A multi-factor correlation method and system for pre-controlling a power load for winter return

Abstract

The present application relates to the technical field of smart grid, disclose a kind of multi-factor associated return home degree winter power load pre-control method and system, method includes: the historical power consumption data obtained in advance is transformed to obtain power load data, and power load data is identified to obtain power consumption problem;The influencing factor is obtained by analyzing power consumption problem, and the strong correlation factor data matrix is obtained by analyzing influencing factor, and the correlation influence coefficient is obtained by analyzing strong correlation factor data matrix;The trained power load prediction model is obtained by iterative training to power load prediction model, and the current power consumption data obtained in advance is input into trained power load prediction model to obtain load prediction data;The red yellow green classification area is obtained by dividing power grid;Power load pre-control scheme is formulated, so as to carry out power consumption prevention and control management.The present application can solve the problem that power grid overload is difficult to prevent and control during return home degree winter and power consumption prediction is not accurate enough.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, and in particular to a method and system for pre-controlling electricity load during the winter travel rush caused by multiple factors. Background Technology

[0002] The power load pre-control system for the winter travel rush mainly includes power load prevention and control schemes and power load prevention and control prediction models. However, traditional prevention and control schemes mainly rely on manual experience and simple threshold settings, lacking intelligent decision support systems. When the load exceeds the preset threshold, the system cannot timely and accurately determine the cause and trend of load growth, making it difficult to formulate targeted and scientifically sound prevention and control strategies. In addition, the collaborative prevention and control capabilities between power systems in different regions are weak, and information transmission is not timely and coordination mechanisms are imperfect during cross-regional power dispatch, making it difficult to achieve efficient load control and ensuring the stability and security of power supply during the winter travel rush.

[0003] Existing power load forecasting models struggle to fully uncover the potential connections between multiple factors and power load, resulting in forecast accuracy that falls short of practical requirements. Traditional forecasting methods primarily rely on historical data for modeling, failing to adequately consider dynamic and uncertain factors such as population movement, leading to significant discrepancies between forecast results and actual load conditions. Furthermore, while meteorological factors have a crucial impact on winter power load, the diversity of meteorological conditions across different regions and the complexity of meteorological changes further complicate accurate forecasting. Summary of the Invention

[0004] This invention provides a method and system for pre-controlling electricity load during the winter travel rush, which is based on multiple factors. Its main purpose is to solve the problems of difficulty in controlling and managing power grid overload and inaccurate electricity consumption forecasting during the winter travel rush.

[0005] To achieve the above objectives, the present invention provides a multi-factor correlation method for pre-controlling electricity load during the winter return travel period, comprising:

[0006] S1. Perform Fourier transform on the pre-acquired historical electricity consumption data to obtain the power load data of the power grid, identify load fluctuations in the power load data, and obtain the electricity consumption problem of the power grid.

[0007] S2. Perform Pareto analysis on the causes of the electricity problem to obtain the influencing factors of the electricity problem. Perform load characteristic cluster analysis on the influencing factors to obtain the data matrix of strongly correlated factors of the electricity problem. Perform quantitative analysis on the data matrix of strongly correlated factors to obtain the correlation influence coefficient of the power grid.

[0008] S3. Based on the correlation influence coefficient, iteratively train the power load prediction model to obtain the trained power load prediction model. Input the pre-acquired current electricity consumption data into the trained power load prediction model to obtain the load prediction data of the power grid.

[0009] S4. Based on the load conditions of the load forecast data and the severity of the power consumption problem, the power grid is classified by color to obtain the red, yellow and green graded transformer areas of the power grid.

[0010] S5. Based on the red, yellow, and green graded distribution areas, formulate a power load pre-control plan for returning home for the winter, and carry out power consumption prevention and control management of the power grid according to the power load pre-control plan.

[0011] In a preferred embodiment, the step of performing a Fourier transform on the pre-acquired historical electricity consumption data to obtain the power load data of the power grid includes:

[0012] The historical electricity consumption data is cleaned of noise to obtain standardized data of the historical electricity consumption data;

[0013] The standardized data is periodically divided to obtain segmented data of the standardized data;

[0014] Perform a Fourier transform on the segmented data to obtain the power data of the segmented data;

[0015] Load characteristics are extracted from the power data to obtain the power load data of the power grid.

[0016] In a preferred embodiment, the step of identifying load fluctuations in the power load data to determine the power grid's electricity consumption issues includes:

[0017] Feature extraction is performed on the time-domain and frequency-domain data of the power load data to obtain the time-frequency features of the power load data;

[0018] Abnormal fluctuations are identified in the time-frequency features to obtain abnormal fluctuation data of the time-frequency features;

[0019] The historical load data is analyzed for statistical characteristics to obtain the load fluctuation threshold of the power grid;

[0020] Based on the abnormal fluctuation data and the load fluctuation threshold, the power grid is diagnosed to obtain the power consumption problems of the power grid.

[0021] In a preferred embodiment, the Pareto analysis of the causes of the electricity problem to obtain the influencing factors of the electricity problem includes:

[0022] The causes of the electricity problems are classified to obtain the causal categories of the electricity problems;

[0023] Based on the aforementioned cause categories, a quantitative statistical analysis of the problems is performed to obtain the number of problems for each cause category.

[0024] A Pareto diagram of the electricity consumption problem is drawn based on the causal category and the number of problems.

[0025] Based on the Pareto chart, the correlation of the electricity consumption problem is verified to obtain the influencing factors of the electricity consumption problem.

[0026] In a preferred embodiment, the step of performing load characteristic clustering analysis on the influencing factors to obtain a data matrix of strongly correlated factors of the electricity consumption problem includes:

[0027] The influencing factors are input into the load characteristic clustering model to obtain the clustering data of the electricity consumption problem;

[0028] The positive impact of the influencing factors is calculated based on the clustering data and the positive impact formula, wherein the positive impact formula is:

[0029]

[0030] In the formula, The historical electricity consumption data is numbered. The number of the clustered data. For the first The first historical electricity consumption data The positive impact of clustering data For the first The first historical electricity consumption data The content of each clustered data, For the first The minimum value of each cluster data point For the first The maximum value of each cluster data point For the first Cluster data;

[0031] The negative impact of the influencing factors is calculated based on the clustering data and the negative impact formula, wherein the negative impact formula is:

[0032]

[0033] In the formula, The historical electricity consumption data is numbered. The number of the clustered data. For the first The first historical electricity consumption data The negative impact of individual clustering data For the first The first historical electricity consumption data The content of each clustered data, For the first The minimum value of each cluster data point For the first The maximum value of each cluster data point For the first Cluster data;

[0034] The data matrix of strongly correlated factors is established based on the positive and negative effects.

[0035] In a preferred embodiment, the step of quantifying and analyzing the strongly correlated factor data matrix to obtain the correlation influence coefficient of the power grid includes:

[0036] Trend analysis is performed on the data matrix of strongly correlated factors to obtain the changing trend of the data matrix of strongly correlated factors;

[0037] The dispersion analysis of the strongly correlated factor data matrix is ​​performed to obtain the degree of dispersion of the strongly correlated factor data matrix;

[0038] Based on the changing trend and the degree of dispersion, the correlation relationship of the strongly correlated factor data matrix is ​​analyzed to obtain the correlation influence coefficient of the power grid.

[0039] In a preferred embodiment, the iterative training of the power load forecasting model based on the correlation influence coefficient to obtain a trained power load forecasting model, and inputting the pre-acquired current electricity consumption data into the trained power load forecasting model to obtain the power grid load forecasting data, includes:

[0040] The power load prediction model is iteratively trained based on the correlation influence coefficient to obtain the iterative power load prediction model.

[0041] The historical electricity consumption data is input into the iterative power load prediction model to calculate the loss value, thereby obtaining the loss value of the iterative power load prediction model. The iterative loss value calculation function is as follows:

[0042]

[0043] In the formula, The loss value, The quantity of the historical electricity consumption data. The number of the correlation influence coefficients, The historical electricity consumption data is numbered. The number is the correlation coefficient. For the first Each correlation coefficient For the first The power load is calculated using a correlation influence coefficient. For the first Historical electricity consumption data, For the first The actual value of historical electricity consumption data;

[0044] The trained power load prediction model is determined based on the loss value;

[0045] The pre-acquired current electricity consumption data is input into the trained power load prediction model to obtain the power grid load prediction data.

[0046] In a preferred embodiment, the step of classifying the power grid into red, yellow, and green graded distribution areas based on the load conditions of the load forecast data and the severity of the electricity consumption problem includes:

[0047] The load forecast data is classified by load grading to obtain the predicted load grading of the power grid;

[0048] The severity of the electricity problems is screened to identify the serious electricity problems in the power grid;

[0049] Establish a quantitative relationship between the predicted load classification and the severe electricity problem;

[0050] The power grid is classified into red, yellow, and green graded substations based on the quantification relationship.

[0051] In a preferred embodiment, the step of formulating a power load pre-control scheme for winter returnees based on the red, yellow, and green graded distribution areas, and carrying out power consumption prevention and control management of the power grid according to the power load pre-control scheme, includes:

[0052] The red-yellow-green graded distribution area is monitored to obtain the real-time load data of the red-yellow-green graded distribution area;

[0053] Based on the real-time load data and the quantitative relationship, demand analysis is performed on the red-yellow-green graded distribution area to obtain the electricity demand of the red-yellow-green graded distribution area.

[0054] Based on the aforementioned electricity demand and pre-acquired distributed resource data, a pre-control scheme for electricity load during the winter return to hometowns is formulated.

[0055] The power grid is managed and controlled according to the power load pre-control scheme.

[0056] To address the aforementioned problems, the present invention also provides a reference information generation system based on artificial intelligence and smart home, the system comprising:

[0057] The electricity problem identification module is used to perform Fourier transform on the pre-acquired historical electricity data to obtain the power load data of the power grid, and to identify the load fluctuations of the power load data to obtain the electricity problem of the power grid.

[0058] The power load influencing factor analysis module is used to perform Pareto analysis on the causes of the power consumption problem, obtain the influencing factors of the power consumption problem, perform load characteristic cluster analysis on the influencing factors, obtain a data matrix of strongly correlated factors of the power consumption problem, and perform quantitative analysis on the data matrix of strongly correlated factors to obtain the correlation influence coefficient of the power grid.

[0059] The power load forecasting module is used to iteratively train the power load forecasting model based on the correlation influence coefficient to obtain the trained power load forecasting model. The pre-acquired current electricity consumption data is input into the trained power load forecasting model to obtain the load forecasting data of the power grid.

[0060] The graded distribution module is used to classify the power grid by color based on the load conditions of the load forecast data and the severity of the power consumption problem, so as to obtain the red, yellow and green graded distribution areas of the power grid.

[0061] The power load pre-control module is used to formulate a power load pre-control plan for returning home for winter based on the red, yellow and green graded transformer areas, and to carry out power consumption prevention and control management of the power grid according to the power load pre-control plan.

[0062] Compared with the prior art, the present invention has the following beneficial effects:

[0063] 1. By analyzing the power load characteristics and environmental conditions of different regions through multi-factor correlation analysis, and combining iterative training to dynamically adjust the prediction model, the efficiency and reliability of the model under different application scenarios are ensured. At the same time, the model can accurately capture complex influencing factors during the winter travel season, such as population flow and weather changes, thereby improving the accuracy of power load prediction.

[0064] 2. Through hierarchical color-coding and real-time monitoring, the system can quickly respond to power load fluctuations, formulate targeted prevention and control plans, and ensure the stable operation of the power grid. At the same time, by building a smart grid, the system can automatically analyze the causes of power consumption problems, optimize power load pre-control strategies, reduce manual intervention, improve decision-making efficiency, and further enhance the timeliness and effectiveness of power load prevention and control. Finally, by integrating distributed resource data, the system can rationally allocate power resources, improve resource utilization efficiency, and reduce energy waste. Attached Figure Description

[0065] Figure 1 This is a flowchart illustrating a multi-factor correlation method for pre-controlling electricity load during the winter return period, as provided in an embodiment of the present invention.

[0066] Figure 2 This is a functional block diagram of a multi-factor correlation power load pre-control system for returning home for winter provided in an embodiment of the present invention;

[0067] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0068] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0069] This application provides a method for pre-controlling electricity load during the winter travel rush based on multiple factors. The execution entity of this method includes, but is not limited to, at least one electronic device that can be configured to execute the method provided in this application, such as a server or a terminal. In other words, the method can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cluster of cloud servers. The server can be an independent server or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0070] Reference Figure 1 The diagram shown is a flowchart illustrating a multi-factor correlation-based method for predicting and controlling electricity load during the winter travel rush, according to an embodiment of the present invention. In this embodiment, the multi-factor correlation-based method for predicting and controlling electricity load during the winter travel rush includes:

[0071] S1. Perform Fourier transform on the pre-acquired historical electricity consumption data to obtain the power load data of the power grid, identify load fluctuations in the power load data, and obtain the electricity consumption problem of the power grid.

[0072] In this embodiment of the invention, the step of performing a Fourier transform on the pre-acquired historical electricity consumption data to obtain the power load data of the power grid includes:

[0073] The historical electricity consumption data is cleaned of noise to obtain standardized data of the historical electricity consumption data;

[0074] The standardized data is periodically divided to obtain segmented data of the standardized data;

[0075] Perform a Fourier transform on the segmented data to obtain the power data of the segmented data;

[0076] Load characteristics are extracted from the power data to obtain the power load data of the power grid.

[0077] The process of identifying load fluctuations in the power load data to determine the power grid's electricity consumption issues includes:

[0078] Feature extraction is performed on the time-domain and frequency-domain data of the power load data to obtain the time-frequency features of the power load data;

[0079] Abnormal fluctuations are identified in the time-frequency features to obtain abnormal fluctuation data of the time-frequency features;

[0080] The historical load data is analyzed for statistical characteristics to obtain the load fluctuation threshold of the power grid;

[0081] Based on the abnormal fluctuation data and the load fluctuation threshold, the power grid is diagnosed to obtain the power consumption problems of the power grid.

[0082] Specifically, firstly, by setting a reasonable data threshold range, data points that deviate from the normal range are identified and removed; secondly, the continuity of the data is checked, and missing values ​​caused by data transmission interruption or record loss are processed; finally, the cleaned data is standardized to unify the units and formats, thereby obtaining standardized data of the historical electricity consumption data.

[0083] Specifically, first, the length of the period is determined based on business needs or data analysis objectives; second, the standardized data is divided into several segments according to the determined period length; finally, each segment of data is checked for completeness to ensure there are no missing or outliers. For incomplete data segments, interpolation or mean-filling methods can be used to supplement them, thereby obtaining the segmented data of the standardized data.

[0084] Specifically, firstly, the segmented data is preprocessed to ensure a uniform data format and no missing values. Secondly, the segmented data is used as input, and the Fourier transform algorithm is applied to calculate frequency domain information. The Fourier transform converts the time-domain signal into a frequency-domain signal, extracting the amplitude and phase information of different frequency components. Finally, key frequency domain features, such as the fundamental frequency, harmonic frequencies, and their corresponding amplitudes, are extracted from the Fourier transform results. These features constitute the power data of the segmented data, which includes the periodic changes in electricity consumption and energy distribution.

[0085] Specifically, first, key features related to power load are selected, such as voltage, current, power, and power frequency; second, load features are extracted from the power data, such as calculating the peak, valley, average, and load change rate of the load, which can describe the fluctuation of the power grid load; finally, the power data is classified according to the extracted load features, such as classifying the load into high load, medium load, and low load categories, to facilitate subsequent analysis and scheduling, and the power load data of the power grid is summarized.

[0086] Specifically, firstly, key features are extracted from the time-domain data, such as the peak, valley, mean, variance, and trend of the load. These features reflect the fluctuation of the power load over time. Secondly, Fourier transform is applied to the power load data to convert the time-domain signal into a frequency-domain signal, and frequency-domain features are extracted, such as the fundamental frequency, harmonic frequencies, and their corresponding amplitudes. These features reflect the frequency distribution characteristics of the load data. Finally, the time-domain features and frequency-domain features are fused to obtain the time-frequency features of the power load data.

[0087] Specifically, firstly, based on historical data or statistical methods, a baseline value (such as mean or standard deviation) of the time-frequency feature is calculated as a reference standard for abnormal fluctuations; secondly, based on the baseline value and a preset fluctuation range (such as ±3 times the standard deviation), an identification threshold for abnormal fluctuations is set to determine whether the data deviates from the normal range; finally, the time-frequency feature data is monitored in real time using algorithms (such as sliding window method, isolated forest, or Z-score detection) to identify abnormal fluctuation points exceeding the threshold range, thereby obtaining the abnormal fluctuation data of the time-frequency feature.

[0088] Specifically, the statistical characteristics of historical load data are calculated, including mean, standard deviation, maximum, minimum and quantile values, to describe the distribution and fluctuation of the load data.

[0089] Furthermore, based on statistical characteristics, fluctuation features of load data, such as peak-to-valley difference, fluctuation amplitude, and trend, are extracted to quantify the intensity and frequency of load fluctuations.

[0090] Furthermore, based on the statistical characteristics of historical data, a baseline value (such as the mean or median) for load fluctuation is determined as a reference standard for fluctuation thresholds. Based on statistical characteristics (such as ±3 standard deviations) or empirical rules, upper and lower thresholds for load fluctuation are set to identify abnormal fluctuations.

[0091] Furthermore, by comparing historical data with the set threshold, the rationality of the threshold is verified, and the threshold range is adjusted according to actual needs to ensure its applicability and accuracy, thereby determining the load fluctuation threshold of the power grid.

[0092] Specifically, firstly, abnormal fluctuation data is compared with load fluctuation thresholds to identify abnormal fluctuation points exceeding the thresholds, marking their time and fluctuation amplitude. Based on the distribution of abnormal fluctuation points, the area in the power grid where the problem occurs, such as a specific node or power supply area, is located, narrowing down the diagnostic scope. Secondly, combined with power grid status data (such as voltage, current, and frequency) and external factors (such as weather and social activities), possible causes of abnormal fluctuations are analyzed. Finally, based on the fluctuation characteristics and causes, the problems are classified, such as overload, voltage instability, frequency deviation, or equipment failure, thereby obtaining the power consumption problems of the power grid.

[0093] In summary, performing Fourier transform on pre-acquired historical electricity consumption data can reveal the periodic characteristics and fluctuation patterns of the power grid's load, thereby identifying anomalies or potential problems in load fluctuations. This analytical method can provide a scientific basis for power grid dispatching, optimize load management strategies, rationally allocate power resources, avoid overload or power shortages, and improve the stability and efficiency of power grid operation. By accurately identifying load fluctuations, it can help the power grid respond to electricity consumption problems more efficiently and reliably.

[0094] S2. Perform Pareto analysis on the causes of the electricity problem to obtain the influencing factors of the electricity problem. Perform load characteristic cluster analysis on the influencing factors to obtain the data matrix of strongly correlated factors of the electricity problem. Perform quantitative analysis on the data matrix of strongly correlated factors to obtain the correlation influence coefficient of the power grid.

[0095] In this embodiment of the invention, the Pareto analysis of the causes of the electricity problem to obtain the influencing factors of the electricity problem includes:

[0096] The causes of the electricity problems are classified to obtain the causal categories of the electricity problems;

[0097] Based on the aforementioned cause categories, a quantitative statistical analysis of the problems is performed to obtain the number of problems for each cause category.

[0098] A Pareto diagram of the electricity consumption problem is drawn based on the causal category and the number of problems.

[0099] Based on the Pareto chart, the correlation of the electricity consumption problem is verified to obtain the influencing factors of the electricity consumption problem.

[0100] The load characteristic clustering analysis of the influencing factors yields a data matrix of strongly correlated factors for the electricity consumption problem, including:

[0101] The influencing factors are input into the load characteristic clustering model to obtain the clustering data of the electricity consumption problem;

[0102] The positive impact of the influencing factors is calculated based on the clustering data and the positive impact formula, wherein the positive impact formula is:

[0103]

[0104] In the formula, The historical electricity consumption data is numbered. The number of the clustered data. For the first The first historical electricity consumption data The positive impact of clustering data For the first The first historical electricity consumption data The content of each clustered data, For the first The minimum value of each cluster data point For the first The maximum value of each cluster data point For the first Cluster data;

[0105] The negative impact of the influencing factors is calculated based on the clustering data and the negative impact formula, wherein the negative impact formula is:

[0106]

[0107] In the formula, The historical electricity consumption data is numbered. The number of the clustered data. For the first The first historical electricity consumption data The negative impact of individual clustering data For the first The first historical electricity consumption data The content of each clustered data, For the first The minimum value of each cluster data point For the first The maximum value of each cluster data point For the first Cluster data;

[0108] The data matrix of strongly correlated factors is established based on the positive and negative effects.

[0109] The quantitative analysis of the strongly correlated factor data matrix to obtain the correlation influence coefficient of the power grid includes:

[0110] Trend analysis is performed on the data matrix of strongly correlated factors to obtain the changing trend of the data matrix of strongly correlated factors;

[0111] The dispersion analysis of the strongly correlated factor data matrix is ​​performed to obtain the degree of dispersion of the strongly correlated factor data matrix;

[0112] Based on the changing trend and the degree of dispersion, the correlation relationship of the strongly correlated factor data matrix is ​​analyzed to obtain the correlation influence coefficient of the power grid.

[0113] Specifically, firstly, key characteristics of electricity consumption problems are extracted, such as fluctuation amplitude, duration, occurrence area, and scope of impact, to provide basic data for causal classification. Secondly, the impact of external environmental factors (such as weather, social activities, and economic conditions) on electricity consumption problems is analyzed to identify possible correlations and obtain external factors. Simultaneously, the internal state data of the power grid (such as voltage, current, and frequency) and equipment operation are examined to determine whether there are equipment failures or system design defects, thus obtaining internal factors. Finally, a correlation model between electricity consumption problems and external and internal factors is established to assess the contribution of each factor to the problem, clarify the main causes, and obtain the causal category of the electricity consumption problem.

[0114] Specifically, firstly, the causes of electricity problems are organized and classified to ensure that the cause category of each problem is clear and specific, providing a basis for subsequent statistics; secondly, the number of electricity problems under each cause category is counted, and the number of problems corresponding to each cause category is recorded to form a preliminary statistical table; finally, the accuracy and completeness of the statistical data are checked, and possible errors or omissions are corrected to ensure the reliability of the statistical results, thereby obtaining the number of problems in the aforementioned cause categories.

[0115] Specifically, first, the causes of electricity problems and their corresponding number of problems are compiled to ensure data completeness and accuracy, providing a basis for drawing a Pareto chart. The cause categories are sorted from highest to lowest number of problems, ensuring that the main problem categories are at the top of the chart. Second, the cumulative percentage of each cause category is calculated to draw the cumulative curve of the Pareto chart. A bar chart is then drawn in the chart to represent the number of problems in each cause category, with the bars arranged in the sorted order. At the same time, a cumulative percentage curve is drawn based on the bar chart to represent the cumulative distribution of the number of problems. Finally, key points in the cumulative curve (such as the 80% cumulative value) are marked to identify the main problem categories and their contribution ratio, thus obtaining the Pareto chart of the electricity problem.

[0116] Specifically, firstly, the main causal categories and their corresponding number of problems are extracted from the Pareto chart as the basic data for correlation verification, and external factors that may affect electricity consumption problems, such as economic factors, social activity factors, and meteorological factors, are identified. Secondly, a preset correlation algorithm (such as correlation coefficient analysis) is used to conduct correlation analysis between electricity consumption problems and external factors to verify their correlation. Finally, based on the correlation analysis results, the influence weight of each external factor is calculated to determine its degree of influence on electricity consumption problems, thereby obtaining correlation verification and identifying the influencing factors of the electricity consumption problems.

[0117] Specifically, first, key features, such as economic factors, social activity factors, and meteorological factors, are extracted from the preprocessed data and used as input to the clustering model. Second, the extracted features are used to train the clustering model, and model parameters (such as the number of clusters or the distance metric) are adjusted to ensure that the model can effectively capture data characteristics. Finally, the influencing factors are input into the trained clustering model to perform cluster analysis, resulting in different categories of electricity consumption problem data and corresponding cluster data.

[0118] In summary, the positive impact formula is used to quantify the positive promoting effect of a factor on a target variable. By assigning positive weights or positive coefficients, the formula can amplify the value of the target variable, thereby strengthening the influence of the factor.

[0119] For example, in the impact of economic factors on grid load, the positive impact formula will increase the weight of economic activities, making them play a more important role in load forecasting.

[0120] In general, the negative impact formula is used to quantify the inhibitory or weakening effect of a factor on a target variable. By assigning negative weights or negative coefficients, this formula can reduce the value of the target variable, thereby weakening the factor's influence.

[0121] For example, in the impact of meteorological factors on power grid load, the negative impact formula reduces the weight of extreme weather conditions to minimize their interference with load forecasting.

[0122] Specifically, the impact factors for positive and negative effects are calculated separately, and formulas are used to quantify the degree of influence of each factor on the target variable. Positive impact factors are usually positive, and negative impact factors are usually negative.

[0123] Furthermore, the weights of each factor are updated based on the impact factors. Factors with larger positive impact factors have higher weights, while factors with larger negative impact factors have lower weights.

[0124] Furthermore, the updated weights are combined with the feature data of each factor to construct a strongly correlated factor data matrix, wherein each row of the matrix represents a factor, and each column represents its weight and feature value at different time points or scenarios.

[0125] Specifically, the data matrix of strongly correlated factors is preprocessed to ensure data consistency and comparability. Standardization (such as Z-score standardization) is used to eliminate dimensional differences between different factors.

[0126] Furthermore, using the standardized data, the covariance matrix of the strongly correlated factor data matrix is ​​calculated.

[0127] In summary, the covariance matrix reflects the linear relationship between various factors and their changing trends, providing a quantitative basis for subsequent analysis.

[0128] Furthermore, by performing eigenvalue decomposition on the covariance matrix, the principal components and their corresponding eigenvectors are extracted.

[0129] In summary, these principal components represent the main direction of data change and can reveal the main trend characteristics of the data matrix of strongly correlated factors.

[0130] Specifically, the data matrix of strongly correlated factors is preprocessed, including removing invalid data and outliers, and eliminating the dimensional differences between different factors through standardization (such as Z-score standardization).

[0131] Furthermore, using the standardized data, the standard deviation of each strongly correlated factor is calculated.

[0132] In general, standard deviation reflects the dispersion of data; a larger value indicates a more dispersed data distribution, while a smaller value indicates a more concentrated data distribution.

[0133] Furthermore, the dispersion of each strongly correlated factor is assessed based on the calculated standard deviation.

[0134] In general, by comparing the standard deviations of different factors, factors with higher dispersion can be identified, which may have a greater impact on the variation of the overall data.

[0135] Specifically, by using a pre-defined correlation algorithm and combining the trend of change and the degree of dispersion, a correlation model between various factors is established.

[0136] In summary, this model reveals the combined effect of factors on grid load by quantifying the interactions between them.

[0137] Furthermore, based on the correlation model, the influence factor of each demand factor is calculated.

[0138] In summary, the influencing factor integrates information on trends and dispersion to measure the strength of its impact on the overall performance of the power grid.

[0139] Furthermore, by combining influencing factors with preset weights, the correlation coefficient of each factor is calculated. The correlation coefficient reflects the relative importance of the factor in power grid dispatching, providing a basis for the formulation of load management strategies.

[0140] In summary, Pareto analysis of the causes of electricity problems can identify the main influencing factors, and load characteristic clustering analysis can screen out a data matrix of strongly correlated factors. Finally, quantitative analysis is used to derive the correlation coefficient of the power grid. This process reveals the core driving factors of electricity problems and their interrelationships, providing a scientific basis for formulating precise load management strategies and optimizing power grid resource scheduling, thereby improving the stability and efficiency of power grid operation. Through data-driven analysis methods, influencing factors can be accurately identified and quantified, assisting the power grid in efficiently solving problems.

[0141] S3. Based on the correlation influence coefficient, iteratively train the power load prediction model to obtain the trained power load prediction model. Input the pre-acquired current electricity consumption data into the trained power load prediction model to obtain the load prediction data of the power grid.

[0142] In this embodiment of the invention, the iterative training of the power load prediction model based on the correlation influence coefficient to obtain the trained power load prediction model, and inputting the pre-acquired current electricity consumption data into the trained power load prediction model to obtain the power grid load prediction data, includes:

[0143] The power load prediction model is iteratively trained based on the correlation influence coefficient to obtain the iterative power load prediction model.

[0144] The historical electricity consumption data is input into the iterative power load prediction model to calculate the loss value, thereby obtaining the loss value of the iterative power load prediction model. The iterative loss value calculation function is as follows:

[0145]

[0146] In the formula, The loss value is... The quantity of the historical electricity consumption data. The number of the correlation influence coefficients, The historical electricity consumption data is numbered. The number is the correlation coefficient. For the first Each correlation coefficient For the first The power load is calculated using a correlation influence coefficient. For the first Historical electricity consumption data, For the first The actual value of historical electricity consumption data;

[0147] The trained power load prediction model is determined based on the loss value;

[0148] The pre-acquired current electricity consumption data is input into the trained power load prediction model to obtain the power grid load prediction data.

[0149] For example, historical electricity consumption data shows that during the Spring Festival travel rush in 2024, Anhui Company experienced 370 transformer overloads (0.01%), a 36.54% decrease compared to 2023 (583 units), mainly concentrated in Fuyang, Lu'an, and Huainan; 1398 transformers experienced low voltage at the outlet (0.45%), a 29.71% decrease compared to 2023 (1989 units), mainly concentrated in Fuyang, Bozhou, and Lu'an; and 500 low voltage work orders were received, a 25.60% decrease compared to 2023 (672 orders), mainly concentrated in Hefei, Lu'an, and Bozhou. It is predicted that during the Spring Festival travel rush in 2025, under extreme conditions, the temperature will be the lowest in nearly 20 years. The number of people returning home reached a near 3-year high of -13 degrees Celsius (24 million), with 305 overloaded red transformer areas (17 of which are special areas such as scenic spots and charging stations); 1078 red transformer areas with low voltage at the outlet (46 of which are special areas such as scenic spots and charging stations); and 100,800 households affected by low voltage at the household level, concentrated in Fuyang, Lu'an, Anqing, and other areas, mainly in rural power grid areas (overload rate 88.85%, outlet low voltage rate 98.14%, and household low voltage rate 95.04%). Verification shows that the actual occurrence data of overload, outlet low voltage, and household low voltage over the past 3 years match the model prediction data by more than 80%, meeting current application needs.

[0150] Specifically, the parameters of the power load forecasting model are initialized based on historical data and preset initial values, including the weights and correlation coefficients of each demand factor.

[0151] Furthermore, based on the aforementioned correlation influence coefficients, the weighted characteristics of each demand factor are calculated. By multiplying the correlation influence coefficients by the characteristic data of the factors, demand weighted characteristics are generated to reflect the actual impact of each factor on the power grid load.

[0152] Furthermore, the demand-weighted features are input into the power load forecasting model to generate load forecast values. The model then uses regression analysis or machine learning algorithms, combined with the weighted features and real-time data, to output the forecast results of the power grid load.

[0153] Furthermore, the predicted values ​​are compared with the actual load data to calculate the prediction error.

[0154] In summary, error reflects the accuracy of the model and provides a basis for subsequent parameter adjustments. By optimizing algorithms (such as gradient descent) to minimize error, model performance can be improved.

[0155] Furthermore, based on the prediction error and the correlation influence coefficient, the parameters of the model are adjusted, and the weights and correlation influence coefficients are updated iteratively to make the model gradually adapt to the dynamic changes of the power grid load, thereby improving the prediction accuracy and finally obtaining the iterated power load prediction model.

[0156] In summary, the iterative loss calculation function quantifies the error between the model's predicted values ​​and the true values, and reflects the contribution of each feature to the prediction results through a weighted approach. This function evaluates the model's prediction accuracy by calculating the squared error between the predicted and true values ​​and combining the weights of each feature. At the same time, it makes the loss values ​​more comparable through normalization (dividing by the total number of samples), providing a clear objective function for model optimization.

[0157] Specifically, the pre-acquired current electricity consumption data is input into the trained power load prediction model as the model's input data. The current electricity consumption data includes real-time power grid operation information such as voltage, current, and power.

[0158] Specifically, the power load forecasting model generates power grid load forecasting data based on the input current electricity consumption data and through internal algorithms (such as regression analysis, neural networks, etc.).

[0159] Specifically, the load forecast data output by the model includes the load change trend of the power grid over a period of time, providing a basis for power grid dispatching decisions.

[0160] In summary, iterative training of the power load forecasting model based on the correlation influence coefficient optimizes the weight allocation of demand factors, thereby improving forecast accuracy. The trained model can use current electricity consumption data as input to accurately predict future load change trends in the power grid, providing a reliable basis for power grid dispatch and ensuring the stability and efficiency of power grid operation.

[0161] S4. Based on the load conditions of the load forecast data and the severity of the power consumption problem, the power grid is classified by color to obtain the red, yellow and green graded transformer areas of the power grid.

[0162] In this embodiment of the invention, the step of classifying the power grid into red, yellow, and green graded distribution areas based on the load conditions of the load forecast data and the severity of the electricity problem includes:

[0163] The load forecast data is classified by load grading to obtain the predicted load grading of the power grid;

[0164] The severity of the electricity problems is screened to identify the serious electricity problems in the power grid;

[0165] Establish a quantitative relationship between the predicted load classification and the severe electricity problem;

[0166] The power grid is classified into red, yellow, and green graded substations based on the quantification relationship.

[0167] Specifically, based on historical load data of the power grid and power load forecasting models, the load is divided into different levels, such as low load, medium load and high load, in order to quantify the severity of the load.

[0168] Furthermore, by analyzing real-time power grid operation data, serious power consumption problems that may lead to power grid overload, voltage instability, or frequency anomalies can be identified and classified.

[0169] Furthermore, by using data-driven methods (such as regression analysis and machine learning), a mapping relationship between predicted load levels and severe power consumption problems can be established, quantifying the impact of different load levels on power grid stability.

[0170] Specifically, the red, yellow, and green distribution areas are classified as follows: red areas are overloaded and low-voltage areas that were not thoroughly addressed in previous years; yellow areas are heavily loaded areas with voltage approaching the boundary value; and green areas are areas with sufficient transformer capacity margin and high voltage quality.

[0171] In summary, based on the load forecast data and the severity of power consumption problems, the power grid is divided into three levels: red, yellow, and green. Red indicates areas with high load or serious power consumption problems, requiring priority resource dispatch; yellow indicates areas with medium load or potential risks, requiring attention and optimized dispatch; and green indicates areas with low load or stable operation, where normal dispatch can be maintained. This color-coded classification can intuitively reflect the operating status of the power grid, help quickly identify problem areas and take targeted measures, and improve the efficiency and stability of power grid dispatch.

[0172] S5. Based on the red, yellow, and green graded distribution areas, formulate a power load pre-control plan for returning home for the winter, and carry out power consumption prevention and control management of the power grid according to the power load pre-control plan.

[0173] In this embodiment of the invention, the step of formulating a power load pre-control scheme for returning home for winter based on the red, yellow, and green graded distribution areas, and carrying out power consumption prevention and control management of the power grid according to the power load pre-control scheme, includes:

[0174] The red-yellow-green graded distribution area is monitored to obtain the real-time load data of the red-yellow-green graded distribution area;

[0175] Based on the real-time load data and the quantitative relationship, demand analysis is performed on the red-yellow-green graded distribution area to obtain the electricity demand of the red-yellow-green graded distribution area.

[0176] Based on the aforementioned electricity demand and pre-acquired distributed resource data, a pre-control scheme for electricity load during the winter return to hometowns is formulated.

[0177] The power grid is managed and controlled according to the power load pre-control scheme.

[0178] Specifically, the power load pre-control scheme includes: red distribution areas are under the key management of the provincial company's "two centers" and are prioritized for resolution through emergency package projects and other channels; yellow distribution areas are tracked and managed by the municipal company, with enhanced monitoring of operational status and temporary handling measures such as reusing existing capacity and increasing capacity.

[0179] For example, in response to problems exposed in the operation of distribution transformers in previous years, Anhui Company has taken the following measures: First, it has improved mechanisms and implemented long-term management, promoting the implementation of the "1+3" power supply quality management mechanism and routinely addressing problems according to the principle of "maintenance first, then technology, and finally engineering." Second, it has implemented comprehensive monitoring and intelligent analysis, deepening voltage monitoring at the point of sale based on real-time monitoring data from HPLC meters, intelligently analyzing the causes of problems, and accurately matching remediation measures. Third, it has implemented tracking and closed-loop management, focusing on rectifying 5,868 repeated low-voltage distribution areas during the 2023 Spring Festival travel rush and the 2024 peak summer season (2,034 have been rectified, and the task is expected to be completed by the end of December), tracking the progress and effectiveness of the remediation daily.

[0180] Furthermore, the power load pre-control scheme also includes: developing a precise pre-control list by dynamically tracking and deepening the application of intelligent auxiliary decision-making, promoting early intervention and targeted management of red power distribution areas, and conducting key monitoring and reasonable reserves for yellow power distribution areas.

[0181] Furthermore, the power load pre-control scheme also includes: precise measures, accelerating the management of power supply quality issues, strengthening the closed-loop mechanism for problem tracking, and relying on the "two centers" to conduct online verification of the management process and confirm the management results, so as to ensure that the problems are truly and effectively resolved.

[0182] Furthermore, the power load pre-control scheme also includes: enhancing power supply service capacity during the Spring Festival through internal and external coordination, deploying power supply guarantee measures in key areas such as scenic spots in advance, and ensuring the power supply reliability of special transformer areas during special periods.

[0183] Furthermore, the power load pre-control scheme also includes: making emergency preparations for severe weather in advance through resource pre-positioning, closely monitoring weather changes, and organizing various units to predict disaster-affected areas and pre-deploy emergency resources in accordance with the provincial and municipal disaster response and disposal mechanisms. At the same time, it strengthens the overall coordination of emergency resources to ensure that teams, materials and equipment can be dispatched across regions in a timely manner.

[0184] For example, to date, 181 red overload transformer areas, 713 low voltage transformer areas at the outlet, and 68,700 households with low voltage at the point of sale have been rectified.

[0185] In summary, a power load pre-control plan for the winter travel rush is formulated based on the red, yellow, and green classification of distribution areas. The plan focuses on strengthening resource allocation and emergency response for red high-load distribution areas, conducting real-time monitoring and optimized scheduling for yellow medium-load distribution areas, and maintaining balanced power supply for green low-load distribution areas. This ensures the stable operation of the power grid during the winter travel rush and peak electricity consumption periods, and prevents regional overload problems and power shortage risks.

[0186] In summary, the power load pre-control scheme achieves precise pre-control through dynamic tracking and intelligent auxiliary decision-making applications, promotes targeted governance of red distribution areas and key monitoring of yellow distribution areas, and optimizes the closed loop of problem governance by relying on the "two centers" to ensure a real and effective improvement in power supply quality. Through internal and external collaboration, it makes advance deployment of key areas and ensures power supply capacity during special periods, and pre-configures emergency resources in the event of severe weather by relying on the provincial and municipal disaster response mechanisms, ultimately comprehensively improving the stability, accuracy and emergency response capabilities of the power grid.

[0187] Compared with the prior art, the present invention has the following beneficial effects:

[0188] 1. By analyzing the power load characteristics and environmental conditions of different regions through multi-factor correlation analysis, and combining iterative training to dynamically adjust the prediction model, the efficiency and reliability of the model under different application scenarios are ensured. At the same time, the model can accurately capture complex influencing factors during the winter travel season, such as population flow and weather changes, thereby improving the accuracy of power load prediction.

[0189] 2. Through hierarchical color-coding and real-time monitoring, the system can quickly respond to power load fluctuations, formulate targeted prevention and control plans, and ensure the stable operation of the power grid. At the same time, by building a smart grid, the system can automatically analyze the causes of power consumption problems, optimize power load pre-control strategies, reduce manual intervention, improve decision-making efficiency, and further enhance the timeliness and effectiveness of power load prevention and control. Finally, by integrating distributed resource data, the system can rationally allocate power resources, improve resource utilization efficiency, and reduce energy waste.

[0190] like Figure 2 The diagram shown is a functional block diagram of a multi-factor correlation power load pre-control system for returning home for winter provided by an embodiment of the present invention.

[0191] The multi-factor correlation-based power load pre-control system 100 for winter return-to-hometown use, as described in this invention, can be installed in an electronic device. Depending on the functions implemented, the multi-factor correlation-based power load pre-control system 100 may include a power problem identification module 101, a power load influencing factor analysis module 102, a power load prediction module 103, a graded transformer substation division module 104, and a power load pre-control module 105. The module described in this invention can also be called a unit, referring to a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, stored in the memory of the electronic device.

[0192] In this embodiment, the functions of each module / unit are as follows:

[0193] The electricity problem identification module 101 is used to perform Fourier transform on the pre-acquired historical electricity data to obtain the power load data of the power grid, and to identify the load fluctuation of the power load data to obtain the electricity problem of the power grid.

[0194] The power load influencing factor analysis module 102 is used to perform Pareto analysis on the causes of the power consumption problem to obtain the influencing factors of the power consumption problem, perform load characteristic cluster analysis on the influencing factors to obtain the data matrix of strongly correlated factors of the power consumption problem, and perform quantitative analysis on the data matrix of strongly correlated factors to obtain the correlation influence coefficient of the power grid.

[0195] The power load prediction module 103 is used to iteratively train the power load prediction model based on the correlation influence coefficient to obtain the trained power load prediction model, and input the pre-acquired current electricity consumption data into the trained power load prediction model to obtain the power grid load prediction data.

[0196] The graded distribution module 104 is used to classify the power grid by color based on the load situation of the load forecast data and the severity of the power consumption problem, so as to obtain the red, yellow and green graded distribution areas of the power grid.

[0197] The power load pre-control module 105 is used to formulate a power load pre-control plan for returning home for winter based on the red, yellow and green graded transformer areas, and to carry out power consumption prevention and control management of the power grid according to the power load pre-control plan.

[0198] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0199] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0200] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0201] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0202] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0203] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for pre-controlling electricity load during the winter travel rush due to multiple factors, characterized in that, The method includes: S1. Perform Fourier transform on the pre-acquired historical electricity consumption data to obtain the power load data of the power grid, identify load fluctuations in the power load data, and obtain the electricity consumption problem of the power grid. S2. Perform a Pareto analysis on the causes of the electricity problem to obtain the influencing factors of the electricity problem, including: The causes of the electricity problems are classified to obtain the causal categories of the electricity problems; Based on the aforementioned cause categories, a quantitative statistical analysis of the problems is performed to obtain the number of problems for each cause category. A Pareto diagram of the electricity consumption problem is drawn based on the causal category and the number of problems. Based on the Pareto chart, the correlation of the electricity consumption problem is verified to obtain the influencing factors of the electricity consumption problem; Load characteristic clustering analysis was performed on the aforementioned influencing factors to obtain a data matrix of strongly correlated factors for the electricity consumption problem, including: The influencing factors are input into the load characteristic clustering model to obtain the clustering data of the electricity consumption problem; The positive impact of the influencing factors is calculated based on the clustering data and the positive impact formula, wherein the positive impact formula is: In the formula, The historical electricity consumption data is numbered. The number of the clustered data. For the first The first historical electricity consumption data The positive impact of clustering data For the first The first historical electricity consumption data The content of each clustered data, For the first The minimum value of each cluster data point For the first The maximum value of each cluster data point For the first Cluster data; The negative impact of the influencing factors is calculated based on the clustering data and the negative impact formula, wherein the negative impact formula is: In the formula, The historical electricity consumption data is numbered. The number of the clustered data. For the first The first historical electricity consumption data The negative impact of individual clustering data For the first The first historical electricity consumption data The content of each clustered data, For the first The minimum value of each cluster data point For the first The maximum value of each cluster data point For the first Cluster data; Establish the data matrix of strongly correlated factors based on the positive and negative effects; The correlation influence coefficient of the power grid is obtained by quantitatively analyzing the data matrix of the strongly correlated factors. S3. Based on the correlation influence coefficient, iteratively train the power load prediction model to obtain the trained power load prediction model. Input the pre-acquired current electricity consumption data into the trained power load prediction model to obtain the load prediction data of the power grid. S4. Based on the load conditions of the load forecast data and the severity of the power consumption problem, the power grid is classified by color to obtain the red, yellow and green graded transformer areas of the power grid. S5. Based on the red, yellow, and green graded distribution areas, formulate a power load pre-control plan for returning home for the winter, and carry out power consumption prevention and control management of the power grid according to the power load pre-control plan.

2. The method for pre-controlling electricity load during the winter travel rush based on multiple factors as described in claim 1, characterized in that, The step of performing a Fourier transform on the pre-acquired historical electricity consumption data to obtain the power load data of the power grid includes: The historical electricity consumption data is cleaned of noise to obtain standardized data of the historical electricity consumption data; The standardized data is periodically divided to obtain segmented data of the standardized data; Perform a Fourier transform on the segmented data to obtain the power data of the segmented data; Load characteristics are extracted from the power data to obtain the power load data of the power grid.

3. The method for pre-controlling electricity load during the winter return-to-hometown period based on multiple factors as described in claim 1, characterized in that, The process of identifying load fluctuations in the power load data to determine the power grid's electricity consumption issues includes: Feature extraction is performed on the time-domain and frequency-domain data of the power load data to obtain the time-frequency features of the power load data; Abnormal fluctuations are identified in the time-frequency features to obtain abnormal fluctuation data of the time-frequency features; Statistical characteristic analysis is performed on the historical electricity consumption data to obtain the load fluctuation threshold of the power grid; Based on the abnormal fluctuation data and the load fluctuation threshold, the power grid is diagnosed to obtain the power consumption problems of the power grid.

4. The method for pre-controlling electricity load related to the return-to-hometown-for-winter holiday as described in claim 1, characterized in that, The quantitative analysis of the strongly correlated factor data matrix to obtain the correlation influence coefficient of the power grid includes: Trend analysis is performed on the data matrix of strongly correlated factors to obtain the changing trend of the data matrix of strongly correlated factors; The dispersion analysis of the strongly correlated factor data matrix is ​​performed to obtain the degree of dispersion of the strongly correlated factor data matrix; Based on the changing trend and the degree of dispersion, the correlation relationship of the strongly correlated factor data matrix is ​​analyzed to obtain the correlation influence coefficient of the power grid.

5. The method for pre-controlling electricity load during the winter return-to-hometown period based on multiple factors as described in claim 1, characterized in that, The power load forecasting model is iteratively trained based on the correlation influence coefficient to obtain a trained power load forecasting model. Pre-acquired current electricity consumption data is then input into the trained power load forecasting model to obtain the power grid load forecasting data, including: The power load prediction model is iteratively trained based on the correlation influence coefficient to obtain the iterative power load prediction model. The historical electricity consumption data is input into the iterative power load prediction model to calculate the loss value, thereby obtaining the loss value of the iterative power load prediction model. The iterative loss value calculation function is as follows: In the formula, The loss value is... The quantity of the historical electricity consumption data. The number of the correlation influence coefficients, The historical electricity consumption data is numbered. The number is the correlation coefficient. For the first Each correlation coefficient For the first The power load is calculated using a correlation influence coefficient. For the first Historical electricity consumption data, For the first The actual value of historical electricity consumption data; The trained power load prediction model is determined based on the loss value; The pre-acquired current electricity consumption data is input into the trained power load prediction model to obtain the power grid load prediction data.

6. The method for pre-controlling electricity load during the winter return travel period based on multiple factors as described in claim 1, characterized in that, The power grid is classified into red, yellow, and green graded distribution areas based on the load forecast data and the severity of the electricity problems, including: The load forecast data is classified by load grading to obtain the predicted load grading of the power grid; The severity of the electricity problems is screened to identify the serious electricity problems in the power grid; Establish a quantitative relationship between the predicted load classification and the severe electricity problem; The power grid is classified into red, yellow, and green graded substations based on the quantification relationship.

7. The method for pre-controlling electricity load related to the return-to-hometown-for-winter holiday season as described in claim 6, characterized in that, The aforementioned power load pre-control scheme for returning home for winter is formulated based on the red, yellow, and green graded distribution areas. Power consumption control and management of the power grid are carried out according to the power load pre-control scheme, including: The red-yellow-green graded distribution area is monitored to obtain the real-time load data of the red-yellow-green graded distribution area; Based on the real-time load data and the quantitative relationship, demand analysis is performed on the red-yellow-green graded distribution area to obtain the electricity demand of the red-yellow-green graded distribution area. Based on the aforementioned electricity demand and pre-acquired distributed resource data, a pre-control scheme for electricity load during the winter return to hometowns is formulated. The power grid is managed and controlled according to the power load pre-control scheme.

8. A multi-factor correlation-based power load pre-control system for people returning home for winter, characterized in that, The system includes: The electricity problem identification module is used to perform Fourier transform on the pre-acquired historical electricity data to obtain the power load data of the power grid, and to identify the load fluctuations of the power load data to obtain the electricity problem of the power grid. The power load influencing factor analysis module is used to perform Pareto analysis on the causes of the power consumption problem to obtain the influencing factors of the power consumption problem, including: The causes of the electricity problems are classified to obtain the causal categories of the electricity problems; Based on the aforementioned cause categories, a quantitative statistical analysis of the problems is performed to obtain the number of problems for each cause category. A Pareto diagram of the electricity consumption problem is drawn based on the causal category and the number of problems. Based on the Pareto chart, the correlation of the electricity consumption problem is verified to obtain the influencing factors of the electricity consumption problem; Load characteristic clustering analysis was performed on the aforementioned influencing factors to obtain a data matrix of strongly correlated factors for the electricity consumption problem, including: The influencing factors are input into the load characteristic clustering model to obtain the clustering data of the electricity consumption problem; The positive impact of the influencing factors is calculated based on the clustering data and the positive impact formula, wherein the positive impact formula is: In the formula, The historical electricity consumption data is numbered. The number of the clustered data. For the first The first historical electricity consumption data The positive impact of clustering data For the first The first historical electricity consumption data The content of each clustered data, For the first The minimum value of each cluster data point For the first The maximum value of each cluster data point For the first Cluster data; The negative impact of the influencing factors is calculated based on the clustering data and the negative impact formula, wherein the negative impact formula is: In the formula, The historical electricity consumption data is numbered. The number of the clustered data. For the first The first historical electricity consumption data The negative impact of individual clustering data For the first The first historical electricity consumption data The content of each clustered data, For the first The minimum value of each cluster data point For the first The maximum value of each cluster data point For the first Cluster data; Establish the data matrix of strongly correlated factors based on the positive and negative effects; The correlation influence coefficient of the power grid is obtained by quantitatively analyzing the data matrix of the strongly correlated factors. The power load forecasting module is used to iteratively train the power load forecasting model based on the correlation influence coefficient to obtain the trained power load forecasting model. The pre-acquired current electricity consumption data is input into the trained power load forecasting model to obtain the load forecasting data of the power grid. The graded distribution module is used to classify the power grid by color based on the load conditions of the load forecast data and the severity of the power consumption problem, so as to obtain the red, yellow and green graded distribution areas of the power grid. The power load pre-control module is used to formulate a power load pre-control plan for returning home for winter based on the red, yellow and green graded transformer areas, and to carry out power consumption prevention and control management of the power grid according to the power load pre-control plan.

Images