CN109816031A - A kind of Transformer State Assessment clustering method based on the unbalanced measurement of data - Google Patents
A kind of Transformer State Assessment clustering method based on the unbalanced measurement of data Download PDFInfo
- Publication number
- CN109816031A CN109816031A CN201910093168.4A CN201910093168A CN109816031A CN 109816031 A CN109816031 A CN 109816031A CN 201910093168 A CN201910093168 A CN 201910093168A CN 109816031 A CN109816031 A CN 109816031A
- Authority
- CN
- China
- Prior art keywords
- cluster
- data
- unbalanced
- sample
- class cluster
- 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.)
- Granted
Links
Abstract
The invention discloses a kind of Transformer State Assessment clustering methods based on the unbalanced measurement of data, it include: according to power transformer most common failure index system, required index parameter corresponding with the analysis of power transformer different types of faults is filtered out from unbalanced monitoring data, and handles the index parameter with the normalized method of ratio;Two groups of data are randomly selected in the index parameter as initial cluster center, and clustering parameter is set;Data in unbalanced monitoring data, the lower aprons collection or borderline region of class cluster are divided to according to Euclidean distance by the Euclidean distance for calculating all types of fault indices parameters and initial cluster center, and calculate the unbalanced degree between class cluster;The unbalanced degree of fusion class cluster measures the degree of membership of monitoring data;According to class cluster data distribution situation, calculating is iterated to class cluster center;Status assessment is finally carried out to power transformer according to cluster result;The present invention effectively increases Condition Assessment for Power Transformer precision.
Description
Technical field
The invention belongs to Condition Assessment for Power Transformer technical fields, specifically relate to a kind of change based on the unbalanced measurement of data
Depressor status assessment clustering method.
Background technique
Power transformer is as important power transmission and transforming equipment, and throughout the key node of entire power transmission and transformation network, safety is steady
Fixed operation is for ensureing that power supply is most important.Heavy economic losses is not only resulted in once breaking down, it can also be to society
Stabilization impacts.Therefore, by analyze electricity transformer monitoring data, to Operation Condition of Power Transformers assessed for
Keep entire safe and stable operation of power system important in inhibiting.Traditional state evaluating method is by analysis by failure wave-recording
The fault data of device record, can preferably diagnose the fault type that transformer has occurred.But with Internet of Things skill
Art is in the extensive use of electric system, and the monitoring device of magnanimity is daily all in the monitoring data for generating magnanimity, mass data
There are the monitoring data under normal operating condition, the monitoring data under the exception/fault state for also having density extremely low, power transformer
Monitoring data constitute the data set of typical unbalanced distribution.How power transformer unbalanced data monitoring data are analyzed,
And the hot spot studied at scholar in recent years has been assessed Power Transformer Condition.
Clustering can divide data into several as classical unsupervised learning algorithm according to the feature between data
Class can be used for analyzing the unbalanced monitoring data of power transformer.But traditional clustering algorithm can be by power transformer unevenness
Heng Jianceshuojuji divides equally, so that the monitoring data of normal operating condition are mistakenly divided into fault data, influences electric power change
Depressor status assessment precision.
Summary of the invention
The present invention problem poor for the existing clustering algorithm processing unbalanced monitoring data collection effect of power transformer,
A kind of Transformer State Assessment clustering method based on the unbalanced measurement of data is given, this method is according to power transformer
Most common failure index system, from filtered out in unbalanced monitoring data it is corresponding with the analysis of power transformer different types of faults needed for
Index parameter, and handle the index parameter with the normalized method of ratio;Finally use unbalanced measurement Rough Fuzzy k-
The Condition Assessment for Power Transformer method of means cluster carries out clustering to unbalanced monitoring data, and obtains corresponding electricity
Power Transformer State Assessment result;Particular technique embodiment is as follows:
A kind of Transformer State Assessment clustering method based on the unbalanced measurement of data, which is characterized in that the side
Method includes:
Step 1 filters out and electric power according to power transformer most common failure index system from unbalanced monitoring data
The corresponding required index parameter of transformer different types of faults analysis, and handle the index with the normalized method of ratio and join
Number;
Step 2 randomly selects in the index parameter two groups of data as initial cluster center, and according to historical data
The clustering parameter of the unbalanced monitoring data is arranged in feature;
Step 3 calculates the Euclidean distance of all types of the fault indices data and the initial cluster center, according to
Data in the unbalanced monitoring data are divided to the lower aprons collection or borderline region of class cluster by the distance, and are calculated between class cluster
Unbalanced degree;
Step 4 merges the unbalanced degree and measures to the degree of membership of the unbalanced monitoring data;
Step 5 is iterated calculating to class cluster center, if in class cluster according to step 3 to the cluster result of data sample
The heart no longer updates, and counts the sample of all kinds of cluster lower aprons collection and borderline region, assesses transformer state;Otherwise, it returns
Step 3.
Further, in the step 1, the unbalanced monitoring data are made of the index parameter.
Further, include: in the step 2
Randomly select two groups of data in the index parameter, the normal class cluster of one group of state as power transformer it is initial
Cluster centre, the initial cluster center of one group of failure classes cluster as power transformer;And it is set according to the historical data feature
Set a Distance Judgment threshold value and fuzzy coefficient.
Further, the Distance Judgment threshold value for comparative sample and the normal class cluster cluster heart Euclidean distance and sample with
The relative size of the Euclidean distance of the failure classes cluster cluster heart, and sample is divided to by class cluster lower aprons collection or boundary according to comparison result
Region;Fuzzy coefficient is a constant coefficient 2 in Rough Fuzzy K-means algorithm.
Further, include: in the step 3
Calculate the corresponding index parameter of different types of faults and the normal normal class cluster cluster heart first it is European away from
From, and the second Euclidean distance with the failure classes cluster cluster heart, judge first Euclidean distance with described second it is European away from
From size, and by the two bigger numerical with compared with the ratio of fractional value;
It, will if the ratio is greater than the Distance Judgment threshold value by the ratio and the Distance Judgment threshold value comparison
The index parameter is divided to the institute of the corresponding class cluster of smaller Euclidean distance in first Euclidean distance and the second Euclidean distance
State lower aprons concentration;Conversely, being then divided to the borderline region;
Calculate separately lower aprons collection sample number described in the normal class cluster and failure classes cluster and the unbalanced prison
The ratio of all lower aprons collection sample numbers in measured data obtains unbalanced between the normal class cluster and the failure classes cluster
Degree.
Further, in the step 4:
When data belong to the lower aprons collection in the normal class cluster or failure classes cluster in the unbalanced monitoring data,
Being subordinate to angle value is 1;When sample data belongs to borderline region, degree of membership is needed through degree of membership formula:It is portrayed, wherein uijIt is sample XjFor being subordinate to for i-th class cluster
Degree;dijIt is sample XjWith the Euclidean distance of the cluster heart;M is fuzzy coefficient;K is the class cluster number of cluster.
Further, include: in the step 5
According to step 3 to the cluster result of data sample, calculating is iterated to class cluster center;If class cluster center is no longer
It updates, count the sample of lower aprons collection and borderline region described in corresponding class cluster, assess transformer state: statistics is normal
Class cluster lower aprons collection sample, these samples, which determine, belongs to normal data, and gives these data markers " 1 ", indicates the data pair
Answer power transformer that the type failure does not occur;Failure classes cluster lower aprons collection sample is counted, the determination of these samples belongs to " failure "
Sample, and these sample labelings " -1 " are given, indicate that the type failure has occurred in the transformer;" borderline region " sample is counted, and
Indicate that the type failure may occur for transformer future to these sample labelings " 0 ";If class cluster center continuous updating, returns
Step 3.
Compared with prior art, the invention has the benefit that monitoring data clustering unbalanced for power transformer
Effect is preferable, by introducing the concept of rough set, monitoring data is divided to determination and belong to normal or failure classes cluster, indicate the prison
Measured data, which determines, belongs to normal or fault data;The data of uncertain classification are divided to borderline region, indicate the monitoring data
Belong to abnormal data, the type failure may occur for future, effectively increase Condition Assessment for Power Transformer precision.
Detailed description of the invention
Fig. 1 is the Transformer State Assessment clustering method based on the unbalanced measurement of data described in the embodiment of the present invention
Flow diagram signal.
Fig. 2 is that the Transformer State Assessment clustering based on the unbalanced measurement of data is realized in the embodiment of the present invention
The general frame figure of method is illustrated.
Fig. 3 is the calculation method flow chart signal of degree of membership described in the embodiment of the present invention.
Fig. 4 is described in the embodiment of the present invention based on unbalanced measurement Rough Fuzzy k-means cluster process diagram meaning.
Fig. 5 is the effect for carrying out status assessment clustering in the embodiment of the present invention to transformer using the method for the present invention
Figure.
Fig. 6 is to carry out status assessment to transformer using classical Rough Fuzzy k-means algorithm in the embodiment of the present invention to gather
The effect picture of alanysis.
Specific embodiment
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described.
In embodiments of the present invention, a kind of Transformer State Assessment clustering based on the unbalanced measurement of data is provided
Method, the method are based on unbalanced measurement Rough Fuzzy k-means cluster and realize to Condition Assessment for Power Transformer, refering to figure
2, the present invention can carry out failure to the power transformer to have broken down by the analysis unbalanced monitoring data of power transformer
Diagnosis can also carry out failure predication to the power transformer that the failure does not occur;Specific method and step sees Fig. 1, from
In it is found that the method for the present invention comprising steps of
Step 1 filters out and electric power according to power transformer most common failure index system from unbalanced monitoring data
The corresponding required index parameter of transformer different types of faults analysis, and index parameter is handled with the normalized method of ratio;Its
In, the unbalanced monitoring data in the present invention are made of index parameter, and it is collected to be primarily referred to as electricity transformer monitoring equipment
Data, such as oil dissolved gas H2、CH4、C2H4、C2H6Deng and voltage, the current parameters of transformer etc.;Preferably, this implementation
Each group of prison in unbalanced monitoring data is expressed as one 1 × 5 matrix, i.e. x by exampleij=[xi1,xi2,xi3,xi4,xi5],i
=1,2,3 ..., 200, then use formulaCarry out ratio normalized;In the present invention
Other embodiments in, each group of prison of unbalanced monitoring data can also be expressed as any type of matrix, however it is not limited to 1
× 5 matrix, concrete condition see the number of the index system of selection.
Step 2 randomly selects in index parameter two groups of data as initial cluster center, and unbalanced monitoring number is arranged
According to clustering parameter;Wherein, the data of initial cluster center, which are selected as, randomly selects, that is, randomly selects unbalanced monitoring number
Two groups of data in, the initial cluster center of the normal class cluster of one group of state as power transformer, one group is used as power transformer
The initial cluster center of the failure classes cluster of device;And it is set according to the historical data feature of the physical record of power transformer
One Distance Judgment threshold value and a fuzzy coefficient;Specifically, choosing a small amount of spy from the state recording data to power transformer
Data setting is levied, characteristic is the record data in the case of power transformer normal operation or failure operation, guarantees to set with this
The Distance Judgment threshold value set meets the actual conditions of power transformer;Wherein, Distance Judgment threshold value is for comparative sample and normal
The relative size of the Euclidean distance of the Euclidean distance and sample and the failure classes cluster cluster heart of the class cluster cluster heart, and according to comparison result by sample
Originally class cluster lower aprons collection or borderline region are divided to;Fuzzy coefficient is a constant coefficient 2 in Rough Fuzzy K-means algorithm.
Preferably, it is assumed that Distance Judgment threshold value is usedIt indicates, Distance Judgment threshold value is set as by the present embodimentCertainly,
This is only the preferred embodiment of this implementation, is not limitation and fixation to the value of Distance Judgment threshold value in the method for the present invention,
It can be set according to the feature of the history index parameter of power transformer.
Step 3 calculates the Euclidean distance of all types of fault indices data and initial cluster center, will not according to distance
Data are divided to the lower aprons collection or borderline region of class cluster in balanced monitoring data, and calculate the unbalanced degree between class cluster: firstly,
Calculate the first Euclidean distance of different types of faults corresponding index parameter and the normal class cluster cluster heart, and with the failure classes cluster cluster heart
The second Euclidean distance, judge the size of the first Euclidean distance and the second Euclidean distance, and obtain the bigger numerical in the two with
Compared with the ratio of fractional value;Then, by obtained ratio and Distance Judgment threshold value comparison, if ratio is greater than Distance Judgment threshold value,
Index parameter is divided to the lower aprons concentration that smaller Euclidean distance in the first Euclidean distance and the second Euclidean distance corresponds to class cluster;
Conversely, being then divided to borderline region;Finally, calculating separately the sample number of lower aprons collection and unevenness in normal class cluster and failure classes cluster
The ratio of all lower aprons collection sample numbers, obtains the unbalanced degree between normal class cluster and failure classes cluster in the monitoring data that weigh;Specifically
, unbalanced degree passes through formulaIt is calculated, whereinExpression falls into the sample of approximate set on i-th of class cluster
Number, CzIndicate data sample xjCross cluster where at present.
Step 4 calculates the degree of membership of data in unbalanced monitoring data based on unbalanced degree, refering to Fig. 3, when unbalanced
When data belong to the lower aprons collection in normal class cluster or failure classes cluster in monitoring data, being subordinate to angle value is 1;When sample data category
When borderline region, degree of membership is needed through degree of membership formula:It carries out
It portrays, wherein uijIt is sample XjFor the degree of membership of i-th of class cluster;dijIt is sample XjWith the Euclidean distance of the cluster heart;M is fuzzy
Coefficient;K is the class cluster number of cluster.
In summary content and combine Fig. 4, entirely cluster in the method for the present invention process include: input first it is to be clustered not
Balanced monitoring data randomly select two data samples as the initial cluster heart;Then each sample is calculated at a distance from the cluster heart, and
According to the Distance Judgment threshold value of setting, each sample is divided to the lower aprons collection or borderline region of class cluster;Count each class cluster
Lower aprons collection contained by number of samples, thus calculate unbalanced degree;And it is calculated with the improved degree of membership calculation formula of the present invention every
The degree of membership of a sample;The cluster heart is updated until the cluster heart no longer changes.
Step 5 is iterated calculating to class cluster center, if in class cluster according to step 3 to the cluster result of data sample
The heart no longer updates, and counts the sample of all kinds of cluster lower aprons collection and borderline region, assesses transformer state;Otherwise, it returns
Step 3: counting normal class cluster lower aprons collection sample, these samples, which determine, belongs to normal data, and gives these data markers " 1 ",
Indicate that the data correspond to power transformer and the type failure does not occur;Count failure classes cluster lower aprons collection sample, these samples
It determines and belongs to " failure " sample, and give these sample labelings " -1 ", indicate that the type failure has occurred in the transformer;Count " side
Battery limit (BL) domain " sample, and indicate that the type failure may occur for transformer future to these sample labelings " 0 ";If class cluster center
Continuous updating, then return step three.
Refering to Fig. 5 and Fig. 6, the surveying to the method for the present invention by taking 4 groups of normal data sets and 16 groups of fault data collection as an example
Examination analysis further relates to the validity of the method for the present invention;Specifically, respectively with Rough Fuzzy k-means algorithm and based on unbalanced
The Condition Assessment for Power Transformer method of measurement Rough Fuzzy k-means cluster handles this group of test data, in figure, " * " and " o "
It indicates to cluster correct sample;" ◇ " indicates that the sample for originally belonging to most class class clusters is divided to minority class class cluster by mistake;It is black
" " of color indicates the sample for being divided to borderline region;According to two kinds of algorithms to the assessment result of 20 groups of test datas from the point of view of,
Condition Assessment for Power Transformer method proposed by the present invention based on unbalanced measurement Rough Fuzzy k-means cluster only comment by mistake
Estimate two groups of data, and Rough Fuzzy k-means algorithmic error assesses six groups of data, that is, illustrates the method for the present invention to power transformer
Status assessment and analysis have better accuracy, can be promoted to Power Transformer Condition more precisely prediction;In this way,
It, can well in advance measure and to power transformer according to the status assessment clustering to power transformer in practical operation
Corresponding maintenance is done, the safety in utilization for promoting power transformer is conducive to.
Compared with prior art, the invention has the benefit that monitoring data clustering unbalanced for power transformer
Effect is preferable, by introducing the concept of rough set, monitoring data is divided to determination and belong to normal or failure classes cluster, indicate the prison
Measured data, which determines, belongs to normal or fault data;The data of uncertain classification are divided to borderline region, indicate the monitoring data
Belong to abnormal data, the type failure may occur for future, effectively increase the precision of Condition Assessment for Power Transformer.
Claims (7)
1. a kind of Transformer State Assessment clustering method based on the unbalanced measurement of data, which is characterized in that the method
Include:
Step 1 filters out and power transformer according to power transformer most common failure index system from unbalanced monitoring data
The corresponding required index parameter of device different types of faults analysis, and the index parameter is handled with the normalized method of ratio;
Step 2 randomly selects in the index parameter two groups of data as initial cluster center, and according to historical data feature
The clustering parameter of the unbalanced monitoring data is set;
Step 3 calculates the Euclidean distance of all types of the fault indices data and the initial cluster center, according to described
Data in the unbalanced monitoring data are divided to the lower aprons collection or borderline region of class cluster by distance, and are calculated between class cluster not
Equilibrium degree;
Step 4 merges the degree of membership that the unbalanced degree calculates the unbalanced monitoring data;
Step 5 is iterated calculating to class cluster center according to step 3 to the cluster result of data sample, if class cluster center is not
It updates again, counts the sample of all kinds of cluster lower aprons collection and borderline region, transformer state is assessed;Otherwise, return step
Three.
2. the Transformer State Assessment clustering method as described in claim 1 based on the unbalanced measurement of data, feature
It is, in the step 1, the unbalanced monitoring data are made of the index parameter.
3. the Transformer State Assessment clustering method as described in claim 1 based on the unbalanced measurement of data, feature
It is, includes: in the step 2
Randomly select two groups of data in the index parameter, the initial clustering of the normal class cluster of one group of state as power transformer
Center, the initial cluster center of one group of failure classes cluster as power transformer;And according to historical data feature setting one
Distance Judgment threshold value and fuzzy coefficient.
4. the Transformer State Assessment clustering method as claimed in claim 2 based on the unbalanced measurement of data, feature
It is, Euclidean distance and sample and the failure classes cluster cluster heart of the Distance Judgment threshold value for comparative sample and the normal class cluster cluster heart
Euclidean distance relative size, and sample is divided to by class cluster lower aprons collection or borderline region according to comparison result;Fuzzy system
Number is a constant coefficient 2 in Rough Fuzzy K-means algorithm.
5. the Transformer State Assessment clustering method as claimed in claim 3 based on the unbalanced measurement of data, feature
It is, includes: in the step 3
The first Euclidean distance of the corresponding index parameter of different types of faults and the normal normal class cluster cluster heart is calculated, with
And the second Euclidean distance with the failure classes cluster cluster heart, judge the big of first Euclidean distance and second Euclidean distance
It is small, and by the bigger numerical in the two and compared with the ratio of fractional value;
It, will be described if the ratio is greater than the Distance Judgment threshold value by the ratio and the Distance Judgment threshold value comparison
Index parameter be divided to the corresponding class cluster of smaller Euclidean distance in first Euclidean distance and the second Euclidean distance it is described under
In approximate set;Conversely, being then divided to the borderline region;
Calculate separately lower aprons collection sample number described in the normal class cluster and failure classes cluster and the unbalanced monitoring number
The ratio of all lower aprons collection sample numbers in, obtains the unbalanced degree between the normal class cluster and the failure classes cluster.
6. the Transformer State Assessment clustering method as described in claim 1 based on the unbalanced measurement of data, feature
It is, in the step 4:
When data belong to the lower aprons collection in the normal class cluster or failure classes cluster in the unbalanced monitoring data, it is subordinate to
Angle value is 1;When sample data belongs to borderline region, degree of membership is needed through degree of membership formula:It is portrayed, wherein uijIt is sample XjFor being subordinate to for i-th class cluster
Degree;dijIt is sample XjWith the Euclidean distance of the cluster heart;M is fuzzy coefficient;K is the class cluster number of cluster.
7. the Condition Assessment for Power Transformer as described in claim 1 based on unbalanced measurement Rough Fuzzy k-means cluster
Method, which is characterized in that include: in the step 5
According to step 3 to the cluster result of data sample, calculating is iterated to class cluster center;If class cluster center no longer updates,
The sample for counting lower aprons collection and borderline region described in corresponding class cluster, assesses transformer state: counting normal class cluster
Lower aprons collection sample, these samples, which determine, belongs to normal data, and gives these data markers " 1 ", indicates the corresponding electricity of the data
The type failure does not occur for power transformer;Failure classes cluster lower aprons collection sample is counted, the determination of these samples belongs to " failure " sample,
And these sample labelings " -1 " are given, indicate that the type failure has occurred in the transformer;" borderline region " sample is counted, and gives these
Sample labeling " 0 " indicates that the type failure may occur for transformer future;If class cluster center continuous updating, return step
Three.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910093168.4A CN109816031B (en) | 2019-01-30 | 2019-01-30 | Transformer state evaluation clustering analysis method based on data imbalance measurement |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910093168.4A CN109816031B (en) | 2019-01-30 | 2019-01-30 | Transformer state evaluation clustering analysis method based on data imbalance measurement |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109816031A true CN109816031A (en) | 2019-05-28 |
CN109816031B CN109816031B (en) | 2022-08-05 |
Family
ID=66606017
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910093168.4A Active CN109816031B (en) | 2019-01-30 | 2019-01-30 | Transformer state evaluation clustering analysis method based on data imbalance measurement |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109816031B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110133410A (en) * | 2019-05-31 | 2019-08-16 | 国网河北省电力有限公司石家庄供电分公司 | Diagnosis Method of Transformer Faults and system based on Fuzzy C-Means Cluster Algorithm |
CN110333414A (en) * | 2019-08-02 | 2019-10-15 | 华北电力大学(保定) | The multi-level state evaluating method of power transformer |
CN111598148A (en) * | 2020-04-29 | 2020-08-28 | 中国电子科技集团公司第二十八研究所 | Capacity evaluation method and device based on historical capacity similarity characteristics |
CN112560900A (en) * | 2020-09-08 | 2021-03-26 | 同济大学 | Multi-disease classifier design method for sample imbalance |
CN112599252A (en) * | 2021-03-04 | 2021-04-02 | 北京冠新医卫软件科技有限公司 | Chronic medical record feature classification processing method and device and terminal |
CN113344347A (en) * | 2021-05-25 | 2021-09-03 | 国网湖南省电力有限公司 | Low-voltage power supply line health state assessment method and system |
CN113537528A (en) * | 2021-07-28 | 2021-10-22 | 贵州电网有限责任公司 | Preprocessing method and system for state monitoring data of power transmission and transformation equipment |
CN113869438A (en) * | 2021-09-30 | 2021-12-31 | 自然资源部第三海洋研究所 | Key defense area dividing method for storm surge disaster based on k-means |
CN116153437A (en) * | 2023-04-19 | 2023-05-23 | 乐百氏(广东)饮用水有限公司 | Water quality safety evaluation and water quality prediction method and system for drinking water source |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130297141A1 (en) * | 2012-05-04 | 2013-11-07 | Chungbuk National University Industry-Academic Cooperation Foundation | Apparatus and method for monitoring abnormal state of vehicle using clustering technique |
CN107656154A (en) * | 2017-09-18 | 2018-02-02 | 杭州安脉盛智能技术有限公司 | Based on the Diagnosis Method of Transformer Faults for improving Fuzzy C-Means Cluster Algorithm |
CN109063734A (en) * | 2018-06-29 | 2018-12-21 | 广东工业大学 | The oil-immersed transformer malfunction appraisal procedure clustered in conjunction with multistage local density |
-
2019
- 2019-01-30 CN CN201910093168.4A patent/CN109816031B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130297141A1 (en) * | 2012-05-04 | 2013-11-07 | Chungbuk National University Industry-Academic Cooperation Foundation | Apparatus and method for monitoring abnormal state of vehicle using clustering technique |
CN107656154A (en) * | 2017-09-18 | 2018-02-02 | 杭州安脉盛智能技术有限公司 | Based on the Diagnosis Method of Transformer Faults for improving Fuzzy C-Means Cluster Algorithm |
CN109063734A (en) * | 2018-06-29 | 2018-12-21 | 广东工业大学 | The oil-immersed transformer malfunction appraisal procedure clustered in conjunction with multistage local density |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110133410A (en) * | 2019-05-31 | 2019-08-16 | 国网河北省电力有限公司石家庄供电分公司 | Diagnosis Method of Transformer Faults and system based on Fuzzy C-Means Cluster Algorithm |
CN110333414A (en) * | 2019-08-02 | 2019-10-15 | 华北电力大学(保定) | The multi-level state evaluating method of power transformer |
CN111598148B (en) * | 2020-04-29 | 2022-09-16 | 中国电子科技集团公司第二十八研究所 | Capacity evaluation method and device based on historical capacity similarity characteristics |
CN111598148A (en) * | 2020-04-29 | 2020-08-28 | 中国电子科技集团公司第二十八研究所 | Capacity evaluation method and device based on historical capacity similarity characteristics |
CN112560900A (en) * | 2020-09-08 | 2021-03-26 | 同济大学 | Multi-disease classifier design method for sample imbalance |
CN112560900B (en) * | 2020-09-08 | 2023-01-20 | 同济大学 | Multi-disease classifier design method for sample imbalance |
CN112599252A (en) * | 2021-03-04 | 2021-04-02 | 北京冠新医卫软件科技有限公司 | Chronic medical record feature classification processing method and device and terminal |
CN113344347A (en) * | 2021-05-25 | 2021-09-03 | 国网湖南省电力有限公司 | Low-voltage power supply line health state assessment method and system |
CN113344347B (en) * | 2021-05-25 | 2023-10-20 | 国网湖南省电力有限公司 | Low-voltage power supply line health state assessment method and system |
CN113537528A (en) * | 2021-07-28 | 2021-10-22 | 贵州电网有限责任公司 | Preprocessing method and system for state monitoring data of power transmission and transformation equipment |
CN113537528B (en) * | 2021-07-28 | 2022-05-03 | 贵州电网有限责任公司 | Preprocessing method and system for state monitoring data of power transmission and transformation equipment |
CN113869438A (en) * | 2021-09-30 | 2021-12-31 | 自然资源部第三海洋研究所 | Key defense area dividing method for storm surge disaster based on k-means |
CN116153437A (en) * | 2023-04-19 | 2023-05-23 | 乐百氏(广东)饮用水有限公司 | Water quality safety evaluation and water quality prediction method and system for drinking water source |
Also Published As
Publication number | Publication date |
---|---|
CN109816031B (en) | 2022-08-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109816031A (en) | A kind of Transformer State Assessment clustering method based on the unbalanced measurement of data | |
WO2022110557A1 (en) | Method and device for diagnosing user-transformer relationship anomaly in transformer area | |
CN109409628B (en) | Acquisition terminal manufacturer evaluation method based on metering big data clustering model | |
CN109273096A (en) | A kind of risk management grading evaluation method based on machine learning | |
CN109389145B (en) | Electric energy meter manufacturer evaluation method based on metering big data clustering model | |
CN116148679B (en) | Battery health state prediction method and related device | |
CN111949939B (en) | Method for evaluating running state of intelligent electric meter based on improved TOPSIS and cluster analysis | |
CN110889441B (en) | Power transformation equipment data anomaly identification method based on distance and point density | |
CN101738998B (en) | System and method for monitoring industrial process based on local discriminatory analysis | |
CN110223193A (en) | The method of discrimination and system of operation of power networks state are used for based on fuzzy clustering and RS-KNN model | |
CN110297207A (en) | Method for diagnosing faults, system and the electronic device of intelligent electric meter | |
CN111339297A (en) | Network asset anomaly detection method, system, medium, and device | |
CN113516228B (en) | Network anomaly detection method based on deep neural network | |
CN109376944A (en) | The construction method and device of intelligent electric meter prediction model | |
CN109542742A (en) | Database server hardware health evaluating method based on expert model | |
CN109947815B (en) | Power theft identification method based on outlier algorithm | |
CN115130578A (en) | Incremental rough clustering-based online evaluation method for state of power distribution equipment | |
CN102945222A (en) | Poor information measurement data gross error discrimination method based on Grey System Theory | |
CN111967717A (en) | Data quality evaluation method based on information entropy | |
CN117274827B (en) | Intelligent environment-friendly remote real-time monitoring and early warning method and system | |
CN110321914A (en) | A kind of Oil Quality Analysis managing and control system | |
CN109784777B (en) | Power grid equipment state evaluation method based on time sequence information fragment cloud similarity measurement | |
CN112345972A (en) | Power failure event-based power distribution network line transformation relation abnormity diagnosis method, device and system | |
CN116522111A (en) | Automatic diagnosis method for remote power failure | |
CN116520068A (en) | Diagnostic method, device, equipment and storage medium for electric power data |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |