CN117034197A - Enterprise power consumption typical mode analysis method based on multidimensional Isolate-detection multi-point detection - Google Patents

Abstract

This invention proposes a method for analyzing typical patterns of enterprise electricity consumption based on multi-dimensional Isolate-Detect multi-change point detection. First, the data is cleaned based on the collected industrial electricity consumption time series data; then the cleaned data is decomposed into seasons and eliminated. Based on the periodic pattern of the data, the residual terms of the data are extracted; multi-dimensional change point detection is performed on the residual terms obtained after decomposition based on multi-dimensional Isolate-Detect, and only a single change point is detected in each divided time interval, so that for The change point detection in the multi-dimensional power consumption curve is more accurate and the calculation efficiency is higher; combined with the mean test statistic and the maximum test statistic, a possible change point set is obtained; multi-sequence fusion is used for the detected change point set Screen the set of possible change points to determine the final change point position of the electricity consumption data; finally, divide the standardized time series data based on the obtained change point positions to extract the typical electricity consumption behavior of electricity consumption.

Description

Analysis of typical patterns of enterprise electricity consumption based on multi-dimensional Isolate-Detect multi-change point detection analysis method

Technical field

The invention relates to the technical field of power data analysis, in particular to a method for analyzing typical patterns of enterprise power consumption based on multi-dimensional Isolate-Detect multi-change point detection.

Background technique

With the continuous improvement of power grid informatization and the rapid development of digitalization and intelligence, the power industry has entered the era of power big data. As an energy system on which economic development and human life depend, the power system generates a large amount of data with rapid growth and rich types during operation. Analyzing electricity consumption behavior can discover patterns, relationships, trends, etc. in electricity big data to understand the characteristics of electricity consumption. Therefore, detecting common change points and extracting typical electricity consumption behaviors in various industries can better analyze and understand electricity consumption, understand electricity consumption characteristics and electricity consumption behaviors, and provide strong support for power supply and management.

However, industrial electricity consumption data shows shortcomings such as large electricity consumption and complex electricity consumption patterns in various industries. At the same time, the data also has periodic and seasonal characteristics, which brings challenges to the analysis of electricity consumption behavior. Today's power data is characterized by high dimensions, multiple types and large volumes, which places higher requirements on data analysis technology.

Contents of the invention

Therefore, in view of the gaps and deficiencies in the existing technology, the present invention proposes a typical enterprise power consumption pattern analysis method based on multi-dimensional Isolate-Detect multi-change point detection, which is used to improve the traditional power consumption behavior analysis model that does not consider timing and To address the shortcomings of phased changes, a multi-dimensional Isolate-Detect model is established based on industrial power consumption time series data to detect common change points in the industry and extract typical power consumption behaviors. Based on historical electricity consumption data, extract one day on weekdays and one day on weekends to detect the difference in electricity consumption behavior between weekdays and weekends; use the multi-dimensional Isolated-Detect algorithm for change point detection, which improves the accuracy of change point detection; finally, based on Based on the detected change point position, the typical power consumption behavior is extracted from the power consumption time series data in segments.

Its main steps include data collection, data cleaning, residual term extraction, change point detection, change point screening, and extraction of typical electricity consumption behaviors. First, the data is cleaned based on the collected industrial electricity consumption time series data; then the cleaned data is decomposed into seasons, the periodic patterns of the data are eliminated, and the residual terms of the data are extracted; based on the multi-dimensional Isolate-Detect pair decomposition, we get The residual term is used for multi-dimensional change point detection, and only a single change point is detected in each divided time interval, making it more accurate and more computationally efficient for change point detection in multi-dimensional power consumption curves. Combining the mean test statistic and the maximum test statistic to obtain the possible change point set; use multi-sequence fusion to filter the possible change point set for the detected change point set, and determine the final change point position of the electricity consumption data; Finally, the standardized time series data is divided based on the obtained change point positions, and the typical electricity consumption behavior of electricity consumption is extracted. The multi-dimensional Isolate-Detect method proposed by the present invention takes into account the sequence correlation between multi-dimensional data, can effectively detect the common change points of multi-dimensional data, and improves the accuracy of multi-dimensional change point detection. In this way, a more realistic and objective time-series power consumption scenario can be obtained, and a more reasonable power consumption strategy can be formulated in subsequent planning.

It specifically adopts the following technical solutions:

A method for analyzing typical patterns of enterprise electricity consumption based on multi-dimensional Isolate-Detect multi-change point detection, which is characterized by: firstly performing data cleaning based on the collected industrial electricity consumption time series data; and then performing seasonal decomposition on the cleaned data. Eliminate the periodic patterns of the data and extract the residual terms of the data; perform multi-dimensional change point detection on the residual terms obtained after decomposition based on multi-dimensional Isolate-Detect, and only detect a single change point in each divided time interval, so that The detection of change points in the multi-dimensional power consumption curve is more accurate and the calculation efficiency is higher; combined with the mean test statistic and the maximum test statistic, a possible change point set is obtained; a multi-sequence method is used for the detected change point set Fusion screens the set of possible change points to determine the final change point position of the electricity consumption data; finally, the standardized time series data is divided based on the obtained change point positions, and the typical electricity consumption behavior of electricity consumption is extracted.

Further, include the following steps:

Step 1: Collect electricity consumption time series data of various industries;

Step 2: Fill in the missing values of the electricity consumption time series data of each industry, extract the electricity consumption data of a certain day on weekdays or a certain day of weekends, use the 3σ principle to eliminate outliers, and then perform standardization processing;

Step 3: Smooth the cleaned data first, then perform seasonal decomposition on the smoothed data, eliminate the periodicity and trend of the data, and obtain the residual term of the smoothed data;

Step 4: Calculate the CUSUM statistic at each time point of the residual term of the electricity consumption data of each industry, and then calculate the mean and maximum value of the CUSUM statistic at each time point, which are recorded as M test statistic and T test statistic respectively. ;

Step 5: Use the Isolate-Detect algorithm to detect common change points in the electricity consumption data of various industries, preset the threshold according to the threshold formula of the Isolate-Detect algorithm, and determine the time point when the M test statistic exceeds the threshold as the change point. The obtained change point Put it into the change point set and record it as the change point set CP_M; determine the time point when the T test statistic exceeds the threshold as the change point, put the obtained change point into the change point set and record it as the change point set CP_T;

Step 6: Record the change point set obtained through data analysis as the change point set CP. Take out the two elements whose absolute difference is less than 3 in the set CP_M and the set CP_T, and select the smaller value to put into the set CP, and get The final set of change points;

Step 7: Segment the standardized electricity consumption time series data based on the obtained change point position, average the electricity consumption of each industry for each period of time series data, and obtain the typical electricity consumption behavior of each period of time series data.

Further, in step 2, after filling in the missing values in the electricity consumption data of each industry, the electricity consumption data of a certain day in the working day and the electricity consumption data of a certain day in the weekend are respectively extracted for subsequent data analysis. Analyze differences in industrial electricity consumption behavior; use the 3σ principle to eliminate outliers to avoid the impact of outliers on subsequent analysis of electricity consumption behavior.

Furthermore, in step 3, the Savitzky-Golay filter is used to smooth the cleaned data. When the loss function obtains the minimum value, the fitting effect of the original data reaches the optimal value; the smoothed original data is obtained through the sliding window. fitting value to effectively reduce the noise of the data.

Further, in step 3, for the smoothed data, an additive model is used for trend decomposition, and the smoothed electricity consumption data is decomposed into the trend part, the periodic part and the residual part to eliminate the periodicity and trend of the data. Sex, expressed as follows:

X _i,t =T _i,t +S _i,t +C _i,t +I _i,t , (7)

Among them, T _i represents the long-term time trend of the i-th industry, S _i represents the seasonal time trend of the i-th industry, C _i represents the cyclic time trend of the i-th industry, and I _i represents the remaining time of the i-th industry. residual term.

Further, in step 4, use the data residual term obtained after decomposition to calculate the mean CUSUM statistic at each time point. The formula is as follows:

Among them, i refers to the i-th industry, s refers to the starting point of the detection interval, e refers to the end point of the detection interval, b refers to the time point of detection, n refers to the total length of the detection interval, I _{i ,t} refers to the smoothed data residual term of the i-th industry at time t, Refers to the CUSUM statistic value of the i-th industry at time point b within the detection interval [s, e], p refers to the total number of data dimensions,/> Refers to the mean CUSUM statistic at time point b within the detection interval [s, e], recorded as

Further, in step 4, use the data residual term obtained after decomposition to calculate the maximum value of the CUSUM statistic at each time point. The formula is as follows:

Among them, i refers to the i-th industry, s refers to the starting point of the detection interval, e refers to the end point of the detection interval, b refers to the time point of detection, n refers to the total length of the detection interval, I _{i ,t} refers to the smoothed data residual term of the i-th industry at time t, Refers to the CUSUM statistic value of the i-th industry at time point b within the detection interval [s, e], p refers to the total number of data dimensions,/> It refers to the maximum value of the CUSUM statistic at time point b within the detection interval [s, e].

Further, in step 5, based on the calculated M test statistic and T test statistic, the Isolate-Detect algorithm is used to detect the common change points of the electricity consumption of each industry:

First create the interval to be detected. For a data sequence of length T, first set a positive constant λ _T , and then create two ordered left and right extension intervals of K = [T/λ _T ]; the jth right extension The interval is R _j =[1,min{jλ _T ,T}], and the jth left extended interval is L _j =[max{1,T-jλ _T +1},T]; in the ordered set S _RL = Collect these intervals in {R ₁ , L ₁ , R ₂ , L ₂ ,..., R _K , L _K }; then identify the point in R ₁ with the largest test statistic value;

Based on the obtained test statistic, the threshold is set, where the threshold is calculated as follows:

In the formula, σ is the standard deviation of the input data, and C is the given parameter value; compare the threshold and the test statistic to determine whether the time point is a change point: if the value of the test statistic exceeds the threshold, the corresponding point is regarded as is the change point; if it does not exceed the threshold, continue to detect the next interval of S _RL .

Further, in step 7, based on the detected change point position, the standardized time series data is divided into segments, the average power consumption of each industry in each segment is calculated, and the typical power consumption scenario of each segment of the time series data is extracted. The calculation formula for the average electricity consumption of the i-th industry is as follows:

In the formula, τ _k refers to the position of the kth change point, It refers to the normalized electricity consumption value of the i-th industry at the k+1th change point position.

Compared with the prior art, the beneficial effects of the present invention and its preferred solutions include at least:

1. Taking into account the difference in electricity consumption behavior between working days and weekends, extract the data of a certain day on the working day and the data of a certain day on the weekend for subsequent data analysis, distinguishing the different electricity consumption characteristics of working days and weekends;

2. The electricity consumption time series data were trend decomposed. Considering that the period may make the model anomaly detection effect unstable, a residual sequence without trend items and period items was used for modeling, which improved the accuracy of subsequent modeling;

3. By calculating the CUSUM mean statistic, the impact of all dimensions on change point detection is considered;

4. To calculate the maximum CUSUM statistic, only the sequence with the largest CUSUM statistic is used, which is not easily affected by outliers;

5. Compared with the traditional change point detection method, the ID method only detects a single change point at a time in an isolated interval, and the calculation efficiency is higher;

6. Considering that the M test statistic is more susceptible to outliers and the T test statistic is not robust, choosing a common change point detected by two test statistics improves the accuracy of change point detection.

9. The standardized time series data is segmented based on the final change point position, and the electricity consumption data of each industry in each segment of time series data is averaged, so that the typical electricity consumption behavior of each segment can be accurately and efficiently extracted.

Description of the drawings

The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments:

Figure 1 is a flow chart of a typical pattern analysis method for enterprise power consumption based on multi-dimensional Isolate-Detect multi-change point detection according to an embodiment of the present invention;

Figure 2 is an original data diagram after filling missing values according to the embodiment of the present invention;

Figure 3 is a time series diagram of electricity consumption data on Monday according to the embodiment of the present invention;

Figure 4 is a time series diagram of Saturday electricity consumption data according to the embodiment of the present invention;

Figure 5 is a time sequence diagram of electricity consumption data of various industries on Monday after data cleaning according to the embodiment of the present invention;

Figure 6 is a time sequence diagram of Saturday electricity consumption data of various industries after data cleaning according to the embodiment of the present invention;

Figure 7 is a smoothing diagram of Monday electricity consumption data of various industries using the Savitzky-Golay filter according to the embodiment of the present invention;

Figure 8 is a smoothing diagram of Monday electricity consumption data of various industries using the Savitzky-Golay filter according to the embodiment of the present invention;

Figure 9 is a residual sequence diagram of electricity consumption data for each industry on Monday after using the additive model according to the embodiment of the present invention;

Figure 10 is a residual sequence diagram of Saturday electricity consumption data of various industries after using the additive model according to the embodiment of the present invention;

Figure 11 is a time series diagram of the M test statistic and the T test statistic of Monday's electricity consumption data according to the embodiment of the present invention;

Figure 12 is a time series diagram of the M test statistic and the T test statistic of the Saturday electricity consumption data according to the embodiment of the present invention;

Figure 13 is the final change point distribution diagram of Monday's electricity consumption detected using the multi-dimensional Isolate-Detect algorithm according to the embodiment of the present invention;

Figure 14 is the final change point distribution diagram of Saturday's electricity consumption detected using the multi-dimensional Isolate-Detect algorithm according to the embodiment of the present invention;

Figure 15 is a segmented mean curve chart of Monday's electricity consumption data according to the embodiment of the present invention;

Figure 16 is a segmented mean curve chart of Saturday electricity consumption data according to the embodiment of the present invention.

Detailed ways

In order to make the features and advantages of this patent more obvious and easy to understand, the following examples are given in detail along with the accompanying drawings:

This embodiment uses electricity consumption data sets from various industries in Fuzhou City, Fujian Province to clearly and completely describe the technical solutions in the embodiments of this application. Figure 1 shows the detailed process of the present invention. Specific application examples for implementing this solution are provided below:

Step 1: Collect electricity consumption time series data for various industries in the industry. The data includes Gregorian calendar date, industry type, and electricity consumption.

Step 2: Fill in missing values for the power consumption time series data. Mark the data with empty and 0 power consumption as missing values. Among them, the data on electricity consumption of electronic machinery and wood processing is missing on July 6, 2020. The electricity consumption data of electronic machinery and wood processing on every Monday from June to September 2020 are averaged and used to interpolate the missing data respectively. value. The power consumption timing diagram is obtained, as shown in Figure 2. Monday’s data and Saturday’s data were extracted from the electricity consumption time series data, as shown in Figure 3 and Figure 4. Then the 3σ principle is used to eliminate outliers, that is, in each dimension of data, data points that deviate from the average value by more than 3 times the standard deviation are identified as abnormal data, and the detected abnormal data are eliminated. Due to the large differences in electricity consumption between various industries, in order to avoid the impact of the large order of magnitude difference in the electricity consumption time series data of each industry on change point detection, the electricity consumption of each industry is standardized separately. The formula is as follows:

Among them, x _i,t represents the electricity consumption of the i-th industry at the t-th time point, μ _i represents the average electricity consumption of the i-th industry, σ _i represents the standard deviation of electricity consumption of the i-th industry, X _{i , t} represents the standardized electricity consumption. After data cleaning, the Monday electricity consumption time series diagram and Saturday power consumption time series diagram of each industry are shown in Figure 5 and Figure 6;

Step 3: First use the Savitzky-Golay filter to smooth the cleaned data. The Savitzky-Golay filter is a filtering method based on local polynomial least squares fitting in the time domain. Assume that the original _time series data is _Xi ={X _{i,1 ,} X _i,2 ,...X _i,t ...,X _i,T }, with With m sample points on the left and right, construct a window array containing 2m+1 sample points, and then construct a p-order polynomial to fit the data in the window. The p-order polynomial is as follows:

Among them -m≤n≤m, p≤2m+1; define the loss function as follows:

When the loss function reaches the minimum value, the fitting effect of the original data is optimal. Obtaining the smoothed fitting value of the original data through the sliding window can effectively reduce the noise of the data. The trend of the smoothed electricity consumption data on Monday is shown in Figure 7, and the trend of the smoothed electricity consumption data on Saturday is shown in Figure 8;

Then the additive model is used to decompose the trend of the smoothed data. The model is expressed as follows:

Among them, T _i represents the long-term time trend of the i-th industry, S _i represents the seasonal time trend of the i-th industry, C _i represents the cyclic time trend of the i-th industry, and I _i represents the remaining time of the i-th industry. residual term. This invention decomposes the smoothed electricity consumption data into the trend part, the period part and the residual part, and obtains the residual sequence of Monday's electricity consumption data and the residual sequence of Saturday's electricity consumption data respectively, as shown in Figure 9 and Figure 10 shown.

Step 4: Calculate the CUSUM statistics of the residual term of electricity consumption in each industry at each time point. The main formula is as follows:

In the formula, i refers to the i-th industry, s refers to the starting point of the detection interval, e refers to the end point of the detection interval, b refers to the time point of detection, n refers to the total length of the detection interval, I _i,t refers to the residual term of the smoothed data of the i-th industry at time t, It refers to the CUSUM statistic value of the i-th industry at time point b within the detection interval [s, e].

Then calculate the mean and maximum CUSUM statistics of each industry residual term at each time point, that is, the M test statistic and the T test statistic. The formula is as follows:

In the formula, p refers to the total number of data dimensions, in this embodiment p=9, Refers to the mean CUSUM statistic at time point b within the detection interval [s, e]. /> It refers to the maximum value of the CUSUM statistic at time point b within the detection interval [s, e]. The obtained M test statistic and T test statistic time series diagram of Monday's electricity consumption data are shown in Figure 11, and the M test statistic and T test statistic time series diagram of Saturday's electricity consumption data are shown in Figure 12.

Step 5: Based on the M test statistic and T test statistic, use the ID method to detect the change point, and output the common change point position of the multi-dimensional electricity consumption data. The ID method mainly screens change points through the threshold method. Its principle is as follows:

First create the interval to be detected. For a data sequence of length T, first set a positive constant λ _T , and then create two ordered left and right extended intervals of K = [T/λ _T ]. The jth right expansion interval is R _j =[1,min{jλ _T ,T}], and the jth left expansion interval is L _j =[max{1,T-jλ _T +1},T]. These intervals are collected in an ordered set S _RL = {R ₁ , L ₁ , R ₂ , L ₂ ,..., R _K , L _K }. ID then identifies the point in _R1 where the test statistic has the largest value. Based on the obtained test statistic, the threshold is set, where the threshold is calculated as follows:

In the formula, ζ _T is the obtained threshold, σ is the standard deviation of the input data, C is the given parameter value, and T is the length of the input data sequence. In this embodiment, select _Saturday =111.

If the value of the test statistic exceeds the threshold, the point is considered a change point. If the threshold is not exceeded, continue to detect the next interval of S _RL . After detection, the ID algorithm starts a new round of detection from the end point (starting point) of the right (or left) extension interval where the detection occurred.

Step 6: Use two test statistics to obtain two change point sets CP_MM and CP_MT of Monday's electricity consumption data. Take out the two elements whose absolute difference is less than 3 in the set CP_MM and the set CP_MT, and select the smaller value. Put it into the set CP_M to get the final change point set. The final change point detection result chart of Monday's electricity consumption data is shown in Figure 13. A total of 11 change points were detected. The same detection method was adopted for Saturday's electricity consumption data, and the final change point set CP_S of Saturday's electricity consumption data was obtained. A total of 10 change points were detected. The final change point detection result is shown in Figure 14.

Step 7: Based on the detected change point position, segment the standardized time series data into segments, average the electricity consumption of each industry in each segment, and extract the typical power consumption scenario of each segment of the time series data. The calculation formula for the average electricity consumption of the i-th industry is as follows:

In the formula, τ _k refers to the position of the kth change point, It refers to the normalized electricity consumption value of the i-th industry at the k+1th change point position. The segmented mean curve of Monday's electricity consumption data is shown in Figure 15, and the segmented mean curve of Saturday's electricity consumption data is shown in Figure 16. At the same time, the average value of each industry in the time series sub-scenario is marked in the figure.

Those skilled in the art will understand that embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment that combines software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each process and/or block in the flowchart illustrations and/or block diagrams, and combinations of processes and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce a use A device for realizing the functions specified in one process or multiple processes of the flowchart and/or one block or multiple blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory that causes a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction means, the instructions The device implements the functions specified in a process or processes of the flowchart and/or a block or blocks of the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device, causing a series of operating steps to be performed on the computer or other programmable device to produce computer-implemented processing, thereby executing on the computer or other programmable device. Instructions provide steps for implementing the functions specified in a process or processes of a flowchart diagram and/or a block or blocks of a block diagram.

The above are only preferred embodiments of the present invention, and are not intended to limit the present invention in other forms. Any skilled person familiar with the art may make changes or modifications to equivalent changes using the technical contents disclosed above. Example. However, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the protection scope of the technical solution of the present invention.

This patent is not limited to the above-mentioned best embodiments. Under the inspiration of this patent, anyone can derive various other forms of analysis methods for typical patterns of enterprise electricity consumption based on multi-dimensional Isolate-Detect multi-change point detection. Anyone based on the present invention Equal changes and modifications made to the scope of the patent application shall be within the scope of this patent.

Claims (9)

1. A method for analyzing typical patterns of enterprise electricity consumption based on multi-dimensional Isolate-Detect multi-change point detection, which is characterized by: first performing data cleaning based on the collected industrial electricity consumption time series data; and then performing seasonal analysis on the cleaned data. Decompose, eliminate the periodic patterns of the data, and extract the residual terms of the data; perform multi-dimensional change point detection on the residual terms obtained after decomposition based on multi-dimensional Isolate-Detect, and only detect a single change point in each divided time interval , which makes the detection of change points in multi-dimensional power consumption curves more accurate and computationally efficient; combined with the mean test statistic and the maximum test statistic, a possible change point set is obtained; for the detected change point set, Multi-sequence fusion screens the set of possible change points to determine the final change point position of the electricity consumption data; finally, the standardized time series data is divided based on the obtained change point positions to extract the typical electricity consumption Behavior. 2. The enterprise power consumption typical pattern analysis method based on multi-dimensional Isolate-Detect multi-change point detection according to claim 1, characterized by: Includes the following steps: Step 1: Collect electricity consumption time series data of various industries; Step 2: Fill in the missing values of the electricity consumption time series data of each industry, extract the electricity consumption data of a certain day on weekdays or a certain day of weekends, use the 3σ principle to eliminate outliers, and then perform standardization processing; Step 3: Smooth the cleaned data first, then perform seasonal decomposition on the smoothed data, eliminate the periodicity and trend of the data, and obtain the residual term of the smoothed data; Step 4: Calculate the CUSUM statistic at each time point of the residual term of the electricity consumption data of each industry, and then calculate the mean and maximum value of the CUSUM statistic at each time point, which are recorded as M test statistic and T test statistic respectively. ; Step 5: Use the Isolate-Detect algorithm to detect common change points in the electricity consumption data of various industries, preset the threshold according to the threshold formula of the Isolate-Detect algorithm, and determine the time point when the M test statistic exceeds the threshold as the change point. The obtained change point Put it into the change point set and record it as the change point set CP_M; determine the time point when the T test statistic exceeds the threshold as the change point, put the obtained change point into the change point set and record it as the change point set CP_T; Step 6: Record the change point set obtained through data analysis as the change point set CP. Take out the two elements whose absolute difference is less than 3 in the set CP_M and the set CP_T, and select the smaller value to put into the set CP, and get The final set of change points; Step 7: Segment the standardized electricity consumption time series data based on the obtained change point position, average the electricity consumption of each industry for each period of time series data, and obtain the typical electricity consumption behavior of each period of time series data. 3. The method for analyzing typical patterns of enterprise electricity consumption based on multi-dimensional Isolate-Detect multi-change point detection according to claim 2, characterized in that: in step 2, after filling in the missing values in the electricity consumption data of each industry, the work is extracted respectively. Conduct subsequent data analysis on the electricity consumption data of a certain day in the middle of the day and the electricity consumption data of a certain day on the weekend, and analyze the differences in industrial electricity consumption behavior between working days and weekends; use the 3σ principle to eliminate outliers to avoid outliers affecting subsequent electricity consumption. Implications for behavioral analysis. 4. The enterprise power consumption typical pattern analysis method based on multi-dimensional Isolate-Detect multi-change point detection according to claim 3, characterized in that: in step 3, the cleaned data is processed using a Savitzky-Golay filter. Smoothing, when the loss function reaches the minimum value, the fitting effect of the original data is optimal; the smoothed fitting value of the original data is obtained through the sliding window to effectively reduce the noise of the data. 5. The enterprise power consumption typical pattern analysis method based on multi-dimensional Isolate-Detect multi-change point detection according to claim 4, characterized by: In step 3, for the smoothed data, the additive model is used for trend decomposition, and the smoothed electricity consumption data is decomposed into the trend part, the periodic part and the residual part to eliminate the periodicity and trend of the data, expressing as follows: X _i,t =T _i,t +S _i,t +C _i,t +I _i,t , (1) Among them, T _i represents the long-term time trend of the i-th industry, S _i represents the seasonal time trend of the i-th industry, C _i represents the cyclic time trend of the i-th industry, and I _i represents the remaining time of the i-th industry. residual term. 6. The enterprise power consumption typical pattern analysis method based on multi-dimensional Isolate-Detect multi-change point detection according to claim 5, characterized by: In step 4, use the data residual term obtained after decomposition to calculate the mean CUSUM statistic at each time point. The formula is as follows: Among them, i refers to the i-th industry, s refers to the starting point of the detection interval, e refers to the end point of the detection interval, b refers to the time point of detection, n refers to the total length of the detection interval, I _{i ,t} refers to the smoothed data residual term of the i-th industry at time t, Refers to the CUSUM statistic value of the i-th industry at time point b within the detection interval [s, e], p refers to the total number of data dimensions,/> Refers to the mean CUSUM statistic at time point b within the detection interval [s, e], recorded as/> 7. The enterprise power consumption typical pattern analysis method based on multi-dimensional Isolate-Detect multi-change point detection according to claim 6, characterized by: In step 4, use the data residual term obtained after decomposition to calculate the maximum value of the CUSUM statistic at each time point. The formula is as follows: Among them, i refers to the i-th industry, s refers to the starting point of the detection interval, e refers to the end point of the detection interval, b refers to the time point of detection, n refers to the total length of the detection interval, I _{i ,t} refers to the smoothed data residual term of the i-th industry at time t, Refers to the CUSUM statistic value of the i-th industry at time point b within the detection interval [s, e], p refers to the total number of data dimensions,/> It refers to the maximum value of the CUSUM statistic at time point b within the detection interval [s, e]. 8. The enterprise power consumption typical pattern analysis method based on multi-dimensional Isolate-Detect multi-change point detection according to claim 7, characterized by: In step 5, based on the calculated M test statistic and T test statistic, the Isolate-Detect algorithm is used to detect the common change points in the electricity consumption of each industry: First create the interval to be detected. For a data sequence of length T, first set a positive constant λ _T , and then create two ordered left and right extension intervals of K = [T/λ _T ]; the jth right extension The interval is R _j =[1,min{jλ _T ,T}], and the jth left extended interval is L _j =[max{1,T-jλ _T +1},T]; in the ordered set S _RL = Collect these intervals in {R ₁ , L ₁ , R ₂ , L ₂ ,..., R _K , L _K }; then identify the point in R ₁ with the largest test statistic value; Based on the obtained test statistic, the threshold is set, where the threshold is calculated as follows: In the formula, σ is the standard deviation of the input data, and C is the given parameter value; compare the threshold and the test statistic to determine whether the time point is a change point: if the value of the test statistic exceeds the threshold, the corresponding point is regarded as is the change point; if it does not exceed the threshold, continue to detect the next interval of S _RL . 9. The enterprise power consumption typical pattern analysis method based on multi-dimensional Isolate-Detect multi-change point detection according to claim 8, characterized by: In step 7, based on the detected change point position, the standardized time series data is divided into segments, the average power consumption of each industry in each segment is calculated, and the typical power consumption scenario of each segment of time series data is extracted, where the i-th The calculation formula for the average electricity consumption of each industry is as follows: In the formula, τ _k refers to the position of the kth change point, It refers to the normalized electricity consumption value of the i-th industry at the k+1th change point position.