real-time observation data of meteorological elements day by day, the daily minimum temperature is obtained; By compiling historical real-time observation data of meteorological elements day by day, the maximum daily wind force can be obtained; By compiling historical observation data of meteorological elements day by day, it is possible to determine whether it rains on a given day.

3. The method for predicting the daily maximum load in winter based on load-sensitive temperature distribution according to claim 2, characterized in that, In step S2, based on the data results, the load correction of the daily maximum load sequence is achieved using the statutory holiday load elimination method and the linear interpolation filling method, including: The historical daily maximum load data from the past few years are arranged in time series and labeled with the weekdays. Based on the specific arrangements of domestic statutory holidays, the daily maximum load corresponding to weekdays and statutory holidays is separated. Remove the daily maximum load corresponding to statutory holidays from the historical daily maximum load data; A linear interpolation method was used to fill in the gaps in the daily maximum load for the excluded dates. The actual daily maximum load was marked as follows: , where j=0……n; the daily maximum load in the sequence numbered j to j+m is removed because of statutory holidays; m is the statutory holiday. Solve for the modified daily maximum load : ； in, The maximum load correction value for the (j+1)th day; —Measured value of maximum load on the j-th day; The measured value of the maximum load on the (j+m)th day; The final result is a revised daily maximum load sequence after removing statutory holidays.

4. The method for predicting the daily maximum load in winter based on load-sensitive temperature distribution according to claim 1, characterized in that, In step S5, constructing the basic load trend change model includes the following steps: The months excluding the analysis period for winter heating load are used as the months for calculating the base load; The method of solving for the arithmetic mean is used to calculate the daily maximum load of the calculation month of the base load each year, which is used as the base load for the winter heating load of that year, and a model of the annual variation trend of the base load is established. Methods for establishing annual variation models of basic loads include: Socioeconomic factors strongly correlated with the base load were selected, and those with high Pearson correlation coefficients were chosen as the dominant factors influencing the base load. A basic load prediction model was established by using the least squares method, linear regression method, and polynomial fitting method, combined with the dominant factors affecting the basic load.

5. The method for predicting the daily maximum load in winter based on load-sensitive temperature distribution according to claim 1, characterized in that, In step S7, calculating the specific range of the load-sensitive temperature isotherm distribution during the analysis period of the winter heating load includes the following steps: Based on the geographical distribution of each meteorological level, and following the principle of proximity, nearby national standard meteorological stations were searched and selected. The isotherm range of load-sensitive temperature is selected to represent meteorological stations, respectively, to characterize the current isotherm position. When more than half of the selected stations have daily minimum temperatures that vary within ±1℃ of the load-sensitive temperature, the isotherm range of load-sensitive temperature is considered to have reached the position it represents.

6. The method for predicting the daily maximum load in winter based on load-sensitive temperature distribution according to claim 4, characterized in that, In step S9, the process of predicting the daily maximum load for the predicted day in the load prediction area includes the following steps: Obtain the year and date of the forecast date for the load forecast area and the forecast values ​​of the dominant factors, including the forecast values ​​of GDP, CPI, and added value of the secondary industry; Obtain the geographical location of the load forecast area, and based on the geographical location, obtain the location of the load-sensitive temperature distribution level in the load forecast area, the rainfall, wind speed, and historical daily maximum load measured values ​​of the stations covered by the level on the forecast day; Inputting the predicted values ​​of the dominant factors into the annual variation trend model of the base load yields the base load L for the predicted date. basic ; The holiday type for the forecast date is determined based on the year and date of the forecast date. This holiday type is then input into the daily maximum load variation model for statutory holidays to obtain the statutory holiday load L for the forecast date. holiday ; The location of the load-sensitive temperature distribution level in the load forecast area, the rainfall, wind speed, and historical daily maximum load measurements of the stations covered by the level are input into the correlation model between heating load and the load-sensitive temperature isotherm level, wind chill, and water chill to obtain the winter heating load L for the forecast day. warm ; Based on the formula for calculating the maximum daily load in winter, the predicted value of the maximum daily load in winter is calculated; the formula for calculating the maximum daily load L in winter is as follows: L= L basic + L warm + L holiday ; Among them, L basic Based on the load, L warm For winter heating load, L holiday The workload is due to statutory holidays.

7. A winter daily maximum load prediction system based on load-sensitive temperature distribution, characterized in that, Including processor and storage media; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1-6.

Images