A summer temperature-sensitive industry power consumption analysis method and system fused with meteorological features

Abstract

The present application belongs to the technical field of power system analysis and prediction, and particularly relates to a summer temperature-sensitive industry power consumption analysis method and system fusing meteorological characteristics. The method first acquires summer multidimensional meteorological data and summer daily power consumption data of each industry, then extracts multidimensional regional adaptive meteorological characteristics from the multidimensional meteorological data to construct a temperature-sensitive recognition dataset; subsequently, the correlation coefficient between the daily power consumption data of each industry and each type of feature in the temperature-sensitive recognition dataset is calculated, and the temperature-sensitive industry is obtained through threshold screening based on the correlation coefficient matrix; finally, the regression curve between the summer daily power consumption data of the temperature-sensitive industry and the daily temperature is fitted, the daily temperature of the prediction day is input into the regression curve, and the daily power consumption of the temperature-sensitive industry on the prediction day is obtained. The present application extracts industry-specific multidimensional regional adaptive meteorological characteristics from multidimensional meteorological data, and identifies the temperature-sensitive industry based on the multidimensional regional adaptive meteorological characteristics, which significantly improves the accuracy of temperature-sensitive industry identification.

Description

Technical Field

[0001] This invention belongs to the field of power system analysis and prediction technology, specifically relating to a method and system for analyzing electricity consumption in temperature-sensitive industries during summer by integrating meteorological characteristics. Background Technology

[0002] During the hot summer months, meteorological factors significantly impact electricity consumption, with the electricity consumption of temperature-sensitive industries being particularly closely linked to weather conditions. Accurately identifying these industries and quantifying the mechanisms by which meteorological characteristics affect their electricity consumption, enabling precise electricity demand forecasting and load allocation, is crucial for ensuring the safe and stable operation of the power grid, optimizing power purchase decisions, and enhancing emergency dispatch capabilities.

[0003] Traditional research and practice in summer electricity consumption analysis commonly suffer from the problems of relying solely on meteorological characteristics and insufficient consideration of industry-specific features. Traditional methods often employ single temperature indicators such as daily average temperature or calculate regional temperatures using spatial averaging, failing to adequately address the differences in the correlation between electricity consumption characteristics and temperature across different industries. This results in a lack of specificity in the matching analysis of industry-specific electricity consumption and meteorological conditions. More importantly, in extreme weather scenarios such as sustained high temperatures, changes in electricity load are not only affected by the daily temperature but also closely related to the duration and intensity of previous temperatures, exhibiting a significant cumulative temperature effect. However, existing mainstream methods such as baseline methods and time-series decomposition methods do not quantify this long-term and short-term cumulative effect, making it difficult to accurately capture the key inflection points of sharp load increases. For example, when there are three or more consecutive days of temperatures exceeding 35°C, air conditioning load often exhibits a non-linear growth trend. Traditional models, unable to characterize the load characteristics driven by this cumulative effect, struggle to accurately predict such load changes. Furthermore, most existing forecasting models take the total social load as the analysis object, without decoupling the load of different industries, and ignore the differences in electricity consumption characteristics between temperature-sensitive and non-temperature-sensitive industries. The electricity consumption curves of temperature-sensitive industries are highly synchronized with temperature fluctuations, while those of non-temperature-sensitive industries show a weak correlation with temperature. This mixed analysis method further reduces the accuracy of electricity consumption analysis.

[0004] To improve the accuracy of summer electricity consumption forecasting, recent research has focused on two main areas: quantifying the cumulative effect of temperature and decoupling industry loads, achieving some progress. Regarding the quantification of the cumulative effect of temperature, some studies have modified traditional forecasting models by introducing accumulated temperature indices or weighting historical temperatures. For example, Granger causality tests are used to determine the accumulated temperature threshold (e.g., two consecutive days above 26°C triggering the cumulative effect), and attenuation coefficients are used to quantify the impact of historical temperatures on current loads. In terms of decoupling industry loads, generalized additive models (GAMs) are mainly used to decompose the temperature-sensitive component of the load, separating base loads from temperature-regulating loads by incorporating social factors such as holidays, or cluster analysis is used to identify the electricity consumption patterns of different industries. However, these studies still have significant shortcomings: the setting of accumulated temperature thresholds and attenuation coefficients lacks a unified standard, limiting the accuracy of quantifying the cumulative effect; in the process of decoupling industry loads, the classification of temperature-sensitive industries relies heavily on manual experience and lacks objective statistical analysis, making it difficult to guarantee the reliability of the decoupling results.

[0005] In summary, existing summer electricity consumption analysis solutions still have several significant shortcomings: First, the integration of meteorological characteristics is insufficient, still relying primarily on a single temperature indicator, and failing to construct a meteorological characteristic system adapted to industry characteristics by considering regional differences in electricity distribution, resulting in insufficient accuracy in identifying temperature-sensitive industries; Second, the consideration of the key meteorological derivative characteristic of long-term and short-term temperature accumulation is still inadequate, with most solutions still relying on the original meteorological data of the day for electricity consumption forecasting, failing to effectively avoid prediction errors caused by cumulative effects; Third, in the process of decomposing basic electricity consumption and air conditioning electricity consumption, the constraints of meteorological steady-state intervals on the selection of typical days are not fully considered, making the decomposition results susceptible to interference from non-steady-state meteorological conditions, and the accuracy is insufficient to meet the needs of practical applications; Fourth, the existing analysis results lack sufficient interpretability in relation to meteorological characteristics, failing to clearly quantify the impact mechanism of meteorological factors on electricity load, and making it difficult to effectively support accurate decision-making in core business scenarios such as power grid agency power purchase and emergency dispatch. Summary of the Invention

[0006] The purpose of this invention is to address the aforementioned problems in the existing technology by providing a summer-based electricity consumption analysis method and system that integrates meteorological characteristics to improve the identification accuracy of temperature-sensitive industries and thus enhance the precision of electricity consumption analysis.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows:

[0008] In a first aspect, the present invention provides a method for analyzing electricity consumption in temperature-sensitive industries during summer by incorporating meteorological characteristics, the method comprising:

[0009] S1. Obtain multi-dimensional meteorological data of the analysis area in summer and daily electricity consumption data of various industries in summer;

[0010] S2. Extract multi-dimensional regional adaptive meteorological features from multi-dimensional meteorological data to construct a temperature-sensitive identification dataset; the multi-dimensional regional adaptive meteorological features include zoned electricity-weighted meteorological features and regional average meteorological features;

[0011] S3. Calculate the correlation coefficients between the daily electricity consumption data of each industry in summer and various features in the temperature-sensitive identification dataset to construct an industry-meteorological feature correlation coefficient matrix; based on the industry-meteorological feature correlation coefficient matrix, perform threshold screening to obtain the temperature-sensitive industries;

[0012] S4. Fit the regression curve between the daily electricity consumption data and daily temperature of the temperature-sensitive industry in summer, obtain the daily temperature of the predicted day and input it into the regression curve to obtain the daily electricity consumption of the temperature-sensitive industry on the predicted day.

[0013] In S2, the zoned electricity weighted meteorological characteristics include daily weighted maximum temperature, daily weighted minimum temperature and daily weighted average temperature; the regional average meteorological characteristics include daily average maximum temperature, daily average minimum temperature and daily average temperature.

[0014] Preferably, the following steps are used to extract the zoned electricity weighted meteorological features from multi-dimensional meteorological data: obtain the proportion of daily electricity consumption of a single industry in each administrative region within the analysis area in the previous year to the total daily electricity consumption of the corresponding industry in the analysis area, and use this proportion as a weight to perform weighted summation of the daily maximum temperature, daily minimum temperature and daily average temperature of each administrative region to obtain the daily weighted maximum temperature, daily weighted minimum temperature and daily weighted average temperature.

[0015] The following steps are used to extract regional average meteorological characteristics from multi-dimensional meteorological data: average the daily maximum temperature, daily minimum temperature and daily average temperature of each administrative region in the analysis area to obtain the daily average maximum temperature, daily average minimum temperature and daily average temperature.

[0016] Preferably, in S4, the daily temperature is corrected using a pre-constructed long-short cumulative effect correction model, and the regression curve between the daily electricity consumption data of the temperature-sensitive industry in summer and the corrected daily temperature is fitted. Based on the regression curve and the corrected daily temperature of the predicted day, the daily electricity consumption of the temperature-sensitive industry on the predicted day is obtained; the daily temperature is the daily average temperature or the daily weighted average temperature.

[0017] Preferably, in S4, the expression for the long-short cumulative effect correction model is:

[0018] ;

[0019] ;

[0020] In the above formula, , These are the daily temperatures before and after the correction, respectively. , These are the short-term meteorological cumulative effect value and the long-term meteorological cumulative effect value, respectively. To and The corresponding weighting coefficients; , , They are respectively with , , The corresponding weighting coefficients; To be consistent with the previous day to the previous The weighting coefficient corresponding to the average daily temperature. ; , , , They were 1 day ago, 2 days ago, and 3 days ago, respectively. The temperature of the day before yesterday.

[0021] Preferably, in S4, the weight coefficients of each item in the long-term and short-term cumulative effect correction model are optimized with the goal of maximizing the correlation coefficient between the daily electricity consumption data of the temperature-sensitive industry in summer and the corrected daily temperature.

[0022] Preferably, the method for analyzing electricity consumption in temperature-sensitive industries during summer further includes:

[0023] S5. Calculate the total monthly electricity consumption of the temperature-sensitive industry based on the predicted daily electricity consumption, obtain the monthly basic electricity consumption of the temperature-sensitive industry, and calculate the meteorological contribution based on the total monthly electricity consumption and the monthly basic electricity consumption of the temperature-sensitive industry. The larger the meteorological contribution, the higher the contribution of meteorology to the summer electricity consumption of the temperature-sensitive industry.

[0024] Preferably, in S5, the formula for calculating the meteorological contribution is:

[0025] ;

[0026] .

[0027] Preferably, in S5, the monthly basic electricity consumption for the temperature-sensitive industry is obtained according to the following steps:

[0028] A daily temperature steady-state range is pre-defined. In spring and autumn, continuous daily series that meet the meteorological steady-state range are selected and named typical daily series. The daily electricity consumption of temperature-sensitive industries in the typical daily series is obtained and the average value is taken. This average value is used as the daily basic electricity consumption of temperature-sensitive industries. The monthly basic electricity consumption of temperature-sensitive industries is calculated based on the daily basic electricity consumption of temperature-sensitive industries.

[0029] Secondly, the present invention provides a summer temperature-sensitive industry power consumption analysis system that integrates meteorological characteristics, the summer temperature-sensitive industry power consumption analysis system comprising:

[0030] The data acquisition module is used to acquire multi-dimensional meteorological data of the analysis area in summer, as well as daily electricity consumption data of various industries in summer.

[0031] The temperature-sensitive feature extraction module is used to extract multi-dimensional regional adaptive meteorological features from multi-dimensional meteorological data to construct a temperature-sensitive identification dataset; the multi-dimensional regional adaptive meteorological features include zoned electricity-weighted meteorological features and regional average meteorological features;

[0032] The temperature-sensitive industry screening module is used to first calculate the correlation coefficient between the daily electricity consumption data of each industry in summer and various features in the temperature-sensitive identification dataset, so as to construct an industry-meteorological feature correlation coefficient matrix. Then, based on the industry-meteorological feature correlation coefficient matrix, threshold screening is performed to obtain the temperature-sensitive industries.

[0033] The daily electricity consumption calculation module is used to fit the regression curve between the daily electricity consumption data and the daily temperature of the temperature-sensitive industry in summer, obtain the daily temperature of the predicted day and input it into the regression curve to obtain the daily electricity consumption of the temperature-sensitive industry on the predicted day.

[0034] Preferably, the zoned electricity-weighted meteorological characteristics include daily weighted maximum temperature, daily weighted minimum temperature, and daily weighted average temperature; the regional average meteorological characteristics include daily average maximum temperature, daily average minimum temperature, and daily average temperature.

[0035] Preferably, the temperature-sensitive feature extraction module is used to extract zoned electricity weighted meteorological features from multi-dimensional meteorological data according to the following steps: obtaining the proportion of daily electricity consumption of a single industry in each administrative region within the analysis area in the previous year to the total daily electricity consumption of the corresponding industry in the analysis area, and using this proportion as a weight to perform weighted summation of the daily maximum temperature, daily minimum temperature and daily average temperature of each administrative region to obtain the daily weighted maximum temperature, daily weighted minimum temperature and daily weighted average temperature.

[0036] The temperature-sensitive feature extraction module is used to extract regional average meteorological features from multi-dimensional meteorological data according to the following steps: average the daily maximum temperature, daily minimum temperature and daily average temperature of each administrative region in the analysis area to obtain the daily average maximum temperature, daily average minimum temperature and daily average temperature.

[0037] Preferably, the daily electricity consumption calculation module includes a correction module and a prediction module. The correction module is used to correct the daily temperature using a pre-built long-short cumulative effect correction model. The prediction module is used to fit the regression curve between the daily electricity consumption data of the temperature-sensitive industry in summer and the corrected daily temperature. Based on the regression curve and the corrected daily temperature of the prediction day, the daily electricity consumption of the temperature-sensitive industry on the prediction day is obtained. The daily temperature is the daily average temperature or the daily weighted average temperature.

[0038] Preferably, the expression for the long-short cumulative effect correction model is:

[0039] ;

[0040] ;

[0041] In the above formula, , These are the daily temperatures before and after the correction, respectively. , These are the short-term meteorological cumulative effect value and the long-term meteorological cumulative effect value, respectively. To and The corresponding weighting coefficients; , , They are respectively with , , The corresponding weighting coefficients; To be consistent with the previous day to the previous The weighting coefficient corresponding to the average daily temperature. ; , , , They were 1 day ago, 2 days ago, and 3 days ago, respectively. The temperature of the day before yesterday.

[0042] Preferably, the correction module is used to optimize the weight coefficients of various components in the long-term and short-term cumulative effect correction model with the goal of maximizing the correlation coefficient between the daily electricity consumption data of the temperature-sensitive industry in summer and the corrected daily temperature.

[0043] Preferably, the summer temperature-sensitive industry power consumption analysis system further includes a meteorological contribution analysis module;

[0044] The meteorological contribution analysis module is used to calculate the total monthly electricity consumption of the temperature-sensitive industry based on the predicted daily electricity consumption, obtain the monthly basic electricity consumption of the temperature-sensitive industry, and calculate the meteorological contribution based on the total monthly electricity consumption and the monthly basic electricity consumption of the temperature-sensitive industry. The larger the meteorological contribution, the higher the meteorological contribution to the summer electricity consumption of the temperature-sensitive industry.

[0045] Preferably, the formula for calculating the meteorological contribution is:

[0046] ;

[0047] .

[0048] Preferably, the meteorological contribution analysis module is used to obtain the monthly basic electricity consumption of the temperature-sensitive industry according to the following steps: pre-setting the daily temperature steady-state range, selecting continuous daily series that meet the meteorological steady-state range in spring and autumn and naming them as typical daily series, obtaining the daily electricity consumption of the temperature-sensitive industry in the typical daily series and taking the average value, using the average value as the daily basic electricity consumption of the temperature-sensitive industry, and calculating the monthly basic electricity consumption of the temperature-sensitive industry based on the daily basic electricity consumption of the temperature-sensitive industry.

[0049] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0050] 1. The summer electricity consumption analysis method for temperature-sensitive industries described in this invention extracts industry-specific multi-dimensional regional adaptive meteorological features from multi-dimensional meteorological data. The multi-dimensional regional adaptive meteorological features include zoned electricity consumption weighted meteorological features and regional average meteorological features. The zoned electricity consumption weighted meteorological features are constructed by the zoned electricity consumption weighting method, and the regional average meteorological features are constructed by the averaging method. This achieves deep integration of meteorological features and electricity consumption data, significantly improving the accuracy and scientific nature of temperature-sensitive industry identification.

[0051] 2. The summer electricity consumption analysis method for temperature-sensitive industries described in this invention incorporates the cumulative effects of long-term and short-term meteorological conditions into the electricity consumption forecast for temperature-sensitive industries. By constructing a correction model for the cumulative effects of long-term and short-term conditions and optimizing the weighting coefficients, the forecasting error caused by the cumulative effects of meteorological conditions is effectively eliminated, improving the accuracy of summer electricity consumption forecasting for temperature-sensitive industries and giving the forecast results stronger meteorological interpretability.

[0052] 3. The summer temperature-sensitive industry electricity consumption analysis method described in this invention, on the one hand, realizes the quantification of meteorological-driven incremental load, and on the other hand, uses the meteorological steady-state range as a constraint to screen typical daily samples, ensuring the objectivity of basic electricity consumption calculation, thereby improving the reliability of air conditioning electricity consumption decomposition.

[0053] 4. The summer temperature-sensitive industry power consumption analysis method described in this invention uses meteorological characteristics as the core link to deeply adapt to multiple core business scenarios of the power grid, realizing the full-link application of meteorological characteristics from the data layer to the decision-making layer, and providing interpretable and accurate technical support for power supply guarantee decisions during the peak summer season. Attached Figure Description

[0054] Figure 1 This is a flowchart of the summer temperature-sensitive industry power consumption analysis method described in this invention.

[0055] Figure 2 Heatmap showing the correlation coefficients between daily electricity consumption in some industries and six types of meteorological characteristics (Summer-1).

[0056] Figure 3 Heatmap showing the correlation coefficients between daily electricity consumption in some industries and six types of meteorological characteristics (Summer-2).

[0057] Figure 4 Heatmap showing the correlation coefficients between daily electricity consumption in some industries and six types of meteorological characteristics (Summer-3).

[0058] Figure 5 Heatmap showing the correlation coefficients between daily electricity consumption in some industries and six types of meteorological characteristics (Summer-4).

[0059] Figure 6 This is a chart showing daily electricity consumption data and meteorological data for urban and rural residents in the summer of 2024.

[0060] Figure 7 A regression curve showing the relationship between daily electricity consumption for urban and rural residents and the corrected daily temperature.

[0061] Figure 8 A graph showing daily electricity consumption data for urban and rural residents in a typical daily sequence.

[0062] Figure 9 This is a structural block diagram of the summer temperature-sensitive industrial power consumption analysis system described in this invention. Detailed Implementation

[0063] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings.

[0064] Example 1:

[0065] See Figure 1 A method for analyzing electricity consumption in temperature-sensitive industries during summer, incorporating meteorological characteristics, is performed in the following steps:

[0066] S1. Obtain multi-dimensional meteorological data and daily electricity consumption data of various industries in the analysis area in recent years.

[0067] First, the analysis area and time range were determined. The electricity consumption analysis time range focused on summer (June-August), while data from spring and autumn (April-May and October-November) were collected simultaneously as a benchmark.

[0068] Then, data collection is carried out. First, daily electricity consumption data of various industries in the region are collected and analyzed, as well as basic attribute information such as industry codes and industry names, to form the original daily electricity consumption dataset. Second, multi-dimensional meteorological data of different administrative regions in the region are collected and analyzed, including daily maximum temperature, daily minimum temperature, and daily average temperature, to form the original meteorological dataset.

[0069] Next, data preprocessing is performed: the collected raw daily electricity consumption dataset and raw meteorological dataset are cleaned (including imputation of missing values ​​using linear interpolation and removal of outliers using the 3σ criterion) to remove missing and outliers; the two datasets are then correlated by date dimension.

[0070] S2. Extract multi-dimensional regional adaptive meteorological features from multi-dimensional meteorological data to construct a temperature-sensitive identification dataset; the multi-dimensional regional adaptive meteorological features include zoned electricity-weighted meteorological features and regional average meteorological features.

[0071] Specifically, the zoned electricity weighted meteorological characteristics include daily weighted maximum temperature, daily weighted minimum temperature, and daily weighted average temperature; the regional average meteorological characteristics include daily average maximum temperature, daily average minimum temperature, and daily average temperature.

[0072] Specifically, the following steps are used to extract the weighted meteorological characteristics of electricity consumption in different regions from multi-dimensional meteorological data: First, obtain the proportion of daily electricity consumption of a single industry in each administrative region within the analysis area in the previous year to the total daily electricity consumption of the corresponding industry in the analysis area. Then, use this proportion as a weight to sum the daily maximum temperature, daily minimum temperature, and daily average temperature of each administrative region to obtain the daily weighted maximum temperature, daily weighted minimum temperature, and daily weighted average temperature. For example, suppose the proportions of daily industrial electricity consumption in cities A, B, and C of a certain province to the total daily industrial electricity consumption of the entire province are 20%, 15%, and 12%, respectively. The daily weighted maximum temperature is calculated as: Daily maximum temperature of city A × 20% + Daily maximum temperature of city B × 15% + Daily maximum temperature of city C × 15%. Similarly, the weighted minimum temperature and weighted average temperature of industry are calculated.

[0073] The following steps are used to extract regional average meteorological characteristics from multi-dimensional meteorological data: average the daily maximum temperature, daily minimum temperature, and daily average temperature of each administrative region within the analysis area to obtain the daily average maximum temperature, daily average minimum temperature, and daily average temperature; for example, the average of the daily maximum temperatures of cities A, B, and C in a certain province is calculated as the provincial average maximum temperature, and the daily average minimum temperature and daily average temperature are obtained in the same way.

[0074] S3. Calculate the correlation coefficients between the daily electricity consumption data of each industry and the various features in the temperature-sensitive identification dataset to construct an industry-meteorological feature correlation coefficient matrix; based on the industry-meteorological feature correlation coefficient matrix, perform threshold screening to obtain the temperature-sensitive industries.

[0075] Specifically, the calculation of the correlation coefficient between the daily electricity consumption data of each industry and the various features in the temperature-sensitive identification dataset refers to: using the Pearson correlation coefficient method to batch calculate the correlation coefficient between the daily electricity consumption data of each industry and the six features in the temperature-sensitive identification dataset, and using the correlation coefficient to quantify the degree of influence of different meteorological features on the electricity consumption of each industry.

[0076] Specifically, the threshold screening based on the industry-meteorological feature correlation coefficient matrix to obtain temperature-sensitive industries refers to: drawing a heat map based on the industry-meteorological feature correlation coefficient matrix to intuitively present the correlation strength between each industry and meteorological features; setting an absolute value of correlation coefficient ≥ 0.6 as a significant correlation threshold, and classifying industries whose daily electricity change trend is basically consistent with the fluctuation of meteorological features as temperature-sensitive industries.

[0077] S4. Fit the regression curve between the daily electricity consumption data and daily temperature of the temperature-sensitive industry in summer, obtain the daily temperature of the predicted day and input it into the regression curve to obtain the daily electricity consumption of the temperature-sensitive industry on the predicted day.

[0078] Specifically, daily temperature has a meteorological cumulative effect on the daily electricity consumption of temperature-sensitive industries. This includes both long-term and short-term meteorological cumulative effects. The long-term effect refers to the overall increase in daily electricity consumption of temperature-sensitive industries over the entire summer period (June-August), synchronized with the overall upward trend in temperature. The short-term effect refers to the sharp increase in daily electricity consumption in temperature-sensitive industries when the maximum temperature exceeds 35°C for several consecutive days and the inter-diurnal temperature difference is less than 1°C. Furthermore, the inflection point of short-term electricity consumption increases or decreases is not strictly synchronized with meteorological changes. To eliminate prediction errors caused by the meteorological cumulative effect and improve prediction accuracy, this invention first uses a pre-constructed long-term and short-term cumulative effect correction model to correct the daily temperature. Then, it fits a regression curve between the daily electricity consumption data of temperature-sensitive industries in summer and the corrected daily temperature. Based on the regression curve and the corrected daily temperature on the predicted day, the daily electricity consumption of temperature-sensitive industries on the predicted day is obtained. The daily temperature is either the daily average temperature or the daily weighted average temperature.

[0079] Specifically, the expression for the long-short cumulative effect correction model is as follows:

[0080] ;

[0081] ;

[0082] In the above formula, , These are the daily temperatures before and after the correction, respectively. , These are the short-term meteorological cumulative effect value and the long-term meteorological cumulative effect value, respectively. To and The corresponding weighting coefficients; , , They are respectively with , , The corresponding weighting coefficients; To be consistent with the previous day to the previous The weighting coefficient corresponding to the average daily temperature. General settings For 30 days; , , , They were 1 day ago, 2 days ago, and 3 days ago, respectively. The temperature of the day before yesterday.

[0083] In the aforementioned long-term and short-term cumulative effect correction model, the short-term meteorological cumulative effect value is determined using the daily temperatures of the most recent three days. The long-term meteorological cumulative effect value was determined using the daily temperature of the most recent month. Based on the original daily temperature information, long-term and short-term cumulative effects are integrated to eliminate prediction errors caused by meteorological cumulative effects.

[0084] The weighting coefficients should meet the following constraints: The initial values ​​of each weight coefficient are set empirically, aiming to maximize the correlation coefficient between the daily electricity consumption data of the temperature-sensitive industry in summer and the corrected daily temperature, and further optimize the weight coefficients in the long-term and short-term cumulative effect correction model. The optimization algorithm can adopt a genetic algorithm. The steps of weight coefficient optimization based on the genetic algorithm include: 1. Extracting the daily electricity consumption data of the temperature-sensitive industry in recent years in summer, and associating it with meteorological data such as the daily temperature of the current day, the temperature of the last 3 days, and the temperature of the last month to form a parameter training dataset; the fields of the parameter training dataset include: date (June-August of the last three years), day type (identifying weekday and weekend), temperature-sensitive industry code, temperature-sensitive industry name, daily electricity consumption (kWh), daily temperature of the current day, daily temperature 1 day ago, daily temperature 2 days ago, daily temperature 3 days ago, and the average daily temperature of the previous 30 days; 2. Using a genetic algorithm to train the long-term and short-term cumulative effect correction model, with the training objective of maximizing the correlation coefficient between the daily electricity consumption data of the temperature-sensitive industry in summer and the corrected daily temperature, to obtain the optimal combination of weight coefficients.

[0085] Specifically, when fitting the regression curve between the daily electricity consumption data of the temperature-sensitive industry in summer and the corrected daily temperature, abnormal fluctuation points (i.e., daily electricity consumption data affected by special events) are removed, and the reliability is verified through R² consistency test. The corrected daily temperature of the predicted day is input into the fitted regression curve to obtain the daily electricity consumption of the temperature-sensitive industry on the predicted day.

[0086] S5. Calculate the total monthly electricity consumption of the temperature-sensitive industry based on the predicted daily electricity consumption, obtain the monthly basic electricity consumption of the temperature-sensitive industry, and calculate the meteorological contribution based on the total monthly electricity consumption and the monthly basic electricity consumption of the temperature-sensitive industry. The larger the meteorological contribution, the higher the contribution of meteorology to the summer electricity consumption of the temperature-sensitive industry.

[0087] Specifically, the monthly basic electricity consumption of the temperature-sensitive industry is obtained according to the following steps: A daily temperature steady-state range is pre-defined, where the daily average temperature or daily weighted average temperature is 10-22℃. Within this temperature range, the impact of meteorological conditions on daily electricity consumption is minimal, indicating a steady state. Then, in spring and autumn, a sliding t-test is used to screen continuous daily series that meet the meteorological steady-state range and name them as typical daily series. The daily electricity consumption of the temperature-sensitive industry within these typical daily series is obtained and averaged. This average is used as the daily basic electricity consumption of the temperature-sensitive industry in spring and autumn, ensuring that the daily basic electricity consumption accurately reflects the industry's normal production and living electricity consumption levels. Finally, the average daily basic electricity consumption of the temperature-sensitive industry in spring and autumn is averaged, and this average is used as the summer average daily basic electricity consumption of the temperature-sensitive industry. Based on the daily basic electricity consumption of the temperature-sensitive industry and the weighting of the number of working days and weekends within a month, the monthly basic electricity consumption of the temperature-sensitive industry is calculated through weighted summation.

[0088] Specifically, the formula for calculating the meteorological contribution is as follows:

[0089] ;

[0090] .

[0091] The electricity consumption analysis results are compiled by summarizing the list of temperature-sensitive industries, corrected daily temperature, daily electricity consumption, monthly total electricity consumption, and monthly air conditioning electricity consumption. This helps to achieve accurate, meteorologically interpretable decision-making for core power grid business scenarios. The decisions include: 1. Optimization of agency electricity purchase decisions: Monthly air conditioning electricity consumption can reflect the incremental load driven by the high-temperature meteorological characteristics of summer, which helps to improve the accuracy of summer electricity purchase demand assessment and effectively reduce the deviation of electricity purchase costs; 2. Electricity consumption measurement monitoring and analysis: The corrected daily temperature is applied to the identification of similar days that considers meteorological similarity, replacing the traditional similar day identification method based on date matching or single raw meteorological data. 1. Intelligent year-on-year and month-on-month calculation to build a meteorologically interpretable power monitoring and analysis benchmark; 2. Dynamic scheduling of backup power: Set early warning trigger thresholds with the corrected daily temperature as the core parameter. When the corrected daily temperature reaches the preset threshold (corresponding to extreme high temperature meteorological conditions), the backup power scheduling early warning mechanism is automatically triggered to improve the power grid's emergency response capability under extreme meteorological scenarios; 3. Precise control of orderly power consumption: Quantify the contribution of meteorological characteristics to the summer power consumption of temperature-sensitive industries by measuring meteorological contribution. Formulate differentiated control strategies for the meteorological response characteristics of different temperature-sensitive industries to improve the effectiveness and pertinence of orderly power consumption.

[0092] Performance verification:

[0093] Daily electricity consumption data for 133 industries (covering primary, secondary, and tertiary industries, as well as urban and rural residential electricity consumption) were obtained from the provincial electricity marketing system, including industry codes, industry names, and other attribute information, forming a raw daily electricity consumption dataset. Daily maximum, minimum, and average temperatures for 14 prefecture-level cities in the province were obtained from the provincial meteorological data disclosure platform, forming a raw meteorological dataset. Using the method described in this invention, a 133×6-dimensional correlation coefficient matrix between daily electricity consumption for each industry and six types of meteorological characteristics was calculated. Heat maps for Summer 1, Summer 2, Summer 3, and Summer 4 were drawn based on the 133×6-dimensional correlation coefficient matrix, as shown below. Figures 2 to 5 As shown in the 133×6-dimensional correlation coefficient matrix, the correlation coefficient between daily electricity consumption for urban and rural residents and the daily weighted average temperature is 0.89, and the correlation coefficient between daily electricity consumption and the daily average temperature is 0.88; the correlation coefficient between daily electricity consumption in the tertiary industry and the daily weighted average temperature is 0.82; the correlation coefficients between daily electricity consumption in the secondary industry and various meteorological characteristics are all less than 0.3. Analysis reveals that the absolute value of the correlation coefficient between the weighted average temperature and daily electricity consumption is greater than that of the other five meteorological characteristics, indicating that the daily weighted average temperature is most sensitive to changes in daily electricity consumption. Based on the meteorological characteristic most sensitive to changes in daily electricity consumption (i.e., daily weighted average temperature), temperature-sensitive industries are screened. Specifically, a correlation coefficient absolute value ≥ 0.6 is set as a significant correlation threshold. Finally, urban and rural residents' electricity consumption, the primary industry, the tertiary industry and its sub-categories such as accommodation and catering, and the construction industry within the secondary industry are classified as temperature-sensitive industries if the absolute value of the correlation coefficient is ≥ 0.6.

[0094] Taking urban and rural residents' electricity consumption as an example, we obtained their daily electricity consumption data and meteorological data for the summer of 2024. Figure 6 As shown, in the long term, the daily temperature from June to August rises from 26℃ to 31℃, and the daily electricity consumption of urban and rural residents increases from 140 million kWh to 292 million kWh, showing a synchronized long-term trend; in the short term ( Figure 6 (The area circled in red) From July 21st to 24th, the daily temperature exceeded 36℃ for four consecutive days, with adjacent daily temperature variations less than 1℃. However, the daily electricity consumption of urban and rural residents surged from 307 million kWh to 372 million kWh. The inflection point and magnitude of the daily electricity consumption increase / decrease showed a short-term discrepancy with the daily temperature changes. A long-short cumulative effect correction model was used to correct for the daily temperature fluctuations, and a genetic algorithm was used to optimize the model's weight coefficients. The genetic algorithm parameters were set as follows: population size 150, number of iterations 700, crossover probability 0.9, and mutation probability 0.4. The optimal combination of weight coefficients was obtained as follows: It is 0.28238. It is 0.34423. It is 0.07448. The value is 0.05157. The value is 0.24734. The regression curve obtained by fitting the relationship between daily electricity consumption for urban and rural residents and the corrected daily temperature is shown below. Figure 7 As shown, the regression equation is In the formula This represents daily electricity consumption data. This indicates that the model has good consistency.

[0095] As shown in Table 1, the following uses late August 2024 as an example to verify the advantage of the corrected daily temperature in reflecting the sensitivity of daily electricity consumption to temperature rise. First, correction for abnormal fluctuations. On August 21st and 23rd, 2024, the temperature rises were 1.2℃ and 0.42℃, respectively, corresponding to daily electricity consumption increases of 0.05 billion kWh and 0.38 billion kWh, respectively. At this time, temperature and electricity consumption changes showed an asymmetrical response. After temperature correction, the temperature rises on the 21st and 23rd were 0.15℃ and 0.77℃, respectively, corresponding to electricity consumption increases of 0.05 billion kWh and 0.38 billion kWh, respectively. At this time, temperature and electricity consumption changes showed a stable growth relationship, accurately matching the trend of electricity consumption changes and revealing the true impact of thermal inertia. Second, verification of model consistency. After temperature correction, the temperature rose from 29.91℃ on the 21st to 30.77℃ on the 22nd, resulting in an increase in electricity consumption of 0.41 billion kWh. This is consistent with the theoretical value of the regression equation (which calculates that a temperature rise of 30℃ to 31℃ corresponds to an increase in daily electricity consumption of 0.411 billion kWh), further verifying the reliability of the model. In summary, this invention, by quantifying the cumulative effects of long and short periods and correcting daily temperature, overcomes the distortion defects of the original temperature in responses to sudden meteorological changes.

[0096] Table 1. Comparison of daily temperatures before and after correction in late August 2024

[0097]

[0098] Continuous daily series satisfying the meteorological steady-state range were selected from the spring and autumn of 2024 and named typical daily series. Daily electricity consumption of urban and rural residents within these typical daily series was obtained. Figure 8As shown, the average daily basic electricity consumption for weekdays in spring (April) of 2024 was 93.396 million kWh, and the average daily basic electricity consumption for weekends was 97.024 million kWh. The average daily basic electricity consumption for weekdays in autumn (October-November) was 95.106 million kWh, and the average daily basic electricity consumption for weekends was 98.741 million kWh. Further calculations show that the average daily basic electricity consumption for weekdays in summer (June) of 2024 was (93.396 + 95.106) / 2 = 94.251 million kWh, and the average daily basic electricity consumption for weekends was (970... 2.4 + 9874.1) / 2 = 9788.25 million kWh. Since June has 19 working days and 11 weekend days (including the Dragon Boat Festival), the monthly base electricity consumption for June 2024 is further calculated as 9425.1 × 19 + 9788.25 × 11 = 286747.4 million kWh. The monthly base electricity consumption for June 2023 was 270168.3 million kWh, with a growth rate of (286747.4 - 270168.3) / 270168.3 × 100% ≈ 6.1%. Further calculations based on the regression curves show that the total monthly electricity consumption for urban and rural residents in June 2024 was 406,198.25 million kWh. The monthly air conditioning electricity consumption was calculated as 406,198.25 - 286,747.4 = 119,450.8 million kWh, and the meteorological contribution rate was approximately 119,450.8 / 406,198.25 × 100% ≈ 29.4%.

[0099] The above calculation results were summarized into electricity consumption analysis results and applied to the core business of the provincial power grid: 1. Agency power purchase decision: Quantify the impact of high temperature on residential electricity consumption by monthly air conditioning consumption to reduce electricity purchase cost losses; 2. Measurement and monitoring: Identify meteorologically similar days in July 2024 and the same period in 2023 based on the corrected daily temperature (corrected meteorological index deviation ≤ 0.5℃), realize intelligent year-on-year and month-on-month comparison of electricity consumption, and solve the monitoring deviation problem caused by meteorological differences in traditional date matching; 3. Backup power dispatch: Set the corrected daily temperature of 38℃ as the warning threshold. On July 20, 2024, the corrected meteorological index was monitored to reach 38.2℃, which automatically triggered the warning and the backup power was started in advance to improve the emergency response efficiency for extreme high temperature; 4. Orderly electricity consumption: For urban and rural residential electricity consumption, the corrected daily temperature changed from 30℃ to 31℃, and the electricity consumption increased by 0.411 billion kWh. Formulate a staggered peak electricity consumption strategy for non-temperature-sensitive industries during high-temperature periods (14:00-16:00) to reduce peak load.

[0100] Example 2:

[0101] See Figure 9A summer-based electricity consumption analysis system for temperature-sensitive industries, incorporating meteorological characteristics, includes a data acquisition module, a temperature-sensitive feature extraction module, a temperature-sensitive industry screening module, a daily electricity consumption calculation module, a meteorological contribution analysis module, and a business application module. The data acquisition module acquires multi-dimensional meteorological data of the analysis area during summer and daily electricity consumption data of various industries during summer. The temperature-sensitive feature extraction module extracts multi-dimensional regionally adapted meteorological features from the multi-dimensional meteorological data to construct a temperature-sensitive identification dataset. Specifically, the multi-dimensional regionally adapted meteorological features include zone-weighted electricity consumption meteorological features and regional average meteorological features. The zone-weighted electricity consumption meteorological features include daily weighted maximum temperature, daily weighted minimum temperature, and daily weighted average temperature. The regional average meteorological features include daily average maximum temperature, daily weighted minimum temperature, and daily weighted average temperature. Specifically, the temperature-sensitive feature extraction module is used to extract zoned electricity-weighted meteorological features from multi-dimensional meteorological data according to the following steps: obtaining the proportion of daily electricity consumption of a single industry in each administrative region within the analysis area in the previous year to the total daily electricity consumption of the corresponding industry in the analysis area, and using this proportion as a weight to perform weighted summation of the daily maximum temperature, daily minimum temperature, and daily average temperature of each administrative region to obtain the daily weighted maximum temperature, daily weighted minimum temperature, and daily weighted average temperature; the temperature-sensitive feature extraction module is used to extract regional average meteorological features from multi-dimensional meteorological data according to the following steps: taking the average of the daily maximum temperature, daily minimum temperature, and daily average temperature of each administrative region within the analysis area to obtain the daily average maximum temperature, daily average minimum temperature, and daily average temperature.

[0102] The temperature-sensitive industry screening module first calculates the correlation coefficients between the daily electricity consumption data of each industry and various features in the temperature-sensitive identification dataset to construct an industry-meteorological feature correlation coefficient matrix. Then, based on the industry-meteorological feature correlation coefficient matrix, threshold screening is performed to obtain temperature-sensitive industries. The daily electricity consumption calculation module is used to fit the regression curve between the daily electricity consumption data of temperature-sensitive industries in summer and the daily temperature, obtain the daily temperature of the predicted day and input it into the regression curve to obtain the daily electricity consumption of the temperature-sensitive industry on the predicted day. Specifically, the daily electricity consumption calculation module includes a correction module and a prediction module. The correction module is used to utilize... A pre-constructed long-short cumulative effect correction model corrects for daily temperature, and optimizes the weight coefficients of various components in the long-short cumulative effect correction model with the goal of maximizing the correlation coefficient between the daily electricity consumption data of temperature-sensitive industries in summer and the corrected daily temperature. The prediction module is used to fit the regression curve between the daily electricity consumption data of temperature-sensitive industries in summer and the corrected daily temperature. Based on the regression curve and the corrected daily temperature on the prediction day, the daily electricity consumption of temperature-sensitive industries on the prediction day is obtained. The daily temperature is the daily average temperature or the daily weighted average temperature. Specifically, the expression of the long-short cumulative effect correction model is as follows:

[0103] ;

[0104] ;

[0105] In the above formula, , These are the daily temperatures before and after the correction, respectively. , These are the short-term meteorological cumulative effect value and the long-term meteorological cumulative effect value, respectively. To and The corresponding weighting coefficients; , , They are respectively with , , The corresponding weighting coefficients; To be consistent with the previous day to the previous The weighting coefficient corresponding to the average daily temperature. ; , , , They were 1 day ago, 2 days ago, and 3 days ago, respectively. The temperature of the day before yesterday.

[0106] The meteorological contribution analysis module is used to calculate the total monthly electricity consumption of temperature-sensitive industries based on the predicted daily electricity consumption, obtain the monthly base electricity consumption of temperature-sensitive industries, and calculate the meteorological contribution rate based on the total monthly electricity consumption and the monthly base electricity consumption of temperature-sensitive industries. The larger the meteorological contribution rate, the higher the meteorological contribution to the summer electricity consumption of temperature-sensitive industries. Specifically, the meteorological contribution analysis module is used to obtain the monthly base electricity consumption of temperature-sensitive industries according to the following steps: pre-setting a daily temperature steady-state range, selecting continuous daily sequences that meet the meteorological steady-state range in spring and autumn and naming them typical daily sequences, obtaining the daily electricity consumption of temperature-sensitive industries in the typical daily sequences and taking the average value, using this average value as the daily base electricity consumption of temperature-sensitive industries, and calculating the monthly base electricity consumption of temperature-sensitive industries based on the daily base electricity consumption of temperature-sensitive industries. Specifically, the calculation formula for the meteorological contribution rate is:

[0107] ;

[0108] .

[0109] The business application module is used to display the calculation results such as the temperature-sensitive industry list, corrected daily temperature, daily electricity consumption, monthly total electricity consumption, and monthly air conditioning electricity consumption through a web-based visualization platform; it connects with the power grid agent purchasing system, dispatch management system, and orderly electricity consumption management system through API interfaces to achieve data sharing and business collaboration, and provides functions such as early warning threshold setting and decision-making scheme output.

Claims

1. A method for analyzing electricity consumption in temperature-sensitive industries during summer, incorporating meteorological characteristics, characterized in that: The method for analyzing electricity consumption in temperature-sensitive industries during summer includes: S1. Obtain multi-dimensional meteorological data of the analysis area in summer and daily electricity consumption data of various industries in summer; S2. Extract multi-dimensional regional adaptive meteorological features from multi-dimensional meteorological data to construct a temperature-sensitive identification dataset; the multi-dimensional regional adaptive meteorological features include zoned electricity-weighted meteorological features and regional average meteorological features; S3. Calculate the correlation coefficients between the daily electricity consumption data of each industry in summer and various features in the temperature-sensitive identification dataset to construct an industry-meteorological feature correlation coefficient matrix; based on the industry-meteorological feature correlation coefficient matrix, perform threshold screening to obtain the temperature-sensitive industries; S4. Fit the regression curve between the daily electricity consumption data and daily temperature of the temperature-sensitive industry in summer, obtain the daily temperature of the predicted day and input it into the regression curve to obtain the daily electricity consumption of the temperature-sensitive industry on the predicted day.

2. The method for analyzing summer temperature-sensitive industry electricity consumption by integrating meteorological characteristics as described in claim 1, characterized in that: In S2, the zoned electricity weighted meteorological characteristics include daily weighted maximum temperature, daily weighted minimum temperature and daily weighted average temperature; the regional average meteorological characteristics include daily average maximum temperature, daily average minimum temperature and daily average temperature.

3. The method for analyzing summer temperature-sensitive industry electricity consumption by integrating meteorological characteristics as described in claim 2, characterized in that: The following steps are used to extract the regional electricity weighted meteorological features from multi-dimensional meteorological data: obtain the proportion of daily electricity consumption of a single industry in each administrative region in the analysis area in the previous year to the total daily electricity consumption of the corresponding industry in the analysis area, and use this proportion as a weight to weight and sum the daily maximum temperature, daily minimum temperature and daily average temperature of each administrative region to obtain the daily weighted maximum temperature, daily weighted minimum temperature and daily weighted average temperature. The following steps are used to extract regional average meteorological characteristics from multi-dimensional meteorological data: average the daily maximum temperature, daily minimum temperature and daily average temperature of each administrative region in the analysis area to obtain the daily average maximum temperature, daily average minimum temperature and daily average temperature.

4. The method for analyzing summer temperature-sensitive industry electricity consumption by integrating meteorological characteristics as described in claim 2, characterized in that: In S4, a pre-constructed long-short cumulative effect correction model is used to correct the daily temperature. The regression curve between the daily electricity consumption data of the temperature-sensitive industry in summer and the corrected daily temperature is fitted. Based on the regression curve and the corrected daily temperature of the predicted day, the daily electricity consumption of the temperature-sensitive industry on the predicted day is obtained. The daily temperature is the daily average temperature or the daily weighted average temperature.

5. The method for analyzing summer temperature-sensitive industry electricity consumption by integrating meteorological characteristics according to claim 4, characterized in that: In S4, the expression for the long-short cumulative effect correction model is: ； ； In the above formula, , These are the daily temperatures before and after the correction, respectively. , These are the short-term meteorological cumulative effect value and the long-term meteorological cumulative effect value, respectively. To and The corresponding weighting coefficients; , , They are respectively with , , The corresponding weighting coefficients; To be consistent with the previous day to the previous The weighting coefficient corresponding to the average daily temperature. ; , , , They were 1 day ago, 2 days ago, and 3 days ago, respectively. The temperature of the day before yesterday.

6. The method for analyzing summer temperature-sensitive industry electricity consumption by integrating meteorological characteristics according to claim 4, characterized in that: In S4, the goal is to maximize the correlation coefficient between the daily electricity consumption data of the temperature-sensitive industry in summer and the corrected daily temperature, and to optimize the various weight coefficients in the long-term and short-term cumulative effect correction model.

7. The method for analyzing summer temperature-sensitive industry electricity consumption by integrating meteorological characteristics according to claim 1, characterized in that: The method for analyzing electricity consumption in temperature-sensitive industries during summer also includes: S5. Calculate the total monthly electricity consumption of the temperature-sensitive industry based on the predicted daily electricity consumption, obtain the monthly basic electricity consumption of the temperature-sensitive industry, and calculate the meteorological contribution based on the total monthly electricity consumption and the monthly basic electricity consumption of the temperature-sensitive industry. The larger the meteorological contribution, the higher the contribution of meteorology to the summer electricity consumption of the temperature-sensitive industry.

8. The method for analyzing summer temperature-sensitive industry electricity consumption by integrating meteorological characteristics as described in claim 7, characterized in that: In S5, the formula for calculating the meteorological contribution is: ； 。 9. The method for analyzing summer temperature-sensitive industry electricity consumption by integrating meteorological characteristics as described in claim 7, characterized in that: In S5, the monthly basic electricity consumption for temperature-sensitive industries is obtained according to the following steps: A daily temperature steady-state range is pre-defined. In spring and autumn, continuous daily series that meet the meteorological steady-state range are selected and named typical daily series. The daily electricity consumption of temperature-sensitive industries in the typical daily series is obtained and the average value is taken. This average value is used as the daily basic electricity consumption of temperature-sensitive industries. The monthly basic electricity consumption of temperature-sensitive industries is calculated based on the daily basic electricity consumption of temperature-sensitive industries.

10. A summer temperature-sensitive industrial power consumption analysis system integrating meteorological characteristics, characterized in that: The summer temperature-sensitive industry power consumption analysis system includes: The data acquisition module is used to acquire multi-dimensional meteorological data of the analysis area in summer, as well as daily electricity consumption data of various industries in summer. The temperature-sensitive feature extraction module is used to extract multi-dimensional regional adaptive meteorological features from multi-dimensional meteorological data to construct a temperature-sensitive identification dataset; the multi-dimensional regional adaptive meteorological features include zoned electricity-weighted meteorological features and regional average meteorological features; The temperature-sensitive industry screening module is used to first calculate the correlation coefficient between the daily electricity consumption data of each industry in summer and various features in the temperature-sensitive identification dataset, so as to construct an industry-meteorological feature correlation coefficient matrix. Then, based on the industry-meteorological feature correlation coefficient matrix, threshold screening is performed to obtain the temperature-sensitive industries. The daily electricity consumption calculation module is used to fit the regression curve between the daily electricity consumption data and the daily temperature of the temperature-sensitive industry in summer, obtain the daily temperature of the predicted day and input it into the regression curve to obtain the daily electricity consumption of the temperature-sensitive industry on the predicted day.

Images