Multi-dimensional electricity utilization monitoring method and system based on key industry electric quantity change characteristics

Abstract

The invention discloses a multi-dimensional electricity utilization monitoring method and system for key industry electric quantity change characteristics, and belongs to the technical field of electric power system monitoring and analysis, and the method comprises the steps: collecting different types of influence factor related data of a certain region, the types of the data comprise electric power, economy, meteorology and industry development, each type comprises influence factors of multiple dimensions; cleaning the collected data, arranging the data into a time sequence data set, and performing missing value repair and standardization processing on the time sequence data set; key industries having important influences on power demands and economic development are screened, and a key industry list is constructed; using an ARIMA model to construct medium and long term power consumption portraits of key industries; constructing multi-dimensional key industry power utilization tags; constructing a medium-and-long-term electric quantity change comprehensive monitoring index system; a medium-and-long-term electric quantity change early warning cube is constructed and used for monitoring electric quantity changes of key industries in real time, carrying out early warning on different abnormal signals and taking corresponding response measures.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system monitoring and analysis, and specifically relates to a multi-dimensional electricity consumption monitoring method and system for electricity consumption change characteristics in key industries. Background Art

[0002] With the rapid development of the economy and society and the continuous advancement of science and technology, the power system, as a key component of modern social infrastructure, carries an increasing and diverse energy demand. Especially in key industries, electricity is an important energy source for production activities, and its consumption and demand changes have a direct impact on the stability and economy of the entire power system. However, with the adjustment of industrial structure and fluctuations in the market environment, traditional electricity management methods can no longer cope with complex changes in power load and the dynamic impact of multiple factors, resulting in severe challenges in supply and demand balance, energy efficiency optimization and power dispatching.

[0003] First, the existing power system monitoring is mostly focused on single-dimensional data collection and analysis, which cannot fully reflect the impact of multi-dimensional factors such as social economy, meteorological conditions, and electricity price fluctuations on power demand. Especially in key industries, factors such as enterprise scale, production technology, and market demand have a significant impact on power use, but there is a lack of comprehensive analysis methods for these factors, resulting in low prediction accuracy of power demand, which in turn affects the efficiency of power dispatching.

[0004] Secondly, facing the increasingly complex electricity consumption environment, the power system still lacks an effective early warning mechanism for medium- and long-term electricity changes. The traditional single indicator monitoring model has been unable to cope with the complex and changing electricity demand, thus increasing the uncertainty and potential risks of power system operation.

[0005] The prior art CN118868404A discloses a smart electricity monitoring system, including: by real-time collection of environmental parameters, power load status and energy consumption data and integration with historical energy consumption records, real-time electricity price information and weather forecast data, a multi-dimensional dynamic energy consumption model is constructed to predict energy consumption trends; based on the multi-dimensional dynamic energy consumption model, energy consumption trends are predicted and dynamic adjustment strategies are formulated; the state of the power grid is dynamically monitored, and the state of the power grid is dynamically analyzed according to the predicted energy consumption trends and real-time power grid conditions, and the optimal load distribution strategy for each power unit is determined; the load balancing strategy is automatically executed through intelligent control equipment; the actual effect after the implementation of the strategy is monitored, and the execution efficiency and energy-saving effect of the strategy are evaluated. The shortcomings of the prior art documents are that they mainly focus on the energy consumption prediction and load regulation of the overall power system, and lack in-depth analysis of the power consumption characteristics of key industries. In addition, the prior art has limited capabilities in dealing with macro variables such as social economy and policy factors, and lacks the ability to dynamically adapt to complex power consumption environments and abnormal warning, making it difficult to cope with the diverse power demand in the industry. Summary of the invention

[0006] The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and to provide a multi-dimensional electricity consumption monitoring method and system for the electricity consumption change characteristics of key industries, so as to improve the safety of electricity consumption in the industry.

[0007] The present invention adopts the following technical solution.

[0008] The first aspect of the present invention provides a multi-dimensional electricity consumption monitoring method for key industry electricity consumption change characteristics, comprising the following steps:

[0009] Collect data on different types of influencing factors in a region, including electricity, economy, meteorology and industry development. Each type contains influencing factors of multiple dimensions.

[0010] Clean the collected data, organize it into a time series data set, and then perform missing value repair and normalization on the time series data set;

[0011] Based on the processed time series data set, key industries with significant impact on electricity demand and economic development are screened, the industries are prioritized, and a list of key industries is constructed;

[0012] Based on the time series data set and the list of key industries, the ARIMA model is used to construct a medium- and long-term electricity consumption profile of key industries;

[0013] Based on the mid- and long-term electricity consumption profiles of key industries in a region, a multi-dimensional electricity consumption label for key industries is constructed;

[0014] Based on the medium- and long-term electricity consumption portraits of key industries and multi-dimensional electricity consumption labels of key industries, a comprehensive monitoring indicator system for medium- and long-term electricity consumption changes is established;

[0015] Based on the comprehensive monitoring indicator system of medium- and long-term electricity quantity changes, a medium- and long-term electricity quantity change early warning cube is constructed to monitor the electricity quantity changes in key industries in real time, issue early warnings for different abnormal signals and take corresponding response measures.

[0016] Optionally, the multi-dimensional influencing factors included in the different types are specifically:

[0017] Electricity data includes electricity consumption, electricity load, and peak-to-valley difference;

[0018] Economic data include industry GDP growth rate, investment level, price index and import and export amount;

[0019] Meteorological data includes temperature, humidity, climate, and season;

[0020] Industry development data includes enterprise capacity expansion, market demand, technological innovation, industry scale, number of new projects and electricity consumption growth rate.

[0021] Optionally, the missing value repair and normalization processing of the time series data set is specifically:

[0022] Establish a regression interpolation model to predict the change of the target variable Y through the independent variables, while considering the impact of the error term on the model results;

[0023] The accuracy of the regression interpolation model is measured using the mean square error. If the calculated mean square error is less than 0.05, it means that the accuracy of the interpolation model has reached the expected requirements. After meeting this standard, the interpolation model is used to interpolate the missing values ​​in the time series data set.

[0024] The imputed time series dataset is standardized by Min-Max standardization.

[0025] Optionally, the use of the ARIMA model to construct a medium- and long-term electricity consumption profile of key industries specifically includes:

[0026] Determine the parameters of the ARIMA model, and input the determined parameters into the ARIMA model for fitting;

[0027] The processed time series data set was used as the model input, and the fitted ARIMA model was validated by Liung-Box residual analysis to check the normality and autocorrelation of the residuals;

[0028] According to the normality and autocorrelation of the residuals, determine whether the residuals are white noise. If the residuals are white noise, a valid ARIMA model is obtained. If the residuals are not white noise, repeat the above steps, adjust the parameters of the ARIMA model and check again whether the residuals are white noise until a valid ARIMA model is obtained.

[0029] The relevant electricity consumption data of key industries are predicted through an effective ARIMA model, and the medium- and long-term electricity consumption portraits of key industries are obtained based on the electricity consumption trends predicted by the ARIMA model.

[0030] Optionally, the medium- and long-term electricity consumption portraits of key industries obtained based on the electricity consumption trends predicted by the ARIMA model include:

[0031] Trend portraits, seasonal portraits and fluctuation characteristic portraits. Among them, trend portraits are used to observe the long-term trend of electricity consumption, seasonal portraits are used to capture seasonal changes in electricity consumption data and identify peaks and troughs, and fluctuation characteristic portraits are used to observe the fluctuation amplitude of electricity consumption and identify demand uncertainty.

[0032] Optionally, the constructing of multi-dimensional key industry electricity consumption labels based on the medium- and long-term electricity consumption portraits of key industries in a region specifically includes:

[0033] Identify the dimensions needed to construct electricity usage labels for key sectors;

[0034] Perform feature engineering on the identified desired dimensions using statistical feature extraction;

[0035] Generate labels for the electricity consumption behaviors of key industries based on the extracted features, use principal component analysis to reduce the dimensionality of high-dimensional features, extract the main influencing factors, and generate concise label types;

[0036] Genetic algorithms are used to optimize the grouping and combination of generated labels to obtain multi-dimensional electricity consumption labels for key industries.

[0037] Optionally, constructing the medium- and long-term electricity quantity change warning cube further includes determining warning dimensions, and the warning dimensions include internal dimensions, external dimensions, time dimensions, and space dimensions.

[0038] Optionally, the constructing of the medium- and long-term electricity quantity change warning cube further includes setting a warning threshold type, and the warning threshold type includes a static threshold, a dynamic threshold and a multi-level threshold.

[0039] Optionally, the constructing of the medium- and long-term electricity quantity change warning cube further includes setting different warning types and their corresponding response measures;

[0040] The warning types include low-level warning, medium-level warning and high-level warning;

[0041] Among them, the response measures corresponding to the low-level warning are to only prompt that the data has fluctuated and no interference is required; the response measures corresponding to the intermediate warning are to notify the dispatcher to pay attention and determine whether the power supply strategy needs to be adjusted; the response measures corresponding to the high-level warning are to activate the emergency plan and notify relevant departments to prepare in advance.

[0042] A second aspect of the present invention provides a multi-dimensional electricity consumption monitoring system for key industry change characteristics, which is used to execute a multi-dimensional electricity consumption monitoring method for key industry change characteristics according to the first aspect of the present invention, and the system comprises:

[0043] Multi-dimensional data collection module, data preprocessing module, key industry screening module, medium- and long-term electricity consumption portrait construction module, monitoring indicator system construction module and early warning module;

[0044] Among them, the multi-dimensional data acquisition module is used to collect data related to different types of influencing factors in a certain area, and its types include electricity, economy, meteorology and industry development, and each type contains influencing factors of multiple dimensions; the data preprocessing module is used to process the data collected by the multi-dimensional data acquisition module and construct a time series data set; the key industry screening module constructs a list of key industries based on the time series data set; the medium- and long-term electricity consumption portrait construction module constructs the medium- and long-term electricity consumption portrait of key industries based on the time series data set and the list of key industries; the monitoring indicator system construction module is used to construct a medium- and long-term electricity consumption change comprehensive monitoring indicator system based on the medium- and long-term electricity consumption portrait of key industries and multi-dimensional key industry electricity consumption labels; the early warning module includes a long-term electricity change early warning cube, which is used to monitor the electricity changes in key industries in real time, issue early warnings for different abnormal signals and take corresponding response measures.

[0045] The beneficial effect of the present invention is that, compared with the prior art, the analysis accuracy of the power consumption characteristics of key industries and the power consumption management efficiency are significantly improved. By introducing time series analysis, regression analysis and machine learning technology, the present invention constructs a detailed power consumption portrait and multi-dimensional label, which can not only comprehensively reflect the power consumption mode of key industries, but also dynamically capture the law of power consumption changes in the industry, and provide a scientific basis for subsequent prediction and optimization. The difference from the prior art is mainly concentrated in the overall power load balance and static strategy adjustment. The present invention incorporates multi-dimensional factors such as social economy, meteorology, and electricity prices into the monitoring system through deep modeling, forming a power consumption management method with high adaptability to industry characteristics. This system not only improves the adaptability of the prediction model to complex environments, but also makes the optimization strategy more in line with the actual needs of the industry. In addition, through real-time monitoring and dynamic regulation, the present invention significantly enhances the abnormal warning capability, can quickly identify the risk of power fluctuations, and provide users with fine-grained power optimization strategies, which improves the flexibility and intelligence level of management. Compared with the existing technology, the present invention shows obvious advantages in prediction accuracy, energy-saving efficiency and real-time response capability, providing industry users with more reliable and comprehensive decision-making support in complex and changeable electricity usage scenarios, and promoting the development of smart energy management from general optimization to precision. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, not all of them. Based on the spirit of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the scope of protection of the present invention.

[0048] Changes in electricity consumption in key industries are often accompanied by the interactive influence of multi-dimensional factors, such as economic growth, investment fluctuations, and meteorological changes.

[0049] Based on the problems existing in the prior art, the present invention proposes a multi-dimensional electricity consumption monitoring method and system for the characteristics of electricity consumption changes in key industries. The system constructs electricity consumption portraits and multi-dimensional labels for key industries by introducing advanced data analysis technologies such as time series analysis, regression analysis, and machine learning. Combining the influence of multiple factors such as social economy, meteorology, and electricity prices, an indicator system is constructed that can monitor, warn, and predict electricity consumption changes in real time. Through this innovative multi-dimensional monitoring method, power companies and government departments can respond to changes in electricity demand in key industries more accurately and promptly, optimize the allocation of power resources, and reduce electricity consumption risks, thereby improving the reliability and operating efficiency of the entire power system.

[0050] To achieve the above object, the present invention provides the following technical solutions.

[0051] In embodiment 1, the present invention provides a multi-dimensional electricity consumption monitoring method for key industries' electricity consumption change characteristics, the method comprising the following steps:

[0052] Step S1: Collect data related to different types of influencing factors in a certain region, including four types: electricity, economy, meteorology, and industry development. Each type contains influencing factors of multiple dimensions;

[0053] In a further preferred but non-limiting embodiment, the multi-dimensional influencing factors included in the different types in step S1 are specifically:

[0054]

[0055] Step S2: Clean the data collected in step S1, organize it into a time series data set, and then repair and normalize the data set to ensure data quality.

[0056] In a further preferred but non-limiting embodiment, in step S2, the specific method of repairing missing values ​​and normalizing the data includes the following:

[0057] Use regression interpolation to interpolate the collected data and establish a regression interpolation model as follows:

[0058] Y=β0+β1X1+β2X2+···+β p X p +ε (1)

[0059] Among them, Y is the interpolation data to be repaired, X1, X2…X p is time, β0, β1, …β p is the model parameter, ε is the error term;

[0060] After interpolation, the mean square error is used to measure the accuracy of the regression interpolation model. The calculation formula of the mean square error is:

[0061]

[0062] Where n is the total number of observations, y i is the ith observation, is the interpolated value of the ith observation;

[0063] Adjust the parameters. When MSE < 0.05, the accuracy of the interpolation model meets the requirements, and the interpolation model is used to interpolate the missing values ​​in the data set.

[0064] The interpolated data set is standardized based on the Min-Max standardization method to ensure that data sets of different dimensions are suitable for unified analysis. The specific formula is as follows:

[0065]

[0066] Where x is the original data value, min(x) is the smallest value in the data set, max(x) is the largest value in the data set, and x ′ is the standardized data value, x ′ The range is [0,1].

[0067] Step S3: Based on the data set processed in step S2, select key industries that have a significant impact on electricity demand and economic development, prioritize the industries through the analytic hierarchy process, and construct a list of key industries.

[0068] In a further preferred but non-limiting embodiment, step S3 specifically comprises:

[0069] S3.1, construct a hierarchical model to decompose the decision problem into the goal layer, criterion layer and solution layer.

[0070] Target layer Select the industry to focus on. The role of this layer is the ultimate goal of decision-making Criteria Layer Criteria or factors that affect the achievement of goals, such as market demand, technological innovation, electricity consumption, etc. Solution layer Specific options to be evaluated

[0071] S3.2, at the criteria and solution levels, the importance of each industry is compared pairwise using expert scoring, and a pairwise comparison matrix is ​​established, where the element a of the comparison matrix is ijIt represents the importance ratio of item i to item j. The commonly used 1-9 scale method represents relative importance:

[0072] 1 Two factors are equally important 3 The former is slightly more important than the latter 5 The former is more important than the latter 7 The former is obviously more important than the latter 9 The former is more important than the latter 2、4、6、8 The median value of the above importance

[0073] S3.3, calculate the weight vector of the influencing factors through the eigenvector method, normalize the eigenvector, and obtain the relative weight of each influencing factor.

[0074] S3.4, conduct consistency test. Due to the subjective bias in pairwise comparison of industry priorities, AHP uses consistency test to verify the rationality of the comparison matrix.

[0075] Calculate the consistency CI:

[0076]

[0077] Among them, λ max is the maximum eigenvalue of the comparison matrix, and n is the order of the matrix.

[0078] Calculate the consistency ratio CR:

[0079]

[0080] Among them, RI is the average random consistency index.

[0081] If CR<0.1, the matrix passes the consistency test, otherwise the comparison matrix needs to be adjusted.

[0082] S3.5. Construct a list of key focus industries based on the comparison matrix that has passed the consistency test.

[0083] Step S4: Based on the time series data set constructed in step S2 and the list of key industries constructed in step S3, the ARIMA model is used to construct a medium- and long-term electricity consumption profile of key industries;

[0084] In a further preferred but non-limiting embodiment, the use of the ARIMA model to construct a medium- and long-term electricity consumption profile of key industries is specifically as follows:

[0085] The autoregressive integrated moving average (ARIMA) model is used to construct electricity consumption profiles of key industries, and the time series data set obtained in step S2 is used as the model input to capture the trend, seasonality and random fluctuation characteristics in the time series, so as to construct electricity consumption profiles of key industries.

[0086] In a further preferred but non-limiting embodiment, step S4 specifically comprises:

[0087] S4.1, determine the parameters (p, d, q) of the ARIMA model. The ARIMA model is controlled by three main parameters (p, d, q). p is the number of autoregressive terms, that is, using the first p lags of autoregression, d is the number of differences, which processes the data stationarity, and q is the number of moving average terms, that is, using the first q error terms. The formula of the ARIMA model is:

[0088]

[0089] Where Y t Output value for the current model, represents the autoregressive coefficient, θ represents the moving average coefficient, ∈ t is the error term, c is the constant term;

[0090] Among them, the number of differences d is determined based on data stationarity detection, usually using first-order differences or second-order differences; the number of autoregressive terms p and the number of moving average terms q are determined using the autocorrelation function (ACF) and partial autocorrelation coefficient (PACF) respectively; the determined parameters (p, d, q) are input into the ARIMA model for fitting;

[0091] S4.2, using the time series data set obtained in step S2 as the model input, verify the ARIMA model fitted in S4.1 through the Liung-Box residual analysis, and check the normality and autocorrelation of the residuals, as follows:

[0092] (1) Assumptions of the Liung-Box test

[0093] Null hypothesis: There is no obvious autocorrelation between residuals (residuals are white noise).

[0094] Alternative hypothesis: There is significant autocorrelation between the residuals.

[0095] (2)Liung-Box test statistic Q

[0096] The calculation formula is as follows:

[0097]

[0098] Where n is the number of samples, m is the number of lags, It represents the k-th order autocorrelation coefficient of the residual. Under large samples, the Q statistic follows a chi-square distribution with m degrees of freedom.

[0099] (3)Liung-Box significance level p

[0100] The p-value is a significance level calculated based on the Q-value, which is used to determine whether the original hypothesis is true, and thus to determine whether the residual has autocorrelation. There is a certain relationship between the p-value and the Q-value. The larger the Q-value, the stronger the autocorrelation of the residual, and the smaller the p-value, the less the residual is white noise; the smaller the Q-value, the less significant the correlation of the residual, and the larger the p-value, the closer the residual is to white noise.

[0101] S4.3, based on the normality and autocorrelation of the residuals calculated in S4.2, determine whether the residuals are white noise. If the residuals are white noise, a valid ARIMA model is obtained. If the residuals are not white noise, repeat S4.1 and S4.2, adjust the parameters of the ARIMA model and check again whether the residuals are white noise until a valid ARIMA model is obtained.

[0102] S4.4, predict the relevant electricity consumption data of key industries through an effective ARIMA model, and obtain the medium- and long-term electricity consumption portrait of key industries based on the electricity consumption trend predicted by the ARIMA model.

[0103] In a further preferred but non-limiting embodiment, in S4.4, the medium- and long-term electricity consumption portrait of key industries obtained based on the electricity consumption trend predicted by the ARIMA model specifically includes:

[0104] (1) Trend profile: observe the long-term trend of electricity consumption;

[0105] (2) Seasonal profiling: capturing seasonal changes in electricity consumption data and identifying peaks and troughs;

[0106] (3) Fluctuation characteristic profile: Observe the fluctuation range of electricity consumption to identify demand uncertainty.

[0107] Step S5: Based on the medium- and long-term electricity consumption portrait of key industries in a region constructed in step S4, construct multi-dimensional electricity consumption labels for key industries.

[0108] In a further preferred but non-limiting embodiment, S5 specifically includes:

[0109] S5.1, identify the key dimensions needed to construct an electricity usage label;

[0110] According to the influencing factor data processed in step S2, the influencing factors required for constructing the label are determined, including: socio-economic factors, meteorological factors, business expansion factors, electricity price factors, etc.

[0111] S5.2, use the statistical feature extraction method to perform feature engineering on the required dimensions determined in S5.1, and construct feature variables for each dimension. The specific steps are as follows:

[0112] First, for the time series data of key dimensions such as social economy, meteorology, industrial expansion, and electricity prices, the relevant data of statistical characteristics are calculated, including mean, median, standard deviation, maximum and minimum variance, skewness, and kurtosis.

[0113] Next, for the time series data of each key dimension, windowed statistical feature extraction is used, the time series is segmented by the sliding window method, and statistical features are extracted within each window to capture the local characteristics of the time series. The relevant parameters of the window are mainly the window size and the sliding step size.

[0114] Finally, the Lasso method is used to extract the correlation between features and target variables and screen out features with higher correlation.

[0115] S5.3, based on the extracted features, generate labels for the electricity consumption behaviors of key industries, use the principal component analysis method to reduce the dimension of high-dimensional features, extract the main influencing factors, and generate concise label types.

[0116] S5.4, use genetic algorithm to optimize the grouping and combination of labels to ensure the discrimination and representativeness of labels, optimize the label combination through selection, crossover and mutation, find the most representative label, and construct multi-dimensional electricity consumption labels for key industries.

[0117] Step S6: Based on the medium- and long-term electricity consumption portraits of key industries and the multi-dimensional electricity consumption labels of key industries generated in steps S4 and S5, a comprehensive monitoring indicator system for medium- and long-term electricity consumption changes is constructed. The details are as follows:

[0118] (1) Internal correlation analysis

[0119] The relationship between electricity consumption and production activities within the industry was analyzed using multiple linear regression methods.

[0120] Y=β0+β1X1+β2X2+···+β p X p +ε (8)

[0121] Among them, Y is the power consumption, X1, X2, ..., X p are influencing factors, β1, β2, …, β p It is a model parameter, which indicates the influence strength of the influencing factor on the object to be predicted. β0 and ε represent the intercept and error respectively.

[0122] (2) External correlation analysis

[0123] The time-lagged correlation coefficient method is used to analyze and study the impact of industry electricity consumption on upstream and downstream industries and external factors such as meteorology.

[0124] (3) Monitoring indicator integration

[0125] The various detection indicators are summarized through the weighted average method to form a comprehensive monitoring indicator system.

[0126] S=ω1Z1+ω2Z2+…+ω n Z n (9)

[0127] Among them, S is the comprehensive monitoring index, Z i is the monitoring indicator, ω i is the corresponding weight.

[0128] Step S7: Based on the comprehensive monitoring indicator system for medium- and long-term electricity changes constructed in step S6, a medium- and long-term electricity change early warning cube is constructed. The electricity changes in key industries are monitored in real time through the medium- and long-term electricity change early warning cube, and corresponding early warnings are issued for different abnormal signals and corresponding response measures are taken.

[0129] In a further preferred but non-limiting embodiment, the S7 is specifically as follows:

[0130] (1) Determine the warning dimension

[0131] Construct a medium- and long-term electricity quantity change warning cube from internal dimension, external dimension, time dimension and space dimension.

[0132] (2) Establishing an early warning indicator system

[0133] Based on the monitoring indicator system in step S6, core indicators that can reflect changes in industry electricity consumption are selected, and warning thresholds are set. The threshold types include:

[0134] Static threshold: set upper and lower limits based on historical mean and standard deviation;

[0135] Dynamic threshold: Dynamically adjust the threshold in combination with time series forecasting settings;

[0136] Multi-level thresholds: According to the different degrees of abnormal situations, the warning thresholds are divided into multiple levels, such as slight abnormalities, significant abnormalities, severe abnormalities, etc.

[0137] (3) Real-time monitoring and anomaly detection

[0138] Input the real-time collected data into the early warning cube, and monitor the changes in the data based on the threshold of each dimension.

[0139] (4) Early warning mechanism and response strategy

[0140] A multi-level early warning method is used to warn different abnormal signals and take different response measures. The warning types are as follows:

[0141] Low-level warning: It only indicates that the data has fluctuated, and no interference is required.

[0142] Intermediate warning: The system notifies the dispatcher to pay attention and determine whether the power supply strategy needs to be adjusted.

[0143] Advanced warning: Immediately activate the emergency plan and notify relevant departments to prepare in advance.

[0144] In Example 2, the present invention provides a multi-dimensional electricity consumption monitoring system for key industry electricity consumption change characteristics, which is used to execute a multi-dimensional electricity consumption monitoring method for key industry electricity consumption change characteristics described in Example 1, and the system includes:

[0145] Multi-dimensional data collection module, data preprocessing module, key industry screening module, medium- and long-term electricity consumption portrait construction module, monitoring indicator system construction module and early warning module;

[0146] Among them, the multi-dimensional data acquisition module is used to collect data related to different types of influencing factors in a certain area, and its types include electricity, economy, meteorology and industry development, and each type contains influencing factors of multiple dimensions; the data preprocessing module is used to process the data collected by the multi-dimensional data acquisition module and construct a time series data set; the key industry screening module constructs a list of key industries based on the time series data set; the medium- and long-term electricity consumption portrait construction module constructs the medium- and long-term electricity consumption portrait of key industries based on the time series data set and the list of key industries; the monitoring indicator system construction module is used to construct a medium- and long-term electricity consumption change comprehensive monitoring indicator system based on the medium- and long-term electricity consumption portrait of key industries and multi-dimensional key industry electricity consumption labels; the early warning module includes a long-term electricity change early warning cube, which is used to monitor the electricity changes in key industries in real time, issue early warnings for different abnormal signals and take corresponding response measures.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A multi-dimensional electricity consumption monitoring method for key industry electricity consumption change characteristics, characterized in that: The steps include: Collect data on different types of influencing factors in a region, including electricity, economy, meteorology and industry development. Each type contains influencing factors of multiple dimensions. Clean the collected data, organize it into a time series data set, and then perform missing value repair and normalization on the time series data set; Based on the processed time series data set, key industries with significant impact on electricity demand and economic development are screened, the industries are prioritized, and a list of key industries is constructed; Based on the time series data set and the list of key industries, the ARIMA model is used to construct a medium- and long-term electricity consumption profile of key industries; Based on the mid- and long-term electricity consumption profiles of key industries in a region, a multi-dimensional electricity consumption label for key industries is constructed; Based on the medium- and long-term electricity consumption portraits of key industries and multi-dimensional electricity consumption labels of key industries, a comprehensive monitoring indicator system for medium- and long-term electricity consumption changes is established; Based on the comprehensive monitoring indicator system of medium- and long-term electricity quantity changes, a medium- and long-term electricity quantity change early warning cube is constructed to monitor the electricity quantity changes in key industries in real time, issue early warnings for different abnormal signals and take corresponding response measures.

2. According to claim 1, a multi-dimensional electricity consumption monitoring method for key industry electricity consumption change characteristics is characterized by: The multi-dimensional influencing factors included in the different types are specifically: Electricity data includes electricity consumption, electricity load, and peak-to-valley difference; Economic data include industry GDP growth rate, investment level, price index and import and export amount; Meteorological data includes temperature, humidity, climate, and season; Industry development data includes enterprise capacity expansion, market demand, technological innovation, industry scale, number of new projects and electricity consumption growth rate.

3. According to claim 1, a multi-dimensional electricity consumption monitoring method for key industry electricity consumption change characteristics is characterized by: The missing value repair and normalization processing of the time series data set is specifically as follows: Establish a regression interpolation model to predict the change of the target variable Y through the independent variables, while considering the impact of the error term on the model results; The accuracy of the regression interpolation model is measured using the mean square error. If the calculated mean square error is less than 0.05, it means that the accuracy of the interpolation model has reached the expected requirements. After meeting this standard, the interpolation model is used to interpolate the missing values ​​in the time series data set. The imputed time series dataset is standardized by Min-Max standardization.

4. According to claim 1, a multi-dimensional electricity consumption monitoring method for key industry electricity consumption change characteristics is characterized by: The use of the ARIMA model to construct a medium- and long-term electricity consumption profile for key industries specifically includes: Determine the parameters of the ARIMA model, and input the determined parameters into the ARIMA model for fitting; The processed time series data set was used as the model input, and the fitted ARIMA model was validated by Liung-Box residual analysis to check the normality and autocorrelation of the residuals; According to the normality and autocorrelation of the residuals, determine whether the residuals are white noise. If the residuals are white noise, a valid ARIMA model is obtained. If the residuals are not white noise, repeat the above steps, adjust the parameters of the ARIMA model and check again whether the residuals are white noise until a valid ARIMA model is obtained. The relevant electricity consumption data of key industries are predicted through an effective ARIMA model, and the medium- and long-term electricity consumption portraits of key industries are obtained based on the electricity consumption trends predicted by the ARIMA model.

5. The multi-dimensional electricity consumption monitoring method of key industry change characteristics according to claim 4 is characterized by: The medium- and long-term electricity consumption portraits of key industries obtained based on the electricity consumption trends predicted by the ARIMA model include: Trend portraits, seasonal portraits and fluctuation characteristic portraits. Among them, trend portraits are used to observe the long-term trend of electricity consumption, seasonal portraits are used to capture seasonal changes in electricity consumption data and identify peaks and troughs, and fluctuation characteristic portraits are used to observe the fluctuation amplitude of electricity consumption and identify demand uncertainty.

6. The multi-dimensional electricity consumption monitoring method of key industry change characteristics according to claim 1 is characterized by: The construction of multi-dimensional key industry electricity consumption labels based on the medium- and long-term electricity consumption portraits of key industries in a region specifically includes: Identify the dimensions needed to construct electricity usage labels for key sectors; Perform feature engineering on the identified desired dimensions using statistical feature extraction; Generate labels for the electricity consumption behaviors of key industries based on the extracted features, use principal component analysis to reduce the dimensionality of high-dimensional features, extract the main influencing factors, and generate concise label types; Genetic algorithms are used to optimize the grouping and combination of generated labels to obtain multi-dimensional electricity consumption labels for key industries.

7. The multi-dimensional electricity consumption monitoring method of key industry change characteristics according to claim 1 is characterized by: The construction of the medium- and long-term electricity quantity change warning cube also includes determining warning dimensions, which include internal dimensions, external dimensions, time dimensions, and space dimensions.

8. The multi-dimensional electricity consumption monitoring method of key industry change characteristics according to claim 1 is characterized by: The construction of the medium- and long-term electricity quantity change warning cube also includes setting a warning threshold type, and the warning threshold type includes a static threshold, a dynamic threshold, and a multi-level threshold.

9. The multi-dimensional electricity consumption monitoring method of key industry change characteristics according to claim 1 is characterized by: The construction of the medium- and long-term electricity quantity change warning cube also includes setting different warning types and their corresponding response measures; The warning types include low-level warning, medium-level warning and high-level warning; Among them, the response measures corresponding to the low-level warning are to only prompt that the data has fluctuated and no interference is required; the response measures corresponding to the intermediate warning are to notify the dispatcher to pay attention and determine whether the power supply strategy needs to be adjusted; the response measures corresponding to the high-level warning are to activate the emergency plan and notify relevant departments to prepare in advance.

10. A multi-dimensional electricity consumption monitoring system for key industry change characteristics, used to execute a multi-dimensional electricity consumption monitoring method for key industry change characteristics according to any one of claims 1 to 9, characterized in that: The system includes: Multi-dimensional data collection module, data preprocessing module, key industry screening module, medium- and long-term electricity consumption portrait construction module, monitoring indicator system construction module and early warning module; Among them, the multi-dimensional data acquisition module is used to collect data related to different types of influencing factors in a certain area, and its types include electricity, economy, meteorology and industry development, and each type contains influencing factors of multiple dimensions; the data preprocessing module is used to process the data collected by the multi-dimensional data acquisition module and construct a time series data set; the key industry screening module constructs a list of key industries based on the time series data set; the medium- and long-term electricity consumption portrait construction module constructs the medium- and long-term electricity consumption portrait of key industries based on the time series data set and the list of key industries; the monitoring indicator system construction module is used to construct a medium- and long-term electricity consumption change comprehensive monitoring indicator system based on the medium- and long-term electricity consumption portrait of key industries and multi-dimensional key industry electricity consumption labels; the early warning module includes a long-term electricity change early warning cube, which is used to monitor the electricity changes in key industries in real time, issue early warnings for different abnormal signals and take corresponding response measures.

Images