CN109242041A - A kind of electric energy meter abnormal deviation data examination method, device, equipment and storage medium - Google Patents
A kind of electric energy meter abnormal deviation data examination method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN109242041A CN109242041A CN201811139036.2A CN201811139036A CN109242041A CN 109242041 A CN109242041 A CN 109242041A CN 201811139036 A CN201811139036 A CN 201811139036A CN 109242041 A CN109242041 A CN 109242041A
- Authority
- CN
- China
- Prior art keywords
- data
- electric energy
- energy meter
- machine learning
- learning algorithm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification
Abstract
The invention discloses a kind of electric energy meter abnormal deviation data examination method, after the initial data when getting electric energy meter work, multiple data sequences are shown with regard to processing;Then the machine learning algorithm model of the sample data building anomaly data detection formed according to the corresponding characteristic value of each data sequence and intensity of anomaly, and the model is trained;Finally electric energy meter data can directly be carried out abnormality detection using trained machine learning algorithm model.This method, the sample data of training machine learning algorithm model is formed by the corresponding characteristic value of each data sequence and intensity of anomaly, so the trained machine learning algorithm model of later-stage utilization can accurately determine which abnormal data be present in period, compared with traditional anomaly data detection mode, it can accurately determine which period abnormal data is present in, improve detection accuracy.In addition, the invention also discloses a kind of electric energy meter anomaly data detection device, equipment and storage medium, effect are as above.
Description
Technical field
The present invention relates to anomaly data detection field, in particular to a kind of electric energy meter abnormal deviation data examination method, is set device
Standby and storage medium.
Background technique
Electric energy meter data it is normal whether not only concerning arrive metering and billing accuracy, but also concerning arrive metering automation main website
The accuracy of upper load forecasting model or Clustering Model.Abnormal data includes event of data loss and due to unplanned interior thing
Part, data acquisition upload the unusual categorical data situation caused situations such as abnormal with storage.It is continuous due to time series
The strong correlation of data, abnormal data will appear in fixed data segment in the form of data set, rather than individual data point.Mesh
When the preceding abnormal data to electric energy meter detects, the root-mean-square value for obtaining data is usually calculated, it then should by judgement
Whether root-mean-square value is more than that the mode of preset value determines whether there is abnormal data, but which can only judge the number obtained
There are abnormal datas in, can not accurately determine which period abnormal data is specifically present in, and then influence abnormal number
According to testing result.
It can be seen that it is quasi- to improve the detection of abnormal data how accurately to determine which abnormal data be present in period
The problem of true property is those skilled in the art's urgent problem to be solved.
Summary of the invention
The embodiment of the present application provides a kind of electric energy meter abnormal deviation data examination method, device, equipment and storage medium, solves
How accurately to determine which abnormal data be present in improve the detection accuracy of abnormal data period in the prior art
The problem of.
In order to solve the above technical problems, the present invention provides a kind of electric energy meter abnormal deviation data examination methods, comprising:
Obtain electric energy meter work when initial data, and according to time sequencing to the initial data carry out segment processing with
Obtain multiple data sequences;
Determine the corresponding characteristic value of each data sequence and intensity of anomaly to form sample data;
It is calculated according to the machine learning algorithm model of sample data building anomaly data detection, and to the machine learning
Method model is trained;
The electric energy meter data are carried out abnormality detection according to the machine learning algorithm model after training.
Preferably, described that segment processing is carried out to the initial data to show that multiple data sequences have according to time sequencing
Body are as follows:
The initial data is divided into according to the time sequencing using sliding pane mode each number of preset length
According to sequence.
Preferably, the corresponding characteristic value of each data sequence of the determination specifically:
Each data sequence is analyzed and processed according to the library tsfresh to obtain the corresponding characteristic value.
Preferably, the characteristic value includes maximum value, minimum value, the degree of bias, average value, median and kurtosis.
Preferably, the machine learning algorithm model is specially non-supervisory formula machine learning algorithm model.
Preferably, the machine learning algorithm model according to after training carries out abnormal inspection to the electric energy meter data
It surveys specifically:
The electric energy meter data are carried out using the k nearest neighbor algorithm in the non-supervisory formula machine learning algorithm model abnormal
Detection.
In order to solve the above technical problems, the present invention also provides a kind of electric energy corresponding with energy table abnormal deviation data examination method
Table anomaly data detection device, comprising:
Module is obtained, for obtaining initial data when electric energy meter work, and according to time sequencing to the initial data
Segment processing is carried out to obtain multiple data sequences;
Determining module, for determining the corresponding characteristic value of each data sequence and intensity of anomaly to form sample data;
Module is constructed, for constructing the machine learning algorithm model of anomaly data detection according to the sample data, and it is right
The machine learning algorithm model is trained;
Abnormality detection module, for being carried out according to the machine learning algorithm model after training to the electric energy meter data
Abnormality detection.
Preferably, the determining module is specifically used for being analyzed and processed each data sequence according to the library tsfresh
To obtain the corresponding characteristic value.
In order to solve the above technical problems, the present invention also provides a kind of electric energy corresponding with energy table abnormal deviation data examination method
Table anomaly data detection equipment, comprising:
Memory, for storing computer program;
Processor, for executing the computer program to realize any one of the above electric energy meter abnormal deviation data examination method
The step of.
In order to solve the above technical problems, the present invention also provides a kind of one kind corresponding with energy table abnormal deviation data examination method
Computer readable storage medium is stored with computer program, the computer program quilt on the computer readable storage medium
Processor executes the step of to realize any one of the above electric energy meter abnormal deviation data examination method.
Compared with the prior art, a kind of electric energy meter abnormal deviation data examination method provided by the present invention, is getting electric energy
After initial data when table works, segment processing is carried out with regard to the initial data that will acquire according to the chronological order of acquisition and is obtained
Multiple data sequences out;Then the sample data structure formed according to the corresponding characteristic value of each data sequence and intensity of anomaly that determine
The machine learning algorithm model of anomaly data detection is built, and machine learning algorithm model is trained;Finally according to training after
Machine learning algorithm model electric energy meter data are carried out abnormality detection, that is to say, that the later period directly utilizes trained engineering
Electric energy meter data can be carried out abnormality detection by practising algorithm model.It can be seen that the abnormal deviation data examination method, because by original
Data are divided into multiple data sequences, the sample data training airplane formed using the corresponding characteristic value of each data sequence and intensity of anomaly
Device learning algorithm model, the trained machine learning algorithm model of later-stage utilization can accurately determine that abnormal data is present in
In which, compared with anomaly data detection mode in the prior art, and then the detection for improving abnormal data is accurate period
Property.In addition, the present invention also provides a kind of electric energy meter anomaly data detection device, equipment and storage medium, effect are as above.
Detailed description of the invention
Fig. 1 is a kind of electric energy meter abnormal deviation data examination method flow chart provided by the embodiment of the present invention;
Fig. 2 is a kind of electric energy meter anomaly data detection device composition schematic diagram provided by the embodiment of the present invention;
Fig. 3 is a kind of electric energy meter anomaly data detection equipment composition schematic diagram provided by the embodiment of the present invention.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present application, technical solutions in the embodiments of the present application carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art without making creative work it is obtained it is all its
Its embodiment, shall fall within the protection scope of the present invention.
Core of the invention is to provide a kind of electric energy meter abnormal deviation data examination method, device, equipment and storage medium, can be with
It is accurate with the detection for improving abnormal data to solve how accurately to determine which abnormal data be present in period in the prior art
The problem of property.
Scheme in order to enable those skilled in the art to better understand the present invention, with reference to the accompanying drawing and specific embodiment party
The present invention is described in further detail for formula.
Fig. 1 is a kind of electric energy meter abnormal deviation data examination method flow chart provided by the embodiment of the present invention, as shown in Figure 1,
The abnormal deviation data examination method includes:
S101: initial data when electric energy meter work is obtained, and segment processing is carried out to initial data according to time sequencing
To obtain multiple data sequences.
It is divided into multiple numbers according to time sequencing when obtaining particularly as initial data when being the electric energy meter operation that will acquire
According to sequence, in order to handle initial data.Initial data may include the voltage value of electric energy meter, current value and power
Etc. data.
In order to improve the detection accuracy of abnormal data, preferably embodiment, according to time sequencing to original number
According to progress segment processing to obtain multiple data sequences specifically: use sliding pane mode by initial data according to time sequencing
It is divided into each data sequence of preset length.
Specifically, the default of the data sequence needed to form can be set out according to the total length of the initial data of acquisition
Length, preset length should be greater than the length of abnormal data to be measured, then will be original using sliding pane mode according to time sequencing
Data are divided into each data sequence of preset length.If the total length of initial data is 5s, preset length 10ms, at this time with regard to every
Secondary sliding 10ms is segmented initial data, and preset length can be determined according to the actual situation.In practical applications, in order to
It determines which period abnormal data is in more accurately, when carrying out segment processing to initial data, each data can be made
The degree of overlapping of sequence reaches 90%, that is, ensures that the degree of overlapping between adjacent data sequence is 90% or more, the length of segment data
The characteristics of degree setting fully considers the continuous feature of electric energy meter abnormal data rather than single exceptional data point, abnormal data is compared in selection
The problems such as length is bigger and the mode of 90% segment superimposition, effectively prevents abnormal data missing inspection, erroneous detection improves abnormal number
According to the accuracy of detection.For example, the initial data D for the time series that length is M is divided into several length using sliding pane mode
Degree is segment sequence (data sequence) Q of Ni:
Wherein, QiFor segment sequence, N is the length of segment sequence, and D is the initial data of time series.M is time series
Initial data length.
Initial data segment processing not only can be improved by the detection of abnormal data using sliding pane mode in the step
Accuracy, it is huge to efficiently solve electric energy meter time series data point, is unfavorable for being directly used in the exception of machine learning
The problem of Data Detection.
S102: determine the corresponding characteristic value of each data sequence and intensity of anomaly to form sample data.
After forming data sequence, determines that the characteristic value in each data sequence and determine the exception of each data sequence
Degree, that is, determine the characteristic value and intensity of anomaly of each data sequence, it is therefore an objective to form sample data when model training.
The intensity of anomaly of each data sequence can obtain according to experiment, particularly as being by testing and the data of standard electric energy meter progress phase
Than the intensity of anomaly in data sequence obtained.In order to obtain data representative in each data sequence in order to different
Regular data is detected, it is preferable that characteristic value includes maximum value, minimum value, the degree of bias, average value, median and kurtosis.In reality
In the application of border, characteristic value can also include flat other than maximum value, minimum value, the degree of bias, average value, median and kurtosis
Fang He, kurtosis, mark it is poor, pass through x-axis number, particular value and with the mean deviation absolute value of the preceding paragraph data sequence, out
Existing percentage etc., in this not go into detail.Preferably embodiment determines that the corresponding characteristic value of each data sequence is specific
Are as follows: each data sequence is analyzed and processed according to the library tsfresh to obtain corresponding characteristic value.Particularly as being to obtain segment
Sequence QiAfterwards, every segment data sequence is analyzed using the library tsfresh obtain corresponding characteristic value, the fast speed that characteristic value determines.
S103: according to the machine learning algorithm model of sample data building anomaly data detection, and to machine learning algorithm
Model is trained.
Particularly as be using the characteristic value for each data sequence determined as the input sample of machine learning algorithm model, will
Output sample of the intensity of anomaly in each data sequence determined as learning algorithm model constructs machine learning algorithm mould
Type using machine learning algorithm model as the model of electric energy meter anomaly data detection, and instructs machine learning algorithm model
Practice.Machine learning algorithm model training process utilizes instruction specifically, sample data is divided into training dataset and test data set
Practice data set to be trained machine learning algorithm model, then utilizes the machine learning algorithm model after test data set training
It is tested to judge whether machine learning algorithm model be trained to accurately, that is, judge machine learning algorithm model
Through reaching requirement.
The length H of the feature value vector of each data sequence obtained will be substantially less that the length N of segment sequence, and then can have
Effect ground reduces the data volume of machine learning algorithm mode input, reduces the time of machine learning algorithm model foundation, eigenvalue cluster
At sequence can be with are as follows:
In formula, f1 (i)Indicate the 1st characteristic value of i-th of segment sequence, fH (i)Indicate the H spy of i-th of segment sequence
Value indicative, FiIndicate the feature vector of i-th of segment training, N is the length of segment sequence, the initial data length of M time series.
S104: electric energy meter data are carried out abnormality detection according to the machine learning algorithm model after training.
In machine learning algorithm model construction and after training, the later period can be by trained machine learning algorithm model
It, can be direct after the raw data associated when getting electric energy meter operation as the detection model of electric energy meter abnormal data
It brings the characteristic value data of each data sequence after classification processing into trained machine learning algorithm model and carries out electric energy meter number
According to abnormality detection, without in the intensity of anomaly for determining compared with the data of standard electric energy meter each data sequence by experiment, into
And determine that each data sequence whether there is abnormal data, it specifically can be by each data sequence that judges the output of machine learning algorithm model
Whether the intensity of anomaly of column is more than threshold value, and then determines the data exception situation of each data sequence, can accurately be determined different
Which regular data be present in period, improves the accuracy of anomaly data detection.
A kind of electric energy meter abnormal deviation data examination method provided by the present invention, the original number when getting electric energy meter work
According to later, segment processing is carried out with regard to the initial data that will acquire according to the chronological order of acquisition and obtains multiple data sequences;
Then the sample data building abnormal data inspection formed according to the corresponding characteristic value of each data sequence and intensity of anomaly determined
The machine learning algorithm model of survey, and machine learning algorithm model is trained;Finally calculated according to the machine learning after training
Method model carries out abnormality detection electric energy meter data, that is to say, that the later period directly utilizes trained machine learning algorithm model just
Electric energy meter data can be carried out abnormality detection.It can be seen that the abnormal deviation data examination method, because initial data is divided into multiple
Data sequence, the sample data training machine learning algorithm mould formed using the corresponding characteristic value of each data sequence and intensity of anomaly
Type, the trained machine learning algorithm model of later-stage utilization can accurately determine which period abnormal data is present in
In, compared with anomaly data detection mode in the prior art, and then improve the detection accuracy of abnormal data.
In order to improve the processing speed to data and the learning efficiency to data structure etc., on the basis of above-described embodiment
On, preferably embodiment, machine learning algorithm model are specially non-supervisory formula machine learning algorithm model.Non-supervisory formula
The target that model is calculated in study is to structure potential in data and distribution modeling, to make further study, institute to data
Data be all no label and the intrinsic structure of algorithm learning data from input data.
On the basis of the above embodiments, preferably embodiment, according to the machine learning algorithm model after training
Electric energy meter data are carried out abnormality detection specifically:
Electric energy meter data are carried out abnormality detection using the k nearest neighbor algorithm in non-supervisory formula machine learning algorithm model.
It specifically, is exactly using a kind of k nearest neighbor algorithm pair under clustering algorithm in non-supervisory formula machine learning algorithm model
Electric energy meter data carry out abnormality detection, and the accuracy in detection of abnormal data can be improved.The feature value vector that can will specifically obtain
Regard the data point of H dimensional space as, i.e., clustering processing is carried out to each feature value vector of acquisition, the number of H dimensional space after cluster
Feature value vector of the strong point A more than or equal to 40 can be regarded as normal data, and A value can be chosen according to the actual situation, this hair
Bright and be not construed as limiting, each data point can indicate corresponding segment sequence (data sequence), using k nearest neighbor algorithm, calculate every
A data point with other A (A is characterized the number of value sequence, and the value of A can be determined according to the actual situation) is a closes on data point
Score value of the average distance as each segment sequence, be in fact exactly to beat the characteristic value sequence of each data sequence
Point, the data exception degree of corresponding characteristic value sequence is obtained, the output pair of non-supervisory formula machine learning algorithm model is then passed through
For the intensity of anomaly answered to carry out abnormality detection to electric energy meter data, specific score value is how much to mean that this feature value sequence is corresponding
Data exception in data sequence can be determined according to the actual situation.In practical application, the length of feature value vector is far small
In the length of segment sequence, the data volume of algorithm model training can reduce, to reduce the model foundation time.
It is described in detail above for a kind of embodiment of electric energy meter abnormal deviation data examination method, is based on above-mentioned reality
The electric energy meter abnormal deviation data examination method for applying example description, the embodiment of the invention also provides a kind of electric energy meters corresponding with this method
Anomaly data detection device.Since the embodiment of device part is corresponded to each other with the embodiment of method part, device part
Embodiment please refer to method part embodiment description, which is not described herein again.
Fig. 2 is a kind of electric energy meter anomaly data detection device composition schematic diagram provided by the embodiment of the present invention, such as Fig. 2 institute
Show, which includes obtaining module 201, and determining module 202 constructs module 203 and abnormality detection module 204.
Obtain module 201, for obtain electric energy meter work when initial data, and according to time sequencing to initial data into
Row segment processing is to obtain multiple data sequences;
Determining module 202, for determining the corresponding characteristic value of each data sequence and intensity of anomaly to form sample data;
Module 203 is constructed, for the machine learning algorithm model according to sample data building anomaly data detection, and to machine
Device learning algorithm model is trained;
Abnormality detection module 204 is abnormal for being carried out according to the machine learning algorithm model after training to electric energy meter data
Detection.
A kind of electric energy meter anomaly data detection device provided by the present invention, the original number when getting electric energy meter work
According to later, segment processing is carried out with regard to the initial data that will acquire according to the chronological order of acquisition and obtains multiple data sequences;
Then the sample data building abnormal data inspection formed according to the corresponding characteristic value of each data sequence and intensity of anomaly determined
The machine learning algorithm model of survey, and machine learning algorithm model is trained;Finally calculated according to the machine learning after training
Method model carries out abnormality detection electric energy meter data, that is to say, that the later period directly utilizes trained machine learning algorithm model just
Electric energy meter data can be carried out abnormality detection.It can be seen that the anomaly data detection device, because initial data is divided into multiple
Data sequence, the sample data training machine learning algorithm mould formed using the corresponding characteristic value of each data sequence and intensity of anomaly
Type, the trained machine learning algorithm model of later-stage utilization can accurately determine which period abnormal data is present in
In, compared with anomaly data detection mode in the prior art, and then improve the detection accuracy of abnormal data.
On the basis of the above embodiments, preferably embodiment, determining module 202 are specifically used for foundation
The library tsfresh is analyzed and processed each data sequence to obtain corresponding characteristic value.
It is described in detail above for a kind of embodiment of electric energy meter abnormal deviation data examination method, is based on above-mentioned reality
The electric energy meter abnormal deviation data examination method for applying example description, the embodiment of the invention also provides a kind of electric energy meters corresponding with this method
Anomaly data detection equipment.Since the embodiment of environment division is corresponded to each other with the embodiment of method part, environment division
Embodiment please refer to method part embodiment description, which is not described herein again.
Fig. 3 is a kind of electric energy meter anomaly data detection equipment composition schematic diagram provided by the embodiment of the present invention, such as Fig. 3 institute
Show, which includes memory 301 and processor 302.
Memory 301, for storing computer program;
Processor 302 realizes that electric energy meter provided by any one above-mentioned embodiment is different for executing computer program
The step of regular data detection method.
A kind of electric energy meter anomaly data detection equipment provided by the present invention, because initial data is divided into multiple data sequences
Column, the sample data training machine learning algorithm model formed using the corresponding characteristic value of each data sequence and intensity of anomaly, after
Phase can accurately determine which abnormal data be present in period using trained machine learning algorithm model, and existing
There is the anomaly data detection mode in technology to compare, can accurately determine which abnormal data be present in period, into
And improve the detection accuracy of abnormal data.
It is described in detail above for a kind of embodiment of electric energy meter abnormal deviation data examination method, is based on above-mentioned reality
The electric energy meter abnormal deviation data examination method for applying example description, the embodiment of the invention also provides a kind of computers corresponding with this method
Readable storage medium storing program for executing.Since the embodiment of computer readable storage medium part is corresponded to each other with the embodiment of method part, because
The embodiment of this computer readable storage medium part please refers to the embodiment description of method part, and which is not described herein again.
A kind of computer readable storage medium is stored with computer program, computer journey on computer readable storage medium
The step of electric energy meter abnormal deviation data examination method that sequence is executed by processor to realize above-mentioned any one embodiment offer.
A kind of computer readable storage medium provided by the present invention, processor can read in readable storage medium storing program for executing and store
Program, it can the electric energy meter abnormal deviation data examination method that above-mentioned any one embodiment provides is realized, because by original number
According to multiple data sequences are divided into, the sample data training machine of the corresponding characteristic value of each data sequence and intensity of anomaly formation is utilized
Learning algorithm model, the trained machine learning algorithm model of later-stage utilization accurately can determine which abnormal data is present in
In a period, compared with anomaly data detection mode in the prior art, it can accurately determine that abnormal data is present in
In which, and then improve the detection accuracy of abnormal data period.
Above to a kind of electric energy meter abnormal deviation data examination method provided by the present invention, device, equipment and storage medium into
It has gone and has been discussed in detail.With several examples, principle and implementation of the present invention are described herein, above embodiments
Explanation, be merely used to help understand method and its core concept of the invention;Meanwhile for the general technology people of this field
Member, according to the thought of the present invention, there will be changes in the specific implementation manner and application range, in conclusion this explanation
Book content should not be construed as limiting the invention, those skilled in the art, under the premise of no creative work, to this hair
Bright made modification, equivalent replacement, improvement etc., should be included in the application.
It should also be noted that, in the present specification, relational terms such as first and second and the like be used merely to by
One operation is distinguished with another operation, without necessarily requiring or implying there are any between these entities or operation
This actual relationship or sequence.Moreover, the similar word such as term " includes ", so that including the unit of a series of elements, equipment
Or system not only includes those elements, but also including other elements that are not explicitly listed, or further includes for this list
Member, equipment or the intrinsic element of system.
Claims (10)
1. a kind of electric energy meter abnormal deviation data examination method characterized by comprising
Initial data when electric energy meter work is obtained, and segment processing is carried out to obtain to the initial data according to time sequencing
Multiple data sequences;
Determine the corresponding characteristic value of each data sequence and intensity of anomaly to form sample data;
According to the machine learning algorithm model of sample data building anomaly data detection, and to the machine learning algorithm mould
Type is trained;
The electric energy meter data are carried out abnormality detection according to the machine learning algorithm model after training.
2. electric energy meter abnormal deviation data examination method according to claim 1, which is characterized in that described according to time sequencing pair
The initial data carries out segment processing to obtain multiple data sequences specifically:
The initial data is divided into according to the time sequencing using sliding pane mode each data sequence of preset length
Column.
3. electric energy meter abnormal deviation data examination method according to claim 2, which is characterized in that each data of determination
The corresponding characteristic value of sequence specifically:
Each data sequence is analyzed and processed according to the library tsfresh to obtain the corresponding characteristic value.
4. electric energy meter abnormal deviation data examination method according to claim 3, which is characterized in that the characteristic value includes maximum
Value, minimum value, the degree of bias, average value, median and kurtosis.
5. electric energy meter abnormal deviation data examination method according to claim 1, which is characterized in that the machine learning algorithm mould
Type is specially non-supervisory formula machine learning algorithm model.
6. electric energy meter abnormal deviation data examination method according to claim 5, which is characterized in that the institute according to after training
Machine learning algorithm model is stated to carry out abnormality detection the electric energy meter data specifically:
Abnormal inspection is carried out to the electric energy meter data using the k nearest neighbor algorithm in the non-supervisory formula machine learning algorithm model
It surveys.
7. a kind of electric energy meter anomaly data detection device characterized by comprising
Module is obtained, the initial data is carried out for obtaining initial data when electric energy meter work, and according to time sequencing
Segment processing is to obtain multiple data sequences;
Determining module, for determining the corresponding characteristic value of each data sequence and intensity of anomaly to form sample data;
Module is constructed, for the machine learning algorithm model according to sample data building anomaly data detection, and to described
Machine learning algorithm model is trained;
Abnormality detection module is abnormal for being carried out according to the machine learning algorithm model after training to the electric energy meter data
Detection.
8. electric energy meter anomaly data detection device according to claim 7, which is characterized in that the determining module is specifically used
Each data sequence is analyzed and processed according to the library tsfresh to obtain the corresponding characteristic value.
9. a kind of electric energy meter anomaly data detection equipment characterized by comprising
Memory, for storing computer program;
Processor, for executing the computer program to realize that the electric energy meter as described in claim 1 to 6 any one is abnormal
The step of data detection method.
10. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium
Program, the computer program are executed by processor to realize the electric energy meter exception number as described in claim 1 to 6 any one
The step of according to detection method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811139036.2A CN109242041A (en) | 2018-09-28 | 2018-09-28 | A kind of electric energy meter abnormal deviation data examination method, device, equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811139036.2A CN109242041A (en) | 2018-09-28 | 2018-09-28 | A kind of electric energy meter abnormal deviation data examination method, device, equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109242041A true CN109242041A (en) | 2019-01-18 |
Family
ID=65057896
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811139036.2A Pending CN109242041A (en) | 2018-09-28 | 2018-09-28 | A kind of electric energy meter abnormal deviation data examination method, device, equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109242041A (en) |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110398375A (en) * | 2019-07-16 | 2019-11-01 | 广州亚美信息科技有限公司 | Monitoring method, device, equipment and the medium of cooling system of vehicle working condition |
CN110764043A (en) * | 2019-11-12 | 2020-02-07 | 国网四川省电力公司电力科学研究院 | Equipment quality condition evaluation method suitable for continuous measurement results |
CN110849617A (en) * | 2019-11-22 | 2020-02-28 | 深圳市通用互联科技有限责任公司 | Conveyor belt fault detection method and device, computer equipment and storage medium |
CN111563521A (en) * | 2019-02-14 | 2020-08-21 | 埃森哲环球解决方案有限公司 | Site-specific anomaly detection |
CN111563776A (en) * | 2020-05-08 | 2020-08-21 | 国网江苏省电力有限公司扬州供电分公司 | Electric quantity decomposition and prediction method based on K neighbor anomaly detection and Prophet model |
US20200320435A1 (en) * | 2019-04-08 | 2020-10-08 | Sri International | Multi-level introspection framework for explainable reinforcement learning agents |
CN112084237A (en) * | 2020-09-09 | 2020-12-15 | 广东电网有限责任公司中山供电局 | Power system abnormity prediction method based on machine learning and big data analysis |
CN112532467A (en) * | 2019-09-17 | 2021-03-19 | 华为技术有限公司 | Method, device and system for realizing fault detection |
CN113011481A (en) * | 2021-03-10 | 2021-06-22 | 广东电网有限责任公司计量中心 | Electric energy meter function abnormity evaluation method and system based on decision tree algorithm |
CN113052195A (en) * | 2019-12-27 | 2021-06-29 | 北京国双科技有限公司 | Logging curve abnormal value marking method and device, computer equipment and storage medium |
CN113612309A (en) * | 2021-09-30 | 2021-11-05 | 北京志翔科技股份有限公司 | Abnormal electricity utilization data identification method and equipment |
CN113626502A (en) * | 2021-08-13 | 2021-11-09 | 南方电网深圳数字电网研究院有限公司 | Power grid data anomaly detection method and device based on ensemble learning |
CN113722300A (en) * | 2021-07-16 | 2021-11-30 | 国网浙江省电力有限公司营销服务中心 | Electric energy meter operation error calculation data cleaning system and method based on electricity utilization characteristics |
CN115293663A (en) * | 2022-10-10 | 2022-11-04 | 国网山东省电力公司滨州供电公司 | Bus unbalance rate abnormity detection method, system and device |
US11597394B2 (en) | 2018-12-17 | 2023-03-07 | Sri International | Explaining behavior by autonomous devices |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106127249A (en) * | 2016-06-24 | 2016-11-16 | 深圳市颐通科技有限公司 | A kind of single time series exception subsequence detection method |
CN108319981A (en) * | 2018-02-05 | 2018-07-24 | 清华大学 | A kind of time series data method for detecting abnormality and device based on density |
US20180239740A1 (en) * | 2017-02-22 | 2018-08-23 | Sas Institute Inc. | Monitoring, detection, and surveillance system using principal component analysis with machine and sensor data |
-
2018
- 2018-09-28 CN CN201811139036.2A patent/CN109242041A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106127249A (en) * | 2016-06-24 | 2016-11-16 | 深圳市颐通科技有限公司 | A kind of single time series exception subsequence detection method |
US20180239740A1 (en) * | 2017-02-22 | 2018-08-23 | Sas Institute Inc. | Monitoring, detection, and surveillance system using principal component analysis with machine and sensor data |
CN108319981A (en) * | 2018-02-05 | 2018-07-24 | 清华大学 | A kind of time series data method for detecting abnormality and device based on density |
Non-Patent Citations (2)
Title |
---|
张力生: ""时间序列重要点分割的异常子序列检测"", 《计算机科学》 * |
沈海涛: ""电力用户用电数据的异常数据审查和分类"", 《电力与能源》 * |
Cited By (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11597394B2 (en) | 2018-12-17 | 2023-03-07 | Sri International | Explaining behavior by autonomous devices |
CN111563521A (en) * | 2019-02-14 | 2020-08-21 | 埃森哲环球解决方案有限公司 | Site-specific anomaly detection |
CN111563521B (en) * | 2019-02-14 | 2023-10-17 | 埃森哲环球解决方案有限公司 | Site-specific anomaly detection |
US20200320435A1 (en) * | 2019-04-08 | 2020-10-08 | Sri International | Multi-level introspection framework for explainable reinforcement learning agents |
CN110398375B (en) * | 2019-07-16 | 2021-10-19 | 广州亚美信息科技有限公司 | Method, device, equipment and medium for monitoring working state of vehicle cooling system |
CN110398375A (en) * | 2019-07-16 | 2019-11-01 | 广州亚美信息科技有限公司 | Monitoring method, device, equipment and the medium of cooling system of vehicle working condition |
CN112532467A (en) * | 2019-09-17 | 2021-03-19 | 华为技术有限公司 | Method, device and system for realizing fault detection |
CN112532467B (en) * | 2019-09-17 | 2022-12-27 | 华为技术有限公司 | Method, device and system for realizing fault detection |
CN110764043A (en) * | 2019-11-12 | 2020-02-07 | 国网四川省电力公司电力科学研究院 | Equipment quality condition evaluation method suitable for continuous measurement results |
CN110849617A (en) * | 2019-11-22 | 2020-02-28 | 深圳市通用互联科技有限责任公司 | Conveyor belt fault detection method and device, computer equipment and storage medium |
CN113052195A (en) * | 2019-12-27 | 2021-06-29 | 北京国双科技有限公司 | Logging curve abnormal value marking method and device, computer equipment and storage medium |
CN111563776B (en) * | 2020-05-08 | 2021-07-16 | 国网江苏省电力有限公司扬州供电分公司 | Electric quantity decomposition and prediction method based on K neighbor anomaly detection and Prophet model |
CN111563776A (en) * | 2020-05-08 | 2020-08-21 | 国网江苏省电力有限公司扬州供电分公司 | Electric quantity decomposition and prediction method based on K neighbor anomaly detection and Prophet model |
CN112084237A (en) * | 2020-09-09 | 2020-12-15 | 广东电网有限责任公司中山供电局 | Power system abnormity prediction method based on machine learning and big data analysis |
CN113011481A (en) * | 2021-03-10 | 2021-06-22 | 广东电网有限责任公司计量中心 | Electric energy meter function abnormity evaluation method and system based on decision tree algorithm |
CN113722300A (en) * | 2021-07-16 | 2021-11-30 | 国网浙江省电力有限公司营销服务中心 | Electric energy meter operation error calculation data cleaning system and method based on electricity utilization characteristics |
CN113626502A (en) * | 2021-08-13 | 2021-11-09 | 南方电网深圳数字电网研究院有限公司 | Power grid data anomaly detection method and device based on ensemble learning |
CN113612309A (en) * | 2021-09-30 | 2021-11-05 | 北京志翔科技股份有限公司 | Abnormal electricity utilization data identification method and equipment |
CN115293663A (en) * | 2022-10-10 | 2022-11-04 | 国网山东省电力公司滨州供电公司 | Bus unbalance rate abnormity detection method, system and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109242041A (en) | A kind of electric energy meter abnormal deviation data examination method, device, equipment and storage medium | |
CN108593990B (en) | Electricity stealing detection method based on electricity consumption behavior mode of electric energy user and application | |
CN103226736B (en) | Based on the long-medium term power load forecasting method of cluster analysis and grey target theory | |
CN103020459B (en) | A kind of cognitive method of various dimensions electricity consumption behavior and system | |
CN104166731A (en) | Discovering system for social network overlapped community and method thereof | |
CN105023066B (en) | A kind of Business Process System analysing and predicting system and method based on seasonal adjustment | |
CN109285087A (en) | A kind of platform area topology identification method accelerated based on NB-IoT and GPU | |
CN102467684A (en) | System and method for evaluating comprehensive performance based on improved radar chart | |
CN103559303A (en) | Evaluation and selection method for data mining algorithm | |
Fidalgo et al. | A new clustering algorithm for load profiling based on billing data | |
CN108596369B (en) | Space load prediction method based on multivariate model and blind number theory | |
CN110400231A (en) | A kind of electric energy measuring equipment crash rate predictor method weighting Nonlinear Bayesian | |
CN106786524A (en) | Load model parameters discrimination method based on noise-like signal and improved differential evolution | |
CN102012522A (en) | Method and device for evaluating seismic acquisition data | |
CN105005711A (en) | Method and device for acquiring statistical line loss | |
Li et al. | Real-time flood simulations using CA model driven by dynamic observation data | |
Meier et al. | Power system data management and analysis using synchrophasor data | |
Boričić et al. | Comprehensive review of short-term voltage stability evaluation methods in modern power systems | |
CN105824987A (en) | Wind field characteristic statistical distributing model building method based on genetic algorithm | |
Lin et al. | Planning life tests with progressively Type-I interval censored data from the lognormal distribution | |
CN102156641A (en) | Prediction method and system for confidence interval of software cost | |
CN106980874A (en) | A kind of multi-time Scales dimension data fusion method towards distribution big data | |
Hiemer et al. | Improved seismicity forecast with spatially varying magnitude distribution | |
Peng et al. | A comprehensive stability evaluation method of multiple salt caverns underground gas storage with interlayers | |
JP6135454B2 (en) | Estimation program, estimation apparatus, and estimation method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190118 |
|
RJ01 | Rejection of invention patent application after publication |