WO2020019403A1

WO2020019403A1 - Electricity consumption abnormality detection method, apparatus and device, and readable storage medium

Info

Publication number: WO2020019403A1
Application number: PCT/CN2018/103220
Authority: WO
Inventors: 郑立颖; 徐亮; 肖京
Original assignee: 平安科技（深圳）有限公司
Priority date: 2018-07-26
Filing date: 2018-08-30
Publication date: 2020-01-30
Also published as: CN109726731A; CN109726731B

Abstract

An electricity consumption abnormality detection method, an electricity consumption abnormality detection apparatus, an electricity consumption abnormality detection device, and a computer-readable storage medium. The electricity consumption abnormality detection method comprises: collecting, according to a preset time interval, a sequence feature value at each target time point from multiple electricity consumption time sequences, so as to generate a feature data sequence corresponding to each of the electricity consumption time sequences (S10); computing abnormality scores of all the feature data sequences according to an isolation forest algorithm (S20); if the abnormality scores are greater than an abnormality warning value, confirming the feature data sequences as normal sequences (S30); and if the abnormality scores are less than or equal to the abnormality warning value, confirming the feature data sequences as abnormal sequences (S40). The accuracy of electricity consumption abnormality detection is improved, the interference of an abnormal value is eliminated, computational complexity is reduced, and detection efficiency is improved.

Description

Method, device, device and readable storage medium for detecting abnormal power consumption Ranch

This application claims the priority of a Chinese patent application filed on July 26, 2018 with the Chinese Patent Office, application number 201810841413.0, and the invention name "Abnormal power consumption detection method, device, device, and readable storage medium". The contents are incorporated in the application by reference.

Technical field

The present application relates to the technical field of abnormality detection, and in particular, to a method for detecting abnormality in power consumption, an abnormality detecting device, an abnormality detecting device, and a computer-readable storage medium.

Background technique

At present, most of the anomaly detection of power consumption time series are from the perspective of power consumption time series (such as year-on-year difference, ring-to-moment difference, moving average, ARIMA prediction, etc.) and distance difference angle (such as comparing sequences within a time window with other sequences Euclidean distance), from the perspective of electricity consumption time series, the disadvantages are: 1. The point of consideration of the year-on-year difference is too one-sided, which is likely to cause misjudgment; 2. The method based on moving average will be greatly affected by outliers, causing a sequence after the outliers to be judged as abnormal; the disadvantage of starting from the perspective of distance difference is that the time series of electricity consumption generally has a wide span and relatively large data. Large, distance calculation is more complicated, time-consuming and inefficient. Therefore, the traditional power consumption anomaly detection method cannot achieve large-scale data detection, and the detection accuracy is not high, it is easy to be disturbed by outliers, the calculation is complicated, and the detection efficiency is low.

Summary of the Invention

The main purpose of the present application is to provide an abnormality detection method, abnormality detection device, abnormality detection device, and computer-readable storage medium, which aims to solve the problem that the traditional abnormality detection method of electrical consumption is not high in accuracy and is easily detected by abnormal values. Interference, and the calculation is complicated, and the detection efficiency is low.

In order to achieve the foregoing objective, an embodiment of the present application provides a method for detecting abnormality in power consumption. The method for detecting abnormality in power consumption includes:

Collecting sequence characteristic values at each target time point from multiple power consumption time series at preset time intervals to generate a characteristic data sequence corresponding to each power consumption time series;

Calculate the abnormal scores of all the characteristic data sequences according to the isolated forest algorithm;

If the abnormal score is greater than the abnormal alert value, confirming the characteristic data sequence as a normal sequence;

If the abnormality score is less than or equal to the abnormality alert value, the characteristic data sequence is confirmed as an abnormality sequence.

The present application also provides an abnormality detection device. The abnormality detection device includes:

A collection module, configured to collect sequence characteristic values at each target time point from multiple power consumption time series at preset time intervals to generate a characteristic data sequence corresponding to each power consumption time series;

A first calculation module, configured to calculate an abnormal score of all the characteristic data sequences according to an isolated forest algorithm;

A first confirmation module, configured to confirm the characteristic data sequence as a normal sequence if the abnormal score is greater than an abnormal alert value;

A second confirmation module is configured to confirm the characteristic data sequence as an abnormal sequence if the abnormal score is less than or equal to an abnormal alert value.

In addition, in order to achieve the above object, the present application further provides an anomaly detection device. The anomaly detection device includes: a memory, a processor, a communication bus, and computer-readable instructions stored on the memory. The communication bus is used for Realize the communication connection between the processor and the memory;

The processor is configured to execute the computer-readable short instruction to implement the following steps:

In addition, in order to achieve the foregoing object, the present application further provides a computer-readable storage medium, where the computer-readable storage medium stores one or more computer-readable instructions, and the one or more computer-readable instructions may be replaced by one Or more than one processor to perform:

The present application collects sequence feature values at each target time point from multiple power consumption time series at preset time intervals to generate a characteristic data sequence corresponding to each power consumption time series; calculates the The abnormal scores of all the characteristic data sequences are described; if the abnormal score is greater than the abnormal alert value, the characteristic data sequence is confirmed as a normal sequence; if the abnormal score is less than or equal to the abnormal alert value, the characteristic data sequence is identified Confirmed as an abnormal sequence. This application uses the isolated forest algorithm to perform anomaly detection on large-scale data in time series. By calculating the abnormal scores in the characteristic data sequence and thresholding the abnormal scores, the normal sequence and the abnormal sequence are determined, and traditional detection is avoided. The method cannot be used for large-scale detection and defects with low accuracy, which improves the accuracy of power consumption abnormality detection, eliminates the interference of abnormal values, reduces the computational complexity, and improves the detection efficiency.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic flowchart of a first embodiment of an abnormality detection method for power consumption in this application;

FIG. 2 is a detailed flowchart of step S20 in FIG. 1; FIG.

3 is a schematic diagram of a device structure of a hardware operating environment involved in a method according to an embodiment of the present application;

FIG. 4 is a schematic diagram of functional modules of the abnormality detection device of the present application.

The implementation, functional features and advantages of the purpose of this application will be further described with reference to the embodiments and the drawings.

detailed description

It should be understood that the specific embodiments described herein are only used to explain the application, and are not used to limit the application.

The present application provides a method for detecting abnormality in power consumption. In a first embodiment of the method for detecting abnormality in power consumption, referring to FIG. 1, the method for detecting abnormality in power consumption includes:

Step S10: Collect sequence characteristic values at each target time point from multiple power consumption time series at preset time intervals to generate a characteristic data sequence corresponding to each power consumption time series;

The power consumption time series represents the distribution data of power consumption at different time periods in a day, and the different power consumption time series represents the power consumption distribution data on different dates. The system can detect historical time series, as well as time series received in real time. The power consumption time series is stored in the database. The system retrieves multiple power consumption time series from the database. Because the data of each power consumption time series is distributed in time units, the system will collect each Some of the eigenvalues in the time series of electricity are used as samples to reduce the computational complexity of the data. The characteristic value of the power consumption time series is distributed according to time units. The system will collect the characteristic value of each power consumption time series by collecting a characteristic data every preset number of hours. Specifically, the preset time interval can be determined according to actual business requirements. For example, if one feature value is collected every one hour, the system can collect 24 feature values from a time series, and the 24 feature values will be used as the power consumption. Characteristic data series of time series. The use of multiple characteristic time series of power consumption time series can effectively avoid the influence of outliers.

Step S20: Calculate the abnormal scores of all the characteristic data sequences according to the isolated forest algorithm;

The characteristic data series can reflect the distribution of the corresponding power consumption time series. In this embodiment, the system will apply the isolated forest algorithm to calculate the abnormal scores of all the characteristic data sequences. Through the isolated forest algorithm, isolated isolated data in the characteristic data sequence can be captured, and the discrete degree of the isolated isolated data can be quantified to a specific value. For example, the system substitutes the feature data sequence into the isolated forest algorithm, and uses the spatial iterative cutting method in the isolated forest algorithm to determine the spatial dispersion of the feature data in the feature data sequence. Specifically, the feature data is spatially cut and The feature data in each space is re-cut until the feature data that is separately cut in the data space is obtained. This process can be embodied in the form of a binary tree layering, that is, all feature data that is cut on the same side of the data space will continue to be iteratively cut, the binary tree will continue to be layered down, and features that are left alone in the data space Since the data will not continue to be cut, it stays at the height of the current binary tree. The isolated forest algorithm will calculate the abnormal score of the feature data sequence according to the height of all discrete feature data.

Step S30: if the abnormal score is greater than the abnormal alert value, confirm the characteristic data sequence as a normal sequence;

In step S40, if the abnormal score is less than or equal to the abnormal alert value, the characteristic data sequence is confirmed as an abnormal sequence.

In this embodiment, an abnormal alert value is set, and the abnormal alert value is a judgment threshold value of the abnormal score. The abnormality of the characteristic data sequence needs to be judged by the abnormal alert value. When the abnormal score is greater than the abnormal alert value, it means that all the characteristic values in the characteristic data sequence corresponding to the abnormal score are relatively concentrated, the discrete offset is small and within a reasonable range, and the corresponding characteristic data sequence belongs to the normal sequence. Conversely, if the system detects that the abnormal score is less than or equal to the abnormal warning value, it indicates that the abnormal score does not reach the standard warning line, and all the characteristic values in the corresponding characteristic data sequence have a large offset. A large number of discrete invalid data cannot reflect the normal power consumption distribution, and the corresponding characteristic data sequence is an abnormal sequence.

The present application collects sequence feature values at each target time point from multiple power consumption time series at preset time intervals to generate a characteristic data sequence corresponding to each power consumption time series; calculates the The abnormal scores of all the characteristic data sequences are described; if the abnormal score is greater than the abnormal alert value, the characteristic data sequence is confirmed as a normal sequence; if the abnormal score is less than or equal to the abnormal alert value, the characteristic data sequence is identified Confirmed as an abnormal sequence. This application uses the isolated forest algorithm to perform anomaly detection on large-scale data in time series. By calculating the abnormal scores in the characteristic data sequence and thresholding the abnormal scores, the normal sequence and the abnormal sequence are determined, and traditional detection is avoided The method cannot be used for large-scale detection and defects with low accuracy, which improves the accuracy of power consumption abnormality detection, eliminates the interference of abnormal values, reduces the computational complexity, and improves the detection efficiency.

Further, based on the first embodiment of the power consumption abnormality detection method of the present application, a second embodiment of the power consumption abnormality detection method of the present application is proposed. Referring to FIG. 2, the difference from the foregoing embodiment is that the step S20 The steps include:

Step A: Determine the positions of the corresponding data points in the isolated forest model space for the sequence feature values at all target time points in each feature data sequence to generate a data point set, and count the total data of the data point set Number of points

It can be understood that each feature data sequence has two types of data: the target time point and the sequence feature value, and these two types of data are mapped to each other. Therefore, each feature data sequence can be worthwhile according to the target time point and the sequence feature. Go to the corresponding data points, and substitute each data point into the isolated forest algorithm model. The model is configured with a model space for inductively placing all data points. That is, the model space is equivalent to a coordinate space. According to the coordinate values of each data point, the system can determine the coordinate positions of all data points in each characteristic data sequence, thereby generating a corresponding data point set in the model space. For example, the current sequence A includes a power consumption value of 5 at 0, a power consumption value of 8 at 6, a power consumption value of 12 at 10, and a power consumption value of 8 at 18. The data points in sequence A therefore include A1 = (0,5), A2 = (6,8), A3 = (12,10), and A4 = (18,8). These data points will be sequentially arranged in the model space according to the coordinates, thereby obtaining the data point set of each data point, and counting the total number of data points of all data points in the data point set.

The above examples are just examples, and do not mean that the data point set includes only the specific values of the above four data points.

Step B: Iteratively space cut all data points in the data point set according to a preset algorithm rule of the isolated forest algorithm until all single data points that are individually cut into a single space are obtained;

The preset algorithm rule of the isolated forest algorithm is to perform iterative space cutting on all data point sets to obtain the cutting space, and calculate the number of data points in each cutting space. The spatial cutting refers to cutting a data point set in an isolated forest model space using a random hyperplane. Assuming that each data point in the data point set is relatively concentrated, it is not easy to have separate data points cut into one space during the space cutting process. If some data points in the data point set are loose or scattered at the edge of the data point set, those scattered data points will be easily cut into a single space. The system cuts through iterative space to obtain all single data points that are individually cut into a single space. It can be understood that when each data point in the data point set is individually cut into a single space, a single data point is generated at this time, and the system will record the single data point. And the number of all single data points is equal to the number of all data points in the data point set.

Step C: Obtain the number of iterations to which each single data point belongs, and obtain a target data point in a preset number of iterations among all the single data points;

Step D: Count the number of data points of all the target data points, and calculate the ratio of the number of data points to the total number of data points;

The system obtains the number of iterations when each single data point is generated. For example, a single data point A is generated during the first spatial cutting, a single data point B, C is generated during the second spatial cutting, a single data point D, E, F, G is generated during the third spatial cutting, etc. The system will count the number of iterations for each single data point. Assuming the preset number of times is 2, the system will obtain the target data points A, B, and C generated in the previous 2 spatial iterations.

The system counts the total number of data points of the target data points as three. Assuming that the total number of data points in the current data point set is 15, then the ratio of the number of data points to the total value is 3/15 = 0.2.

In step E, the percentage value is set as an abnormal score.

The system will set the percentage value to the abnormal score as a reference value for subsequent numerical comparisons.

Further, based on the second embodiment of the power consumption abnormality detection method of the present application, a third embodiment of the power consumption abnormality detection method of the present application is proposed. The difference from the foregoing embodiment is that the method further includes:

Step S50: Generate corresponding normal distribution curves and abnormal distribution curves from the respective normal sequences and the respective abnormal sequences;

The significance of visualizing each sequence as a curve is that users can intuitively observe and identify which time series of power consumption deviate from the normal distribution situation.

In this embodiment, in order to facilitate the user to intuitively view and analyze the difference in sequence feature values between the normal sequence and the abnormal sequence, the system will convert the normal sequence and the abnormal sequence into corresponding normal distribution curves after determining the normal sequence and the abnormal sequence, respectively. And anomalous distribution curves. Because the sequence feature values in the normal sequence and the abnormal sequence are sorted in chronological order, the chronological system can be based on the target time point and their corresponding The series eigenvalues generate normal distribution curves and abnormal distribution curves.

In step S60, the normal distribution curve and the abnormal distribution curve are respectively displayed and marked in a preset coordinate system for user identification.

After obtaining all the normal and abnormal distribution curves, in order to more intuitively reflect the difference between the normal and abnormal distribution curves, the system will display the normal and abnormal distribution curves in a preset coordinate system. In the preset coordinate system, in order to improve the recognition degree, the normal distribution curve and the abnormal distribution curve will be displayed through different marking forms, for example, the normal distribution curve is marked as a green curve, and the abnormal distribution curve is marked as a red curve. The normal and abnormal sequences respectively corresponding to the normal distribution curve and the abnormal distribution curve are shown to facilitate user identification and analysis.

Further, on the basis of the third embodiment of the power consumption abnormality detection method of the present application, a fourth embodiment of the power consumption abnormality detection method of the present application is proposed. The difference from the foregoing embodiment is that the isolated forest algorithm stores the The abnormal eigenvalues in all characteristic data sequences, after step S60, further include:

Step S70: Obtain a target data point corresponding to the abnormal sequence in the isolated forest model space, and obtain an abnormal time point and an abnormal feature value of the target data point;

The anomaly sequence includes free anomalous eigenvalues in the model space, but also includes normal eigenvalues. In this embodiment, because the target data point is cut out a preset number of times before, the system has located a target data point with a large offset. Therefore, the abnormal time points and abnormal eigenvalues of the target data points of the abnormal sequence in the isolated forest model space can be obtained. The abnormal eigenvalues at these abnormal time points will be displayed in the abnormal distribution curve, but they are not clearly marked to increase the degree of recognition. When the number of anomalous distribution curves is large, it will disturb the user's analysis and judgment. In order to improve the recognition of abnormal eigenvalues, this embodiment obtains the target data points of the abnormal sequence in the model space of the isolated forest algorithm. Since the target data points are established abnormal data points and the abnormal sequence has been established, the system can Obtain the target data points' abnormal time points and abnormal eigenvalues from the previous calculation process of the isolated forest algorithm. The offset of the abnormal feature value in the isolated forest algorithm is much larger than other feature values.

Step S80: Obtain an abnormal distribution curve corresponding to the abnormal sequence, locate the target data point on the abnormal distribution curve, and display the abnormal time point and abnormal characteristic value of the target data point correspondingly.

The system obtains the corresponding abnormal distribution curve according to the abnormal sequence, and locates the target data point from the abnormal distribution curve. For example, if the target data point is point C, then the point C in the abnormal distribution curve is obtained, and according to the abnormal time of the data point Points and outlier eigenvalues in outliers It is marked in the distribution curve, and the data point is displayed as the target data point, and its abnormal time point and abnormal characteristic value are displayed, thereby improving the degree of data identification.

Further, on the basis of the fourth embodiment of the power consumption abnormality detection method of the present application, a fifth embodiment of the power consumption abnormality detection method of the present application is proposed. The difference from the foregoing embodiment lies in each target in the normal sequence. There are corresponding normal eigenvalues at time points. After step S60, the method further includes:

Step S90: Collect the corresponding normal feature values in each normal sequence, and perform an average calculation on all the normal feature values at the same target time point in all the normal sequences to obtain the feature average value of each target time point;

All normal sequences represent the distribution of the power consumption time series under normal conditions. This embodiment will provide an ideal distribution curve in a preset coordinate system based on the data of all normal sequences. Specifically, the system collects the feature values of all normal sequences at different time points. Assuming that there are n feature data sequences currently determined to be normal sequences, the system extracts feature values at the same time point from the n normal sequences. There are n in total. The system will calculate the mean of all the feature values at the same time point to get the feature mean value at that time point. The feature mean value can represent the ideal level of normal feature values at that time point. For example, there are a total of five normal sequences currently collected, and the five characteristic values of the time points in the five normal sequences are three, five, four, 3.5, and 4.5. Then the average feature value at 18 o'clock is (3 + 5 + 4 + 3.5 + 4.5) / 5 = 4. The mean value of 4 can represent the average distribution of the target time points in the five normal sequences.

In step S100, the mean values of all the features are converted into an average distribution curve in the order of the target time points, and the average distribution curve is displayed and marked in the preset coordinate system.

After obtaining the feature mean at each time point, the system can sort the feature mean according to the time point order and convert it into an average distribution curve. Since different feature mean values represent the average distribution at the corresponding time points, all feature mean values are sorted in order of time points to obtain an average distribution curve representing the average distribution situation at different time points as a whole. The system will display the average distribution curve in a preset coordinate system, and mark the average distribution curve to facilitate user identification and analysis.

Further, based on the fifth embodiment of the power consumption abnormality detection method of the present application, a sixth embodiment of the power consumption abnormality detection method of the present application is proposed. The difference from the foregoing embodiment is that after step S80, the method further includes:

Step S110: performing a difference calculation on the abnormal characteristic values at each abnormal time point on the abnormal distribution curve and the average characteristic value corresponding to each abnormal time point in the average distribution curve to obtain the abnormal distribution curve and the The characteristic offset value corresponding to the average distribution curve at an abnormal time point;

Further, displaying the abnormal distribution curve and the average distribution curve in a preset coordinate system can be conveniently viewed, but the user may need to accurately analyze the results. For example, the landlord needs to make a detailed analysis of the abnormal power consumption of a tenant, or the user needs to investigate the specific details of the abnormal power consumption. In order to facilitate the user's comparison analysis, the system will be based on the abnormal time points in the abnormal distribution curve. Difference calculation is performed between the abnormal eigenvalue and the average eigenvalue at the corresponding abnormal time point on the average distribution curve to obtain the characteristic offset value of the abnormal distribution curve and the average distribution curve at each abnormal time point. For example, when the abnormal characteristic value is 5 on the abnormal distribution curve and the average characteristic value is 5 on the average distribution curve, the characteristic offset value is 10-5 = 5.

In step S120, if a trigger operation based on the abnormal distribution curve is detected, the characteristic offset values at the target time points of the abnormal distribution curve are displayed in the preset coordinate system for user analysis.

When the user triggers the abnormal distribution curve (such as clicking to check the data in the abnormal distribution curve), the system will directly display the characteristic offset value of each target time point of the abnormal distribution curve in the preset coordinate system. In this way, according to the characteristic offset value, the user can know the change trend of power consumption at different target time points, and can analyze the cause of the abnormality through the change trend analysis, such as a short circuit of the circuit, a failure of the meter, and the like.

Further, based on the sixth embodiment of the power consumption abnormality detection method of the present application, a seventh embodiment of the power consumption abnormality detection method of the present application is proposed. The difference from the foregoing embodiment is that the method further includes:

Step a: When an abnormal control ratio is received, obtain a target alert value mapped to the abnormal control ratio from a preset mapping table;

In a real scenario, the abnormality judgment criteria for anomaly detection may change. For example, if the power supply is limited or cut off for a certain period of time, the power consumption will change drastically, and the system will judge that it is abnormal. In fact, this change is not caused by abnormalities, but caused by known and controllable reasons. Therefore, this abnormal situation needs to be eliminated. In other words, this application can adjust the abnormality judgment standard according to actual business requirements. For example, if there is a need to limit power within the current month, the number of corresponding abnormal sequences will increase. In order to exclude such abnormal sequences, the system can change the judgment criteria, and the determination of the judgment criteria is related to the preset value. It can be understood that the proportion of abnormal control refers to filtering all abnormal sequences in proportion.

In this embodiment, the system stores a preset mapping table. The system receives the abnormal control ratio input from the outside, and finds the target preset value mapped to the ratio in the preset mapping table. For example, in the case of power limitation or power failure, it is known that 5% of the normal sequences in all characteristic data sequences are determined to be abnormal sequences. Then, by adjusting the exception control ratio, this part of the original sequence is abnormal. Sequence exclusion. In the preset mapping table in this embodiment, there is a one-to-one correspondence between the abnormal control ratio and the alert value. It can be understood that the setting of the actual distribution can be determined by the abnormal alert value, that is, the abnormal alert value can be customized to adjust the judgment standard of the abnormal sequence.

In step b, the current default abnormality alert value is adjusted to the target alert value, so as to adjust the judgment criteria of the normal sequence and the abnormal sequence.

It can be known from the above that the current system's default abnormal alert value will no longer meet the current actual business needs. After obtaining the target alert value, the system will directly adjust the current default abnormal alert value to the target alert value, thereby adjusting the normal sequence and abnormality. The criterion of the sequence.

Referring to FIG. 3, FIG. 3 is a schematic diagram of a device structure of a hardware operating environment involved in the method according to the embodiment of the present application.

The device in this embodiment of the present application may be a PC, or a smart phone, a tablet computer, an e-book reader, or MP3 (Moving Picture). Experts Group Audio Layer III, standard video layer 3) player, MP4 (Moving Picture Experts Group Audio Layer IV, motion picture expert compression standard audio layer 4) Terminal equipment such as players, portable computers.

As shown in FIG. 3, the abnormality detection device may include: a processor 1001, such as a CPU, a memory 1005, and a communication bus 1002. The communication bus 1002 is used to implement connection and communication between the processor 1001 and the memory 1005. The memory 1005 may be a high-speed RAM memory or a non-volatile memory. memory), such as disk storage. The memory 1005 may optionally be a storage device independent of the foregoing processor 1001.

Optionally, the abnormality detection device may further include a user interface, a network interface, a camera, and an RF (Radio Frequency) circuits, sensors, audio circuits, WiFi modules, and more. The user interface may include a display, an input unit such as a keyboard, and the optional user interface may further include a standard wired interface and a wireless interface. The network interface can optionally include a standard wired interface and a wireless interface (such as a WI-FI interface).

Those skilled in the art can understand that the structure of the abnormality detection device shown in FIG. 3 does not constitute a limitation on the abnormality detection device, and may include more or fewer components than shown in the figure, or combine some components or different components Layout.

As shown in FIG. 3, the memory 1005 as a computer storage medium may include an operating system, a network communication module, and computer-readable instructions. The operating system is a program that manages and controls the hardware and software resources of the anomaly detection device, and supports the execution of computer-readable instructions and other software and / or programs. The network communication module is used to implement communication between components in the memory 1005 and to communicate with other hardware and software in the abnormality detection device.

In the abnormality detection device shown in FIG. 3, the processor 1001 is configured to execute computer-readable instructions stored in the memory 1005 to implement the following steps:

The specific implementation manners of the abnormality detection device of the present application are basically the same as the embodiments of the foregoing abnormality detection method for power consumption, and are not repeated here.

Referring to FIG. 4, the present application provides an abnormality detection device. The abnormality detection device includes:

Further, the first calculation module includes:

A generating unit, configured to determine the positions of the corresponding data points in the isolated forest model space of the sequence feature values at all target time points in each of the feature data sequences to generate a data point set, and count the Total number of data points;

A cutting unit, configured to perform iterative space cutting on all data points in the data point set according to a preset algorithm rule of an isolated forest algorithm until all single data points that are individually cut into a single space are obtained;

An obtaining unit, configured to obtain the number of iterations to which each single data point belongs, and to obtain a target data point of a preset number of iterations among all the single data points;

A calculation unit, configured to count the number of data points of all the target data points, and calculate a ratio of the number of data points to the total number of data points;

A setting unit, configured to set the ratio value as an abnormal score.

Further, the abnormality detection device further includes:

A generating module, configured to generate corresponding normal distribution curves and abnormal distribution curves from the respective normal sequences and the respective abnormal sequences;

A first display module is configured to display and mark the normal distribution curve and the abnormal distribution curve respectively in a preset coordinate system for user identification.

Further, the isolated forest algorithm stores abnormal characteristic values in all characteristic data sequences, and the abnormality detection device further includes:

A first obtaining module, configured to obtain a target data point corresponding to the abnormal sequence in an isolated forest model space, and obtain an abnormal time point and an abnormal characteristic value of the target data point;

A second display module is configured to obtain an abnormal distribution curve corresponding to the abnormal sequence, locate the target data point on the abnormal distribution curve, and display the abnormal time point and abnormal characteristic value of the target data point correspondingly.

Further, each target time point in the normal sequence has a corresponding normal feature value, and the abnormality detection device further includes:

A second calculation module is configured to collect the corresponding normal feature values in each normal sequence, and perform an average calculation on all the normal feature values at the same target time point in all the normal sequences to obtain the target time points. Feature mean

A conversion module, configured to convert the mean values of all features into an average distribution curve in the order of target time points, and display and mark the average distribution curve in the preset coordinate system.

Further, the abnormality detection device further includes:

A third calculation module, configured to perform difference calculation on the abnormal characteristic values at each abnormal time point on the abnormal distribution curve and the average characteristic value corresponding to each abnormal time point in the average distribution curve to obtain the abnormal distribution A characteristic offset value at an abnormal time point corresponding to the curve and the average distribution curve;

A third display module, configured to display, if a trigger operation based on the abnormal distribution curve is detected, the characteristic offset value at each target time point of the abnormal distribution curve in the preset coordinate system, For user analysis.

Further, the abnormality detection device further includes:

A second obtaining module, configured to obtain a target alert value mapped to the abnormal control ratio from a preset mapping table when the abnormal control ratio is received;

An adjustment module is configured to adjust a current default abnormal alert value to the target alert value, so as to adjust the judgment criteria of the normal sequence and the abnormal sequence.

It should be noted that a person of ordinary skill in the art may understand that all or part of the steps for implementing the foregoing embodiments may be completed by hardware, or computer-readable instructions may be used to instruct related hardware to complete. The computer-readable instructions may be stored in In a computer-readable storage medium, the aforementioned storage medium may be a read-only memory, a magnetic disk, or an optical disk.

This application also provides a computer-readable storage medium, where the computer-readable storage medium stores one or more computer-readable instructions, and the one or more computer-readable instructions can also be processed by one or more The device executes steps for implementing the method for detecting an abnormality in power consumption according to any one of the foregoing. The computer-readable storage medium may be a non-volatile readable storage medium, such as a RAM, a magnetic disk, an optical disk, or the like.

The specific implementation manner of the readable storage medium in this application is basically the same as each embodiment of the foregoing method for detecting an abnormality in power consumption, and details are not described herein again.

It should be noted that, in this article, the terms "including", "including" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, It also includes other elements not explicitly listed, or elements inherent to such a process, method, article, or device. Without more restrictions, an element limited by the sentence "including a ..." does not exclude that there are other identical elements in the process, method, article, or device that includes the element.

The above-mentioned serial numbers of the embodiments of the present application are merely for description, and do not represent the superiority or inferiority of the embodiments.

Through the description of the above embodiments, those skilled in the art can clearly understand that the methods in the above embodiments can be implemented by means of software plus a necessary universal hardware platform, and of course, also by hardware, but in many cases the former is better. Implementation. Based on such an understanding, the technical solution of this application that is essentially or contributes to the existing technology can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, The optical disc) includes several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in the embodiments of the present application.

The above are only preferred embodiments of the present application, and thus do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the description and drawings of the present application, or directly or indirectly used in other related technical fields Are included in the scope of patent protection of this application.

Claims

An abnormality detection method of power consumption, characterized in that the abnormality detection method of power consumption includes:

Collecting sequence characteristic values at each target time point from multiple power consumption time series at preset time intervals to generate a characteristic data sequence corresponding to each power consumption time series;

Calculate the abnormal scores of all the characteristic data sequences according to the isolated forest algorithm;

If the abnormal score is greater than the abnormal alert value, confirming the characteristic data sequence as a normal sequence;

If the abnormality score is less than or equal to the abnormality alert value, the characteristic data sequence is confirmed as an abnormality sequence.
The method for detecting abnormal power consumption according to claim 1, wherein the step of calculating the abnormal scores of all the characteristic data sequences according to the isolated forest algorithm comprises:

Determining the position of the corresponding data points of the sequence feature values at all target time points in each feature data sequence in the isolated forest model space to generate a data point set, and counting the total number of data points in the data point set ;

Performing iterative space cutting on all data points in the data point set according to a preset algorithm rule of the isolated forest algorithm until all single data points that are individually cut into a single space are obtained;

Obtaining the number of iterations to which each single data point belongs, and obtaining the target data point in the preset number of iterations among all the single data points;

Counting the number of data points of all the target data points, and calculating the ratio of the number of data points to the total number of data points;

The percentage value is set as an abnormal score.
The method for detecting abnormal power consumption according to claim 2, wherein the method further comprises:

Generating corresponding normal distribution curves and abnormal distribution curves from the respective normal sequences and the respective abnormal sequences;

The normal distribution curve and the abnormal distribution curve are respectively displayed and marked in a preset coordinate system for user identification.
The method for detecting abnormal power consumption according to claim 3, wherein the isolated forest algorithm stores abnormal characteristic values in all characteristic data sequences, and the normal distribution curve and the abnormal distribution curve are stored in the isolated forest algorithm. After the steps of displaying and marking in the preset coordinate system for user identification, the method further includes:

Acquiring a target data point corresponding to the abnormal sequence in the isolated forest model space, and acquiring an abnormal time point and an abnormal characteristic value of the target data point;

Obtain the abnormal distribution curve corresponding to the abnormal sequence, locate the target data point on the abnormal distribution curve, and display the abnormal time point and abnormal characteristic value of the target data point correspondingly.
The method according to claim 3, wherein each target time point in the normal sequence has a corresponding normal characteristic value, and the normal distribution curve and the abnormal distribution curve are After the steps of displaying and marking in the preset coordinate system for user identification, the method further includes:

Collecting corresponding normal feature values in each normal sequence, and performing average calculation on all the normal feature values at the same target time point in all normal sequences to obtain the feature average value of each target time point;

Converting all the mean values of the features into an average distribution curve according to the sequence of the target time point, and displaying and marking the average distribution curve in the preset coordinate system.
The method for detecting abnormal power consumption according to claim 5, wherein the average values of all the features are converted into an average distribution curve in the order of the target time points, and the average distribution curve is displayed and marked in After the step of presetting the coordinate system, the method further includes:

Performing difference calculation on the abnormal characteristic values at each abnormal time point on the abnormal distribution curve and the average characteristic value corresponding to each abnormal time point in the average distribution curve to obtain the abnormal distribution curve and the average distribution curve Corresponds to the characteristic offset value at the abnormal time point;

If a trigger operation based on the abnormal distribution curve is detected, the characteristic offset values at the target time points of the abnormal distribution curve are displayed in the preset coordinate system for user analysis.
The method for detecting an abnormal power consumption according to claim 1, further comprising:

When an abnormal control ratio is received, obtaining a target alert value mapped to the abnormal control ratio from a preset mapping table;

Adjusting the current default abnormal alert value to the target alert value to adjust the normal sequence and the judgment criterion of the abnormal sequence.
An abnormality detection device, characterized in that the abnormality detection device includes:

A collection module, configured to collect sequence characteristic values at each target time point from multiple power consumption time series at preset time intervals to generate a characteristic data sequence corresponding to each power consumption time series;

A first calculation module, configured to calculate an abnormal score of all the characteristic data sequences according to an isolated forest algorithm;

A first confirmation module, configured to confirm the characteristic data sequence as a normal sequence if the abnormal score is greater than an abnormal alert value;

A second confirmation module is configured to confirm the characteristic data sequence as an abnormal sequence if the abnormal score is less than or equal to an abnormal alert value.
An abnormality detection device, characterized in that the abnormality detection device includes: a memory, a processor, a communication bus, and computer-readable instructions stored on the memory, and the processor is configured to execute the computer-readable instructions, To achieve the following steps:

Collecting sequence characteristic values at each target time point from multiple power consumption time series at preset time intervals to generate a characteristic data sequence corresponding to each power consumption time series;

Calculate the abnormal scores of all the characteristic data sequences according to the isolated forest algorithm;

If the abnormal score is greater than the abnormal alert value, confirming the characteristic data sequence as a normal sequence;

If the abnormality score is less than or equal to the abnormality alert value, the characteristic data sequence is confirmed as an abnormality sequence.
The abnormality detection device according to claim 9, wherein the step of calculating the abnormality scores of all the characteristic data sequences according to the isolated forest algorithm comprises:

Determining the position of the corresponding data points of the sequence feature values at all target time points in each feature data sequence in the isolated forest model space to generate a data point set, and counting the total number of data points in the data point set ;

Performing iterative space cutting on all data points in the data point set according to a preset algorithm rule of the isolated forest algorithm until all single data points that are individually cut into a single space are obtained;

Obtaining the number of iterations to which each single data point belongs, and obtaining the target data point in the preset number of iterations among all the single data points;

Counting the number of data points of all the target data points, and calculating the ratio of the number of data points to the total number of data points;

The percentage value is set as an abnormal score.
The abnormality detection device according to claim 10, wherein the processor is configured to execute the computer-readable instructions, and further to implement the following steps:

Generating corresponding normal distribution curves and abnormal distribution curves from the respective normal sequences and the respective abnormal sequences;

The normal distribution curve and the abnormal distribution curve are respectively displayed and marked in a preset coordinate system for user identification.
The abnormality detection device according to claim 11, wherein the isolated forest algorithm stores abnormal characteristic values in all characteristic data sequences, and the normal distribution curve and the abnormal distribution curve are separately displayed and The steps marked in the preset coordinate system for user identification also include:

Acquiring a target data point corresponding to the abnormal sequence in the isolated forest model space, and acquiring an abnormal time point and an abnormal characteristic value of the target data point;

Obtain the abnormal distribution curve corresponding to the abnormal sequence, locate the target data point on the abnormal distribution curve, and display the abnormal time point and abnormal characteristic value of the target data point correspondingly.
The abnormality detection device according to claim 11, wherein each target time point in the normal sequence has a corresponding normal eigenvalue, and the normal distribution curve and the abnormal distribution curve are separately displayed and displayed. The steps marked in the preset coordinate system for user identification also include:

Collecting corresponding normal feature values in each normal sequence, and performing average calculation on all the normal feature values at the same target time point in all normal sequences to obtain the feature average value of each target time point;

Converting all the mean values of the features into an average distribution curve according to the sequence of the target time point, and displaying and marking the average distribution curve in the preset coordinate system.
The abnormality detection device according to claim 13, wherein the average value of all the features is converted into an average distribution curve in the order of the target time points, and the average distribution curve is displayed and marked in the pre- The steps of setting the coordinate system also include:

Performing difference calculation on the abnormal characteristic values at each abnormal time point on the abnormal distribution curve and the average characteristic value corresponding to each abnormal time point in the average distribution curve to obtain the abnormal distribution curve and the average distribution curve Corresponds to the characteristic offset value at the abnormal time point;

If a trigger operation based on the abnormal distribution curve is detected, the characteristic offset values at the target time points of the abnormal distribution curve are displayed in the preset coordinate system for user analysis.
A computer-readable storage medium is characterized in that computer-readable instructions are stored on the computer-readable storage medium, and the computer-readable instructions are executed by a processor to implement the following steps:

Collecting sequence characteristic values at each target time point from multiple power consumption time series at preset time intervals to generate a characteristic data sequence corresponding to each power consumption time series;

Calculate the abnormal scores of all the characteristic data sequences according to the isolated forest algorithm;

If the abnormal score is greater than the abnormal alert value, confirming the characteristic data sequence as a normal sequence;

If the abnormality score is less than or equal to the abnormality alert value, the characteristic data sequence is confirmed as an abnormality sequence.
The computer-readable storage medium of claim 15, wherein the step of calculating an abnormal score of all the characteristic data sequences according to an isolated forest algorithm comprises:

Determining the position of the corresponding data points of the sequence feature values at all target time points in each feature data sequence in the isolated forest model space to generate a data point set, and counting the total number of data points in the data point set ;

Performing iterative space cutting on all data points in the data point set according to a preset algorithm rule of the isolated forest algorithm until all single data points that are individually cut into a single space are obtained;

Obtaining the number of iterations to which each single data point belongs, and obtaining the target data point in the preset number of iterations among all the single data points;

Counting the number of data points of all the target data points, and calculating the ratio of the number of data points to the total number of data points;

The percentage value is set as an abnormal score.
The computer-readable storage medium of claim 16, wherein the computer-readable instructions are executed by a processor to further implement the following steps:

Generating corresponding normal distribution curves and abnormal distribution curves from the respective normal sequences and the respective abnormal sequences;

The normal distribution curve and the abnormal distribution curve are respectively displayed and marked in a preset coordinate system for user identification.
The computer-readable storage medium of claim 17, wherein the isolated forest algorithm stores abnormal characteristic values in all characteristic data sequences, and the normal distribution curve and the abnormal distribution curve are respectively stored in the isolated forest algorithm. After the step of displaying and marking in a preset coordinate system for identification by the user, the method further includes:

Acquiring a target data point corresponding to the abnormal sequence in the isolated forest model space, and acquiring an abnormal time point and an abnormal characteristic value of the target data point;

Obtain the abnormal distribution curve corresponding to the abnormal sequence, locate the target data point on the abnormal distribution curve, and display the abnormal time point and abnormal characteristic value of the target data point correspondingly.
The computer-readable storage medium of claim 17, wherein each target time point in the normal sequence has a corresponding normal characteristic value, and the normal distribution curve and the abnormal distribution curve are respectively After the step of displaying and marking in a preset coordinate system for identification by the user, the method further includes:

Collecting corresponding normal feature values in each normal sequence, and performing average calculation on all the normal feature values at the same target time point in all normal sequences to obtain the feature average value of each target time point;

Converting all the mean values of the features into an average distribution curve according to the sequence of the target time point, and displaying and marking the average distribution curve in the preset coordinate system.
The computer-readable storage medium of claim 19, wherein the average values of all the features are converted into an average distribution curve in the order of the target time points, and the average distribution curve is displayed and marked on all After the steps of describing the preset coordinate system, the method further includes:

Performing difference calculation on the abnormal characteristic values at each abnormal time point on the abnormal distribution curve and the average characteristic value corresponding to each abnormal time point in the average distribution curve to obtain the abnormal distribution curve and the average distribution curve Corresponds to the characteristic offset value at the abnormal time point;

If a trigger operation based on the abnormal distribution curve is detected, the characteristic offset values at the target time points of the abnormal distribution curve are displayed in the preset coordinate system for user analysis. Ranch