A system for collecting and analyzing residents' electricity consumption behavior based on autoencoders

Abstract

The present invention provides a residential electricity consumption behavior collection and analysis system based on an autoencoder, which relates to the technical field of residential electricity consumption analysis, including collecting relevant data and analyzing and processing the data, and outputting a confidence level of abnormal electricity consumption. The present invention collects the number of unplanned openings of meter cabinets of residential users in the previous year and electricity parameters, and analyzes them from three aspects: the abnormal opening rate of the meter cabinets, the decreasing trend of user electricity use, and the abnormal amount of user electricity loss. The collected data is processed for correlation, and three intermediate quantities related to the abnormal electricity consumption of residential users, namely, an annual trend index of electricity consumption, an evaluation index of electricity loss, and an unplanned opening rate of meter cabinets, are output. A comprehensive abnormal electricity consumption index for residential users that reflects the degree of abnormal electricity consumption of residential users is generated, and the comprehensive abnormal electricity consumption index for residential users is compared with an abnormal threshold value, and the confidence level of abnormal electricity consumption for residents is output, so as to more accurately judge whether the residential electricity consumption is abnormal or not.

Description

Technical Field

[0001] The present invention relates to the technical field of residential electricity consumption analysis, and in particular to a residential electricity consumption behavior collection and analysis system based on an autoencoder. Background Art

[0002] Residential electricity analysis refers to the process of studying and evaluating the electricity consumption patterns of households or residential areas. This analysis is usually aimed at understanding electricity usage habits, identifying energy saving potential, optimizing energy use, and reducing electricity bills. As the coverage of the power network expands, obvious abnormal power consumption behavior will appear at individual power nodes, which is also the main reason for the emergence of abnormal electricity usage data. Residential electricity analysis detection aims to identify and detect illegal or unauthorized electricity use behavior. As an important device for the collection and transmission of electricity usage data, the working principle of the autoencoder is based on its unique network structure, which consists of two main parts: encoder and decoder. The encoder compresses the input data into a smaller representation (called latent space or encoding), while the decoder reconstructs this compressed representation back to the original data. The autoencoder can learn a compact representation of the data for data dimensionality reduction, which is very useful when processing high-dimensional data.

[0003] Existing electricity consumption behavior detection collects and analyzes instantaneous changes in residents' electricity consumption, and it is difficult to analyze and confirm abnormal residents' electricity consumption that is relatively stable over a long period of time.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0005] The purpose of the present invention is to provide a system for collecting and analyzing residents' electricity consumption behavior based on an autoencoder to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A residential electricity consumption behavior collection and analysis system based on an autoencoder, comprising:

[0008] A meter cabinet switch module, the meter cabinet switch module is used to collect the number of unplanned opening times of meter cabinets of m residential users on the selected line in the previous year;

[0009] A switch frequency analysis module, which is used to analyze the number of unplanned openings of the meter cabinets of residential users, and generate an unplanned opening rate of the meter cabinets of residential users on the selected line in the previous year, and the unplanned opening rate of the meter cabinets is used to reflect the abnormal opening rate of the meter cabinets of residential users;

[0010] An electricity loss module, the electricity loss module is used to collect electricity parameters of m residential users on the selected line in the previous year;

[0011] An electric power loss analysis module, which is used to perform correlation analysis on electric power parameters, generate an electric power loss expression, and perform correlation analysis on the electric power loss expression to generate the average amount of electric power loss of residential users in the previous two days and the average amount of electric power loss in the next x days, and generate an electric power loss evaluation index, which is used to reflect the abnormal rate of electric power loss of residential users in the previous year;

[0012] An electricity usage analysis module, which is used to perform correlation analysis on the electricity parameters collected by the electricity loss module, generate an average amount of electricity usage, and generate an electricity usage trend. By preprocessing the electricity usage trend, an annual electricity trend index is generated. The annual electricity trend index is used to reflect the annual electricity usage trend decreasing index;

[0013] A comprehensive analysis module, which is used to perform correlation analysis on the annual trend index of electricity consumption, the evaluation index of electricity loss and the unplanned opening rate of the metering cabinet, and generate a comprehensive abnormal electricity consumption index for residential users;

[0014] The comparison module compares the residential user's comprehensive power consumption anomaly index with the anomaly threshold, and outputs the residential user's anomaly confidence level, where the anomaly confidence level is used to reflect the degree of the residential user's power consumption anomaly.

[0015] Furthermore, the meter cabinet switch module uses door and window switch sensors to record the number of unplanned openings of the meter cabinet of each resident user.

[0016] Furthermore, the number of unplanned openings of the meter cabinet of the pth residential user is expressed as C p It means the number of unplanned openings of the meter cabinet of the pth residential user C p Perform correlation analysis to generate the unplanned opening rate FKP of the meter cabinet of the pth residential user p , the formula based on is:

[0017]

[0018] Among them, γ is the planned opening weight factor, and the unplanned opening rate FKP of the metering cabinet is p Used to reflect the abnormal opening rate of residential users' metering cabinets.

[0019] Furthermore, the power parameters include the power supply to the pth residential user on the jth day of the previous year on the selected line. And the electricity consumption of the pth residential user on the jth day y(j) p, perform correlation analysis on the power parameters and generate the power loss expression L(j) of the pth residential user on the jth day, as follows:

[0020]

[0021] The power loss expression L(j) is used to reflect the power loss rate of the pth residential user. The power loss expression L(j) of the jth day is analyzed to generate the average power loss of the tth day and the previous two days. and the average power consumption in the next x days The formula is:

[0022]

[0023] The average power consumption of the tth day and the previous two days and the average power consumption in the next x days Perform correlation analysis to generate the power loss index DSZ on day t t , the formula based on is:

[0024]

[0025] Among them, the value of x is 7, ε is the power loss rate ratio factor, and the power loss index DSZ t A scalar that reflects the abnormal power consumption of the p-th user every day in the previous year.

[0026] Furthermore, for the power loss index DSZ on day t, t Perform correlation analysis to generate the power loss evaluation index DSP of the pth user of the selected line, based on the formula:

[0027]

[0028] Wherein, β is the power loss index factor, and the power loss evaluation index DSP is used to reflect the abnormal power loss rate of the p-th user in the previous year.

[0029] Furthermore, for the electricity consumption y(i) of the pth residential user on the i-th day p And the power usage of the x days before and after the i-th day is analyzed to generate the average power usage The formula is:

[0030]

[0031] Among them, y(d) p represents the electricity usage of the pth residential user on the dth day as the independent variable, and the average electricity usage Used to reflect the average power usage of 2x+1 days.

[0032] Furthermore, the average amount of electricity used and the electricity usage of the pth residential user on the i-th day as the independent variable y(i) p Perform correlation analysis to generate the power usage trend DQ of the pth resident on the i-th day i , the formula based on is:

[0033]

[0034] DQ for power usage trend i Perform correlation analysis to generate the power consumption decreasing trend indicator DSQ i , the formula based on is:

[0035]

[0036] Use the trend indicator DSQ for decreasing power i Correlation analysis is performed to generate the annual trend indicator DYZ for electricity consumption, based on the following formula:

[0037]

[0038] Among them, the power usage trend DQ i Used to reflect the power usage trend ratio of the i-th day, the power decreasing usage trend indicator DSQ i It is used to reflect whether the electricity usage on the i-th day is decreasing compared with the previous day. The annual trend index of electricity usage DYZ is used to reflect the decreasing index of the annual electricity usage trend, and α is the decreasing weight factor of electricity usage.

[0039] Furthermore, the annual trend index of electricity consumption DYZ, the electricity loss evaluation index DSP and the unplanned opening rate of the meter cabinet FKP are analyzed. p Correlation analysis is performed to generate the comprehensive abnormal power consumption index ZYP for residential users, based on the following formula:

[0040]

[0041] The comprehensive abnormal power consumption index of residential users is used to reflect the abnormal power consumption ratio of the p-th residential user.

[0042] Furthermore, the abnormal threshold is θ. When ZYP ≥ θ, the abnormal confidence level of the output residential user is level 3, and the abnormal confidence is 90%; when When , the abnormal confidence level of the output residential users is level 2, and the abnormal confidence is 65%; when When , the abnormal confidence level of the output residential user is level 1, and the abnormal confidence is 20%.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] The present invention collects the number of unplanned opening times and electricity parameters of the meter cabinets of residential users in the previous year, and analyzes them from three aspects: the abnormal opening rate of the meter cabinets, the decreasing trend of users' electricity usage, and the abnormal amount of users' electricity loss. The collected data is correlated and processed to output three intermediate quantities related to the abnormal electricity consumption of residential users, namely, the annual trend index of electricity consumption, the electricity loss evaluation index and the unplanned opening rate of the meter cabinet. The three intermediate quantities are subjected to correlation analysis to generate a comprehensive abnormal electricity consumption index for residential users that reflects the degree of abnormal electricity consumption of residential users, and the comprehensive abnormal electricity consumption index for residential users is compared with the abnormal threshold value to output the confidence level of abnormal electricity consumption of residents, so as to more accurately judge whether the residential electricity consumption is abnormal or not. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the overall module flow of the present invention. DETAILED DESCRIPTION

[0046] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0047] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0048] Example:

[0049] See also Figure 1 , the present invention provides a technical solution:

[0050] A residential electricity consumption behavior collection and analysis system based on an autoencoder. The data collected by the present invention is compressed and decoded by an autoencoder. The working principle of the autoencoder is based on its unique network structure, which includes two main parts: an encoder and a decoder. The encoder compresses the input data into a smaller representation (called a latent space or encoding), and the decoder reconstructs this compressed representation back to the original data. The autoencoder can learn a compact representation of the data for data dimensionality reduction, which is very useful when processing high-dimensional data. The present invention collects the electricity consumption behavior of residential users in a digitized manner, and performs relevant processing on these data, thereby outputting an evaluation of the abnormal confidence of the residential users, and classifying and judging the electricity consumption behavior of the residential users, including:

[0051] A meter cabinet switch module, the meter cabinet switch module is used to collect the number of unplanned opening times of meter cabinets of m residential users on the selected line in the previous year;

[0052] The metering cabinet is a closed structural cabinet used to install and fix power metering equipment. Its main function is to provide a safe, reliable and generally compliant environment with relevant electrical specifications to accommodate different types of metering equipment, such as electric meters, protective fuses, circuit breakers, etc.

[0053] The meter cabinet switch module adopts door and window switch sensors to record the number of unplanned openings of the meter cabinet of each resident user, the opening of the meter cabinet within the grid plan, and the inspection, maintenance and debugging of the equipment in the meter cabinet by the staff. These are all planned openings of the meter cabinet. The unplanned openings of the meter cabinet are the abnormal number of openings of the meter cabinet that are not planned by the grid. The number of openings of the meter cabinet can be recorded by the door and window switch sensors. Before the abnormal electricity consumption of residents, the facilities of these meter cabinets are often debugged. Therefore, the unplanned openings of the meter cabinet are related to the abnormal electricity consumption of residents.

[0054] A switch frequency analysis module, which is used to analyze the number of unplanned openings of the meter cabinets of residential users, and generate an unplanned opening rate of the meter cabinets of residential users on the selected line in the previous year, and the unplanned opening rate of the meter cabinets is used to reflect the abnormal opening rate of the meter cabinets of residential users;

[0055] The number of unplanned openings of the metering cabinet of the pth residential user is C p It means the number of unplanned openings of the meter cabinet of the pth residential user C p Perform correlation analysis to generate the unplanned opening rate FKP of the meter cabinet of the pth residential user p , the formula based on is:

[0056]

[0057] Among them, γ is the planned opening weight factor, and the value range of γ is [1,15]. The unplanned opening rate FKP of the metering cabinet is p Used to reflect the abnormal opening rate of the metering cabinet of residential users, FKP p The higher the value, the higher the probability of abnormal electricity consumption among residents.

[0058] An electricity loss module, the electricity loss module is used to collect electricity parameters of m residential users on the selected line in the previous year;

[0059] The power parameters include the power supply to the pth residential user on the jth day of the previous year for the selected line. And the electricity consumption of the pth residential user on the jth day y(j) p , perform correlation analysis on the power parameters and generate the power loss expression L(j) of the pth residential user on the jth day, as follows:

[0060]

[0061] The power loss expression L(j) is used to reflect the power loss rate of the pth residential user. The power loss expression L(j) of the jth day is analyzed to generate the average power loss of the tth day and the previous two days. and the average power consumption in the next x days The formula is:

[0062]

[0063] The value of X is 7, which is a cycle of working and rest days.

[0064] The average power consumption of the tth day and the previous two days and the average power consumption in the next x days Perform correlation analysis to generate the power loss index DSZ on day t t , the formula based on is:

[0065]

[0066] Among them, the value of x is 7, ε is the power loss rate ratio factor, and the value range of ε is The power loss indicator DSZ t It is used to reflect the abnormal power consumption of the p-th user every day in the previous year, where: It means that the power loss exceeds ε. The days exceeding the limit are counted as 1, and the days not exceeding the limit are counted as 0, which is convenient for counting the days exceeding the limit.

[0067] An electric power loss analysis module, which is used to perform correlation analysis on electric power parameters, generate an electric power loss expression, and perform correlation analysis on the electric power loss expression to generate the average amount of electric power loss of residential users in the previous two days and the average amount of electric power loss in the next x days, and generate an electric power loss evaluation index, which is used to reflect the abnormal rate of electric power loss of residential users in the previous year;

[0068] If the electricity consumption of residential users is abnormal, the proportion of the user's electricity loss on that day will expand and show an upward trend. Considering that the electricity consumption of residential users will fluctuate on the same day, the error will be too large, so the electricity loss in the previous and next few days should be considered, and then it is determined whether the growth rate of electricity loss is greater than the prescribed ratio.

[0069] Specifically, for the power loss index DSZ on day t t Perform correlation analysis to generate the power loss evaluation index DSP of the pth user of the selected line, based on the formula:

[0070]

[0071] Among them, β is the power loss index factor, and the value range of β is [126,365]. t The electricity calculation for the previous two days and the next seven days is already included, so the sum is counted from 3 to 358. For a leap year, the data for the last day of the year is discarded. The number of days with abnormal electricity loss in the previous year and the electricity loss index factor are comprehensively analyzed to output the electricity loss evaluation index DSP. The electricity loss evaluation index DSP is used to reflect the abnormal electricity loss rate of the p-th user in the previous year. The electricity loss evaluation index DSP is used to evaluate the abnormal electricity loss rate of residential users in the previous year. The electricity loss evaluation index DSP is positively correlated with the abnormal electricity loss rate in the previous year.

[0072] An electricity usage analysis module, which is used to perform correlation analysis on the electricity parameters collected by the electricity loss module, generate an average amount of electricity usage, and generate an electricity usage trend. By preprocessing the electricity usage trend, an annual electricity trend index is generated. The annual electricity trend index is used to reflect the annual electricity usage trend decreasing index;

[0073] Specifically, the electricity consumption y(i) of the pth residential user on the i-th day is p And the power usage of the x days before and after the i-th day is analyzed to generate the average power usage The formula is:

[0074]

[0075] Among them, y(d) prepresents the electricity usage of the pth residential user on the dth day as the independent variable, and the average electricity usage Used to reflect the average power usage of 2x+1 days.

[0076] A comprehensive analysis module, which is used to perform correlation analysis on the annual trend index of electricity consumption, the evaluation index of electricity loss and the unplanned opening rate of the metering cabinet, and generate a comprehensive abnormal electricity consumption index for residential users;

[0077] Average power usage and the electricity usage of the pth residential user on the i-th day as the independent variable y(i) p Perform correlation analysis to generate the power usage trend DQ of the pth resident on the i-th day i , the formula based on is:

[0078]

[0079] DQ for power usage trend i Perform correlation analysis to generate the power consumption decreasing trend indicator DSQ i , the formula based on is:

[0080]

[0081] Among them, the power consumption decreasing trend indicator DSQ i Days with a value of 1 indicate that the electricity usage trend on those days is decreasing. If there are more days with a value of 1, it means that the user's electricity usage trend is gradually decreasing throughout the year, and the user is considered to be classified as an abnormal electricity user.

[0082] Use the trend indicator DSQ for decreasing power i Correlation analysis is performed to generate the annual trend indicator DYZ for electricity consumption, based on the following formula:

[0083]

[0084] Among them, the power usage trend DQ i Used to reflect the power usage trend ratio of the i-th day, the power decreasing usage trend indicator DSQ i It is used to reflect whether the electricity usage on the i-th day is decreasing compared with the previous day. The annual trend index of electricity usage DYZ is used to reflect the decreasing index of the annual electricity usage trend. Since the calculation is performed on the seven days before and after, i starts counting from 8 to 358. α is the decreasing weight factor of electricity usage, and the value range of α is [150,365].

[0085] The annual trend index of electricity consumption DYZ, the electricity loss evaluation index DSP and the unplanned opening rate of the meter cabinet FKP pCorrelation analysis is performed to generate the comprehensive abnormal power consumption index ZYP for residential users, based on the following formula:

[0086]

[0087] Among them, the residential users' comprehensive power consumption abnormality index ZYP is used to reflect the abnormal power consumption ratio of residential users. If the residential users' power consumption is abnormal, the proportion of the user's power loss on that day will expand and show an upward trend. Considering that the residential users' power consumption on that day will fluctuate, the error will be too large, so the power loss in the previous and next few days is considered, and then it is judged whether the power loss growth rate is greater than the specified ratio. Therefore, the power loss evaluation index DSP is positively correlated with the residential users' comprehensive power consumption abnormality index ZYP.

[0088] The comparison module compares the residential user's comprehensive power consumption anomaly index with the anomaly threshold, and outputs the residential user's anomaly confidence level, where the anomaly confidence level is used to reflect the degree of the residential user's power consumption anomaly.

[0089] The abnormal threshold is θ, and the value of θ is 1.69. When ZYP ≥ θ, the abnormal confidence level of the output residential users is level 3, and the abnormal confidence is 90%; when When , the abnormal confidence level of the output residential users is level 2, and the abnormal confidence is 65%; when When the abnormal confidence level of residential users is output as level one, the abnormal confidence level is 20%. The higher the abnormal confidence level, the higher the probability of abnormal electricity consumption of residential users. For residential users with an abnormal confidence level of 90%, on-site investigation and testing can be carried out to determine whether the electricity consumption is normal.

[0090] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0091] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0092] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0093] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be covered by the protection scope of the present application.

Claims

1. A residential electricity consumption behavior collection and analysis system based on an autoencoder, characterized in that: include: A meter cabinet switch module, the meter cabinet switch module is used to collect the number of unplanned openings of the meter cabinets of m residential users on the selected line in the previous year; the meter cabinet switch module uses door and window switch sensors to record the number of unplanned openings of the meter cabinet of each residential user; A switch frequency analysis module, which is used to analyze the number of unplanned openings of the meter cabinets of residential users, and generate an unplanned opening rate of the meter cabinets of residential users on the selected line in the previous year, and the unplanned opening rate of the meter cabinets is used to reflect the abnormal opening rate of the meter cabinets of residential users; An electricity loss module, the electricity loss module is used to collect electricity parameters of m residential users on the selected line in the previous year; An electric power loss analysis module, which is used to perform correlation analysis on electric power parameters, generate an electric power loss expression, and perform correlation analysis on the electric power loss expression to generate the average amount of electric power loss of residential users in the previous two days and the average amount of electric power loss in the next x days, and generate an electric power loss evaluation index, which is used to reflect the abnormal rate of electric power loss of residential users in the previous year; An electricity usage analysis module, which is used to perform correlation analysis on the electricity parameters collected by the electricity loss module, generate an average amount of electricity usage, and generate an electricity usage trend. By preprocessing the electricity usage trend, an annual electricity trend index is generated. The annual electricity trend index is used to reflect the annual electricity usage trend decreasing index; A comprehensive analysis module, which is used to perform correlation analysis on the annual trend index of electricity consumption, the evaluation index of electricity loss and the unplanned opening rate of the metering cabinet, and generate a comprehensive abnormal electricity consumption index for residential users; The comparison module compares the residential user's comprehensive power consumption anomaly index with the anomaly threshold, and outputs the residential user's anomaly confidence level, where the anomaly confidence level is used to reflect the degree of the residential user's power consumption anomaly.

2. According to claim 1, a residential electricity consumption behavior collection and analysis system based on an autoencoder is characterized by: The number of unplanned openings of the metering cabinet of the pth residential user is C p It means the number of unplanned openings of the meter cabinet of the pth residential user C p Perform correlation analysis to generate the unplanned opening rate FKP of the meter cabinet of the pth residential user p , the formula based on is: Among them, γ is the planned opening weight factor, and the unplanned opening rate FKP of the metering cabinet is p Used to reflect the abnormal opening rate of residential users' metering cabinets.

3. According to claim 1, a residential electricity consumption behavior collection and analysis system based on an autoencoder is characterized by: The power parameters include the power supply to the pth residential user on the jth day of the previous year for the selected line. And the electricity consumption of the pth residential user on the jth day y(j) p , perform correlation analysis on the power parameters and generate the power loss expression L( / ) of the pth residential user on the jth day, as follows: The power loss expression L(j) is used to reflect the power loss rate of the pth residential user. The power loss expression L(j) of the jth day is analyzed to generate the average power loss of the tth day and the previous two days. and the average power consumption in the next x days The formula is: The average power consumption of the tth day and the previous two days and the average power consumption in the next x days Perform correlation analysis to generate the power loss index DSZ on day t t , the formula based on is: Among them, the value of x is 7, ε is the power loss rate ratio factor, and the power loss index DSZ t A scalar that reflects the abnormal power consumption of the p-th user every day in the previous year.

4. According to claim 3, a residential electricity consumption behavior collection and analysis system based on an autoencoder is characterized in that: For the power loss index DSZ on day t t Perform correlation analysis to generate the power loss evaluation index DSP of the pth user of the selected line, based on the formula: Wherein, β is the power loss index factor, and the power loss evaluation index DSP is used to reflect the abnormal power loss rate of the p-th user in the previous year.

5. According to claim 4, a residential electricity consumption behavior collection and analysis system based on an autoencoder is characterized in that: The electricity consumption of the pth residential user on the i-th day is y(i) p And the power usage of the x days before and after the i-th day is analyzed to generate the average power usage The formula is: Where y(d)p represents the power usage of the pth residential user on the dth day as the independent variable, and the average power usage is Used to reflect the average power usage of 2x+1 days.

6. The system for collecting and analyzing residents' electricity consumption behavior based on an autoencoder according to claim 5 is characterized in that: Average power usage The correlation analysis is performed with the power usage y(i)p of the pth resident user with the i-th day as the independent variable to generate the power usage trend DQ of the pth resident on the i-th day i ; DQ for power usage trend i Perform correlation analysis to generate the power consumption decreasing trend indicator DSQ i , the formula based on is: Use the trend indicator DSQ for decreasing power i Correlation analysis is performed to generate the annual trend indicator DYZ for electricity consumption, based on the following formula: Among them, the power usage trend DQ i Used to reflect the power usage trend ratio of the i-th day, the power decreasing usage trend indicator DSQ i It is used to reflect whether the electricity usage on the i-th day is decreasing compared with the previous day. The annual trend index of electricity usage DYZ is used to reflect the decreasing index of the annual electricity usage trend, and α is the decreasing weight factor of electricity usage.

7. The system for collecting and analyzing residents' electricity consumption behavior based on an autoencoder according to claim 6 is characterized in that: The annual trend index of electricity consumption DYZ, the electricity loss evaluation index DSP and the unplanned opening rate of the meter cabinet FKP p Correlation analysis is performed to generate the comprehensive abnormal power consumption index ZYP for residential users, based on the following formula: The comprehensive abnormal power consumption index of residential users is used to reflect the abnormal power consumption ratio of the p-th residential user.

8. The system for collecting and analyzing residents' electricity consumption behavior based on an autoencoder according to claim 7 is characterized in that: The abnormal threshold is θ. When ZYP ≥ θ, the abnormal confidence level of the output residential users is level 3, and the abnormal confidence is 90%. When , the abnormal confidence level of the output residential users is level 2, and the abnormal confidence is 65%; when When , the abnormal confidence level of the output residential user is level 1, and the abnormal confidence is 20%.

Images