CN113139610A - Abnormity detection method and device for transformer monitoring data - Google Patents
Abnormity detection method and device for transformer monitoring data Download PDFInfo
- Publication number
- CN113139610A CN113139610A CN202110475703.XA CN202110475703A CN113139610A CN 113139610 A CN113139610 A CN 113139610A CN 202110475703 A CN202110475703 A CN 202110475703A CN 113139610 A CN113139610 A CN 113139610A
- Authority
- CN
- China
- Prior art keywords
- abnormal
- sequence
- mode
- adopting
- type
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 59
- 238000001514 detection method Methods 0.000 title claims abstract description 25
- 230000002159 abnormal effect Effects 0.000 claims abstract description 265
- 238000000034 method Methods 0.000 claims abstract description 73
- 239000013598 vector Substances 0.000 claims abstract description 55
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 40
- 238000004458 analytical method Methods 0.000 claims abstract description 25
- 238000012795 verification Methods 0.000 claims abstract description 13
- 230000011218 segmentation Effects 0.000 claims description 15
- 230000008569 process Effects 0.000 claims description 14
- 230000008859 change Effects 0.000 claims description 12
- 238000010219 correlation analysis Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 5
- 238000011161 development Methods 0.000 claims description 4
- 230000018109 developmental process Effects 0.000 claims description 4
- 238000013508 migration Methods 0.000 claims description 4
- 230000005012 migration Effects 0.000 claims description 4
- 230000000877 morphologic effect Effects 0.000 claims description 4
- 238000004445 quantitative analysis Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 3
- 230000005856 abnormality Effects 0.000 abstract description 6
- 230000006870 function Effects 0.000 description 6
- 238000000354 decomposition reaction Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 238000001228 spectrum Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000013519 translation Methods 0.000 description 3
- 238000012098 association analyses Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000004140 cleaning Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000002547 anomalous effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000021615 conjugation Effects 0.000 description 1
- 238000001739 density measurement Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000011895 specific detection Methods 0.000 description 1
- 238000013112 stability test Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000002759 z-score normalization Methods 0.000 description 1
Images
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
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the specification discloses an abnormality detection method and device for transformer monitoring data. The method comprises the following steps: acquiring a sequence to be detected of transformer online monitoring data; constructing an abnormal data identification model by adopting time sequence modeling and an isolated forest algorithm; constructing an abnormal type identification mode by adopting an improved multi-dimensional SAX vector representation method; identifying abnormal data of the sequence to be detected by adopting the abnormal data identification model; determining the abnormal type of the abnormal data by adopting the abnormal type identification mode, wherein the abnormal type comprises an invalid abnormal mode and an effective abnormal mode; and when the exception type is the invalid exception mode, performing relevance verification on the exception type by adopting sequence relevance analysis. According to the scheme, the abnormal mode can be deeply analyzed on the basis of effectively identifying abnormal data information, and the effective abnormal points and the ineffective abnormal points can be accurately distinguished.
Description
Technical Field
The application relates to the technical field of computers, in particular to an anomaly detection method and device for transformer monitoring data.
Background
With the wide application of the technology based on big data and the internet of things in the state perception and the operation and maintenance of the power transformer, the scale of the monitoring data of the transformer shows an exponential growth trend, and an important data basis is provided for the comprehensive state evaluation and prediction of equipment. However, under the influence of various emergencies, the equipment online monitoring system inevitably generates part of abnormal data. The reliable identification of abnormal data and the effective distinguishing of the modes thereof are important foundations for realizing the efficient cleaning of online monitoring data and the accurate grasping of the running state of equipment. Existing anomaly detection studies involve clustering-based, classification-based, statistics-based methods, and the like.
The selection of an initial clustering center in the conventional abnormal recognition algorithm such as a clustering algorithm can cause great influence on the clustering convergence effect; the classification algorithm is suitable for a data set of a large number of abnormal samples, and in most scenes, abnormal data are few parts; the statistical-based method is greatly interfered by the selected sample set, and the identification effect is reduced when the sample data fluctuates seriously. In addition, the existing methods are not deep enough to study how to effectively distinguish different types of abnormal patterns, and further study needs to be carried out.
Disclosure of Invention
In view of this, the embodiment of the present application provides an anomaly detection method and apparatus for transformer monitoring data, and the method and apparatus can deeply analyze an anomaly mode on the basis of effectively identifying anomalous data information, and accurately distinguish between an effective anomaly point and an invalid anomaly point.
In order to solve the above technical problem, the embodiments of the present specification are implemented as follows:
an abnormality detection method for transformer monitoring data provided by an embodiment of the present specification includes:
acquiring a sequence to be detected of transformer online monitoring data;
constructing an abnormal data identification model by adopting time sequence modeling and an isolated forest algorithm;
constructing an abnormal type identification mode by adopting an improved multi-dimensional SAX vector representation method;
identifying abnormal data of the sequence to be detected by adopting the abnormal data identification model;
determining the abnormal type of the abnormal data by adopting the abnormal type identification mode, wherein the abnormal type comprises an invalid abnormal mode and an effective abnormal mode;
and when the exception type is the invalid exception mode, performing relevance verification on the exception type by adopting sequence relevance analysis.
Optionally, the identifying the abnormal data of the sequence to be detected by using the abnormal data identification model specifically includes:
adaptively decomposing the sequence to be detected into time sequence components with different frequencies by using an empirical wavelet transform theory;
respectively carrying out time sequence modeling on the time sequence components through a differential autoregressive moving average model, and reconstructing the predicted value of each component to obtain the predicted value of the monitoring sequence;
calculating a difference value between the predicted value and the measured value to obtain a residual sequence;
and carrying out abnormity identification on the residual sequence by using an isolated forest algorithm to realize effective extraction of abnormal information in the sequence to be detected.
Optionally, the improved multidimensional SAX-based vector representation method selects a feature value vector composed of a mean, a slope, and sample entropy from the viewpoint of statistical characteristics, morphological characteristics, and entropy characteristics of the time series, respectively, to completely represent the sequence characteristics.
Optionally, the performing relevance verification on the abnormal type by using sequence relevance analysis specifically includes:
and performing relevance verification on the abnormal type by adopting sequence relevance analysis of a grey relevance analysis algorithm.
Optionally, the gray correlation analysis algorithm judges the strength of the correlation degree between the parameters according to the similarity degree of the geometric shape of the change curve of each parameter, and the gray correlation analysis algorithm performs quantitative analysis on the development situation of the dynamic process to compare the geometric relationship of the time series related statistical data and calculate the correlation degree between the parameters.
Optionally, the invalid abnormal mode includes noise points and missing values, an observed value of the invalid abnormal mode at an abnormal occurrence time may deviate from an expected value seriously, and time sequences before and after the abnormal occurrence time may keep relatively consistent characteristics; the effective abnormal mode refers to that the time series characteristics before and after the abnormal occurrence time show larger difference due to the horizontal migration and trend change of the monitoring data caused by the abnormal change of the equipment state.
Optionally, the method for constructing the abnormal type identification mode by using the improved multidimensional SAX vector representation method specifically includes:
normalizing the time series with different magnitudes by adopting zero-mean normalization (z-score);
carrying out equidistant segmentation on the time sequence after the standardization treatment, and constructing a characteristic value vector representing the characteristics of the time sequence by selecting the average value, the slope and the sample entropy as the characteristic values of the time sequence;
and performing symbolization processing on the characteristic value vector.
Optionally, the determining the abnormal type of the abnormal data by using the abnormal type identification mode specifically includes:
inputting position information of the abnormal data;
segmenting the abnormal data according to the position information to generate a segmentation sequence;
carrying out multi-dimensional symbolization vector representation on the segmented sequence by improving a multi-dimensional SAX vector representation method, and calculating the correlation coefficient of symbol vectors at two sides of each abnormal point;
and judging whether the correlation coefficient is higher than a preset threshold value, and if not, determining that the abnormal type is an effective abnormal mode.
Optionally, the preset threshold is 0.7-0.8, and more preferably 0.75.
An anomaly detection device for transformer monitoring data provided by the embodiments of the present specification, the device includes:
the sequence to be detected acquisition module is used for acquiring a sequence to be detected of the transformer online monitoring data;
the abnormal data identification model building module is used for building an abnormal data identification model by adopting time sequence modeling and an isolated forest algorithm;
the abnormal type identification mode construction module is used for constructing an abnormal type identification mode by adopting an improved multidimensional SAX vector representation method;
the abnormal data identification module is used for identifying the abnormal data of the sequence to be detected by adopting the abnormal data identification model;
the abnormal type determining module is used for determining the abnormal type of the abnormal data by adopting the abnormal type identification mode, wherein the abnormal type comprises an invalid abnormal mode and an effective abnormal mode;
and the relevance checking module is used for performing relevance checking on the abnormal type by adopting sequence relevance analysis when the abnormal type is the invalid abnormal mode.
An anomaly detection device for transformer monitoring data provided by the embodiments of the present specification includes:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to:
acquiring a sequence to be detected of transformer online monitoring data;
constructing an abnormal data identification model by adopting time sequence modeling and an isolated forest algorithm;
constructing an abnormal type identification mode by adopting an improved multi-dimensional SAX vector representation method;
identifying abnormal data of the sequence to be detected by adopting the abnormal data identification model;
determining the abnormal type of the abnormal data by adopting the abnormal type identification mode, wherein the abnormal type comprises an invalid abnormal mode and an effective abnormal mode;
and when the exception type is the invalid exception mode, performing relevance verification on the exception type by adopting sequence relevance analysis.
The embodiment of the specification adopts at least one technical scheme which can achieve the following beneficial effects:
1. on the basis of effectively identifying abnormal data information, an abnormal mode is deeply analyzed, and effective abnormal points and ineffective abnormal points are accurately distinguished.
2. And modeling the time sequence relation in the online monitoring data by combining an EWT theory and an ARIMA model to obtain a residual sequence capable of reflecting abnormal characteristics of the monitoring data, and further utilizing an iForest algorithm to realize efficient extraction of abnormal information in the residual sequence.
3. On the basis of deep analysis of mode difference of invalid abnormal data and effective abnormal data, an improved multi-dimensional SAX vector representation method is introduced to symbolize a time sequence, the characteristic difference of segmentation sequences at two sides of an abnormal point is measured by similarity score of a symbol vector, and effective distinguishing of abnormal modes is realized by combining a judgment threshold.
4. The grey correlation analysis algorithm is utilized to accurately measure the correlation degree between the monitored sequences, and the abnormal mode judgment result is further verified on the basis of considering the time sequence correlation, so that the limitation of the judgment threshold setting is effectively avoided.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
fig. 1 is a schematic flowchart of an anomaly detection method for transformer monitoring data according to an embodiment of the present disclosure;
FIG. 2 is a schematic diagram of a moving process of a target template;
FIG. 3 is a flow chart of anomaly detection;
fig. 4 is a schematic structural diagram of an abnormality detection apparatus for transformer monitoring data, corresponding to fig. 1, provided in an embodiment of the present disclosure.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Aiming at the defects of the existing abnormal data detection method, the invention provides a diagnosis strategy for reliably distinguishing the effective abnormal points from the ineffective abnormal points on the basis of effectively identifying the abnormal data points.
The invalid abnormal point refers to an abnormal point with a large difference between an observed value and an expected value at a certain moment of the time series, such as a missing value, a noise point and the like; the effective abnormal point refers to an abnormal point where the behaviors of the previous and subsequent sequences show large differences at a certain time of the time sequence, such as abnormal values of the running state of the equipment, and different analysis methods should be adopted for different abnormal values.
The technical solutions provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.
Fig. 1 is a schematic flowchart of an anomaly detection method for transformer monitoring data according to an embodiment of the present disclosure. From the viewpoint of a program, the execution subject of the flow may be a program installed in an application server or an application client.
As shown in fig. 1, the process may include the following steps:
step 102: acquiring a sequence to be detected of transformer online monitoring data;
step 104: constructing an abnormal data identification model by adopting time sequence modeling and an isolated forest algorithm;
step 106: constructing an abnormal type identification mode by adopting an improved multi-dimensional SAX vector representation method;
step 108: identifying abnormal data of the sequence to be detected by adopting the abnormal data identification model;
step 110: determining the abnormal type of the abnormal data by adopting the abnormal type identification mode, wherein the abnormal type comprises an invalid abnormal mode and an effective abnormal mode;
step 112: and when the exception type is the invalid exception mode, performing relevance verification on the exception type by adopting sequence relevance analysis.
Optionally, the identifying the abnormal data of the sequence to be detected by using the abnormal data identification model specifically includes:
adaptively decomposing the sequence to be detected into time sequence components with different frequencies by using an empirical wavelet transform theory;
respectively carrying out time sequence modeling on the time sequence components through a differential autoregressive moving average model, and reconstructing the predicted value of each component to obtain the predicted value of the monitoring sequence;
calculating a difference value between the predicted value and the measured value to obtain a residual sequence;
and carrying out abnormity identification on the residual sequence by using an isolated forest algorithm to realize effective extraction of abnormal information in the sequence to be detected.
Optionally, the improved multidimensional SAX-based vector representation method selects a feature value vector composed of a mean, a slope, and sample entropy from the viewpoint of statistical characteristics, morphological characteristics, and entropy characteristics of the time series, respectively, to completely represent the sequence characteristics.
Optionally, the performing relevance verification on the abnormal type by using sequence relevance analysis specifically includes:
and performing relevance verification on the abnormal type by adopting sequence relevance analysis of a grey relevance analysis algorithm.
Optionally, the gray correlation analysis algorithm judges the strength of the correlation degree between the parameters according to the similarity degree of the geometric shape of the change curve of each parameter, and the gray correlation analysis algorithm performs quantitative analysis on the development situation of the dynamic process to compare the geometric relationship of the time series related statistical data and calculate the correlation degree between the parameters.
Optionally, the invalid abnormal mode includes noise points and missing values, an observed value of the invalid abnormal mode at an abnormal occurrence time may deviate from an expected value seriously, and time sequences before and after the abnormal occurrence time may keep relatively consistent characteristics; the effective abnormal mode refers to that the time series characteristics before and after the abnormal occurrence time show larger difference due to the horizontal migration and trend change of the monitoring data caused by the abnormal change of the equipment state.
Optionally, the method for constructing the abnormal type identification mode by using the improved multidimensional SAX vector representation method specifically includes:
the time series of different magnitudes are normalized by using z-score (zero-mean normalization);
carrying out equidistant segmentation on the time sequence after the standardization treatment, and constructing a characteristic value vector representing the characteristics of the time sequence by selecting the average value, the slope and the sample entropy as the characteristic values of the time sequence;
and performing symbolization processing on the characteristic value vector.
Optionally, the determining the abnormal type of the abnormal data by using the abnormal type identification mode specifically includes:
inputting position information of the abnormal data;
segmenting the abnormal data according to the position information to generate a segmentation sequence;
carrying out multi-dimensional symbolization vector representation on the segmented sequence by improving a multi-dimensional SAX vector representation method, and calculating the correlation coefficient of symbol vectors at two sides of each abnormal point;
and judging whether the correlation coefficient is higher than a preset threshold value, and if not, determining that the abnormal type is an effective abnormal mode.
Optionally, the preset threshold is 0.7-0.8, and more preferably 0.75.
Based on the method of fig. 1, the embodiments of the present specification also provide some specific implementations of the method, which are described below.
Aiming at the defects of the existing abnormal data detection method, the embodiment of the specification provides a diagnosis strategy for reliably distinguishing the effective abnormal value from the invalid abnormal value on the basis of effectively identifying the abnormal data point.
The invalid abnormal point refers to an abnormal point with a large difference between an observed value and an expected value at a certain moment of the time series, such as a missing value, a noise point and the like; the effective abnormal point refers to an abnormal point where the behaviors of the previous and subsequent sequences show large differences at a certain time of the time sequence, such as abnormal values of the running state of the equipment, and different analysis methods should be adopted for different abnormal values.
Abnormal data identification based on time sequence modeling and isolated forest algorithm
Aiming at the abnormal identification of the on-line monitoring data of the transformer, firstly, an original sequence is decomposed into time sequence components with different frequencies in a self-adaptive manner by using an empirical wavelet transform theory so as to weaken the mutual influence among information with different scales; secondly, respectively carrying out time sequence modeling on the time sequence components through a differential autoregressive moving average model, and reconstructing predicted values of the components to obtain predicted values of the monitoring sequence; on the basis, calculating the difference value between the measured value and the abnormal data to obtain a residual sequence, wherein the abnormal data characteristics can be obviously represented in the residual sequence; and finally, carrying out abnormity identification on the residual sequence by using an isolated forest algorithm, and realizing effective extraction of abnormal information in the monitoring sequence.
1.1 EWT
Empirical Wavelet Transform (EWT) is a method of adaptive analysis of signals. The core idea is that a proper orthogonal wavelet filter bank is constructed by adaptively dividing the Fourier frequency spectrum of an original signal so as to extract the amplitude modulation-frequency modulation component of the original signal. Taking a time domain discrete signal as an example, the specific steps of empirical wavelet transform are as follows:
1) the input signal F (t) is fourier transformed to obtain its fourier spectrum F (ω), ω being defined in the range [0, π ].
2) Adaptively partitioning the Fourier spectrum of a signal into N segments, omegan(N-1, 2, …, N) indicates the boundary of the segment number.
3) N empirical wavelets are constructed according to Fourier spectrum segmentation, calculation formulas of an empirical wavelet function and an empirical scale function are respectively shown as formulas (1) and (2), wherein values of beta and gamma are shown as formulas (3) and (4).
4) Constructing empirical wavelet transform, and respectively carrying out inner product operation on an original signal, an empirical wavelet function and an empirical scale function to obtain a detail coefficient and an approximate coefficient:
in the formula:
and
are respectively phi1(t) and psin(t) complex conjugation.
5) Reconstructing the original signal according to equation (7) and obtaining therefrom an original signal decomposition result f0(t)、 fk(t)。
1.2 differential autoregressive moving average model
A differential autoregressive moving average (ARIMA) model, commonly referred to as ARIMA (p, d, q). The basic idea is to perform d-order difference on a non-stationary time sequence to form a stationary time sequence, then use an autoregressive moving average model (ARMA) to model the stationary sequence, and then obtain an original sequence through inverse transformation.
Firstly, stability test is carried out on an input time sequence, and the value of a difference order is determined. The method selects the structural test statistic to carry out hypothesis test, and judges the stationarity of the input time sequenceThe difference processing needs to be repeated for the non-stationary time series until the processed time series is stationary. For some non-stationary time series { xtThe difference processing procedure of (10) is shown in the formula.
In the formula, B is a delay operator;
an ordered difference operator; d represents the difference order.
Non-stationary time series x by differential processingtThe conversion into a stationary time series ytOn the basis, establishing an ARMA (p, q) model for the cells:
in the formula,
a predicted value representing time t; p and q respectively represent the orders of an autoregressive term and a moving average term in the model;
coefficients representing the ith auto-regressive term; thetajA coefficient representing a jth moving average term; { εtDenotes a white noise sequence following an independent normal distribution.
The ARMA model construction process comprises model order determination and parameter estimation. On the basis of estimating model parameters by adopting a maximum likelihood method, an order combination which minimizes an AIC value is selected as a model order determination result by limiting the value ranges of p and q according to an Akaike Information Criterion (AIC).
And carrying out multi-scale decomposition on the transformer monitoring sequence through an EWT theory, and constructing an ARIMA prediction model by aiming at the modal component obtained by decomposition through the steps. In order to ensure the prediction accuracy of the ARIMA model, the single-step prediction is only carried out on the component values, and a complete prediction sequence related to the modal component can be obtained by sliding a fitting window and a prediction window to the right along with time; and reconstructing the prediction results of the components to obtain a complete prediction sequence related to the monitoring data.
1.3 iForest theory
After a predicted value of the monitoring index is obtained by applying the EWT and ARIMA models, the predicted value is subtracted from an actual measured value, and a residual error item at a corresponding moment is obtained, as shown in a formula (12). The residual sequence is numerically eliminated from the influence of periodicity and tendency of the original sequence in the change process, so that residual terms fluctuate around a zero value. Therefore, abnormal data caused by various types of emergencies can be more obviously represented in the residual sequence in the form of outliers.
The isolated forest (iForest) algorithm is an unsupervised anomaly detection method suitable for continuous data. Different from the distance-based and density-based anomaly detection method, the isolated forest algorithm does not depend on any distance or density measurement, so that the operation cost is greatly reduced, and the method has the advantages of high precision and high calculation efficiency. Meanwhile, the outliers in the residual sequence are consistent with the definition of abnormal data in the isolated forest algorithm, i.e., the data points which are sparsely distributed and far away from the high-density population. Therefore, the method selects to use the isolated forest algorithm to perform anomaly identification on the residual sequence, and realizes effective extraction of anomaly information in the monitoring sequence.
Second, abnormal mode distinguishing technology for monitoring data flow
On the basis of the effective extraction of the abnormal data information in the foregoing, the accurate judgment of the abnormal mode is realized. The invalid abnormal mode mainly comprises noise points and missing values, the observed value of the invalid abnormal mode can be seriously deviated from an expected value at the abnormal occurrence moment, and time sequences before and after the moment can keep relatively consistent characteristics; the effective abnormal mode mainly refers to that the time series characteristics before and after the abnormal occurrence time show larger difference due to the horizontal migration and trend change of the monitoring data caused by the abnormal change of the equipment state. Therefore, on the basis of dividing an original sequence by taking an abnormal point as a segmentation boundary, the method introduces an improved multidimensional SAX vector representation method to carry out multidimensional symbolization vector representation on a segmentation subsequence, distinguishes different abnormal modes by calculating the similarity score of two adjacent symbol vectors and combining a decision threshold value, and further verifies a mode decision result by utilizing sequence relevance analysis.
2.1 abnormal Pattern determination based on improved multidimensional SAX vector representation method
The SAX algorithm can represent a continuous time sequence as a discretized symbol sequence, divides intervals by utilizing the characteristic of standard normal distribution, and respectively represents numerical values in the intervals by different symbols, thereby realizing the conversion of the numerical value sequence into the symbol sequence.
In the conventional SAX method, on the basis of dividing a time series, a numerical average of each time series is used as an expression feature of the time series. In consideration of the great limitation of the mean value expression method, the improved multidimensional SAX vector expression method adopted by the patent takes the statistical characteristics, morphological characteristics and entropy characteristics of the time series into consideration, and selects the characteristic value vector consisting of the mean value, the slope and the sample entropy to completely express the sequence characteristics, and the specific flow comprises the following steps:
1) z-score normalization processing of time series
The z-score standardization can convert data of different magnitudes into scores of uniform measurement for measurement so as to ensure comparability among the data.
2) Segmentation of time series at equal intervals and characterization
And carrying out equidistant segmentation on the time sequence after the standardization treatment, and constructing a characteristic value vector capable of completely representing the characteristics of the time sequence by selecting the mean value, the slope and the sample entropy as the characteristic values of the time sequence so as to improve the accuracy of subsequent similarity retrieval and query.
3) Symbolic vector representation of a time series
According to the value distribution of the inter-sequence characteristic values, the value space of each type of characteristic values is divided with equal probability, and different characters are used for representing the divided value subspace area, such as the letter set { A, B, C, D, E, … }. And (3) recording the scale parameter of the set as alpha, wherein the larger the value of the alpha is, the finer the granularity of the average numerical value space is, and the higher the distinguishing precision is. In general, α is in the range of [3,20 ]]. The character sequences representing the features of mean, slope and sample entropy are respectively recorded as
Therefore, each sub-segment characteristic of the time series can be represented by a symbol vector in a three-dimensional space, namely, the symbol vector can be used
To characterize the i-th sub-segment of the time series.
When the abnormal point belongs to an effective abnormal mode, the characteristics of the subsequences on the left side and the right side of the abnormal point have larger difference; when the abnormal point belongs to the invalid abnormal mode, the subsequences at the left and right sides of the abnormal point keep more consistent characteristics. Therefore, the abnormal mode is accurately judged by calculating the symbol vector similarity values of the subsequences at both sides of the abnormal point, and the specific flow is as follows:
1) and comparing the lengths of the multidimensional symbolic vectors of the subsequences on two sides of a certain segment boundary point. Multi-dimensional symbolizing vector sequence of long sequence L
As the sequence to be matched, the multi-dimensional symbolic vector sequence of the short sequence Q
As a target template sequence.
2) Sequence of target templates
In waiting to matchSequence of
Up-translation as shown in fig. 1. And calculating the similarity value of the two at each position in the translation process, as formulas (13) and (14). And on the basis of obtaining the similarity score value set generated in the translation process, selecting the minimum value of the similarity score value set as the abnormal mode judgment score of the segmentation point.
Wherein w represents the length of the target template sequence; dist () represents a measurement function of the character distance, and the distance between any two characters can be found by looking up a table. And the similarity is calculated on the basis, and compared with numerical calculation without symbolic representation, the judgment conclusion is consistent, but the calculation efficiency is effectively improved.
3) Setting a threshold value T for mode judgment, and judging that the abnormal point belongs to an effective abnormal mode if the score is greater than T; if the score is smaller than T, the abnormal point is judged to belong to an invalid abnormal mode.
4) And repeating the steps until all abnormal points in the monitoring sequence are judged.
Based on the example, the similarity search tests which are performed for multiple times in advance find that the similarity values of two groups of sequences with relatively consistent patterns are stable below 0.5, so the threshold value for determining the patterns is set to 0.5. However, given the limitation of threshold setting, sequence association analysis is introduced to further verify the determination result on the basis of distinguishing abnormal modes by using the threshold.
2.2 sequence Association analysis based on Gray Association Algorithm
And judging the strength of the correlation degree among the parameters by a gray correlation analysis algorithm according to the similarity degree of the geometric shapes of the parameter change curves. The algorithm completes the comparison of the geometric relation of time series related statistical data through the quantitative analysis of the development situation of the dynamic process, and solves the association degree among all parameters.
Reference sequences are herein denoted
And assuming that there are m sets of comparison sequences, each of which is designated as
Wherein i is 1,2,3, …, m. Since the physical meanings of the monitoring indexes are different, the dimensions of the data are not necessarily the same, and therefore, the sequence needs to be subjected to non-dimensionalization:
on the basis, calculating the association coefficient of each comparison sequence and the corresponding element of the reference sequence:
wherein rho is a resolution coefficient and is generally 0.5, n is a two-stage minimum difference, and m is a two-stage maximum difference.
By integrating the gray correlation coefficients at each time point, the gray correlation degree between the reference sequence and the ith group of comparison sequences can be obtained:
rithe larger the sequence is, the stronger the relevance between the comparison sequence and the sequence to be detected is, and the relevance threshold value r is setm0.75, and a comparison sequence greater than 0.75 is taken as the cognate sequence.
Third, monitoring data anomaly detection technical framework
By combining the above contents, a transformer monitoring data anomaly detection technology framework including an anomaly identification and mode determination function module is constructed, as shown in fig. 3, a specific detection flow is summarized as follows:
1) measuring the degree of correlation between the sequence to be detected and the rest of the monitoring sequences by a grey correlation analysis algorithm, and if the correlation sequences exist, reserving a checking link in the abnormal mode judging process; and if the correlation sequence does not exist, removing the verification link.
2) And performing multi-scale decomposition on the monitoring sequence by utilizing an EWT theory, respectively establishing an ARIMA prediction model aiming at modal components obtained by decomposition, and reconstructing prediction results of all components on the basis to obtain a prediction sequence related to the monitoring index.
3) And calculating a difference value between the predicted value and the actual value to obtain a residual sequence, and carrying out abnormal value identification on the residual sequence by combining an iForest algorithm, so as to segment the original monitoring sequence by taking the abnormal point as a segmentation boundary.
4) The segmented sequence is subjected to multi-dimensional symbolized vector representation by improving a multi-dimensional SAX vector representation method, and similarity scores of symbol vectors on two sides of each abnormal point are calculated, so that different abnormal modes are distinguished by combining with a judgment threshold value.
5) From the perspective of guaranteeing safe and stable operation of the equipment, when a certain abnormal point of the monitoring sequence is determined as an invalid abnormal mode, the determination result needs to be checked in combination with the associated sequence. If no abnormal point appears in the correlation sequence of the monitoring sequence at the same or adjacent time, the abnormal point can be judged to belong to an invalid abnormal mode; if abnormal points appear in the association sequence at the same or adjacent moments, the abnormal points are classified as effective abnormal modes, and the abnormal reasons may be abnormal changes of the operation state of the power transformer or interference of external factors on the related monitored quantities in the measurement or transmission process, and further intervention judgment by related operation and maintenance personnel is needed.
The scheme has the following advantages:
1. according to the scheme, the modeling is performed on the time sequence relation in the online monitoring data by combining the EWT theory and the ARIMA model, a residual sequence capable of reflecting the abnormal characteristics of the monitoring data is obtained, and the high-efficiency extraction of the abnormal information in the residual sequence is further realized by using the iForest algorithm.
2. On the basis of deep analysis of mode difference of invalid abnormal data and effective abnormal data, an improved multi-dimensional SAX vector representation method is introduced to symbolize a time sequence, the characteristic difference of segmentation sequences at two sides of an abnormal point is measured by similarity score of a symbol vector, and effective distinguishing of abnormal modes is realized by combining a judgment threshold.
3. The grey correlation analysis algorithm is utilized to accurately measure the correlation degree between the monitored sequences, and the abnormal mode judgment result is further verified on the basis of considering the time sequence correlation, so that the limitation of the judgment threshold setting is effectively avoided.
4. On the basis of effectively identifying abnormal data information, the abnormal mode is deeply analyzed, and effective abnormality and invalid abnormality are accurately distinguished.
5) The anomaly detection technical framework constructed by the method can provide key technical support for efficient cleaning of the on-line monitoring data of the power transformer and accurate grasping of the running state of the equipment.
Based on the same idea, the embodiment of the present specification further provides a device corresponding to the above method. Fig. 4 is a schematic structural diagram of an abnormality detection apparatus for transformer monitoring data, corresponding to fig. 1, provided in an embodiment of the present disclosure. As shown in fig. 4, the apparatus may include:
a to-be-detected sequence acquisition module 402, configured to acquire a to-be-detected sequence of transformer online monitoring data;
an abnormal data identification model construction module 404, configured to construct an abnormal data identification model by using time-series modeling and an isolated forest algorithm;
an abnormal type identification mode construction module 406, configured to construct an abnormal type identification mode by using an improved multidimensional SAX vector representation-based method;
an abnormal data identification module 408, configured to identify abnormal data of the sequence to be detected by using the abnormal data identification model;
an exception type determining module 410, configured to determine an exception type of the exception data using the exception type identification pattern, where the exception type includes an invalid exception pattern and a valid exception pattern;
and the relevance checking module 412 is configured to, when the exception type is the invalid exception mode, perform relevance checking on the exception type by using sequence relevance analysis.
An embodiment of the present specification further provides an anomaly detection device for transformer monitoring data, including:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to:
acquiring a sequence to be detected of transformer online monitoring data;
constructing an abnormal data identification model by adopting time sequence modeling and an isolated forest algorithm;
constructing an abnormal type identification mode by adopting an improved multi-dimensional SAX vector representation method;
identifying abnormal data of the sequence to be detected by adopting the abnormal data identification model;
determining the abnormal type of the abnormal data by adopting the abnormal type identification mode, wherein the abnormal type comprises an invalid abnormal mode and an effective abnormal mode;
and when the exception type is the invalid exception mode, performing relevance verification on the exception type by adopting sequence relevance analysis.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (11)
1. An anomaly detection method for transformer monitoring data, characterized in that the method comprises:
acquiring a sequence to be detected of transformer online monitoring data;
constructing an abnormal data identification model by adopting time sequence modeling and an isolated forest algorithm;
constructing an abnormal type identification mode by adopting an improved multi-dimensional SAX vector representation method;
identifying abnormal data of the sequence to be detected by adopting the abnormal data identification model;
determining the abnormal type of the abnormal data by adopting the abnormal type identification mode, wherein the abnormal type comprises an invalid abnormal mode and an effective abnormal mode;
and when the exception type is the invalid exception mode, performing relevance verification on the exception type by adopting sequence relevance analysis.
2. The method according to claim 1, wherein the identifying the abnormal data of the sequence to be detected by using the abnormal data identification model specifically comprises:
adaptively decomposing the sequence to be detected into time sequence components with different frequencies by using an empirical wavelet transform theory;
respectively carrying out time sequence modeling on the time sequence components through a differential autoregressive moving average model, and reconstructing the predicted value of each component to obtain the predicted value of the monitoring sequence;
calculating a difference value between the predicted value and the measured value to obtain a residual sequence;
and carrying out abnormity identification on the residual sequence by using an isolated forest algorithm to realize effective extraction of abnormal information in the sequence to be detected.
3. The method of claim 1, wherein the improved multidimensional SAX-based vector representation method selects a feature value vector consisting of a mean, a slope and a sample entropy to completely represent the sequence characteristics in view of statistical characteristics, morphological characteristics and entropy characteristics of the time sequence, respectively.
4. The method of claim 1, wherein the performing relevance checking on the anomaly type by using sequence relevance analysis specifically comprises:
and performing relevance verification on the abnormal type by adopting sequence relevance analysis of a grey relevance analysis algorithm.
5. The method as claimed in claim 4, wherein the grey correlation analysis algorithm judges the strength of the correlation between the parameters according to the similarity of the geometric shapes of the variation curves of the parameters, and the grey correlation analysis algorithm performs comparison of the geometric relationships of the time series related statistical data through quantitative analysis of the development situation of the dynamic process and obtains the correlation between the parameters.
6. The method of claim 1, wherein the invalid anomaly patterns include noise points and missing values, and wherein the observed values at the time of occurrence of an anomaly deviate significantly from expected values, and wherein the time series around the time of occurrence of the anomaly maintain relatively consistent characteristics; the effective abnormal mode refers to that the time series characteristics before and after the abnormal occurrence time show larger difference due to the horizontal migration and trend change of the monitoring data caused by the abnormal change of the equipment state.
7. The method of claim 3, wherein constructing the abnormal type recognition pattern by using an improved multidimensional SAX vector representation method specifically comprises:
adopting zero-mean normalization to carry out standardization processing on time sequences with different magnitudes;
carrying out equidistant segmentation on the time sequence after the standardization treatment, and constructing a characteristic value vector representing the characteristics of the time sequence by selecting the average value, the slope and the sample entropy as the characteristic values of the time sequence;
and performing symbolization processing on the characteristic value vector.
8. The method according to claim 1, wherein the determining the anomaly type of the anomaly data using the anomaly type recognition mode specifically includes:
inputting position information of the abnormal data;
segmenting the abnormal data according to the position information to generate a segmentation sequence;
carrying out multi-dimensional symbolization vector representation on the segmented sequence by improving a multi-dimensional SAX vector representation method, and calculating the correlation coefficient of symbol vectors at two sides of each abnormal point;
and judging whether the correlation coefficient is higher than a preset threshold value, and if not, determining that the abnormal type is an effective abnormal mode.
9. The method of claim 8, wherein the predetermined threshold is 0.75.
10. An anomaly detection apparatus for transformer monitoring data, the apparatus comprising:
the sequence to be detected acquisition module is used for acquiring a sequence to be detected of the transformer online monitoring data;
the abnormal data identification model building module is used for building an abnormal data identification model by adopting time sequence modeling and an isolated forest algorithm;
the abnormal type identification mode construction module is used for constructing an abnormal type identification mode by adopting an improved multidimensional SAX vector representation method;
the abnormal data identification module is used for identifying the abnormal data of the sequence to be detected by adopting the abnormal data identification model;
the abnormal type determining module is used for determining the abnormal type of the abnormal data by adopting the abnormal type identification mode, wherein the abnormal type comprises an invalid abnormal mode and an effective abnormal mode;
and the relevance checking module is used for performing relevance checking on the abnormal type by adopting sequence relevance analysis when the abnormal type is the invalid abnormal mode.
11. An anomaly detection device for transformer monitoring data, comprising:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to:
acquiring a sequence to be detected of transformer online monitoring data;
constructing an abnormal data identification model by adopting time sequence modeling and an isolated forest algorithm;
constructing an abnormal type identification mode by adopting an improved multi-dimensional SAX vector representation method;
identifying abnormal data of the sequence to be detected by adopting the abnormal data identification model;
determining the abnormal type of the abnormal data by adopting the abnormal type identification mode, wherein the abnormal type comprises an invalid abnormal mode and an effective abnormal mode;
and when the exception type is the invalid exception mode, performing relevance verification on the exception type by adopting sequence relevance analysis.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110475703.XA CN113139610A (en) | 2021-04-29 | 2021-04-29 | Abnormity detection method and device for transformer monitoring data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110475703.XA CN113139610A (en) | 2021-04-29 | 2021-04-29 | Abnormity detection method and device for transformer monitoring data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113139610A true CN113139610A (en) | 2021-07-20 |
Family
ID=76816426
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110475703.XA Pending CN113139610A (en) | 2021-04-29 | 2021-04-29 | Abnormity detection method and device for transformer monitoring data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113139610A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113591401A (en) * | 2021-08-24 | 2021-11-02 | 华北电力大学（保定） | Power transformer data cleaning method based on time series decomposition |
| CN114254695A (en) * | 2021-11-18 | 2022-03-29 | 中国空间技术研究院 | Spacecraft telemetry data self-adaptive anomaly detection method and device |
| CN114912852A (en) * | 2022-07-18 | 2022-08-16 | 山东鑫亚装备制造有限公司 | Abnormity monitoring and early warning system for storage tank of refrigeration bin |
| CN115496644A (en) * | 2022-11-18 | 2022-12-20 | 山东超华环保智能装备有限公司 | Solid waste treatment equipment monitoring method based on data identification |
| CN116702081A (en) * | 2023-08-07 | 2023-09-05 | 西安格蒂电力有限公司 | Intelligent inspection method for power distribution equipment based on artificial intelligence |
| CN116955895A (en) * | 2023-09-21 | 2023-10-27 | 山东博诚电气有限公司 | Transformer operation abnormity monitoring method and system |
| CN117007135A (en) * | 2023-10-07 | 2023-11-07 | 东莞百舜机器人技术有限公司 | Hydraulic fan automatic assembly line monitoring system based on internet of things data |
| CN117131369A (en) * | 2023-10-27 | 2023-11-28 | 福建福昇消防服务集团有限公司 | Data processing method and system of intelligent safety management and emergency rescue integrated station |
| CN117932523A (en) * | 2024-03-25 | 2024-04-26 | 山东诚祥建设集团股份有限公司 | Construction engineering construction data processing method and system |
| CN118013304A (en) * | 2024-02-19 | 2024-05-10 | 国网湖北省电力有限公司襄阳供电公司 | Transformer fault positioning method and system based on clustering algorithm |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080270088A1 (en) * | 2007-04-30 | 2008-10-30 | International Business Machines Corporation | Method and system for causal modeling and outlier detection |
| US20140379301A1 (en) * | 2013-06-24 | 2014-12-25 | Infosys Limited | Systems and methods for data-driven anomaly detection |
| CN105138413A (en) * | 2015-08-19 | 2015-12-09 | 中国电力科学研究院 | Distribution network online data anomaly detection method |
| US20170055913A1 (en) * | 2015-08-27 | 2017-03-02 | Tata Consultancy Services Limited | System and method for determining information and outliers from sensor data |
| CN109919186A (en) * | 2019-01-28 | 2019-06-21 | 平安科技（深圳）有限公司 | Abnormal point ratio optimization method, apparatus, computer equipment and storage medium |
| CN110929800A (en) * | 2019-11-29 | 2020-03-27 | 四川万益能源科技有限公司 | Business body abnormal electricity utilization detection method based on sax algorithm |
| CN111507376A (en) * | 2020-03-20 | 2020-08-07 | 厦门大学 | Single index abnormality detection method based on fusion of multiple unsupervised methods |
| CN111522804A (en) * | 2020-04-15 | 2020-08-11 | 国网江苏省电力有限公司 | Cleaning method for abnormal data of transformer equipment state monitoring |
| CN112597209A (en) * | 2020-12-15 | 2021-04-02 | 深圳前海微众银行股份有限公司 | Data verification method, device and system and computer readable storage medium |
-
2021
- 2021-04-29 CN CN202110475703.XA patent/CN113139610A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080270088A1 (en) * | 2007-04-30 | 2008-10-30 | International Business Machines Corporation | Method and system for causal modeling and outlier detection |
| US20140379301A1 (en) * | 2013-06-24 | 2014-12-25 | Infosys Limited | Systems and methods for data-driven anomaly detection |
| CN105138413A (en) * | 2015-08-19 | 2015-12-09 | 中国电力科学研究院 | Distribution network online data anomaly detection method |
| US20170055913A1 (en) * | 2015-08-27 | 2017-03-02 | Tata Consultancy Services Limited | System and method for determining information and outliers from sensor data |
| CN109919186A (en) * | 2019-01-28 | 2019-06-21 | 平安科技（深圳）有限公司 | Abnormal point ratio optimization method, apparatus, computer equipment and storage medium |
| CN110929800A (en) * | 2019-11-29 | 2020-03-27 | 四川万益能源科技有限公司 | Business body abnormal electricity utilization detection method based on sax algorithm |
| CN111507376A (en) * | 2020-03-20 | 2020-08-07 | 厦门大学 | Single index abnormality detection method based on fusion of multiple unsupervised methods |
| CN111522804A (en) * | 2020-04-15 | 2020-08-11 | 国网江苏省电力有限公司 | Cleaning method for abnormal data of transformer equipment state monitoring |
| CN112597209A (en) * | 2020-12-15 | 2021-04-02 | 深圳前海微众银行股份有限公司 | Data verification method, device and system and computer readable storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 李明: "无线传感器网络中离群点检测算法研究", 中国优秀硕士学位论文全文数据库信息辑, pages 4 * |
| 王超: "云环境中时序数据的预测和异常检测算法的研究", 中国优秀硕士学位论文全文数据库信息辑, pages 2 - 4 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113591401B (en) * | 2021-08-24 | 2023-10-20 | 华北电力大学（保定） | Power transformer data cleaning method based on time sequence decomposition |
| CN113591401A (en) * | 2021-08-24 | 2021-11-02 | 华北电力大学（保定） | Power transformer data cleaning method based on time series decomposition |
| CN114254695A (en) * | 2021-11-18 | 2022-03-29 | 中国空间技术研究院 | Spacecraft telemetry data self-adaptive anomaly detection method and device |
| CN114254695B (en) * | 2021-11-18 | 2023-04-07 | 中国空间技术研究院 | Spacecraft telemetry data self-adaptive anomaly detection method and device |
| CN114912852A (en) * | 2022-07-18 | 2022-08-16 | 山东鑫亚装备制造有限公司 | Abnormity monitoring and early warning system for storage tank of refrigeration bin |
| CN114912852B (en) * | 2022-07-18 | 2022-09-23 | 山东鑫亚装备制造有限公司 | Abnormity monitoring and early warning system for storage tank of refrigeration bin |
| CN115496644A (en) * | 2022-11-18 | 2022-12-20 | 山东超华环保智能装备有限公司 | Solid waste treatment equipment monitoring method based on data identification |
| CN115496644B (en) * | 2022-11-18 | 2023-09-26 | 南通万达能源动力科技有限公司 | Solid waste treatment equipment monitoring method based on data identification |
| CN116702081A (en) * | 2023-08-07 | 2023-09-05 | 西安格蒂电力有限公司 | Intelligent inspection method for power distribution equipment based on artificial intelligence |
| CN116702081B (en) * | 2023-08-07 | 2023-10-24 | 西安格蒂电力有限公司 | Intelligent inspection method for power distribution equipment based on artificial intelligence |
| CN116955895A (en) * | 2023-09-21 | 2023-10-27 | 山东博诚电气有限公司 | Transformer operation abnormity monitoring method and system |
| CN116955895B (en) * | 2023-09-21 | 2023-12-15 | 山东博诚电气有限公司 | Transformer operation abnormity monitoring method and system |
| CN117007135A (en) * | 2023-10-07 | 2023-11-07 | 东莞百舜机器人技术有限公司 | Hydraulic fan automatic assembly line monitoring system based on internet of things data |
| CN117007135B (en) * | 2023-10-07 | 2023-12-12 | 东莞百舜机器人技术有限公司 | Hydraulic fan automatic assembly line monitoring system based on internet of things data |
| CN117131369A (en) * | 2023-10-27 | 2023-11-28 | 福建福昇消防服务集团有限公司 | Data processing method and system of intelligent safety management and emergency rescue integrated station |
| CN117131369B (en) * | 2023-10-27 | 2023-12-22 | 福建福昇消防服务集团有限公司 | Data processing method and system of intelligent safety management and emergency rescue integrated station |
| CN118013304A (en) * | 2024-02-19 | 2024-05-10 | 国网湖北省电力有限公司襄阳供电公司 | Transformer fault positioning method and system based on clustering algorithm |
| CN117932523A (en) * | 2024-03-25 | 2024-04-26 | 山东诚祥建设集团股份有限公司 | Construction engineering construction data processing method and system |
| CN117932523B (en) * | 2024-03-25 | 2024-06-04 | 山东诚祥建设集团股份有限公司 | Construction engineering construction data processing method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113139610A (en) | Abnormity detection method and device for transformer monitoring data | |
| CN116861354A (en) | Abnormality identification and mode distinguishing method and system for monitoring data of power transformer | |
| CN110995508B (en) | KPI mutation-based adaptive unsupervised online network anomaly detection method | |
| Kalpakis et al. | Distance measures for effective clustering of ARIMA time-series | |
| US7849124B2 (en) | Method and system for detecting difference between plural observed results | |
| CN111898639B (en) | Dimension reduction-based hierarchical time memory industrial anomaly detection method and device | |
| CN105275833A (en) | CEEMD (Complementary Empirical Mode Decomposition)-STFT (Short-Time Fourier Transform) time-frequency information entropy and multi-SVM (Support Vector Machine) based fault diagnosis method for centrifugal pump | |
| Qiu et al. | A piecewise method for bearing remaining useful life estimation using temporal convolutional networks | |
| Mechefske et al. | Fault detection and diagnosis in low speed rolling element bearings Part II: The use of nearest neighbour classification | |
| CN108399434B (en) | Analysis and prediction method of high-dimensional time series data based on feature extraction | |
| Dervilis et al. | Robust methods for outlier detection and regression for SHM applications | |
| CN112116198A (en) | Data-driven process industrial state perception network key node screening method | |
| CN113704201A (en) | Log anomaly detection method and device and server | |
| CN116597939A (en) | Medicine quality control management analysis system and method based on big data | |
| CN115409066A (en) | Method and device for detecting abnormality of time series data, and computer storage medium | |
| CN116047164A (en) | Detection method and detection device for insulation resistance abnormality of electric automobile | |
| CN118194186A (en) | Snail meat feed production management method and system based on Internet of things | |
| CN112069621B (en) | Method for predicting residual service life of rolling bearing based on linear reliability index | |
| CN118013191A (en) | Denoising method based on LSM filtering and CEEMDAN combined wavelet detection insulator | |
| CN117093944A (en) | Time sequence data template self-adaptive abnormal mode identification method and system | |
| CN115587009A (en) | Cloud platform time sequence data anomaly detection method, system, equipment and medium | |
| CN115545104A (en) | KPI (Key Performance indicator) anomaly detection method, system and medium based on functional data analysis | |
| Zhuang et al. | An autoregressive model-based degradation trend prognosis considering health indicators with multiscale attention information | |
| Reger et al. | Pattern recognition in load profiles of electric drives in manufacturing plants | |
| Zhao et al. | Anomaly detection and pattern differentiation in monitoring data from power transformers |
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: 20210720 |
|
| RJ01 | Rejection of invention patent application after publication |