CN107844540A - A kind of time series method for digging for electric power data - Google Patents
A kind of time series method for digging for electric power data Download PDFInfo
- Publication number
- CN107844540A CN107844540A CN201711007576.0A CN201711007576A CN107844540A CN 107844540 A CN107844540 A CN 107844540A CN 201711007576 A CN201711007576 A CN 201711007576A CN 107844540 A CN107844540 A CN 107844540A
- Authority
- CN
- China
- Prior art keywords
- sequence
- subevent
- frequent
- library
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2477—Temporal data queries
Abstract
The invention discloses a kind of time series method for digging for electric power data, on the basis of being pre-processed in original alarm data, split by feature extraction and suitable time window function pair database, formation sequence set of patterns, prefix projection sequence mining algorithm generation frequent mode is recycled, and template library is weighted by the iterative construction of limited number of time;The present invention is simple with algorithm, and is based entirely on historical data and is excavated, and makes full use of time parameter, and weighting each rule-like in template library has significant otherness, there is important reference to electric network failure diagnosis;Secondly, compared with traditional algorithm, the present invention improves the speed excavated to alarm signal time series, can complete the excavation of large-scale dataset, and the characteristic of generation weights template library, improves the overall performance of time series mining algorithm.
Description
Technical field
The invention belongs to power system control technique field, more specifically, be related to it is a kind of for electric power data when
Between sequential mining method.
Background technology
Using caused remote measurement, remote signalling and event sequence information etc. during electric power system fault, failure judgement element and
Fault type, protection and the breaker of incorrect operation, auxiliary dispatching or operations staff's handling failure are identified, to shorten at accident
The time is managed, prevents fault spread, acceleration system is recovered.When power system is broken down, data with alert amount that control centre receives
Very huge, dispatcher's artificial treatment magnanimity data with alert is extremely difficult, it is necessary to introduces intellectualized technology.
Such as document " power system failure diagnostic rule diggings of Yang Yidong, Sun Zhi the brightness based on time series " describes
Utilize the theoretical application for carrying out data mining technology in power system failure diagnostic of Rough collection.Document " in rainbow, the great of Huang Yan
The associative classification method based on data mining technology is applied in stable operation Rules extraction method based on time series association analysis "
Operation rule extraction is carried out, rule caused by method provides the influence factor information with power system stability operation strong correlation.
Such scheme provides effective support in fault location and crash analysis, but above two scheme belongs to point
Class algorithm, the known accident pattern of timed sample sequence data and positioning, need substantial amounts of sample data to make in learning process
Training effect is more excellent, and this partial data needs advanced pedestrian's work to define;With increasing for sample data dimension and data volume, attribute
The difficulty of yojan and feature extraction rises therewith.
Document " Zhong Jinyuan, the transmission system method for diagnosing faults that a rock is matched based on Time Series Similarity " describes
Time Series Similarity matching process, by constructing corresponding time series models, solved using subsequence matching querying method
Carry out fault diagnosis.This method needs protecting electrical power system configuration universal model support, have to the integrality of experts database more by force according to
Rely.
In document " Fan Xihui, application of a flame sequential mode mining in the processing of power system warning information ", it is proposed that
A kind of Sequential Pattern Mining Algorithm for being based on time window, and by its application and power system warning information Intelligent treatment.
The selection of this method passage time window completes the construction to frequent mode, and FP-Growth algorithms are warned applied to power system
Report information processing is made that discussion.
Because FP-Growth algorithms require that identical element can not be contained in frequent mode, and power system physical fault
There is the failure of a large amount of identical elements in positioning, FP-Growth is had significant limitation in actual signal processing procedure.The calculation
Method need to be to scan database three times, and space-time execution efficiency need to be improved.
The content of the invention
It is an object of the invention to overcome the deficiencies of the prior art and provide a kind of time series excavation for electric power data
Method, to the electric power alarm signal in actual production by successive ignition, realize the extraction of alarm regulation, so as to fast positioning in
The fault diagnosis of power system.
For achieving the above object, a kind of time series method for digging for electric power data of the invention, its feature exist
In comprising the following steps:
(1), alarm data is pre-processed;
Alarm data is cleaned, removes noise and unrelated alarm data, then by the alarm data and data after cleaning
Source associates, and completes data integration, then carries out distributed treatment according to time-sequencing to the alarm data after integrating, obtains pre- place
Alarm data after reason;
(2), pretreated alarm data is handled using time window, obtains subevent sequence library D;
Pretreated alarm data is formed into sequence of events S, then a time window is a sub- sequence of events in S
Sw, Sw=(w, ts,te), SwContain te≤t≤tsAll alarm datas in period, wherein, w=te-ts, represent the time
Window width, teRepresent the initial time of subevent sequence, tsRepresent the end time of subevent sequence;
After sequence of events S is divided according to time window width w, the subevent sequence being made up of multiple subevent sequences is obtained
Database D;
(3), set masterplate and excavate iteration total degree as F, set support, recycle prefixspan algorithms to subevent
Sequence library is excavated;
(3.1), make masterplate excavate iteration time f=1, set a length as l sequence α, find out in the sequence library D of subevent
Sequence number comprising sequence α, the total number that these are found out are labeled as supports of the sequence α in the sequence library D of subevent
Degree, is designated as Support (α);
(3.2), the support of support threshold min-sup, sequence α in the sequence library D of subevent is given to be not less than
Support threshold, i.e. Support (α)>Min-sup, then sequence α is labeled as frequent episode α;
(3.3), iteration total degree l is set;
Iterations t=1 is made, in frequent episode α, it is prefix to take l=1, every height in the sequence library D of subevent
In sequence of events, all frequent episodes using l=1 as prefix are found out, are added to labeled as frequent 1, then all frequent 1
In template library;
Iterations t=2 is made, in frequent episode α, it is prefix to take l=2, every height in the sequence library D of subevent
In sequence of events, all frequent episodes using l=2 as prefix are found out, are added to labeled as frequent 2, then all frequent 2
In template library;
Similarly, as iterations t=l or when producing without frequent episode, iteration terminates;
Using caused frequent episode in each iterative process in template library as the Result after f=1 iteration, while handle
All frequent episodes caused by f=1 are deleted in the sequence library D of subevent, the subevent sequence number after being updated
According to storehouse D;
F=2 is made again, and repeat step (3.1)-(3.3), when as f=F or subevent sequence library D does not update
When, iteration terminates, and will finally give final Result of the template library as sequence.
What the goal of the invention of the present invention was realized in:
A kind of time series method for digging for electric power data of the invention, by being pre-processed in original alarm data
On the basis of, split by feature extraction and suitable time window function pair database, formation sequence set of patterns, recycled
Prefix projection sequence mining algorithm generates frequent mode, and weights template library by the iterative construction of limited number of time;The present invention has
Algorithm is simple, and is based entirely on historical data and is excavated, and makes full use of time parameter, and weighting each rule-like in template library has
Significant otherness, there is important reference to electric network failure diagnosis;Secondly, compared with traditional algorithm, the present invention improves pair
The speed that alarm signal time series is excavated, the excavation of large-scale dataset can be completed, the characteristic weighting template library of generation, is carried
The overall performance of time series mining algorithm is risen.
Brief description of the drawings
Fig. 1 is a kind of time series method for digging flow chart for electric power data of invention;
Fig. 2 is sequence length and quantitative comparison under time windows selection;
Fig. 3 is the time overhead comparison diagram of three kinds of Mining Algorithms of Frequent Patterns;
Fig. 4 is the memory cost comparison diagram of three kinds of Mining Algorithms of Frequent Patterns;
Fig. 5 is that frequent mode quantity length compares under each iterations.
Embodiment
The embodiment of the present invention is described below in conjunction with the accompanying drawings, so as to those skilled in the art preferably
Understand the present invention.Requiring particular attention is that in the following description, when known function and the detailed description of design perhaps
When can desalinate the main contents of the present invention, these descriptions will be ignored herein.
Embodiment
Fig. 1 is a kind of time series method for digging flow chart for electric power data of invention.
In the present embodiment, as shown in figure 1, a kind of time series method for digging for electric power data of the invention, it is special
Sign is, comprises the following steps:
S1, alarm data is pre-processed;
Including data cleansing:Remove noise and unrelated alarm data, for example, the serious unreasonable signal of telemetry value, information type
Signal is not inconsistent it with regulation, be listed signal etc.;
Data integration:Alarm data after cleaning is associated with data source, complete data integration, such as by alarm signal with
Unit account of plant information association is got up;
Hough transformation:Distributed treatment is carried out according to time-sequencing to the alarm data after integrated, obtained pretreated
Alarm data.
S2, using time window pretreated alarm data is handled, obtain subevent sequence library D;
In the present embodiment, it is necessary to select the suitable pretreated alarm data of time window function pair to be handled, such as
Shown in Fig. 2, the increase at window function interval over time, sequence length increases therewith in alarm data, the sequence number of each length
Decline, total sequence number is also on a declining curve, it is therefore desirable to sets suitable time window function according to the actual requirements;
Pretreated alarm data is formed into sequence of events S, then a time window is a sub- sequence of events in S
Sw, Sw=(w, ts,te), SwContain te≤t≤tsAll alarm datas in period, wherein, w=te-ts, represent the time
Window width, teRepresent the initial time of subevent sequence, tsRepresent the end time of subevent sequence;
After sequence of events S is divided according to time window width w, the subevent sequence being made up of multiple subevent sequences is obtained
Database D, as shown in table 1;
Table 1 is subevent sequence library D;
Serial ID | Sequence |
1 | <a(abc)(ac)d(cf)> |
2 | <(ad)c(bc)(ae)> |
3 | <(ef)(ab)(df)cb> |
4 | <eg(af)cbc> |
Table 1
S3, masterplate excavation iteration total degree is set as F, set support, recycle prefixspan algorithms to subevent
Sequence library is excavated;
S3.1, make masterplate excavate iteration time f=1, set a length as l sequence α, find out in the sequence library D of subevent
Sequence number comprising sequence α, the total number that these are found out are labeled as supports of the sequence α in the sequence library D of subevent
Degree, is designated as Support (α);
S3.2, supports of given the support threshold min-sup, sequence α in the sequence library D of subevent are not less than branch
Degree of holding threshold value, i.e. Support (α)>Min-sup, then sequence α is labeled as frequent episode α;
S3.3, iteration total degree l is set;
Iterations t=1 is made, in frequent episode α, it is prefix to take l=1, every height in the sequence library D of subevent
In sequence of events, all frequent episodes using l=1 as prefix are found out, are added to labeled as frequent 1, then all frequent 1
In template library;As shown in table 2, if support is arranged to 50%, frequent 1 is { a, b, c, d, e, f };
Table 2 is the frequent episode using l=1 as prefix;
a | b | c | d | e | f | g |
4 | 4 | 4 | 3 | 3 | 3 | 1 |
Table 2
Iterations t=2 is made, in frequent episode α, it is prefix to take l=2, every height in the sequence library D of subevent
In sequence of events, all frequent episodes using l=2 as prefix are found out, are added to labeled as frequent 2, then all frequent 2
In template library;
By taking frequent 1 ' d ' as an example, suffix corresponding to ' d ' is as shown in table 3, is counted, obtained with the degree of being supported of support 50%
To { a:1,b:2,c:3,e:1,f:1,_f:1 }, then frequent 2 of prefix d for<db>,<dc>}.
Table 3 is frequent 2 of prefix d;
<d> |
<(cf)> |
<c(bc)(ae)> |
<(_f)cb> |
Table 3
Similarly, as iterations t=l or when producing without frequent episode, iteration terminates;
Using caused frequent episode in each iterative process in template library as the Result after f=1 iteration, while handle
All frequent episodes caused by f=1 are deleted in the sequence library D of subevent, the subevent sequence number after being updated
According to storehouse D;
F=2, repeat step S3.1-S3.3 are made again, when as f=F or when subevent sequence library D does not update,
Iteration terminates, and will finally give final Result of the template library as sequence.
Simulating, verifying
In the present embodiment, from tri- kinds of Mining Algorithms of Frequent Patterns of PrefixSpan, Spade and Gsp, as shown in Figure 3
Understand in three kinds of Mining Algorithms of Frequent Patterns, the time overhead of the PrefixSpan algorithms of use under each support is superior to
Other algorithms, especially under support reduced levels, time overhead has significant advantage.
As shown in Figure 4 in three kinds of Mining Algorithms of Frequent Patterns, the PrefixSpan algorithms used herein are in each support
Under memory consumption it is more stable and memory cost is in reduced levels.
With the increase of iterations it can be seen from Fig. 5, sequence pattern par is in logarithm downward trend, with
The increase iteration effect of iterations is by unobvious.The average length of frequent mode is as the increase of iterations is in after falling before
The situation of rising, the power network of frequent mode major embodiment caused by first time iteration normally inform alarm signal sequence, quantity
More but sequence length is small.Sequence pattern part after successive ignition is accident class sequence, and such sequence is due to frequency
Seldom, support is small but sequence length substantially rises.
It shown below is in certain sometime interior specific iterative process each time of standing of southwest, excavate the process of failure.
4~table of table 6 is illustrated using above-mentioned electric power data time series method for digging, by secondary iteration, including to one
Original sequence data storehouse, two renewal masterplate frequent-item three times result.As a result each iteration Result is embodied
Length characteristic, quantitative characteristics and the equipment associate feature with power system, alarm sequence is in device class and fault type
There is obvious differentiation, have important references value to electric power system fault modeling.
Every table is specifically described for we below.
Table 4 illustrates certain frequent mode caused by certain first time in month iteration of standing of southwest and concentrates each representational frequency of length
Numerous schema entry, under the support, incidence relation between the tranformer protection signal composite signal represented in frequent mode.
First set protection XX-T2 low-pressure sides back-up protection outlet | 2# transformers |
First set protection XX-T2 medium voltage sides back-up protection outlet | 3# transformers |
Frequent 2 examples | Frequency:656 |
First set protection XX-T2 low-pressure sides back-up protection outlet | 2# transformers |
The side outlet of first set protection XX-T2 malfunctioning switches shunt tripping three | 2# transformers |
First set protection XX-T2 medium voltage sides back-up protection outlet | 2# transformers |
The side outlet of first set protection XX-T2 malfunctioning switches shunt tripping three | 3# transformers |
First set protection XX-T2 medium voltage sides back-up protection outlet | 3# transformers |
Frequent 5 examples | Frequency:486 |
First set protection XX-T2 low-pressure sides back-up protection outlet | 2# transformers |
The side outlet of first set protection XX-T2 malfunctioning switches shunt tripping three | 2# transformers |
First set protection XX-T2 medium voltage sides back-up protection outlet | 2# transformers |
First set protection XX-T2TV broken strings | 3# transformers |
First set protection XX-T2 low-pressure sides back-up protection outlet | 3# transformers |
The side outlet of first set protection XX-T2 malfunctioning switches shunt tripping three | 3# transformers |
Frequent 6 examples | Frequency:305 |
Table 4
Table 5 illustrates certain frequent mode caused by second of iteration of certain month of standing of southwest and concentrates each representational frequency of length
Numerous schema entry, reflect the frequent starting of the period failure wave-recording equipment.
Table 5
Table 6 illustrates certain frequent mode caused by certain third time iteration of standing of southwest in month and concentrates representational frequent mode
, the frequent mode number is 1, sequence length 31, reflects the intelligent terminal and combining unit alarm sequence of certain circuit.
Table 6
Although the illustrative embodiment of the present invention is described above, in order to the technology of the art
Personnel understand the present invention, it should be apparent that the invention is not restricted to the scope of embodiment, to the common skill of the art
For art personnel, if various change in the spirit and scope of the present invention that appended claim limits and determines, these
Change is it will be apparent that all utilize the innovation and creation of present inventive concept in the row of protection.
Claims (1)
1. a kind of time series method for digging for electric power data, it is characterised in that comprise the following steps:
(1), alarm data is pre-processed;
Alarm data is cleaned, removes noise and unrelated alarm data, then the alarm data after cleaning and data source are closed
Connection, data integration is completed, distributed treatment then is carried out according to time-sequencing to the alarm data after integrating, after obtaining pretreatment
Alarm data;
(2), pretreated alarm data handle using time window, obtain subevent sequence library D;
Pretreated alarm data is formed into sequence of events S, then a time window is a sub- sequence of events S in Sw, Sw=
(w,ts,te), SwContain te≤t≤tsAll alarm datas in period, wherein, w=te-ts, time window width is represented,
tsRepresent the end time of subevent sequence;
After sequence of events S is divided according to time window width w, the subevent sequence data being made up of multiple subevent sequences is obtained
Storehouse D;
(3), set masterplate and excavate iteration total degree as F, set support, recycle prefixspan algorithms to sub- sequence of events
Database is excavated;
(3.1), make masterplate excavate iteration time f=1, set a length as l sequence α, find out in the sequence library D of subevent and include
Sequence α sequence number, the total number that these are found out are labeled as supports of the sequence α in the sequence library D of subevent, note
For Support (α);
(3.2) support of support threshold min-sup, sequence α in the sequence library D of subevent, is given not less than support
Spend threshold value, i.e. Support (α)>Min-sup, then sequence α is labeled as frequent episode α;
(3.3), iteration total degree l is set;
Iterations t=1 is made, in frequent episode α, it is prefix to take l=1, each subevent in the sequence library D of subevent
In sequence, all frequent episodes using l=1 as prefix are found out, masterplate is added to labeled as frequent 1, then all frequent 1
In storehouse;
Iterations t=2 is made, in frequent episode α, it is prefix to take l=2, each subevent in the sequence library D of subevent
In sequence, all frequent episodes using l=2 as prefix are found out, masterplate is added to labeled as frequent 2, then all frequent 2
In storehouse;
Similarly, as iterations t=l or when producing without frequent episode, iteration terminates;
Using caused frequent episode in each iterative process in template library as the Result after f=1 iteration, while f=1
Caused all frequent episodes are deleted in the sequence library D of subevent, the subevent sequence library after being updated
D;
F=2 is made again, repeat step (3.1)-(3.3), when as f=F or when subevent sequence library D does not update, is changed
In generation, terminates, and will finally give final Result of the template library as sequence.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711007576.0A CN107844540A (en) | 2017-10-25 | 2017-10-25 | A kind of time series method for digging for electric power data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711007576.0A CN107844540A (en) | 2017-10-25 | 2017-10-25 | A kind of time series method for digging for electric power data |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107844540A true CN107844540A (en) | 2018-03-27 |
Family
ID=61663125
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711007576.0A Pending CN107844540A (en) | 2017-10-25 | 2017-10-25 | A kind of time series method for digging for electric power data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107844540A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108667648A (en) * | 2018-04-03 | 2018-10-16 | 南方电网调峰调频发电有限公司 | A kind of alarm sequential pattern mining method based on network and time-constrain |
CN109471885A (en) * | 2018-09-30 | 2019-03-15 | 齐鲁工业大学 | Based on the data analysing method and system for weighting positive and negative sequence pattern |
CN110108980A (en) * | 2019-04-29 | 2019-08-09 | 国网宁夏电力有限公司电力科学研究院 | A kind of recognition methods of the anomalous event of electric system and device |
CN112734261A (en) * | 2021-01-18 | 2021-04-30 | 国网山东省电力公司菏泽供电公司 | Power distribution network operation index sequence correlation analysis method and system |
CN114091625A (en) * | 2022-01-18 | 2022-02-25 | 岚图汽车科技有限公司 | Vehicle part failure prediction method and system based on fault code sequence |
CN114390086A (en) * | 2021-06-30 | 2022-04-22 | 国网新疆电力有限公司信息通信公司 | Method suitable for power communication operation analysis |
CN116780781A (en) * | 2023-08-24 | 2023-09-19 | 滇恒能源技术(云南)有限公司 | Power management method for smart grid access |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102809965A (en) * | 2012-07-30 | 2012-12-05 | 燕山大学 | Fault early warning method for hydraulic equipment based on fault frequent pattern |
CN104537025A (en) * | 2014-12-19 | 2015-04-22 | 北京邮电大学 | Frequent sequence mining method |
CN105138916A (en) * | 2015-08-21 | 2015-12-09 | 中国人民解放军信息工程大学 | Multi-track malicious program feature detecting method based on data mining |
-
2017
- 2017-10-25 CN CN201711007576.0A patent/CN107844540A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102809965A (en) * | 2012-07-30 | 2012-12-05 | 燕山大学 | Fault early warning method for hydraulic equipment based on fault frequent pattern |
CN104537025A (en) * | 2014-12-19 | 2015-04-22 | 北京邮电大学 | Frequent sequence mining method |
CN105138916A (en) * | 2015-08-21 | 2015-12-09 | 中国人民解放军信息工程大学 | Multi-track malicious program feature detecting method based on data mining |
Non-Patent Citations (2)
Title |
---|
汪林林,范军: "基于PrefixSpan 的序列模式挖掘改进算法", 《计算机工程》 * |
闫伟,张浩,陆剑峰: "基于设备故障监控的时间序列模式研究应用", 《计算机应用》 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108667648A (en) * | 2018-04-03 | 2018-10-16 | 南方电网调峰调频发电有限公司 | A kind of alarm sequential pattern mining method based on network and time-constrain |
CN109471885A (en) * | 2018-09-30 | 2019-03-15 | 齐鲁工业大学 | Based on the data analysing method and system for weighting positive and negative sequence pattern |
CN109471885B (en) * | 2018-09-30 | 2022-05-31 | 齐鲁工业大学 | Data analysis method and system based on weighted positive and negative sequence mode |
CN110108980A (en) * | 2019-04-29 | 2019-08-09 | 国网宁夏电力有限公司电力科学研究院 | A kind of recognition methods of the anomalous event of electric system and device |
CN110108980B (en) * | 2019-04-29 | 2021-08-17 | 国网宁夏电力有限公司电力科学研究院 | Method and device for identifying abnormal event of power system |
CN112734261A (en) * | 2021-01-18 | 2021-04-30 | 国网山东省电力公司菏泽供电公司 | Power distribution network operation index sequence correlation analysis method and system |
CN112734261B (en) * | 2021-01-18 | 2023-05-16 | 国网山东省电力公司菏泽供电公司 | Power distribution network operation index sequence association analysis method and system |
CN114390086A (en) * | 2021-06-30 | 2022-04-22 | 国网新疆电力有限公司信息通信公司 | Method suitable for power communication operation analysis |
CN114390086B (en) * | 2021-06-30 | 2023-11-17 | 国网新疆电力有限公司信息通信公司 | Method suitable for power communication operation analysis |
CN114091625A (en) * | 2022-01-18 | 2022-02-25 | 岚图汽车科技有限公司 | Vehicle part failure prediction method and system based on fault code sequence |
CN116780781A (en) * | 2023-08-24 | 2023-09-19 | 滇恒能源技术(云南)有限公司 | Power management method for smart grid access |
CN116780781B (en) * | 2023-08-24 | 2023-11-10 | 滇恒能源技术(云南)有限公司 | Power management method for smart grid access |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107844540A (en) | A kind of time series method for digging for electric power data | |
CN102098175B (en) | Alarm association rule obtaining method of mobile internet | |
CN103488802A (en) | EHV (Extra-High Voltage) power grid fault rule mining method based on rough set association rule | |
CN105372557A (en) | Power grid resource fault diagnosis method based on association rules | |
CN103135014A (en) | Transformer fault diagnosis method based on case-based reasoning | |
CN102403718B (en) | Generating method for power grid topological relationship based on Arcgis | |
CN107704610A (en) | A kind of power distribution network operation data event correlation analysis system and analysis method | |
CN105184394A (en) | On-line data mining optimized control method based on cyber physical system (CPS) of power distribution network | |
CN107784392A (en) | A kind of the defects of transmission line of electricity based on machine learning Forecasting Methodology and device | |
CN105139031A (en) | Data processing method based on subspace clustering | |
CN106155985A (en) | A kind of shortage of data fill method based on adjacent data feature | |
CN104317794A (en) | Chinese feature word association pattern mining method based on dynamic project weight and system thereof | |
CN113268370B (en) | Root cause alarm analysis method, system, equipment and storage medium | |
CN110765582B (en) | Self-organization center K-means microgrid scene division method based on Markov chain | |
CN105529714A (en) | Normal distribution combination characteristic-based rapid probabilistic power flow calculation method | |
CN114153821A (en) | Electric quantity graph database construction and search method based on graph theory | |
CN112181715A (en) | Visual backup and comparison method based on distribution network automation system model | |
CN116662412B (en) | Data mining method for big data of power grid distribution and utilization | |
CN116415002B (en) | Power grid fault recovery error-proof checking method based on graph matching | |
Vo et al. | Parallel method for mining high utility itemsets from vertically partitioned distributed databases | |
CN104821580A (en) | Three-phase reactive power control partitioning method based on reactive power source clustering analysis | |
CN105306075B (en) | A kind of three value FPRM circuit power consumption optimum polarity search methods | |
CN113536508B (en) | Manufacturing network node classification method and system | |
CN112711947B (en) | Text vectorization-based fault power failure emergency repair handling reference method | |
CN114036319A (en) | Power knowledge extraction method, system, device and storage medium |
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: 20180327 |
|
RJ01 | Rejection of invention patent application after publication |