CN117828507A

CN117828507A - Enterprise power consumption detection method, enterprise power consumption detection device, computer equipment and storage medium

Info

Publication number: CN117828507A
Application number: CN202311796262.9A
Authority: CN
Inventors: 彭文钦; 杨光; 李伟鹏; 李九兰
Original assignee: China Southern Power Grid Digital Platform Technology Guangdong Co ltd
Current assignee: China Southern Power Grid Digital Platform Technology Guangdong Co ltd
Priority date: 2023-12-25
Filing date: 2023-12-25
Publication date: 2024-04-05

Abstract

The application relates to an enterprise electricity usage detection method, apparatus, computer device, storage medium and computer program product. The method comprises the following steps: the method comprises the steps of obtaining real-time electricity utilization data, real-time environment data and electricity utilization factor of a target enterprise, carrying out data preprocessing on the real-time electricity utilization data, the real-time environment data and the electricity utilization factor to obtain target electricity utilization data and target environment data, extracting electricity utilization characteristic data from the target electricity utilization data, extracting environment characteristic data from the target environment data, and calling a trained enterprise electricity utilization detection model to carry out electricity utilization abnormality detection by taking the electricity utilization characteristic data, the environment characteristic data and the target electricity utilization factor as input to obtain an enterprise electricity utilization detection result. By adopting the method, abnormal electricity utilization behavior can be detected more accurately, and the accuracy of the enterprise electricity utilization detection result is improved.

Description

Enterprise power consumption detection method, enterprise power consumption detection device, computer equipment and storage medium

Technical Field

The present application relates to the field of artificial intelligence and power technology, and in particular, to an enterprise power usage detection method, apparatus, computer device, storage medium, and computer program product.

Background

With the rise of industries such as big data, artificial intelligence, internet of things and the like, the production and living modes of people are greatly changed. People can build a model by means of data to predict, classify or evaluate certain things, so that the change rule of the things can be accurately drawn, and the specific problem can be solved.

Aiming at the electricity safety analysis of dangerous enterprises, the method mainly analyzes the aspects of electric load, type of electric equipment, power requirement and the like, and simultaneously considers the characteristics and the technological requirements of dangerous chemicals to determine the special electric safety requirements.

At present, the electrical safety analysis of dangerous chemical enterprises mainly comprises the steps of modeling and simulation analysis of an electrical system of the dangerous chemical enterprises by using electrical system simulation software, wherein the operation conditions of the electrical system can be simulated by the simulation software, including calculation and analysis of parameters such as current, voltage and power, so as to evaluate the stability, the load capacity and potential safety hazards of the system.

However, the above-mentioned scheme of electricity safety analysis still has a problem that the analysis result is not accurate enough.

Disclosure of Invention

In view of the foregoing, it is desirable to provide an enterprise electricity usage detection method, apparatus, computer device, computer readable storage medium, and computer program product that improve the accuracy of electricity usage security analysis.

In a first aspect, the present application provides a method for enterprise electricity usage detection. The method comprises the following steps:

acquiring real-time electricity utilization data, real-time environment data and electricity utilization factor of a target enterprise;

carrying out data preprocessing on the real-time electricity utilization data, the real-time environment data and the electricity utilization factor to obtain target electricity utilization data and target environment data;

extracting electricity utilization characteristic data from the target electricity utilization data, and extracting environmental characteristic data from the target environmental data;

taking electricity characteristic data, environment characteristic data and target electricity dynamic factors as input, and calling a trained enterprise electricity detection model to detect electricity anomalies so as to obtain an enterprise electricity detection result;

the enterprise electricity consumption detection model is obtained by training the classification prediction model based on historical electricity consumption data, historical environment data and historical electricity consumption dynamic factors of a target enterprise in a historical time period.

In one embodiment, the electricity usage characteristic data includes electricity usage load data, electricity usage amount, and electricity usage time;

the enterprise electricity detection model performs the following steps when invoked:

counting the change rule of the power consumption load data by a window counting method to obtain power consumption load change data;

Based on the existing historical electricity consumption data, analyzing the change rule of electricity consumption and electricity consumption time to obtain electricity consumption change data;

and carrying out power consumption abnormality detection on the power consumption load change data, the power consumption change data and the environmental characteristic data according to a preset abnormal power consumption judgment rule to obtain an enterprise power consumption detection result.

In one embodiment, before invoking the trained enterprise electricity detection model to perform electricity safety detection, the method further includes:

acquiring historical electricity consumption data, historical environment data and historical electricity consumption dynamic factors of a target enterprise in a historical time period;

carrying out data preprocessing on the historical electricity consumption data, the historical environment data and the historical electricity consumption dynamic factors to obtain target historical electricity consumption data, target historical environment data and target historical electricity consumption dynamic factors;

extracting historical electricity utilization characteristic data in the target historical electricity utilization data and historical environment characteristic data in the target historical environment data;

constructing a model training set based on the historical electricity utilization characteristic data, the historical environment characteristic data and the target historical electricity utilization electric wave factor;

based on the model training set, training an initial enterprise electricity detection model by a gradient descent method, and determining model parameters of the initial enterprise electricity detection model to obtain a trained enterprise electricity detection model;

The initial enterprise electricity detection model is constructed based on a logistic regression analysis algorithm.

In one embodiment, after obtaining the trained enterprise electricity usage detection model, the method further comprises:

cross-verifying the trained enterprise electricity detection model, and calculating the model performance index after each cross-verification to obtain a plurality of model performance indexes;

and obtaining a model performance evaluation result according to the model performance indexes.

In one embodiment, performing data preprocessing on the real-time electricity consumption data, the real-time environment data and the electricity consumption motioning element to obtain target electricity consumption data, target environment data and target electricity consumption motioning element includes:

and performing at least one of data cleaning, missing value filling, outlier processing and normalization processing on the real-time electricity data, the real-time environment data and the electricity consumption dynamic factors to obtain target electricity consumption data, target environment data and target electricity dynamic factors.

In one embodiment, after obtaining the enterprise electricity consumption detection result, the method further includes:

and pushing the early warning message under the condition that the enterprise electricity utilization detection result represents that the target enterprise has abnormal electricity utilization behavior.

In a second aspect, the present application further provides an enterprise electricity usage detection apparatus. The device comprises:

the data acquisition module is used for acquiring real-time electricity utilization data, real-time environment data and electricity utilization factor of a target enterprise;

the data preprocessing module is used for preprocessing the real-time power consumption data, the real-time environment data and the power consumption dynamic factors to obtain target power consumption data and target environment data;

the data extraction module is used for extracting electricity utilization characteristic data from the target electricity utilization data and extracting environment characteristic data from the target environment data;

the data detection module is used for calling a trained enterprise electricity detection model to detect electricity utilization abnormality by taking electricity utilization characteristic data, environment characteristic data and target electricity utilization dynamic factors as inputs to obtain an enterprise electricity utilization detection result;

In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor which when executing the computer program performs the steps of:

In a fourth aspect, the present application also provides a computer-readable storage medium. The computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:

In a fifth aspect, the present application also provides a computer program product. The computer program product comprises a computer program which, when executed by a processor, implements the steps of:

According to the enterprise electricity consumption detection method, device, computer equipment, storage medium and computer program product, the relation between the electricity consumption behavior and the environmental factors is captured through the historical electricity consumption data, the historical environmental data and the historical electricity consumption factor in advance to train an enterprise electricity consumption detection model, the accuracy and generalization capability of the model on electricity consumption abnormality detection tasks can be improved, under the actual electricity consumption detection scene, the current electricity consumption behavior and the environmental conditions can be reflected more accurately through acquiring the real-time electricity consumption data, the real-time environmental data and the electricity consumption factor of a target enterprise, the real-time property and the accuracy of electricity consumption abnormality detection are improved, the obtained target electricity consumption data, the target environmental data and the target electricity consumption factor are more standard through preprocessing the real-time electricity consumption data, the real-time environmental data and the real-time electricity consumption factor, the obtained target electricity consumption characteristic data and the model detection effect is improved, the electricity consumption characteristic data and environmental characteristic data are respectively extracted from the target electricity consumption data, the current electricity consumption characteristic data and the target electricity consumption factor is called as the current electricity consumption characteristic data, the current electricity consumption characteristic data of the enterprise electricity consumption characteristic data is accurately, the electricity consumption abnormality detection result of the enterprise electricity consumption abnormality detection is accurately called, and the enterprise electricity consumption abnormality detection efficiency is accurately detected, and the enterprise electricity consumption abnormality is accurately detected. Further, the scheme can be well suitable for electricity detection of dangerous enterprises, so that the electricity safety management level of the dangerous enterprises is improved.

Drawings

FIG. 1 is a diagram of an application environment for an enterprise electricity usage detection method in one embodiment;

FIG. 2 is a flow chart of an enterprise electricity usage detection method in one embodiment;

FIG. 3 is a flow chart of a step of detecting abnormal electricity consumption in one embodiment;

FIG. 4 is a flow chart of a step of training an enterprise electricity test model in one embodiment;

FIG. 5 is a flowchart illustrating a step of training an enterprise electricity test model in accordance with another embodiment;

FIG. 6 is a detailed flow diagram of an enterprise electricity usage detection method in one embodiment;

FIG. 7 is a block diagram of an enterprise electricity usage detection apparatus in one embodiment;

FIG. 8 is a block diagram of an enterprise electricity usage detection apparatus in another embodiment;

fig. 9 is an internal structural diagram of a computer device in one embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.

The enterprise electricity consumption detection method provided by the embodiment of the application can be applied to an application environment shown in fig. 1. Wherein the data acquisition terminal 102 communicates with the server 104 via a network. The data storage system may store data that the server 104 needs to process. The data storage system may be integrated on the server 104 or may be located on a cloud or other network server. Server 104 builds a trained enterprise electricity usage detection model. Specifically, the data acquisition terminal 102 may monitor and acquire the power consumption data, the surrounding environment data and the current power consumption factor of the target enterprise in real time, upload the acquired real-time power consumption data, real-time environment data and power consumption factor to the server 104 in real time, the server 104 acquires the real-time power consumption data, real-time environment data and power consumption factor of the target enterprise, then perform data preprocessing on the real-time power consumption data, the real-time environment data and the power consumption factor to obtain the target power consumption data, the target environment data and the target power consumption factor, then extract the power consumption characteristic data from the target power consumption data, extract the environment characteristic data from the target environment data, and finally call the trained power consumption detection model of the enterprise to perform power consumption abnormality detection by taking the power consumption characteristic data, the environment characteristic data and the target power consumption factor as inputs, so as to obtain the power consumption detection result of the enterprise.

The terminal 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, internet of things devices, and portable wearable devices, where the internet of things devices may be smart speakers, smart televisions, smart air conditioners, smart vehicle devices, and the like. The portable wearable device may be a smart watch, smart bracelet, headset, or the like. The server 104 may be implemented as a stand-alone server or as a server cluster of multiple servers.

In one embodiment, as shown in fig. 2, an enterprise electricity usage detection method is provided, and the method is applied to the server 104 in fig. 1 for illustration, and includes the following steps:

s200, acquiring real-time electricity utilization data, real-time environment data and electricity utilization factors of a target enterprise.

In this embodiment, the target enterprise refers to an enterprise to be detected whether abnormal electricity usage exists, and may be a hazardous enterprise or other type of enterprise. The hazardous chemical enterprises refer to enterprises engaged in activities such as dangerous chemical production, operation, storage, use and the like, and concretely comprise chemical enterprises, petrochemical enterprises, pharmaceutical enterprises, coating enterprises, dangerous article storage enterprises and the like. The real-time electricity consumption data comprise real-time electric energy consumption data, real-time electricity consumption load, real-time voltage, real-time power, real-time electricity consumption time and the like. The real-time environmental data includes temperature, humidity, wind speed, pressure, air quality, and illumination intensity of the current environment. The radio-active factors include, but are not limited to, holiday information (weekends, cases where holiday holidays lead to reduced or stopped production), hot events (hot news or events affect production), market changes, seasonal factors, and the like. Hot events generally refer to news or events that are currently social heat or have a relatively large impact, and may have an impact on the production of an enterprise.

In specific implementation, taking the target enterprise as a dangerous enterprise as an example, the power consumption of the dangerous enterprise can be recorded, measured and measured in real time through power sensor equipment, an intelligent electric energy meter or an electronic electric energy meter and the like, the data including power consumption, power, voltage, current, power load and the like are included, and the measured real-time power consumption data are transmitted to the server 104 in real time. By deploying sensors such as temperature sensors, humidity sensors, pressure sensors, light sensors, etc. at different locations around the hazardous chemical enterprise, real-time environmental data around the hazardous chemical enterprise is collected and uploaded to server 104. Or a special environment monitoring system can be used, and the system can be connected to various environment sensors through a network, collect and record environment data around the hazardous chemical enterprises in real time, upload the real-time environment data to the server 104, and provide data reporting and analysis functions.

S400, carrying out data preprocessing on the real-time electricity utilization data, the real-time environment data and the electricity utilization factor to obtain target electricity utilization data, target environment data and target electricity utilization factor.

After the real-time electricity data and the real-time environment data of the hazardous chemical enterprises are received, the data contains a large amount of noise data and invalid information, so that the data preprocessing, such as data cleaning and data normalization, can be performed on the real-time electricity data and the real-time environment data, and the real-time electricity data and the real-time environment data are processed into standard data which can be identified by a model, so that the target electricity data and the target environment data are obtained. For preprocessing with electric wave dynamic factors, holiday information can be converted into binary or numerical representation, so that a model can be better understood and processed. For example, holidays are defined as 1, non-holidays are defined as 0, or different holidays are assigned different numerical codes. The hotspot event may be converted to a binary variable, indicating whether there is a hotspot event impact. For example, 1 when there is a hot spot event, and 0 when there is no hot spot event.

S600, extracting electricity utilization characteristic data from the target electricity utilization data and extracting environment characteristic data from the target environment data.

The electricity consumption characteristic data is characteristic data which can reflect the electricity consumption change condition most, and comprises data such as electricity consumption load, electricity consumption amount, electricity consumption time, average electricity consumption power and the like. The environmental characteristic data comprise temperature, temperature average value, temperature variation amplitude, humidity, wind speed, humidity variation amplitude and the like.

Illustratively, the electricity usage characteristic data including the electricity usage load, the electricity usage time, the electricity usage amount, and the like may be extracted from the target electricity usage data by a statistical algorithm such as a statistical average, a variance, and the like, a frequency domain analysis such as fourier transform, and the like), and a time domain analysis such as a sliding window method, a difference method. And meanwhile, extracting environmental characteristic data such as temperature, temperature change amplitude, humidity, wind speed and the like from the target environmental data. Specifically, the extracted electricity characteristic data and environment characteristic data can be represented in the form of vectors, matrixes and the like so as to facilitate subsequent data analysis and application.

S800, calling a trained enterprise electricity detection model to detect electricity utilization abnormality by taking electricity utilization characteristic data, environment characteristic data and target electricity utilization dynamic factors as inputs, and obtaining an enterprise electricity utilization detection result.

The enterprise electricity consumption detection result comprises data such as electricity consumption abnormality identification, abnormality type, abnormality reason and the like. Specifically, the electricity usage abnormality identification indicates whether there is a binary identification of an electricity usage abnormality (abnormality/normal) or a probability value of whether the electricity usage abnormality. The anomaly type may be overload, power failure, voltage anomalies, etc. The abnormal reasons can be electric leakage of the electric facilities and fault shutdown of production equipment. The classification prediction model may be a model constructed by an algorithm such as a K nearest neighbor (K Nearest Neighbors, KNN) algorithm, a decision tree, or logistic regression.

In practical applications, basic information of a hazardous chemical enterprise, historical electricity data (including the result of whether abnormal electricity exists) and surrounding environment data of the hazardous chemical enterprise in half a year in a certain period of time, such as the last half year, can be collected from a marketing system and a metering system in advance. Meanwhile, electricity consumption wave dynamic factors, such as holiday information, production conditions affected by hot events and the like, are collected within half a year. Then, based on historical electricity consumption data, historical environment data and electricity consumption dynamic factors of the dangerous enterprise, training a classification prediction model, capturing and learning correlations among electricity consumption behaviors, environment data, electricity consumption dynamic data and electricity consumption data, obtaining a trained enterprise electricity consumption detection model, accurately judging whether the dangerous enterprise has abnormal electricity consumption behaviors according to the input electricity consumption data, environment data and electricity consumption dynamic factors, and outputting an enterprise electricity consumption detection result.

After the electricity characteristic data and the environment characteristic data are obtained, the electricity characteristic data, the environment characteristic data and the target electricity consumption dynamic factor are taken as input, a trained enterprise electricity detection model is called to detect electricity consumption abnormality, the model captures the association between electricity consumption behaviors and the environment data, the electricity consumption dynamic data and the electricity consumption data, whether dangerous enterprises have abnormal electricity consumption behaviors is judged, and an enterprise electricity consumption detection result is output. The enterprise electricity consumption detection result comprises binary identification (abnormal/normal) of electricity consumption abnormality or data such as probability value, abnormality type, abnormality cause and the like of the electricity consumption abnormality.

According to the enterprise electricity consumption detection method, the relation between the electricity consumption behavior and the environmental factors is captured by training the enterprise electricity consumption detection model through the historical electricity consumption data, the historical environmental data and the historical electricity consumption factor in advance, the accuracy and generalization capability of the model on electricity consumption abnormality detection tasks can be improved, under the actual electricity consumption detection scene, the current electricity consumption behavior and the environmental condition can be reflected more accurately by acquiring the real-time electricity consumption data, the real-time environmental data and the electricity consumption factor of a target enterprise, the timeliness and the accuracy of the electricity consumption abnormality detection are improved, the obtained target electricity consumption data, the target environmental data and the target electricity consumption factor are more standard by preprocessing the real-time electricity consumption data, the target environmental data and the electricity consumption factor, the effect of subsequent feature extraction and model detection is improved, the trained enterprise electricity consumption detection is called by taking the real-time electricity consumption feature data, the environment feature data and the target electricity consumption factor as input, whether the electricity consumption behavior of the different enterprises can be judged efficiently, and the electricity consumption detection result of the enterprise can be improved. Further, the scheme can be well suitable for electricity detection of dangerous enterprises, so that the electricity safety management level of the dangerous enterprises is improved.

As shown in FIG. 3, in one embodiment, the enterprise electricity usage detection model, when invoked, performs the following steps:

s820, counting the change rule of the power load data by a window counting method to obtain the power load change data.

And S840, analyzing the change rule of the electricity consumption and the electricity consumption time based on the existing historical electricity consumption data to obtain electricity consumption change data.

And S860, carrying out power consumption abnormality detection on power consumption load change data, power consumption change data and environmental characteristic data according to a preset abnormal power consumption judgment rule, and obtaining an enterprise power consumption detection result.

In practical application, the electricity consumption characteristic data can reflect whether the dangerous enterprise has abnormal electricity consumption behaviors or not most intuitively, so that the electricity consumption characteristic data can be mainly considered in electricity consumption abnormality detection. In this embodiment, the electricity consumption characteristic data includes electricity consumption load data, electricity consumption time, and electricity consumption amount. The preset abnormal electricity consumption judgment rule may be a preset set of rules for detecting abnormal electricity consumption, such as setting a threshold value of load fluctuation, a threshold value of electricity consumption change, and a common electricity consumption peak period and electricity consumption valley period.

After the electricity characteristic data is extracted, the electricity load data can be counted according to a certain time window, for example, the time window can be in dimensions of every minute, every hour and the like. And obtaining the change data of the electric load by calculating indexes such as average, maximum or minimum value and the like of the electric load in each time window. The historical electricity consumption and the real-time electricity consumption are compared, the electricity consumption time is compared, for example, the electricity consumption peak period or the electricity consumption valley period is analyzed, and the data are analyzed and processed to obtain the change data of the electricity consumption and the electricity consumption time. Specifically, statistical methods, time series analysis or other related techniques may be employed to obtain a change rule such as a trend, periodicity or seasonality of the electricity consumption.

After the electricity load change data, the electricity consumption change data and the electricity consumption time change data are obtained, the electricity consumption abnormality detection is carried out on the electricity consumption load change data, the electricity consumption change data and the environmental characteristic data according to a preset abnormal electricity consumption judgment rule, so that an enterprise electricity consumption detection result is obtained.

Specifically, the useful electric load change threshold may be set based on statistical information or expert knowledge of the history data. And setting a power consumption change threshold according to analysis and rules of the historical data. The environmental characteristic change threshold value, for example, the degree of abnormal fluctuation of temperature, humidity, etc., is set in consideration of the influence of the environmental characteristic on the electricity load and the electricity consumption amount. Then, based on the threshold value of the power consumption load change, the power consumption amount change threshold value, and the environmental characteristic change threshold value, a plurality of abnormal power consumption judgment rules are set, for example, the power consumption is determined to be abnormal when the load fluctuation exceeds a certain percentage or absolute value, for example, the power consumption is determined to be abnormal when the power consumption exceeds or falls below a certain percentage or absolute value. Specifically, the real-time load data of the dangerous chemical enterprises can be analyzed, if the load suddenly increases, the problem of electric leakage of electric facilities (electric energy meters, lines and the like) can be caused, and if the load suddenly decreases, the problem of fault shutdown of production equipment can be caused. And analyzing the electricity consumption and the electricity consumption time in a near period of time, comparing the electricity consumption with the history synchronization, and if the electricity consumption is greatly increased, possibly judging whether the enterprise has illegal production or not, and the like. And through the set abnormal electricity utilization detection rule, the model is analyzed and predicted, and finally a detection result of whether the electricity utilization of the dangerous enterprise is abnormal is output.

In other embodiments, the distribution characteristics of the electricity load and the electricity consumption, such as the mean, variance, kurtosis, skewness and other statistical indexes, may be analyzed by a statistical method, and the comparison with the historical data is used to determine whether an abnormal situation exists. Or modeling and analyzing the time series data of the electricity load and the electricity consumption, and detecting the trend, periodicity or seasonality of the abnormality. Alternatively, outliers in the power load and the power consumption may be detected by applying an outlier detection algorithm, such as an isolated Forest (Isolation Forest), a local outlier factor (Local Outlier Factor), or the like. It can be understood that the method is specifically used comprehensively according to actual requirements and data characteristics, and a proper mode or a combination mode is selected to detect the power utilization abnormality so as to realize more accurate and reliable results.

In this embodiment, abnormal electricity consumption judgment rules are used to detect electricity consumption load change data, electricity consumption change data and environmental characteristic data, so that electricity consumption abnormality judgment can be quickly and accurately performed, and rules support expansion and modification, so that comprehensive and real-time electricity consumption safety monitoring can be realized.

As shown in FIG. 4, in one embodiment, before invoking the trained enterprise electricity detection model to perform electricity safety detection, with the electricity characteristic data as input, the method further comprises:

S700, acquiring historical electricity consumption data, historical environment data and historical electricity consumption factors of a target enterprise in a historical time period.

S720, carrying out data preprocessing on the historical electricity consumption data, the historical environment data and the historical electricity consumption electric wave factor to obtain target historical electricity consumption data, target historical environment data and target historical electricity consumption electric wave factor.

S740, extracting historical electricity utilization characteristic data in the target historical electricity utilization data and historical environment characteristic data in the target historical environment data.

S760, a model training set is constructed based on the historical electricity utilization characteristic data, the historical environment characteristic data and the target historical electricity utilization factor.

S780, training an initial enterprise electricity detection model based on the model training set through a gradient descent method, and determining model parameters of the initial enterprise electricity detection model to obtain a trained enterprise electricity detection model.

In this embodiment, a training process of the enterprise electricity detection model is described. The method mainly adopts a mathematical modeling mode, and calculates whether the electricity consumption of the dangerous enterprise is safe or not through a model algorithm. Firstly, basic information of a hazardous chemical enterprise is collected, wherein the basic information comprises data such as an electricity consumption address, installed capacity, a transformer, a line and the like of the hazardous chemical enterprise, and historical electricity consumption data, historical environment data and electricity consumption dynamic data of a historical time period of the hazardous chemical enterprise are also collected. Likewise, the historical electricity usage data includes data of the electricity consumption, electricity load, voltage, power, and electricity usage time of the target enterprise over a historical period of time, such as half a year. The historical environmental data includes temperature, humidity, wind speed, pressure, air quality, and light intensity over a historical period of time (e.g., half a year). The electric wave factor includes but is not limited to holiday information, hot spot event, etc. In this embodiment, the initial enterprise electricity detection model is constructed based on a logistic regression analysis algorithm.

In specific implementation, taking a target enterprise as a dangerous enterprise as an example, an initial enterprise electricity detection model can be constructed based on a logistic regression algorithm such as a K nearest neighbor method. After the historical electricity data, the historical environment data and the historical electricity consumption moving factors of the dangerous chemical enterprises are collected in the historical time period, the historical electricity data, the historical environment data and the historical electricity consumption moving factors can be subjected to data preprocessing, including cleaning data: and processing missing values, abnormal values, repeated data and the like, and ensuring the integrity and the accuracy of the data. And then, screening out characteristic variables related to electricity safety analysis according to domain knowledge and data analysis. For example, electricity usage characteristic data such as average load, peak-to-valley difference, load fluctuation, and the like are extracted from the target historical electricity usage data. Environmental characteristic data, such as average temperature, humidity change, weather conditions, etc., are extracted from the target historical environmental data. And then, carrying out normalization or normalization processing on the characteristic data to ensure that the data have the same scale. After the above processing is completed, the target historical electricity consumption data, the target historical environment data and the target historical electricity consumption factor are obtained.

Then, a model training set is constructed based on the historical electricity utilization characteristic data, the historical environment characteristic data and the target historical electricity utilization factor. And inputting the data in the model training set into the initial enterprise electricity detection model in batches, and continuously adjusting model parameters of the initial enterprise electricity detection model through a gradient descent algorithm so that the model can more accurately predict electricity fluctuation conditions until the iteration times are reached, stopping training, and obtaining a trained enterprise electricity detection model.

In the embodiment, an initial enterprise electricity safety analysis model is built through a logistic regression analysis algorithm, and optimization and adjustment of the model are performed according to historical electricity characteristic data, historical environment characteristic data and target historical electricity dynamic factors, so that accuracy and reliability of the model are improved.

As shown in fig. 5, in one embodiment, after S780, the method further includes:

s790, performing cross verification on the trained enterprise electricity detection model, and calculating the model performance index after each cross verification to obtain a plurality of model performance indexes.

S792, obtaining a model performance evaluation result according to the model performance indexes.

After the enterprise electricity analysis detection model is obtained through training, the model performance can be further verified and evaluated. Specifically, a data set for cross-validation may be constructed based on the historical electricity usage characteristic data, the historical environmental characteristic data, and the target historical electricity usage factor, ensuring that the data set has sufficient sample coverage and representativeness. The data set is then divided into K subsets, where K is the cross-validated fold. Several values commonly used, such as 5, 10, etc., can be chosen to ensure that each subset is similar in size and that the samples are evenly distributed. And then sequentially selecting one subset as a verification set and the rest as training sets to perform multiple training and verification. For each cross-validation, the model is evaluated using the validation set and a series of performance metrics such as accuracy, precision, recall, F1 score, etc. are calculated. These indicators can reflect the behavior of the model on different data sets. For each cross-validation, a set of model performance metrics is obtained. Each index is recorded to form a plurality of model performance indices for subsequent evaluation and analysis. And then comprehensively considering the performance indexes of the multiple models, and comprehensively evaluating the trained enterprise electricity utilization detection model. Weighted averages, composite scores, or other methods may be employed to take into account the importance of the different indicators to obtain the final model performance assessment results. Specifically, the average performance indexes such as average accuracy, average recall, average F1 value and the like can be selected and calculated, and the weight adjustment can be performed on each index to obtain the comprehensive performance evaluation result by combining the service requirement and the domain knowledge.

In this embodiment, by cross-verifying the trained enterprise electricity detection model and evaluating the performance results thereof, the performance of the model on different data subsets can be better understood, and reliable performance evaluation results can be provided to guide the subsequent model improvement and optimization work.

As shown in fig. 6, in one embodiment, S400 includes: s420, performing at least one of data cleaning, missing value filling, outlier processing and normalization processing on the real-time electricity consumption data, the real-time environment data and the electricity consumption dynamic factors to obtain target electricity consumption data, target environment data and target electricity consumption dynamic factors.

In practical application, the acquired real-time electricity data, real-time environment data and electricity consumption factors have non-uniform data formats and include more worthless data, so that data preprocessing is required. For example, at least one of data cleansing, missing value filling, outlier processing, and normalization processing may be performed on the data.

Specifically, the data cleansing may be: removing invalid data: whether invalid or erroneous records exist in the data, such as duplicate data, data in a wrong format, etc., are checked and deleted or corrected. Noise removal data: by observing and analyzing the data, noise data which may exist is identified and cleaned or repaired by a suitable method. The missing value padding may be: whether missing values exist in the data set is analyzed, and the missing values can be detected by observing the missing mode of the data or using a statistical tool.

For missing values, interpolation methods (e.g., linear interpolation, polynomial interpolation, etc.) or statistical-based methods (e.g., mean, median, mode fill, etc.) may be used to fill in to preserve the integrity of the data. Regarding outlier handling may be: using statistical analysis or data visualization methods, outliers in the data are identified, which may be due to data acquisition errors, instrument failure, or other abnormal conditions. For detected outliers, appropriate processing strategies may be employed, such as deleting outliers, replacing with missing values, smoothing, etc., to ensure that the model is not affected by outliers during training and prediction. Regarding the normalization process: and selecting a proper normalization method according to the characteristics and service requirements of the data. And carrying out normalization processing on the real-time electricity utilization data and the real-time environment data, and converting the real-time electricity utilization data and the real-time environment data into the same scale range so as to eliminate the dimensional differences among different features or dimensions. Common normalization methods include maximum-minimum normalization, normalization (Z-score) normalization, and the like. For preprocessing with electric wave dynamic factors, holiday information can be converted into binary or numerical representation, so that a model can be better understood and processed. For example, holidays are defined as 1, non-holidays are defined as 0, or different holidays are assigned different numerical codes. The hotspot event may be converted to a binary variable, indicating whether there is a hotspot event impact. For example, 1 when there is a hot spot event, and 0 when there is no hot spot event. And selecting a proper normalization method according to the data characteristics and the service requirements.

In this embodiment, at least one mode of data cleaning, missing value filling, abnormal value processing and normalization processing is performed on the data, so that the normalization and quality of the data can be effectively improved, and subsequent abnormal electricity detection is facilitated.

As shown in fig. 6, in one embodiment, after obtaining the power consumption detection result of the enterprise, the method further includes: s900, pushing the early warning message under the condition that the enterprise electricity utilization detection result represents that abnormal electricity utilization exists in the target enterprise.

In practical application, if the enterprise electricity detection result represents that the target enterprise has abnormal electricity behavior, the abnormal electricity situation can be organized into clear and understandable early warning messages according to the enterprise electricity detection result, and the early warning messages are sent to related personnel through proper communication channels (such as short messages, mails, application program (APP), pushing and the like), including dangerous enterprise responsible personnel, energy management personnel and the like. To inform the relevant responsible person to determine whether the problem of abnormal electricity consumption is caused by human operation. If not, the professional needs to go to check whether the problem is caused by the power facility fault, and inform fire protection and related departments of the related conditions for early warning, and intervene in the enterprise in advance to do guarantee work, so that the possibility of safety accidents is reduced. If the enterprise electricity utilization detection result represents that the dangerous enterprise does not have abnormal electricity utilization behaviors, recording the detection data, and continuing the subsequent abnormal electricity utilization detection.

In this embodiment, when the enterprise electricity detection result represents that the target enterprise has abnormal electricity consumption, the early warning message is pushed, so that potential safety hazards can be found in time, and the possibility of occurrence of safety accidents is reduced.

In order to make a clearer description of the method for detecting power consumption of an enterprise provided by the present application, a specific embodiment is described below with reference to fig. 6, where the specific embodiment includes the following steps:

S420, performing at least one of data cleaning, missing value filling, outlier processing and normalization processing on the real-time electricity consumption data, the real-time environment data and the electricity consumption dynamic factors to obtain target electricity consumption data, target environment data and target electricity consumption dynamic factors.

S900, pushing the early warning message under the condition that the enterprise electricity utilization detection result represents that abnormal electricity utilization exists in the target enterprise.

It should be understood that, although the steps in the flowcharts related to the embodiments described above are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.

Based on the same inventive concept, the embodiment of the application also provides an enterprise electricity detection device for realizing the enterprise electricity detection method. The implementation of the solution provided by the device is similar to the implementation described in the above method, so the specific limitation in the embodiments of the enterprise electricity detection device provided below may be referred to the limitation of the enterprise electricity detection method hereinabove, and will not be repeated herein.

In one embodiment, as shown in FIG. 7, there is provided an enterprise electricity usage detection apparatus 700 comprising: a data acquisition module 710, a data preprocessing module 720, a data extraction module 730, and a data detection module 740, wherein:

the data acquisition module 710 is configured to acquire real-time electricity consumption data, real-time environment data, and electricity consumption factors of a target enterprise.

The data preprocessing module 720 is configured to perform data preprocessing on the real-time power consumption data, the real-time environment data and the power consumption factor, so as to obtain target power consumption data and target environment data.

The data extraction module 730 is configured to extract electricity usage characteristic data from the target electricity usage data, and extract environmental characteristic data from the target environmental data.

The data detection module 740 is configured to call the trained power consumption detection model for power consumption anomaly detection to obtain a power consumption detection result of the enterprise, with the power consumption feature data, the environmental feature data and the target power consumption dynamic factor as input.

the data detection module 740 is further configured to calculate a change rule of the electricity load data by using a window statistical method, obtain electricity load change data, analyze a change rule of electricity consumption and electricity consumption time based on existing historical electricity consumption data, obtain electricity consumption change data, and perform electricity consumption anomaly detection on the electricity load change data, the electricity consumption change data and environmental feature data according to a preset anomaly electricity consumption judgment rule, so as to obtain an enterprise electricity consumption detection result.

As shown in fig. 8, in one embodiment, the apparatus further includes a model training module 702 configured to obtain historical electricity data, historical environment data, and historical electricity consumption factors of the target enterprise in a historical time period, perform data preprocessing on the historical electricity data, the historical environment data, and the historical electricity consumption factors to obtain target historical electricity data, target historical environment data, and target historical electricity consumption factors, extract historical electricity feature data in the target historical electricity data, and historical environment feature data in the target historical environment data, construct a model training set based on the historical electricity feature data, the historical environment feature data, and the target historical electricity consumption factors, train an initial enterprise electricity detection model based on the model training set by a gradient descent method, determine model parameters of the initial enterprise electricity detection model, and obtain a trained enterprise electricity detection model, where the initial enterprise electricity detection model is constructed based on a logistic regression analysis algorithm.

In one embodiment, the apparatus further includes a model verification module 704, configured to cross-verify the trained power utilization detection model of the enterprise, calculate a model performance index after each cross-verification, obtain a plurality of model performance indexes, and obtain a model performance evaluation result according to the plurality of model performance indexes.

In one embodiment, the data preprocessing module 720 is further configured to perform at least one of data cleaning, missing value filling, outlier processing, and normalization processing on the real-time electricity data, the real-time environment data, and the electricity consumption factor to obtain the target electricity data, the target environment data, and the target electricity consumption factor.

As shown in fig. 8, in one embodiment, the apparatus further includes a data early warning module 750, configured to push an early warning message when the enterprise electricity detection result indicates that the target enterprise has abnormal electricity consumption behavior.

The modules in the enterprise electricity detection apparatus may be implemented in whole or in part by software, hardware, or a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 9. The computer device includes a processor, a memory, an Input/Output interface (I/O) and a communication interface. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer equipment is used for storing real-time electricity utilization data, real-time environment data, electricity utilization factor and other data. The input/output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used for communicating with an external terminal through a network connection. The computer program, when executed by a processor, implements an enterprise power usage detection method.

It will be appreciated by those skilled in the art that the structure shown in fig. 9 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the computer device to which the present application applies, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.

In one embodiment, a computer device is provided, including a memory and a processor, where the memory stores a computer program, and the processor implements the steps of the embodiments of the enterprise electricity usage detection method described above when the computer program is executed.

In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, which when executed by a processor, implements the steps of the enterprise electricity usage detection method embodiments described above.

In one embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the enterprise electricity usage detection method embodiments described above.

It should be noted that, the user information (including but not limited to user equipment information, user basic information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) referred to in the present application are information and data authorized by the user or sufficiently authorized by each party, and the collection, use, and processing of the related data need to comply with the related laws and regulations and standards of the related country and region.

Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the various embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the various embodiments provided herein may include at least one of relational databases and non-relational databases. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic units, quantum computing-based data processing logic units, etc., without being limited thereto.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The above examples only represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the present application. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application shall be subject to the appended claims.

Claims

1. An enterprise electricity usage detection method, the method comprising:

carrying out data preprocessing on the real-time electricity utilization data, the real-time environment data and the electricity utilization factor to obtain target electricity utilization data, target environment data and target electricity utilization factor;

the electricity utilization characteristic data, the environment characteristic data and the target electricity utilization factor are used as input, a trained enterprise electricity utilization detection model is called to conduct electricity utilization abnormality detection, and an enterprise electricity utilization detection result is obtained;

the enterprise electricity consumption detection model is obtained by training a classification prediction model based on historical electricity consumption data, historical environment data and historical electricity consumption dynamic factors of a target enterprise in a historical time period.

2. The method of claim 1, wherein the electricity usage characteristic data includes electricity usage load data, electricity usage amount, and electricity usage time;

counting the change rule of the electricity load data by a window counting method to obtain electricity load change data;

based on the existing historical electricity consumption data, analyzing the electricity consumption and the change rule of the electricity consumption time to obtain electricity consumption change data;

and carrying out power consumption abnormality detection on the power consumption load change data, the power consumption quantity change data and the environment characteristic data according to a preset abnormal power consumption judgment rule to obtain an enterprise power consumption detection result.

3. The method of claim 1, wherein the step of invoking a trained enterprise electricity usage detection model for electricity usage security detection is preceded by the step of taking the electricity usage feature data as input, the method further comprising:

acquiring historical electricity consumption data, historical environment data and historical electricity consumption motioning factors of the target enterprise in a historical time period;

performing data preprocessing on the historical electricity consumption data, the historical environment data and the historical electricity consumption dynamic factors to obtain target historical electricity consumption data, target historical environment data and target historical electricity consumption dynamic factors;

constructing a model training set based on the historical electricity utilization characteristic data, the historical environment characteristic data and the target historical electricity utilization electric wave dynamic factors;

the initial enterprise electricity utilization detection model is constructed based on a logistic regression analysis algorithm.

4. The method of claim 3, wherein after the obtaining the trained business electrical inspection model, the method further comprises:

performing cross verification on the trained enterprise power utilization detection model, and calculating the model performance index after each cross verification to obtain a plurality of model performance indexes;

5. The method according to any one of claims 1 to 4, wherein the performing data preprocessing on the real-time electricity data to obtain target electricity data includes:

and performing at least one of data cleaning, missing value filling, abnormal value processing and normalization processing on the real-time power consumption data.

6. The method according to any one of claims 1 to 4, wherein after the obtaining the power consumption detection result of the enterprise, the method further comprises:

and pushing an early warning message under the condition that the enterprise electricity utilization detection result represents that the target enterprise has abnormal electricity utilization behavior.

7. An enterprise electricity usage detection apparatus, the apparatus comprising:

the data acquisition module is used for acquiring real-time electricity utilization data, real-time environment data and electricity utilization factor of an enterprise;

The data preprocessing module is used for preprocessing the real-time electricity utilization data and the real-time environment data to obtain target electricity utilization data and target environment data;

the data detection module is used for calling a trained enterprise electricity detection model to detect electricity utilization abnormality by taking the electricity utilization characteristic data, the environment characteristic data and the target electricity utilization dynamic factor as inputs to obtain an enterprise electricity utilization detection result;

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.

9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.