CN117436017A - Abnormality early warning method and device for electric energy quality monitoring device - Google Patents
Abnormality early warning method and device for electric energy quality monitoring device Download PDFInfo
- Publication number
- CN117436017A CN117436017A CN202311413186.9A CN202311413186A CN117436017A CN 117436017 A CN117436017 A CN 117436017A CN 202311413186 A CN202311413186 A CN 202311413186A CN 117436017 A CN117436017 A CN 117436017A
- Authority
- CN
- China
- Prior art keywords
- feature
- differential
- matrix
- data
- map
- 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
- 230000005856 abnormality Effects 0.000 title claims abstract description 57
- 238000012806 monitoring device Methods 0.000 title claims abstract description 52
- 238000000034 method Methods 0.000 title claims abstract description 41
- 239000011159 matrix material Substances 0.000 claims abstract description 104
- 238000012544 monitoring process Methods 0.000 claims abstract description 74
- 238000001228 spectrum Methods 0.000 claims abstract description 74
- 239000013598 vector Substances 0.000 claims abstract description 52
- 238000000605 extraction Methods 0.000 claims description 44
- 230000004927 fusion Effects 0.000 claims description 28
- 238000011176 pooling Methods 0.000 claims description 13
- 230000035772 mutation Effects 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 7
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 230000000694 effects Effects 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 230000002159 abnormal effect Effects 0.000 claims 1
- 238000012423 maintenance Methods 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 9
- 230000006870 function Effects 0.000 description 7
- 238000004590 computer program Methods 0.000 description 6
- 238000007405 data analysis Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 230000036541 health Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- G06F18/211—Selection of the most significant subset of features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2323—Non-hierarchical techniques based on graph theory, e.g. minimum spanning trees [MST] or graph cuts
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Economics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- Entrepreneurship & Innovation (AREA)
- Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Discrete Mathematics (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention provides an abnormality early warning method and device for an electric energy quality monitoring device, and belongs to the technical field of electric energy monitoring equipment maintenance. The method comprises the steps of firstly, acquiring a plurality of monitoring data sets; then extracting a plurality of feature vectors according to the plurality of monitoring data sets; then constructing a feature matrix according to the feature vectors, and extracting a first feature map according to the feature matrix; and finally classifying the first characteristic spectrum, and determining an abnormality index of the power quality monitoring device according to the position of the first characteristic spectrum in the class. The embodiment of the invention finds out the position of the characteristic map relative to the whole class in a classifying mode, determines the abnormality index based on the position deviation, and outputs the early warning message when the abnormality index exceeds the threshold value.
Description
Technical Field
The invention belongs to the technical field of maintenance of electric energy monitoring equipment, and particularly relates to an abnormality early warning method and device for an electric energy quality monitoring device.
Background
The power quality monitoring device is a special device for measuring power quality related parameters by introducing voltage and current signals, and is generally used for measuring and analyzing the alternating current power quality supplied by a public power grid to a user side, and comprises the following components: frequency deviation, voltage fluctuation and flicker, and three-phase voltage allow for unbalance, grid harmonics.
The power quality monitoring device generally needs to be periodically calibrated to ensure the stability and reliability of its operation and the accuracy of data acquisition. In routine maintenance, the accuracy of the power quality monitoring device data is only noticeable if a significant deviation is found. In other words, through general daily checks, early anomalies in the power quality monitoring device are not easily detected.
In the prior art, one technical scheme is to mine early abnormality of the power quality monitoring device through data analysis. Application number: CN202010677313.6, patent name: the invention discloses an intelligent grading and automatic early warning method for the health condition of power equipment, which is established based on an electric energy quality monitoring platform. Based on the power quality monitoring platform, the multi-dimensional mapping of the equipment monitoring data is realized by comprehensively utilizing the data of each monitoring device in the system from the global angle through a correlation analysis method, the health rating of the running state of the power equipment is carried out and early warning information is formed by combining with the essential correlation reasoning technology of the power equipment, and finally the integrated platform construction of information acquisition, data analysis, relation reasoning and power equipment health diagnosis of the running state of the power equipment is realized, so that the problem that the health state evaluation technology of the power equipment lacks information mining of the power quality monitoring data and cannot meet the current power grid development requirements is solved.
The patent needs to build a general mathematical model for evaluating the health condition of the power equipment, the building process is complex, the application range is narrow, and the application scene is limited.
Based on the above, it is necessary to develop an abnormality early warning method for the power quality monitoring device.
Disclosure of Invention
The embodiment of the invention provides an abnormality early warning method and device for an electric energy quality monitoring device, which are used for solving the problem that the method for mining early abnormality of the electric energy quality monitoring device by data analysis in the prior art is narrow in application range.
In a first aspect, an embodiment of the present invention provides an anomaly early warning method for a power quality monitoring device, including:
acquiring a plurality of monitoring data sets, wherein the monitoring data sets are constructed based on the monitoring data acquired by the plurality of time nodes;
extracting a plurality of feature vectors according to the plurality of monitoring data sets, wherein the feature vectors represent the fluctuation of the monitoring data;
constructing a feature matrix according to the plurality of feature vectors, and extracting a first feature map according to the feature matrix, wherein the first feature map represents data mutation contained in the feature matrix;
classifying the first characteristic spectrum, and determining an abnormality index of the power quality monitoring device according to the position of the first characteristic spectrum in the class.
In one possible implementation manner, the extracting a plurality of feature vectors according to the plurality of monitoring data sets includes:
for each of the plurality of monitored data sets, performing the steps of:
acquiring a basic period, wherein the basic period is determined according to a basic activity rule of a monitored point;
determining a plurality of frequency domain features according to the basic period and the monitoring data set, wherein the frequency domain features represent feature values of frequency characteristics contained in the monitoring data set;
constructing a first feature vector according to the plurality of frequency domain features;
and taking the unit vector of the first feature vector as a feature vector.
In a possible implementation manner, the determining a plurality of frequency domain features according to the basic period and the monitoring data set includes:
determining a plurality of frequency domain features according to a first formula, the fundamental period and a monitoring dataset, wherein the first formula is:
wherein fs (k) is the kth sinusoidal feature, idata (iN) is the iN-th data of the monitored dataset, iN is the total number of data iN the monitored dataset, k is the frequency domain number, ω 0 For the frequency corresponding to the fundamental period, tn is the sampling number of the monitored data in the fundamental period, fc (k) is the kth cosine feature, sin () is a sine function, cos () is a cosine function, and f (k) is the kth frequency domain feature.
In one possible implementation manner, the constructing a feature matrix according to the feature vectors, and extracting a first feature map according to the feature matrix includes:
obtaining a plurality of differential feature extraction operator templates;
constructing the plurality of feature vectors into a feature matrix according to a preset monitoring quantity sequence;
performing differential feature extraction on the feature matrix according to the plurality of differential feature extraction operator templates to obtain a plurality of differential feature maps;
fusing the differential feature patterns to obtain a fused feature pattern;
and pooling the fusion characteristic spectrum to obtain the first characteristic spectrum.
In one possible implementation manner, the performing differential feature extraction on the feature matrix according to the plurality of differential feature extraction operator templates to obtain a plurality of differential feature maps includes:
for each of the plurality of differential feature extraction operator templates, performing the steps of:
acquiring a position indication;
according to the position indication, a data block with the same type as the difference feature extraction operator template is taken out from the feature matrix;
extracting differential features according to the second formula, the differential feature extraction operator template and the data block, wherein the second formula is as follows:
wherein FDF is differential feature, mobile rn,ln Extracting operator templates for differential features line-first and column-secondElement, dblock rn,ln rN is the number of rows of the differential feature extraction operator templates for the elements of the first row and the first column of the data block, and lN is the number of columns of the differential feature extraction operator templates;
according to the position indication, the differential features are put into an atlas matrix;
if the position indication does not reach the end of the feature matrix, shifting the position indication according to a preset sequence;
otherwise, taking the map matrix as a differential characteristic map.
In one possible implementation manner, the fusing the plurality of differential feature maps to obtain a fused feature map includes:
setting the numerical values smaller than a differential threshold value in the differential characteristic maps to zero;
calculating the sum spectrum of the differential characteristic spectrums after zero setting as a primary fusion spectrum;
taking a plurality of non-zero data with the distance from the neighbor data being greater than an island distance threshold value in the initial fusion map as a plurality of island data, wherein the neighbor data is the non-zero data with the nearest distance;
deleting the island data from the primary fusion map, and taking the primary fusion map after deleting the island data as a fusion characteristic map.
In one possible implementation, classifying the first feature map includes:
obtaining a plurality of feature map samples, wherein the plurality of feature map samples are classified into a plurality of classes based on a kmeans clustering algorithm;
according to a third formula, calculating distances between the first characteristic spectrum and the plurality of characteristic spectrum samples respectively, wherein the third formula is as follows:
wherein DIS is the distance between the first characteristic spectrum and the characteristic spectrum sample, SM mrn,mln FM is the element in mln th column of mrn th row of characteristic spectrum sample mrn,mln The elements of the first characteristic spectrum in the mrn th row and mlN th column are taken as elements, mrN is the number of rows of the first characteristic spectrum, and mlN is the number of columns of the first characteristic spectrum;
selecting a characteristic spectrum sample with the smallest distance from the first characteristic spectrum as a target sample;
and classifying the first characteristic map into the class where the target sample is located.
In a possible implementation manner, the determining the abnormality index of the power quality monitoring device according to the position of the first feature map in the class includes:
obtaining a plurality of class samples, wherein the class samples are characteristic spectrum samples of the class in which the first characteristic spectrum is located;
respectively calculating differences between the plurality of class samples and the first characteristic map to obtain a plurality of difference matrixes;
calculating the sum of the plurality of difference matrixes to obtain a sum matrix;
determining an offset coefficient matrix according to the sum matrix, the first characteristic map and a fourth formula, wherein the fourth formula is as follows:
in BM mrn,mln AM for shifting elements of row mrn, column mln of the coefficient matrix mrn,min Mln elements of row mrn and column mln of the matrix;
and determining an abnormality index of the power quality monitoring device according to the sum matrix.
In one possible implementation, determining an abnormality index of the power quality monitoring device according to the sum matrix includes:
acquiring the maximum value of the sum matrix median as a deviation extremum;
calculating the average value of a plurality of elements in the sum matrix as an average value;
and determining an abnormality index of the power quality monitoring device according to the deviation extreme value and the average value.
In a second aspect, an embodiment of the present invention provides an abnormality early warning device for a power quality monitoring device, including:
the data acquisition module is used for acquiring a plurality of monitoring data sets, wherein the monitoring data sets are constructed based on the monitoring data acquired by the plurality of time nodes;
the feature extraction module is used for extracting a plurality of feature vectors according to the plurality of monitoring data sets, wherein the feature vectors represent the fluctuation of the monitoring data;
the map construction module is used for constructing a feature matrix according to the plurality of feature vectors and extracting a first feature map according to the feature matrix, wherein the first feature map represents data mutation contained in the feature matrix;
the method comprises the steps of,
the abnormality index determining module is used for classifying the first characteristic patterns and determining abnormality indexes of the power quality monitoring device according to the positions of the first characteristic patterns in the class.
Compared with the prior art, the invention has the beneficial effects that:
the embodiment of the invention discloses an abnormality early warning method and device of an electric energy quality monitoring device, which comprises the steps of firstly acquiring a plurality of monitoring data sets, wherein the monitoring data sets are constructed based on monitoring data acquired by a plurality of time nodes; then extracting a plurality of feature vectors according to the plurality of monitoring data sets, wherein the feature vectors represent the fluctuation of the monitoring data; then constructing a feature matrix according to the feature vectors, and extracting a first feature map according to the feature matrix, wherein the first feature map represents data mutation contained in the feature matrix; and finally classifying the first characteristic spectrum, and determining an abnormality index of the power quality monitoring device according to the position of the first characteristic spectrum in the class.
According to the embodiment of the invention, the data analysis is performed based on the characteristics of the monitoring data set, compared with a mode of analyzing batch data, the data calculation cost is low, the characteristics are obvious, and the data abnormality can be found conveniently.
According to the embodiment of the invention, through differential extraction, fusion, island data removal and pooling operations, the constructed feature matrix is converted into a matrix with various differential feature fusion, further remarkable feature boundaries and singular data removal, and after pooling operations, the data size is further reduced, so that the subsequent data processing process is simplified.
The embodiment of the invention finds out the position of the characteristic map relative to the whole class in a classifying mode, determines the abnormality index based on the position deviation, and outputs the early warning message when the abnormality index exceeds the threshold value.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments or the description of the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of an implementation of an abnormality early warning method of an electric energy quality monitoring device according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a process for extracting a data block from a feature matrix according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a feature map classification process according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an abnormality early warning device of an electric energy quality monitoring device according to an embodiment of the present invention.
Detailed Description
In the following description, for purposes of explanation and not limitation, specific details are set forth such as the particular system architecture, techniques, etc., in order to provide a thorough understanding of the embodiments of the present invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present invention with unnecessary detail.
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the following description will be made by way of specific embodiments with reference to the accompanying drawings.
Fig. 1 is a flowchart of an implementation of an anomaly early warning method of an electric energy quality monitoring device according to an embodiment of the present invention, which is described in detail below:
in step 101, a plurality of monitoring data sets are acquired, wherein the monitoring data sets are constructed based on the monitoring data acquired by the plurality of time nodes.
Step 102, extracting a plurality of feature vectors according to the plurality of monitoring data sets, wherein the feature vectors represent the fluctuation of the monitoring data.
In some embodiments, step 102 comprises:
for each of the plurality of monitored data sets, performing the steps of:
acquiring a basic period, wherein the basic period is determined according to a basic activity rule of a monitored point;
determining a plurality of frequency domain features according to the basic period and the monitoring data set, wherein the frequency domain features represent feature values of frequency characteristics contained in the monitoring data set;
constructing a first feature vector according to the plurality of frequency domain features;
and taking the unit vector of the first feature vector as a feature vector.
In some embodiments, the determining a plurality of frequency domain features from the fundamental period and the monitoring dataset includes:
determining a plurality of frequency domain features according to a first formula, the fundamental period and a monitoring dataset, wherein the first formula is:
where fs (k) is the kth sinusoidal feature and idata (in) is the monitored datasetiN data, iN is total data iN the monitoring data set, k is frequency domain times and omega 0 For the frequency corresponding to the fundamental period, tn is the sampling number of the monitored data in the fundamental period, fc (k) is the kth cosine feature, sin () is a sine function, cos () is a cosine function, and f (k) is the kth frequency domain feature.
Illustratively, the power quality monitoring device is configured to detect a parameter related to power quality, and includes: frequency deviation, voltage fluctuation and flicker, three-phase voltage allowable unbalance, grid harmonics, which data, if collected according to a predetermined time node, can be formed into data sequences, e.g. frequency deviation data sequences, voltage deviation data sequences, etc., which data sequences are cut out according to a predetermined time length, e.g. data cut out at a length of each month, a data set, e.g. frequency deviation data set, voltage deviation data set, etc., is obtained.
Because the density of the electric energy quality monitoring data acquisition is high, the data obtained in a preset time period are large and messy, and the embodiment of the invention adopts a mode of extracting the characteristic vector of the monitoring data set to analyze the monitoring data set.
First, a basic period is determined, which is related to the activity law of the monitored object, for example, for electricity for life, the fluctuation law thereof is generally related to the natural day, and thus, the period can be determined as the length of the natural day.
Then, according to this basic period, a plurality of frequency domain features are extracted, a plurality of frequency domain feature construction vectors are constructed, and finally, unit vectors of this construction vector are utilized as feature vectors.
In the embodiment of the invention, two features are extracted by adopting Fourier transformation: the sine feature and the cosine feature are fused, the final feature is obtained, and a first formula is applied:
where fs (k) is the kth sinusoidal feature,idata (iN) is the iN-th data of the monitoring dataset, iN is the total number of data iN the monitoring dataset, k is the frequency domain number, ω 0 For the frequency corresponding to the fundamental period, tn is the sampling number of the monitored data in the fundamental period, fc (k) is the kth cosine feature, sin () is a sine function, cos () is a cosine function, and f (k) is the kth frequency domain feature.
The method and the device perform data analysis based on the characteristics of the monitoring data set, have low data calculation cost and obvious characteristics compared with a mode of analyzing batch data, and are convenient for finding out the abnormality of the data.
And 103, constructing a feature matrix according to the feature vectors, and extracting a first feature map according to the feature matrix, wherein the first feature map represents the data mutation contained in the feature matrix.
In some embodiments, step 103 comprises:
obtaining a plurality of differential feature extraction operator templates;
constructing the plurality of feature vectors into a feature matrix according to a preset monitoring quantity sequence;
performing differential feature extraction on the feature matrix according to the plurality of differential feature extraction operator templates to obtain a plurality of differential feature maps;
fusing the differential feature patterns to obtain a fused feature pattern;
and pooling the fusion characteristic spectrum to obtain the first characteristic spectrum.
In some embodiments, the extracting the differential features of the feature matrix according to the plurality of differential feature extraction operator templates to obtain a plurality of differential feature maps includes:
for each of the plurality of differential feature extraction operator templates, performing the steps of:
acquiring a position indication;
according to the position indication, a data block with the same type as the difference feature extraction operator template is taken out from the feature matrix;
extracting differential features according to the second formula, the differential feature extraction operator template and the data block, wherein the second formula is as follows:
wherein FDF is differential feature, mobile rn,ln Extracting elements of a first row and a first column of an operator template for differential features, and Block rn,ln rN is the number of rows of the differential feature extraction operator templates for the elements of the first row and the first column of the data block, and lN is the number of columns of the differential feature extraction operator templates;
according to the position indication, the differential features are put into an atlas matrix;
if the position indication does not reach the end of the feature matrix, shifting the position indication according to a preset sequence;
otherwise, taking the map matrix as a differential characteristic map.
In some embodiments, the fusing the plurality of differential feature maps to obtain a fused feature map includes:
setting the numerical values smaller than a differential threshold value in the differential characteristic maps to zero;
calculating the sum spectrum of the differential characteristic spectrums after zero setting as a primary fusion spectrum;
taking a plurality of non-zero data with the distance from the neighbor data being greater than an island distance threshold value in the initial fusion map as a plurality of island data, wherein the neighbor data is the non-zero data with the nearest distance;
deleting the island data from the primary fusion map, and taking the primary fusion map after deleting the island data as a fusion characteristic map.
For example, the feature vectors obtained in the embodiment of the present invention correspond to a plurality of monitoring parameters, such as frequency deviation, voltage fluctuation and flicker, three-phase voltage allowable unbalance, and grid harmonics, and in order to organically combine these parameters, a feature matrix is constructed by combining these feature vectors, and abrupt change data in the feature matrix is extracted through a differential operator template to perform the salification.
In some embodiments, the operator templates employ differential operator templates, e.g., horizontal differential operator templates:
vertical differential operator templates:
the templates perform differential feature extraction on the feature matrix, and the obtained atlas is further fused and pooled to obtain the expected feature atlas.
The differential feature extraction process is to extract a data block from the feature matrix, operate with the differential feature matrix, extract differential features, and apply the formula to express:
wherein FDF is differential feature, mobile rn,ln Extracting elements of a first row and a first column of an operator template for differential features, and Dblock rn,ln And as for the elements of the first row and the first column of the data block, rN is the number of rows of the differential feature extraction operator template, and lN is the number of columns of the differential feature extraction operator template.
Fig. 2 shows a process of extracting data from the feature matrix, in fig. 2, starting from the upper left corner of the feature matrix 201, extracting a data block 205, calculating by the above formula, extracting a differential feature, and placing the differential feature into the map matrix, and then shifting the extraction position 203 to the right of the home position 202 (if the rightmost position is reached, the line is fed to the leftmost position of the next line) in a manner of shifting one data unit at a time, and continuing to extract a new data block until the extraction position 203 reaches the lower right corner position of the feature matrix 201.
As for the differential feature patterns, as described above, a plurality of differential feature patterns are obtained according to the types and the numbers of the differential templates, the patterns are firstly subjected to boundary saliency, namely, the numerical value lower than the differential threshold value is set to zero, then, the differential feature patterns after zero setting are summed to obtain primary fusion patterns, the primary fusion patterns comprise some singular data, and island data are deleted in an island discrimination mode. Island data refers to data which is farther from other non-zero data at the position of the matrix among the non-zero data. And deleting the primary fusion map after island data to obtain a fusion characteristic map.
Pooling refers to reducing and making remarkable an atlas matrix, wherein a data block is extracted from the atlas matrix, the maximum value or the average value of the data block is selected as a pooling value, the pooling value is added into the pooled atlas matrix, and the operations of extraction and value extraction are repeated, so that the pooling operation is completed.
After the differential extraction, fusion, island data removal and pooling operation, the constructed feature matrix is converted into a matrix with various differential feature fusion, further remarkable feature boundaries and singular data removal, and after the pooling operation, the data size is further reduced, so that the subsequent data processing process is simplified.
Step 104, classifying the first characteristic spectrum, and determining an abnormality index of the power quality monitoring device according to the position of the first characteristic spectrum in the class.
In some embodiments, a plurality of feature map samples are obtained, wherein the plurality of feature map samples are classified into a plurality of classes based on a kmeans clustering algorithm;
according to a third formula, calculating distances between the first characteristic spectrum and the plurality of characteristic spectrum samples respectively, wherein the third formula is as follows:
wherein DIS is a first characteristic spectrum anddistance of feature pattern sample, SM mrn,mln FM is the element in mln th column of mrn th row of characteristic spectrum sample mrn,mln The elements of the first characteristic spectrum in the mrn th row and mlN th column are taken as elements, mrN is the number of rows of the first characteristic spectrum, and mlN is the number of columns of the first characteristic spectrum;
selecting a characteristic spectrum sample with the smallest distance from the first characteristic spectrum as a target sample;
and classifying the first characteristic map into the class where the target sample is located.
In some embodiments, the determining the abnormality index of the power quality monitoring device according to the position of the first feature map in the class includes:
obtaining a plurality of class samples, wherein the class samples are characteristic spectrum samples of the class in which the first characteristic spectrum is located;
respectively calculating differences between the plurality of class samples and the first characteristic map to obtain a plurality of difference matrixes;
calculating the sum of the plurality of difference matrixes to obtain a sum matrix;
determining an offset coefficient matrix according to the sum matrix, the first characteristic map and a fourth formula, wherein the fourth formula is as follows:
in BM mrn,mln AM for shifting elements of row mrn, column mln of the coefficient matrix mrn,mln Mln elements of row mrn and column mln of the matrix;
and determining an abnormality index of the power quality monitoring device according to the sum matrix.
In some embodiments, the determining the abnormality index of the power quality monitoring device based on the sum matrix includes:
acquiring the maximum value of the sum matrix median as a deviation extremum;
calculating the average value of a plurality of elements in the sum matrix as an average value;
and determining an abnormality index of the power quality monitoring device according to the deviation extreme value and the average value.
Illustratively, for the atlas categorization aspect, embodiments of the present invention first obtain a plurality of feature atlas samples. As shown in fig. 3, these feature pattern samples 301 are divided into a plurality of classes 302 by kmeans clustering algorithm, and feature patterns obtained by the signing step find the feature pattern sample closest to the feature pattern sample through the distance, and add the feature pattern to the class where the feature pattern sample closest to the feature pattern sample is located. There are various distance calculation modes, and in the embodiment of the invention, a calculation method shown by a third formula is adopted:
wherein DIS is the distance between the first characteristic spectrum and the characteristic spectrum sample, SM mrn,mln FM is the element in mln th column of mrn th row of characteristic spectrum sample mrn,mln The elements in the mlN th column of the mrn th row of the first characteristic spectrum, mrN is the number of rows of the first characteristic spectrum, and mlN is the number of columns of the first characteristic spectrum.
After classifying the first feature map, determining the position of the first feature map in the class, and outputting an abnormality early warning index according to the position of the first feature map in the class.
In one embodiment, the location of the first feature map includes two aspects: the position of the first feature pattern relative to the class whole and the maximum offset of a certain dimension of the first feature pattern relative to the whole.
The position of the first characteristic spectrum relative to the whole class is to calculate the difference between the first characteristic spectrum and other class samples, add the differences to obtain a sum matrix, and finally calculate an offset coefficient by using a fourth formula to obtain an offset coefficient matrix, wherein the fourth formula is as follows:
in BM mrn,mln For a matrix of offset coefficientsMrn, mln column element, AM mrn,mln Is the element of mln th column of the matrix mrn th row.
The maximum value in the offset coefficient matrix is the deviation extremum, that is, the maximum offset of a certain dimension of the first characteristic spectrum relative to the whole. And the average value of all elements in the offset coefficient matrix reflects the position of the first characteristic map relative to the integral class.
Using the mean and the deviation extremum, an abnormality index is determined, for example, in some scenarios the output index is according to the following formula:
wherein IND is an abnormality index, AVG is a deviation average value, EXT is a deviation extremum.
Firstly, acquiring a plurality of monitoring data sets, wherein the monitoring data sets are constructed based on monitoring data acquired by a plurality of time nodes; then extracting a plurality of feature vectors according to the plurality of monitoring data sets, wherein the feature vectors represent the fluctuation of the monitoring data; then constructing a feature matrix according to the feature vectors, and extracting a first feature map according to the feature matrix, wherein the first feature map represents data mutation contained in the feature matrix; and finally classifying the first characteristic spectrum, and determining an abnormality index of the power quality monitoring device according to the position of the first characteristic spectrum in the class.
According to the embodiment of the invention, the data analysis is performed based on the characteristics of the monitoring data set, compared with a mode of analyzing batch data, the data calculation cost is low, the characteristics are obvious, and the data abnormality can be found conveniently.
According to the embodiment of the invention, through differential extraction, fusion, island data removal and pooling operations, the constructed feature matrix is converted into a matrix with various differential feature fusion, further remarkable feature boundaries and singular data removal, and after pooling operations, the data size is further reduced, so that the subsequent data processing process is simplified.
The embodiment of the invention finds out the position of the characteristic map relative to the whole class in a classifying mode, determines the abnormality index based on the position deviation, and outputs the early warning message when the abnormality index exceeds the threshold value.
It should be understood that the sequence number of each step in the foregoing embodiment does not mean that the execution sequence of each process should be determined by the function and the internal logic, and should not limit the implementation process of the embodiment of the present invention.
The following are device embodiments of the invention, for details not described in detail therein, reference may be made to the corresponding method embodiments described above.
Fig. 4 is a schematic structural diagram of an abnormality pre-warning device for a power quality monitoring device according to an embodiment of the present invention, and for convenience of explanation, only a portion related to the embodiment of the present invention is shown, which is described in detail below:
as shown in fig. 4, the power quality monitoring device abnormality warning device 4 includes: a data acquisition module 401, a feature extraction module 402, a map construction module 403, and an abnormality index determination module 404, wherein:
a data acquisition module 401, configured to acquire a plurality of monitoring data sets, where the monitoring data sets are constructed based on monitoring data acquired by a plurality of time nodes;
a feature extraction module 402, configured to extract a plurality of feature vectors according to the plurality of monitoring data sets, where the feature vectors characterize volatility of the monitoring data;
a spectrum construction module 403, configured to construct a feature matrix according to the plurality of feature vectors, and extract a first feature spectrum according to the feature matrix, where the first feature spectrum characterizes a data mutation contained in the feature matrix;
the abnormality index determination module 404 is configured to categorize the first feature map, and determine an abnormality index of the power quality monitoring device according to a position of the first feature map in the category.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments.
Those of ordinary skill in the art will appreciate that the templates, elements, and algorithm steps described in connection with the embodiments disclosed herein can be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
The modules/units, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. Based on such understanding, the present invention may implement all or part of the procedures in the above-described embodiments of the method, or may be implemented by instructing the relevant hardware by a computer program, where the computer program may be stored in a computer readable storage medium, and the computer program may implement the steps of the above-described embodiments of the abnormality warning method for the power quality monitoring device when executed by a processor. Wherein the computer program comprises computer program code which may be in source code form, object code form, executable file or some intermediate form etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash disk, a removable hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, an electrical carrier wave signal, a telecommunication signal, a software distribution medium, and so forth.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention, and are intended to be included in the scope of the present invention.
Claims (10)
1. An anomaly early warning method for a power quality monitoring device is characterized by comprising the following steps:
acquiring a plurality of monitoring data sets, wherein the monitoring data sets are constructed based on the monitoring data acquired by the plurality of time nodes;
extracting a plurality of feature vectors according to the plurality of monitoring data sets, wherein the feature vectors represent the fluctuation of the monitoring data;
constructing a feature matrix according to the plurality of feature vectors, and extracting a first feature map according to the feature matrix, wherein the first feature map represents data mutation contained in the feature matrix;
classifying the first characteristic spectrum, and determining an abnormality index of the power quality monitoring device according to the position of the first characteristic spectrum in the class.
2. The method for pre-warning of anomalies in a power quality monitoring device of claim 1, wherein extracting a plurality of feature vectors from the plurality of monitored data sets comprises:
for each of the plurality of monitored data sets, performing the steps of:
acquiring a basic period, wherein the basic period is determined according to a basic activity rule of a monitored point;
determining a plurality of frequency domain features according to the basic period and the monitoring data set, wherein the frequency domain features represent feature values of frequency characteristics contained in the monitoring data set;
constructing a first feature vector according to the plurality of frequency domain features;
and taking the unit vector of the first feature vector as a feature vector.
3. The method of claim 2, wherein determining a plurality of frequency domain features from the base cycle and the monitoring data set comprises:
determining a plurality of frequency domain features according to a first formula, the fundamental period and a monitoring dataset, wherein the first formula is:
wherein fs (k) is the kth sinusoidal feature, idata (iN) is the iN-th data of the monitored dataset, iN is the total number of data iN the monitored dataset, k is the frequency domain number, ω 0 For the frequency corresponding to the fundamental period, tn is the sampling number of the monitored data in the fundamental period, fc (k) is the kth cosine feature, sin () is a sine function, cos () is a cosine function, and f (k) is the kth frequency domain feature.
4. The method for pre-warning of abnormality of a power quality monitoring device according to claim 1, wherein the constructing a feature matrix from the plurality of feature vectors and extracting a first feature map from the feature matrix includes:
obtaining a plurality of differential feature extraction operator templates;
constructing the plurality of feature vectors into a feature matrix according to a preset monitoring quantity sequence;
performing differential feature extraction on the feature matrix according to the plurality of differential feature extraction operator templates to obtain a plurality of differential feature maps;
fusing the differential feature patterns to obtain a fused feature pattern;
and pooling the fusion characteristic spectrum to obtain the first characteristic spectrum.
5. The method for pre-warning of abnormality of a power quality monitoring device according to claim 4, wherein the performing differential feature extraction on the feature matrix according to the plurality of differential feature extraction operator templates to obtain a plurality of differential feature maps includes:
for each of the plurality of differential feature extraction operator templates, performing the steps of:
acquiring a position indication;
according to the position indication, a data block with the same type as the difference feature extraction operator template is taken out from the feature matrix;
extracting differential features according to the second formula, the differential feature extraction operator template and the data block, wherein the second formula is as follows:
wherein FDF is differential feature, mobile rn,ln Extracting elements of a first row and a first column of an operator template for differential features, and Dblock rn,ln rN is the number of rows of the differential feature extraction operator templates for the elements of the first row and the first column of the data block, and lN is the number of columns of the differential feature extraction operator templates;
according to the position indication, the differential features are put into an atlas matrix;
if the position indication does not reach the end of the feature matrix, shifting the position indication according to a preset sequence;
otherwise, taking the map matrix as a differential characteristic map.
6. The method for pre-warning of abnormality of a power quality monitoring device according to claim 4, wherein the fusing the plurality of differential feature maps to obtain a fused feature map includes:
setting the numerical values smaller than a differential threshold value in the differential characteristic maps to zero;
calculating the sum spectrum of the differential characteristic spectrums after zero setting as a primary fusion spectrum;
taking a plurality of non-zero data with the distance from the neighbor data being greater than an island distance threshold value in the initial fusion map as a plurality of island data, wherein the neighbor data is the non-zero data with the nearest distance;
deleting the island data from the primary fusion map, and taking the primary fusion map after deleting the island data as a fusion characteristic map.
7. The method of any one of claims 1-6, wherein classifying the first signature comprises:
obtaining a plurality of feature map samples, wherein the plurality of feature map samples are classified into a plurality of classes based on a kmeans clustering algorithm;
according to a third formula, calculating distances between the first characteristic spectrum and the plurality of characteristic spectrum samples respectively, wherein the third formula is as follows:
wherein DIS is the distance between the first characteristic spectrum and the characteristic spectrum sample, SM mrn,mln FM is the element in mln th column of mrn th row of characteristic spectrum sample mrn,mln The elements of the first characteristic spectrum in the mrn th row and mlN th column are taken as elements, mrN is the number of rows of the first characteristic spectrum, and mlN is the number of columns of the first characteristic spectrum;
selecting a characteristic spectrum sample with the smallest distance from the first characteristic spectrum as a target sample;
and classifying the first characteristic map into the class where the target sample is located.
8. The method for pre-warning of abnormality of a power quality monitoring device according to claim 7, wherein determining an abnormality index of the power quality monitoring device based on the position of the first characteristic map in the class includes:
obtaining a plurality of class samples, wherein the class samples are characteristic spectrum samples of the class in which the first characteristic spectrum is located;
respectively calculating differences between the plurality of class samples and the first characteristic map to obtain a plurality of difference matrixes;
calculating the sum of the plurality of difference matrixes to obtain a sum matrix;
determining an offset coefficient matrix according to the sum matrix, the first characteristic map and a fourth formula, wherein the fourth formula is as follows:
in BM mrn,mln AM for shifting elements of row mrn, column mln of the coefficient matrix mrn,mln Mln elements of row mrn and column mln of the matrix;
and determining an abnormality index of the power quality monitoring device according to the sum matrix.
9. The power quality monitoring device abnormality pre-warning method according to claim 8, characterized in that said determining an abnormality index of the power quality monitoring device from the sum matrix includes:
acquiring the maximum value of the sum matrix median as a deviation extremum;
calculating the average value of a plurality of elements in the sum matrix as an average value;
and determining an abnormality index of the power quality monitoring device according to the deviation extreme value and the average value.
10. An abnormal early warning device of an electric energy quality monitoring device, which is characterized by comprising:
the data acquisition module is used for acquiring a plurality of monitoring data sets, wherein the monitoring data sets are constructed based on the monitoring data acquired by the plurality of time nodes;
the feature extraction module is used for extracting a plurality of feature vectors according to the plurality of monitoring data sets, wherein the feature vectors represent the fluctuation of the monitoring data;
the map construction module is used for constructing a feature matrix according to the plurality of feature vectors and extracting a first feature map according to the feature matrix, wherein the first feature map represents data mutation contained in the feature matrix;
the method comprises the steps of,
the abnormality index determining module is used for classifying the first characteristic patterns and determining abnormality indexes of the power quality monitoring device according to the positions of the first characteristic patterns in the class.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311413186.9A CN117436017A (en) | 2023-10-27 | 2023-10-27 | Abnormality early warning method and device for electric energy quality monitoring device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311413186.9A CN117436017A (en) | 2023-10-27 | 2023-10-27 | Abnormality early warning method and device for electric energy quality monitoring device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117436017A true CN117436017A (en) | 2024-01-23 |
Family
ID=89547607
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311413186.9A Pending CN117436017A (en) | 2023-10-27 | 2023-10-27 | Abnormality early warning method and device for electric energy quality monitoring device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117436017A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117648670A (en) * | 2024-01-24 | 2024-03-05 | 润泰救援装备科技河北有限公司 | Rescue data fusion method, electronic equipment, storage medium and rescue fire truck |
-
2023
- 2023-10-27 CN CN202311413186.9A patent/CN117436017A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117648670A (en) * | 2024-01-24 | 2024-03-05 | 润泰救援装备科技河北有限公司 | Rescue data fusion method, electronic equipment, storage medium and rescue fire truck |
| CN117648670B (en) * | 2024-01-24 | 2024-04-12 | 润泰救援装备科技河北有限公司 | Rescue data fusion method, electronic equipment, storage medium and rescue fire truck |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115578015B (en) | Sewage treatment whole process supervision method, system and storage medium based on Internet of things | |
| CN116150676B (en) | Equipment fault diagnosis and identification method and device based on artificial intelligence | |
| CN117436017A (en) | Abnormality early warning method and device for electric energy quality monitoring device | |
| CN110632546B (en) | Electronic transformer credibility evaluation method and device based on whole-network-domain evidence set | |
| CN113011481A (en) | Electric energy meter function abnormity evaluation method and system based on decision tree algorithm | |
| CN116401532B (en) | Method and system for recognizing frequency instability of power system after disturbance | |
| CN106935038B (en) | Parking detection system and detection method | |
| CN112816881A (en) | Battery differential pressure abnormality detection method, battery differential pressure abnormality detection device and computer storage medium | |
| CN117748507B (en) | Distribution network harmonic access uncertainty assessment method based on Gaussian regression model | |
| CN114879070A (en) | Battery state evaluation method and related equipment | |
| CN114676749A (en) | Power distribution network operation data abnormity judgment method based on data mining | |
| CN108574290B (en) | Method, device, terminal and readable storage medium for positioning oscillation source of forced oscillation | |
| CN112215482B (en) | Household transformer relation recognition method and device | |
| CN117434372A (en) | Electromagnetic compatibility immunity test method and system for electronic product | |
| CN116520068B (en) | Diagnostic method, device, equipment and storage medium for electric power data | |
| CN112345972A (en) | Power failure event-based power distribution network line transformation relation abnormity diagnosis method, device and system | |
| CN115494431A (en) | Transformer fault warning method, terminal equipment and computer readable storage medium | |
| Zhou et al. | A novel prediction method about single components of analog circuits based on complex field modeling | |
| CN112924743B (en) | Instrument state detection method based on current data | |
| CN114782211A (en) | Method and system for acquiring information of sea and mountain distribution range | |
| Zhao et al. | Study on evaluation model of rate grade of soil corrosion for pipeline | |
| CN109871998B (en) | Power distribution network line loss rate prediction method and device based on expert sample library | |
| CN113554079A (en) | Electric power load abnormal data detection method and system based on secondary detection method | |
| CN111368782A (en) | Training method of coal fire area identification model, and coal fire area identification method and device | |
| CN112073932A (en) | Network data anomaly detection and positioning method of wireless sensor network |
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 |