CN107301243A - Switchgear fault signature extracting method based on big data platform - Google Patents
Switchgear fault signature extracting method based on big data platform Download PDFInfo
- Publication number
- CN107301243A CN107301243A CN201710550324.6A CN201710550324A CN107301243A CN 107301243 A CN107301243 A CN 107301243A CN 201710550324 A CN201710550324 A CN 201710550324A CN 107301243 A CN107301243 A CN 107301243A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- data
- msubsup
- mfrac
- 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
- G06F16/2453—Query optimisation
- G06F16/24532—Query optimisation of parallel queries
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/327—Testing of circuit interrupters, switches or circuit-breakers
- G01R31/3271—Testing of circuit interrupters, switches or circuit-breakers of high voltage or medium voltage devices
- G01R31/3275—Fault detection or status indication
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/17—Details of further file system functions
- G06F16/172—Caching, prefetching or hoarding of files
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/17—Details of further file system functions
- G06F16/176—Support for shared access to files; File sharing support
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/18—File system types
- G06F16/182—Distributed file systems
-
- 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
- G06F16/2471—Distributed queries
-
- 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
- G06F16/248—Presentation of query results
-
- 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
- G06F16/25—Integrating or interfacing systems involving database management systems
- G06F16/258—Data format conversion from or to a database
Abstract
The present invention proposes a kind of feature extracting method of the switchgear failure based on big data platform, mainly solves prior art in the data in face of magnanimity switchgear failure, it is impossible to the problem of efficiently and accurately carrying out feature extraction to various fault types.Its implementation is：Build Hadoop sub-platforms and carry out Data Collection, storage and data prediction；Build SparkR platforms and carry out multivariable multi-scale entropy MMSE Distributed Calculation, and result of calculation is saved in distributed file system HDFS；Result of calculation is downloaded from HDFS, the multivariate sample entropy curve of each failure of R Software on Drawing switchgears is utilized；According to the multivariate sample entropy curve of each failure, the multivariate sample entropy for choosing correspondence scale factor scope is used as the characteristic parameter of each failure.Whole conceptual design of the invention is rigorous, complete, possesses mass data storage and distributed computation ability, and the efficiency and accuracy that fault signature is extracted are high, can provide foundation for diagnosis and anticipation switchgear failure in time.
Description
Technical field
The invention belongs to industrial big data processing technology field, specifically a kind of switchgear fault signature extracting method,
It can be applied to the feature extraction to the various failures of enterprise switches equipment.
Background technology
Switchgear is born in power system as one of power system terminal device and controls and protect dual
Business, its reliability and intelligent level will produce far-reaching influence to the stabilization and automaticity of power system.Switchgear
The statistical analysis of accident shows that the reason for causing primary cut-out failure mainly has abnormal operating mechanism, SF6 leakages, assisted parts
Part is damaged and critical piece deterioration.The influence factor of switchgear failure mainly have the use time of switchgear, annual load factor,
Environment Operation class, temperature, number of operations and current times etc..Fault characteristic parameters are extracted by switching devices, are timely
Diagnosis and anticipation switchgear failure provide foundation, reduce its O＆M cost.
Traditional feature extracting method is typically that feature extraction is carried out under unit, serial mode.Traditional characteristic extraction side
The treatable data volume of method is smaller, and this accuracy to feature extraction has considerable influence.When in face of mass data, data
Storage and processing can expose that poor fault tolerance, speed is slow, the low problem of efficiency.With switchgear failure influence factor not
Disconnected to increase, the continuous expansion of equipment fault data scale is difficult to store mass data under unit serial mode, can not be significantly
Improve data processing speed；Simultaneously under unit serial mode, traditional characteristic extracting method can only handle small sample amount data, enter
And the accuracy of feature extraction can be reduced.When fault impact factor constantly expands, traditional characteristic extracting method is also difficult to handle
The data set of multivariable.
In summary, traditional feature extracting method is typically only capable to handle small sample amount data.When the variable of data set increases
When many, traditional feature extracting method is also difficult to the data set for handling multivariable gradually.The data volume of the every kind of failure of switchgear
Not only big, the influence factor of every kind of failure is more than use time, load factor, environment Operation class, temperature, number of operations and opened
This six factors of disconnected number of times, future, fault impact factor can be on the increase.Therefore, traditional feature extracting method is difficult to face simultaneously
Continuous expansion and influence factor to fault data amount are on the increase, and then the speed and the degree of accuracy extracted to fault signature are caused
Influence.
The content of the invention
It is an object of the invention to for above-mentioned problem of the prior art, propose a kind of opening based on big data processing platform
Equipment fault feature extracting method is closed, to improve the degree of accuracy and the extraction rate of the extraction of switchgear fault signature.
The present invention technical thought be：By the processing to big-sample data amount, introduce multivariable multi-scale entropy MMSE and calculate
Method, to solve the problem of fault impact factor is on the increase, improves the degree of accuracy that switchgear fault signature is extracted indirectly；Pass through
Distributed Parallel Computing of the MMSE algorithms on SparkR platforms, improves the speed that switchgear fault signature is extracted.It is realized
Scheme includes as follows：
Switchgear fault signature extracting method proposed by the present invention based on big data platform, step includes as follows：
(1) SparkR big data platforms are built：
(1a) installs linux system, Hadoop open source softwares and Spark open source softwares；
(1b) determines the node number of platform cluster according to existing fault data scale, and to be processed according to subsequently needing
Fault data scale, can extend or reduce to the node number；
Each node of (1c) configuration platform cluster, i.e., regard any 1 node as host node from the nodes of determination
Master, remaining is as from node Slave；
(1d) it is determined that host node Master and all from node Slave, configuration server process SSH (Secure
Shell) and carry out without password authentification, and Java software, configuration Java context, configuration Hadoop core documents and Spark are installed
Core document；
(2) Data Collection and storage：Host node Master is adopted from platform exterior by Hadoop Sqoop component technologys
Collect the fault data of relationship type；The fault data of file type is gathered by Flume component technologys, and these data of collection are deposited
In the distributed file system HDFS for storing up Hadoop, host node Master and all these data are shared from node Slave；
(3) data prediction：Carry out conversion successively to the fault data in distributed file system HDFS and normalized
Pretreatment, quality data is provided for subsequent data analysis；
(4) data distribution formula is calculated：
On local host, the multivariable multi-scale entropy MMSE that can only be run on unit is rewritten into energy using R softwares
The distributed algorithm run on big data platform SparkR；
Host node Master calls MMSE's by big data platform SparkR SparkR api interfaces from local host
Distributed algorithm, is deployed to each from node Slave, and is used as using pretreated data the input of the algorithm；
From the multivariate sample entropy of each failure of node Slave parallel computations, and result of calculation is saved in Hadoop's
In distributed file system HDFS；
(5) visual presentation：Under stand-alone environment, local host is from the distributed file system HDFS of big data platform
Result data is downloaded, the multivariate sample entropy song of the various failures of switchgear is then drawn using the drawing function for R softwares of increasing income
Line；
(6) feature extraction：According to the multivariate sample entropy curve of each failure, each damage curve of selection is all shallower, and respectively
The multivariate sample entropy of failure correspondence scale factor differs larger scale factor scope each other, and by the scale factor scope
Multivariate sample entropy as each failure characteristic parameter.
The present invention compared with prior art, with advantages below：
1) present invention uses big data platform SparkR, will increase income R softwares and Spark softwares of increasing income are combined by force, can be with
Spark elasticity distribution formula data set RDD and DataFrameAPI is seamlessly used in R softwares, by Spark internal memory meters
The advantage of a variety of computation models is supported in calculation, unified software stack, distributed data calculating and analysis is efficiently carried out, big advise is solved
The challenge that mould data set is brought.
2) though existing multivariable multi-scale entropy MMSE is applied in the ambits such as physics, physiology, opening
Equipment fault analysis field is closed, MMSE is not employed also.Multivariable multi-scale entropy MMSE algorithms are applied to switch by the present invention
Equipment fault analysis field, solves the problem of fault impact factor increases, and the extraction of switchgear fault signature is improved indirectly
The degree of accuracy；Distributed Parallel Computing of the MMSE algorithms on SparkR platforms is realized, the speed of feature extraction is improved.
3) present invention can become apparent from, more intuitively due to introducing scale factor in multivariable multi-scale entropy MMSE
Distinguish several malfunction types of switchgear.
Brief description of the drawings
Fig. 1 be the present invention realize general flow chart；
Fig. 2 is interior joint configuration flow figure of the present invention
Fig. 3 is SparkR integrated stand compositions in the present invention
Fig. 4 is the flow chart of feature extraction algorithm MMSE in the present invention
Fig. 5 is the multivariate sample entropy curve map of 4 kinds of failures
Embodiment
The present invention is elaborated with reference to the accompanying drawings and detailed description.
When the feature extracting method of traditional switchgear failure faces magnanimity fault impact factor data, do not possess big rule
Mould data storage and disposal ability, are all that feature extraction is carried out under unit, serial mode, speed is slow, efficiency is low and security
Difference, directly influences the efficiency and accuracy of fault signature extraction.
Hadoop big data processing platforms, its HDFS distributed file systems and MapReduce programming modes is relatively good
Ground solves the problem of mass data distributed storage and processing.Compared with Hadoop, Spark provides distributed data collection
Abstract, programming model is more flexible and efficient, internal memory can be made full use of to carry out improving performance.Spark can solve iteration well
Computing and interactive computing, it introduces elasticity distribution formula data set RDD, there is a fault tolerant mechanism, and data acquisition system can be by simultaneously
Row operation, can be cached in internal memory, without reloading data from HDFS every time as MapReduce.
Pretreated data set is created into RDD by data calculation process, Spark, be cached to internal memory, and then is performed parallel by multiple
Task is reused.R softwares possess powerful function of statistic analysis and abundant third party's expanding packet, but the core fortune of R softwares at present
Row environment is single thread, and treatable data volume is limited to the memory size of unit, the mass data processing in big data epoch
The challenge to R software sharings.SparkR will increase income R softwares and Spark softwares of increasing income are combined by force, can be seamless in R softwares
Ground uses Spark RDD and DataFrameAPI, is calculated by Spark internal memories, unifies in software stack to support a variety of computation models
Advantage, efficiently carry out distributed data calculating and analysis, solve the challenge that large-scale dataset is brought.Therefore, it is of the invention
SparkR platforms are introduced, by the processing to big-sample data amount, multivariable multi-scale entropy MMSE algorithms are introduced, to solve failure
The problem of influence factor is on the increase, improves the degree of accuracy that switchgear fault signature is extracted indirectly；Existed by MMSE algorithms
Distributed Parallel Computing on SparkR platforms, improves the speed that switchgear fault signature is extracted.
Reference picture 1, step is as follows for of the invention realizing：
Step 1, SparkR big data platforms are built.
(1a) installs the CentOS-6.3 versions of linux system, the Hadoop- for Hadoop softwares of increasing income on local host
2.6.0 version, the Spark-1.4.0 versions for Spark softwares of increasing income.
With reference to table 1, the correlation technique component needed for mounting platform SparkR sub-platform Hadoop, including Flume,
Sqoop。
Technology component needed for the sub-platform Hadoop of table 1
Wherein：Core：Represent distributed file system and general purpose I/O components and interface；
Avro：Represent to provide efficiently, across language RPC data sequence system, perdurable data storage；
HDFS：Distributed file system is represented, for realizing that the piecemeal of large-scale data is stored；
MapReduce：Represent distributed data processing framework and performing environment；
Zookeeper：Represent the distributed coordination service of high availability；
Pig：Data-flow language and running environment are represented, to retrieve large-scale dataset；
Chukwa：The collector of data storage in operation HDFS is represented, analysis report is generated using MapReduce；
Mahout：Represent machine learning algorithm storehouse；
Flume：Represent result collection system；
Sqoop：Data syn-chronization instrument is represented, for transmitting data between traditional data and Hadoop；
(1b), according to existing fault data scale, the node number for determining platform cluster is 4；And need place according to follow-up
The fault data scale of reason, can extend or reduce to the node number；
(1c) with reference to table 2, each node of configuration platform cluster, i.e., from the nodes of determination using any 1 node as
Host node Master, remaining is connected as from LAN between node Slave, node；
The Master nodes mainly configure name manager NameNode and task manager JobTracker role, bear
Blame the execution of house steward's distributed data and task resolution；Host node Master attribute is NameNode, its as master server,
Operated for managing the access of the NameSpace and client of file system to file system；
This 3 are transported actuator from node Salve1, Slave2 and Slave3 configuration data memory DataNode and task
TaskTracker, is responsible for the execution of Distributed Storage and task.Attribute from node Slave is DataNode, and it is led
Want the data that function is management storage.
The platform cluster node structure of table 2
| Namespace node | Ip addresses | Attribute |
| Master | 192.168.137.2 | NameNode |
| Slave1 | 192.168.137.3 | DataNode |
| Slave2 | 192.168.137.4 | DataNode |
| Slave3 | 192.168.137.5 | DateNode |
(1d) installs related software in host node and three from node and configures associated documents：
Reference picture 2, it is determined that host node Master and three from node lave1, Slave2 and Slave3, configuration clothes
Be engaged in device process SSH simultaneously carry out without password authentification, and install Java software, configuration Java context, configuration Hadoop core documents and
Spark core documents；Wherein Hadoop core documents include core-site.xml, hdfs-site.xml, mapred-
Site.xml and yarn-site.xml；Spark core documents include Spark-env.sh, slaves and profile.
After the completion of above-mentioned (1a)~(1d) steps, platform SparkR overall architectures are obtained, as is shown with reference to figure 3.
Reference picture 3, the SparkR platforms that this example is built, including each node of cluster and virtual machine JVM rear ends two parts.
SparkR provides elasticity distribution formula data set RDD and data frame application programming interfaces DataFrame for the operation of R softwares
API.SparkR API are operated in R softwares, and Core is operated in virtual machine JVM.JVM rear ends are a groups in Core
Part there is provided the bridging functionality between R softwares and virtual machine JVM, can allow R software programmings code establishing java class reality
Example, call the case method of Java object or the static method of java class.SparkR DataFrame API need not be passed
Enter, the data in data frame DataFrame are entirely to be stored with JVM data type.DataFrame API further comprises one
Part RDD API.
During work, DataFrame is first converted into elasticity distribution formula data set RDD, elasticity distribution formula data are then called
Collect RDD packet, polymerization and repartition operation, launching process RWorker carries out MMSE Distributed Calculation.By using
The customized simple efficient binary protocol socket in family, the R softwares after host node RDD partition data, serializing are compiled
The algorithm routine and other information write are transmitted to process Rworker, the partition data that process Rworker unserializings are received and
The algorithm routine of R software programmings, the algorithm routine of R software programmings is applied on partition data, then result data is serialized
JVM ends are passed back into byte arrays.
Step 2, Data Collection and storage.
Host node Master gathers the fault data of relationship type from platform exterior by Hadoop Sqoop component technologys；
The fault data of file type is gathered by Flume component technologys, and by these data Cun Chudao Hadoop of collection distribution
In file system HDFS, host node Master and all these data are shared from node Slave.
The fault data that described Sqoop component technologys and Flume component technologys are collected is respectively 5000, such as 3~table of table 6
It is shown.
The fault category of table 3 is every influence factor data of " operating mechanism is abnormal "
The fault category of table 4 is every influence factor data of " SF6 leakages "
The fault category of table 5 is every influence factor data of " accessory damage "
The fault category of table 6 is the influence factor data of " critical piece deterioration "
Step 3, to being stored in the fault data in distributed file system HDFS conversion is carried out successively and normalized pre-
Processing.
(3a) in data set with the data conversion that represents of interval into corresponding single number：
" less than 40% " in influence factor " annual load factor " is converted into 0.25, " 40%~60% " is converted into
0.5, " 60%~80% " is converted into 0.75, and " more than 80% " is converted into 0.9.
It is interval that attribute in data set is normalized to [0,1] by (3b)：
Wherein, x is the actual value of the influence factor of each failure, xmax、xminMaximum and minimum respectively in actual value
Value, y is the value after normalization.
Step 4, data distribution formula is calculated.
On local host, the multivariable multi-scale entropy MMSE that can only be run on unit is rewritten into energy using R softwares
The distributed algorithm run on big data platform SparkR；
Host node Master calls MMSE point by big data platform SparkR SparkRAPI interfaces from local host
Cloth algorithm, is deployed to each from node Slave, and is used as using pretreated data the input of the algorithm；From node
The multivariate sample entropy of each failure of Slave parallel computations, and result of calculation is saved in Hadoop distributed file system
In HDFS.
Reference picture 4, fault signature extraction algorithm MMSE flow is as follows：
(4a) determines embedded dimension m=(2,2,2,2,2,2), delay vector τ=(1,1,1,1,1,1), threshold value r=0.2*
Sd, sd are the standard deviation of each variable, scale factor ε=1,2 ..., 20；Determined according to the number of the influence factor of failure
First variable p=6, according to fault data bar, several numbers determine the second variable N=5000；
(4b) builds length as N using pretreated data and includes the data set { x of p variablek,i, wherein i=1,
2,...,N；K=1,2 ..., p；
(4c) is to multivariate data collectionThick-breakpoint processing is carried out based on scale factor ε, is obtained
It is to new data set：
To each scale factor ε=1,2 ..., 20, it is N and the multivariate data collection for including p variable that length is sought respectivelyMultivariate sample entropy：
(4d) builds N-n m dimension composite delay vectors Xm(i)∈Rm, i=1,2 ..., N-n, n=max { M } × max
(τ), wherein M=[m1,m2,...,mp]∈Rp, wherein m1,m2,...,mpAll it is positive integer, embedded dimension vectorDelay vector τ=[τ1,τ2,...,τp], wherein τ1,τ2,...,τpAll it is positive integer, then Mixed Delay is vectorial
Xm(i) it can be expressed as：
(4e) definition vector XmAnd X (i)m(j) distance between is the poor maximum of its corresponding element, i.e.,：
(4f) is to each composite delay vector Xm(i) itself and other vector distances, are asked respectively, and statistical distance is less than given
Threshold value r number PiAnd PiThe probability of appearance
Pi={ d [Xm(i),Xm(j)]≤r,i≠j}；
(4g) calculates probabilityAverage value Bm(r)：
Composite delay vector in (4d) is expanded to m+1 dimensions by (4h) from m dimensions, and vector M includes p element, has p kinds real
Existing method, i.e. M=[m1,m2,...,mk+1,...,mp], k=1,2 ..., p, the individual Mixed Delay vector X of construction p × (N-n)m+1
(i)∈Rm+1；
(4i) defines two vector Xm+1And X (i)m+1(j) distance between is the maximum of its corresponding element difference, seeks Vector Groups
Xm+1(i) distance between any two, and statistical distance is less than given threshold value r number QiAnd QiThe probability of appearance
Qi={ d [Xm+1(i),Xm+1(j)]≤r,i≠j}；
(4j) is calculatedAverage value B under m+1 dimensionsm+1(r)：
(4k) is according to step (4f) result of calculationWith the result of calculation B of step (4j)m+1(r) multivariable sample, is calculated
This entropy MSampEn：
After calculating terminates, multivariate sample entropy of 4 kinds of failures on scale factor ε=1,2 ..., 20 is obtained, and will
As a result it is saved in distributed file storage system HDFS.
Step 5, visual presentation.
Under stand-alone environment, local host downloads result data from the distributed file system HDFS of big data platform,
The visualization bag enriched using R softwares, draws multivariate sample entropy curve of each failure of switchgear in 20 scale factors,
As shown in Figure 5.
From figure 5 it can be seen that the multivariate sample entropy curve of 4 kinds of failures, in addition to scale factor 1, the multivariable sample of 4 kinds of failures
This entropy curve does not all intersect, classifying quality highly significant.
Step 6, feature extraction.
The multivariate sample entropy curve of 4 kinds of failures according to Fig. 5, the curve of 4 kinds of failures is in scale factor 10~20
In the range of all than shallower, and the multivariate sample entropy of 4 kinds of failures correspondence scale factors differs larger each other, so choosing yardstick
The multivariate sample entropy of the scope of the factor 10~20, as the characteristic parameter of 4 kinds of failures, is timely diagnosis and the event of anticipation switchgear
Barrier provides foundation.
Above description is only example of the present invention, does not constitute any limitation of the invention, it is clear that for this
, all may be without departing substantially from principle of the invention structure after present invention and principle has been understood for the professional in field
In the case of, the various modifications and changes in form and details are carried out, but these modifications and variations based on inventive concept are still
Within the claims of the present invention.
Claims (3)
1. a kind of switchgear fault signature extracting method based on big data platform, including:
(1) SparkR big data platforms are built：
(1a) installs linux system, Hadoop open source softwares and Spark open source softwares；
(1b) determines the node number of platform cluster according to existing fault data scale, and according to subsequently needing failure to be processed
Data scale, can extend or reduce to the node number；
Each node of (1c) configuration platform cluster, i.e., using any 1 node as host node Master from the nodes of determination,
Remaining is as from node Slave；
(1d) it is determined that host node Master and all from node Slave, configuration server process SSH (Secure
Shell) and carry out without password authentification, and Java software, configuration Java context, configuration Hadoop core documents and Spark are installed
Core document；
(2) Data Collection and storage：Host node Master is gathered by Hadoop Sqoop component technologys and closed from platform exterior
It is the fault data of type；The fault data of file type is gathered by Flume component technologys, and by these data Cun Chudao of collection
In Hadoop distributed file system HDFS, host node Master and all these data are shared from node Slave；
(3) data prediction：The fault data being stored in distributed file system HDFS is changed and normalized successively
Pretreatment, provide quality data for subsequent data analysis；
(4) data distribution formula is calculated：
On local host, being rewritten into the multivariable multi-scale entropy MMSE that can only be run on unit using R softwares can be big
The distributed algorithm run on data platform SparkR；
Host node Master calls MMSE distribution by big data platform SparkR SparkR api interfaces from local host
Formula algorithm, is deployed to each from node Slave, and is used as using pretreated data the input of the algorithm；
From the multivariate sample entropy of each failure of node Slave parallel computations, and result of calculation is saved in Hadoop distribution
In file system HDFS；
(5) visual presentation：Under stand-alone environment, local host is downloaded from the distributed file system HDFS of big data platform
Result data, then draws the multivariate sample entropy curve of the various failures of switchgear using the drawing function for R softwares of increasing income；
(6) feature extraction：According to the multivariate sample entropy curve of each failure, each damage curve of selection is all shallower, and each failure
The multivariate sample entropy of correspondence scale factor differs larger scale factor scope each other, and by many of the scale factor scope
Variable sample entropy as each failure characteristic parameter.
2. according to the method described in claim 1, it is characterised in that to the event in distributed file system HDFS in step (3)
Barrier data carry out conversion and normalized pretreatment successively, be first in data set with the interval data conversion represented into corresponding
Single number；The attribute in data set is finally normalized into [0,1] interval.
3. according to the method described in claim 1, it is characterised in that from each failure of node Slave parallel computations in step (4)
Multivariate sample entropy, is carried out as follows：
(3a) determines that embedded dimension m, delay vector τ, threshold value r=0.2*sd, sd are the standard deviation of each variable, scale factor
ε；First variable p is determined according to the number of the influence factor of failure, several numbers determine the second variable N according to fault data bar；
(3b) data using after pretreatment build length as N and include the data set { x of p variablek,i, wherein i=1,2 ..., N；
K=1,2 ..., p；
(3c) is to multivariate data collectionThick-breakpoint processing is carried out based on scale factor ε, obtains new
Data set be：
<mrow>
<msubsup>
<mi>y</mi>
<mrow>
<mi>k</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
<mi>&epsiv;</mi>
</msubsup>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>&epsiv;</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mi>&epsiv;</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>j</mi>
<mi>&epsiv;</mi>
</mrow>
</munderover>
<msub>
<mi>x</mi>
<mrow>
<mi>k</mi>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>,</mo>
<mn>1</mn>
<mo>&le;</mo>
<mi>j</mi>
<mo>&le;</mo>
<mfrac>
<mi>N</mi>
<mi>&epsiv;</mi>
</mfrac>
<mo>,</mo>
<mi>k</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mi>p</mi>
<mo>;</mo>
</mrow>
To each scale factor ε, it is N and the multivariate data collection for including p variable that length is sought respectively
Multivariate sample entropy：
(3d) builds N-n m dimension composite delay vectors Xm(i)∈Rm, i=1,2 ..., N-n, n=max { M } × max (τ), its
Middle M=[m1,m2,...,mp]∈Rp, wherein m1,m2,...,mpAll it is positive integer, embedded dimension vectorDelay
Vectorial τ=[τ1,τ2,...,τp], wherein τ1,τ2,...,τpAll be positive integer, then Mixed Delay vector Xm(i) it can be expressed as：
<mfenced open = "" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>X</mi>
<mi>m</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mo>&lsqb;</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<msub>
<mi>&tau;</mi>
<mn>1</mn>
</msub>
</mrow>
</msub>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>m</mi>
<mn>1</mn>
</msub>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<msub>
<mi>&tau;</mi>
<mn>1</mn>
</msub>
</mrow>
</msub>
<mo>,</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>x</mi>
<mrow>
<mn>2</mn>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>2</mn>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<msub>
<mi>&tau;</mi>
<mn>2</mn>
</msub>
</mrow>
</msub>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>2</mn>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>m</mi>
<mn>2</mn>
</msub>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<msub>
<mi>&tau;</mi>
<mn>2</mn>
</msub>
</mrow>
</msub>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>x</mi>
<mrow>
<mi>p</mi>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mi>p</mi>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<msub>
<mi>&tau;</mi>
<mi>p</mi>
</msub>
</mrow>
</msub>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mi>p</mi>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>m</mi>
<mi>p</mi>
</msub>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<msub>
<mi>&tau;</mi>
<mi>p</mi>
</msub>
</mrow>
</msub>
<mo>&rsqb;</mo>
<mo>;</mo>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
(3e) definition vector XmAnd X (i)m(j) distance between is the poor maximum of its corresponding element, i.e.,：
<mrow>
<mi>d</mi>
<mo>&lsqb;</mo>
<msub>
<mi>X</mi>
<mi>m</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>,</mo>
<msub>
<mi>X</mi>
<mi>m</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>&rsqb;</mo>
<mo>=</mo>
<munder>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
</mrow>
<mrow>
<mi>l</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mi>m</mi>
</mrow>
</munder>
<mo>{</mo>
<mo>|</mo>
<mi>x</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>+</mo>
<mi>l</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mi>x</mi>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>+</mo>
<mi>l</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>|</mo>
<mo>}</mo>
<mo>;</mo>
</mrow>
(3f) is to each Mixed Delay vector Xm(i) itself and other vector distances, are asked respectively, and statistical distance is less than given threshold value r
Number PiAnd PiThe probability of appearance
Pi={ d [Xm(i),Xm(j)]≤r,i≠j}；
<mrow>
<msubsup>
<mi>B</mi>
<mi>i</mi>
<mi>m</mi>
</msubsup>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mi>N</mi>
<mo>-</mo>
<mi>n</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</mfrac>
<msub>
<mi>P</mi>
<mi>i</mi>
</msub>
<mo>;</mo>
</mrow>
(3g) calculates probabilityAverage value Bm(r)：
<mrow>
<msup>
<mi>B</mi>
<mi>m</mi>
</msup>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mi>N</mi>
<mo>-</mo>
<mi>n</mi>
</mrow>
</mfrac>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>N</mi>
<mo>-</mo>
<mi>n</mi>
</mrow>
</msubsup>
<msubsup>
<mi>B</mi>
<mi>i</mi>
<mi>m</mi>
</msubsup>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
Composite delay vector in (3d) is expanded to m+1 dimensions by (3h) from m dimensions, and vector M includes p element, has p kinds realization side
Method, i.e. M=[m1,m2,...,mk+1,...,mp], k=1,2 ..., p, the individual Mixed Delay vector X of construction p × (N-n)m+1(i)
∈Rm+1：
<mfenced open = "" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>X</mi>
<mrow>
<mi>m</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mo>&lsqb;</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<msub>
<mi>&tau;</mi>
<mn>1</mn>
</msub>
</mrow>
</msub>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>m</mi>
<mn>1</mn>
</msub>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<msub>
<mi>&tau;</mi>
<mn>1</mn>
</msub>
</mrow>
</msub>
<mo>,</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>x</mi>
<mrow>
<mn>2</mn>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>2</mn>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<msub>
<mi>&tau;</mi>
<mn>2</mn>
</msub>
</mrow>
</msub>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>2</mn>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>m</mi>
<mn>2</mn>
</msub>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<msub>
<mi>&tau;</mi>
<mn>2</mn>
</msub>
</mrow>
</msub>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>x</mi>
<mrow>
<mi>k</mi>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mi>k</mi>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<msub>
<mi>&tau;</mi>
<mi>k</mi>
</msub>
</mrow>
</msub>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mi>k</mi>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<mo>&lsqb;</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>m</mi>
<mi>k</mi>
</msub>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mn>1</mn>
<mo>&rsqb;</mo>
<msub>
<mi>&tau;</mi>
<mi>k</mi>
</msub>
</mrow>
</msub>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>x</mi>
<mrow>
<mi>p</mi>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mi>p</mi>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<msub>
<mi>&tau;</mi>
<mi>p</mi>
</msub>
</mrow>
</msub>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mi>p</mi>
<mo>,</mo>
<mi>i</mi>
<mo>+</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>m</mi>
<mi>p</mi>
</msub>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<msub>
<mi>&tau;</mi>
<mi>p</mi>
</msub>
</mrow>
</msub>
<mo>&rsqb;</mo>
<mo>;</mo>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
(3i) defines two vector Xm+1And X (i)m+1(j) distance between is the maximum of its corresponding element difference, seeks Vector Groups Xm+1
(i) distance between any two, and statistical distance is less than given threshold value r number QiAnd QiThe probability of appearance
Qi={ d [Xm+1(i),Xm+1(j)]≤r,i≠j}；
<mrow>
<msubsup>
<mi>B</mi>
<mi>i</mi>
<mrow>
<mi>m</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msubsup>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>N</mi>
<mo>-</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</mfrac>
<msub>
<mi>Q</mi>
<mi>i</mi>
</msub>
<mo>;</mo>
</mrow>
(3j) is calculatedAverage value B under m+1 dimensionsm+1(r)：
<mrow>
<msup>
<mi>B</mi>
<mrow>
<mi>m</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>N</mi>
<mo>-</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>N</mi>
<mo>-</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
</mrow>
</msubsup>
<msubsup>
<mi>B</mi>
<mi>i</mi>
<mrow>
<mi>m</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msubsup>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
(3k) is according to step (3f) result of calculationWith the result of calculation B of step (3j)m+1(r) multivariate sample entropy, is calculated
MSampEn：
<mrow>
<mi>M</mi>
<mi>S</mi>
<mi>a</mi>
<mi>m</mi>
<mi>p</mi>
<mi>E</mi>
<mi>n</mi>
<mo>=</mo>
<mo>-</mo>
<mi>l</mi>
<mi>n</mi>
<mo>&lsqb;</mo>
<mfrac>
<mrow>
<msup>
<mi>B</mi>
<mrow>
<mi>m</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msup>
<mi>B</mi>
<mi>m</mi>
</msup>
<mrow>
<mo>(</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&rsqb;</mo>
<mo>.</mo>
</mrow>
3
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710550324.6A CN107301243A (en) | 2017-07-07 | 2017-07-07 | Switchgear fault signature extracting method based on big data platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710550324.6A CN107301243A (en) | 2017-07-07 | 2017-07-07 | Switchgear fault signature extracting method based on big data platform |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107301243A true CN107301243A (en) | 2017-10-27 |
Family
ID=60133884
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710550324.6A Pending CN107301243A (en) | 2017-07-07 | 2017-07-07 | Switchgear fault signature extracting method based on big data platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107301243A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108762959A (en) * | 2018-04-02 | 2018-11-06 | 阿里巴巴集团控股有限公司 | A kind of method, apparatus and equipment of selecting system parameter |
| CN111289888A (en) * | 2019-12-11 | 2020-06-16 | 嘉兴恒创电力集团有限公司博创物资分公司 | High-voltage circuit breaker state detection and fault diagnosis method based on big data technology |
| CN111308336A (en) * | 2020-03-24 | 2020-06-19 | 广西电网有限责任公司电力科学研究院 | High-voltage circuit breaker fast overhaul method and device based on big data |
| CN112357771A (en) * | 2020-11-19 | 2021-02-12 | 中船重工(青岛)海洋装备研究院有限责任公司 | Ship-shore integrated equipment state monitoring system and method |
| CN112666451A (en) * | 2021-03-15 | 2021-04-16 | 南京邮电大学 | Integrated circuit scanning test vector generation method |
| CN114236374A (en) * | 2021-12-13 | 2022-03-25 | 中国矿业大学 | Real-time diagnosis method for open circuit fault of rectifier |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104699985A (en) * | 2015-03-26 | 2015-06-10 | 西安电子科技大学 | Medical big-data acquisition and analysis system and method |
-
2017
- 2017-07-07 CN CN201710550324.6A patent/CN107301243A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104699985A (en) * | 2015-03-26 | 2015-06-10 | 西安电子科技大学 | Medical big-data acquisition and analysis system and method |
Non-Patent Citations (3)
| Title |
|---|
| 孔宪光等: "面向复杂工业大数据的实时特征提取方法", 《西安电子科技大学学报（自然科学版）》 * |
| 宋亚奇: "云平台下电力设备监测大数据存储优化与并行处理技术研究", 《中国博士学位论文全文数据库（电子期刊）工程科技Ⅱ辑》 * |
| 曲朝阳等: "基于Spark的电力设备在线监测数据可视化方法", 《电工电能新技术》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108762959A (en) * | 2018-04-02 | 2018-11-06 | 阿里巴巴集团控股有限公司 | A kind of method, apparatus and equipment of selecting system parameter |
| CN108762959B (en) * | 2018-04-02 | 2021-07-06 | 创新先进技术有限公司 | Method, device and equipment for selecting system parameters |
| CN111289888A (en) * | 2019-12-11 | 2020-06-16 | 嘉兴恒创电力集团有限公司博创物资分公司 | High-voltage circuit breaker state detection and fault diagnosis method based on big data technology |
| CN111289888B (en) * | 2019-12-11 | 2022-04-08 | 嘉兴恒创电力集团有限公司博创物资分公司 | High-voltage circuit breaker state detection and fault diagnosis method based on big data technology |
| CN111308336A (en) * | 2020-03-24 | 2020-06-19 | 广西电网有限责任公司电力科学研究院 | High-voltage circuit breaker fast overhaul method and device based on big data |
| CN112357771A (en) * | 2020-11-19 | 2021-02-12 | 中船重工(青岛)海洋装备研究院有限责任公司 | Ship-shore integrated equipment state monitoring system and method |
| CN112666451A (en) * | 2021-03-15 | 2021-04-16 | 南京邮电大学 | Integrated circuit scanning test vector generation method |
| CN112666451B (en) * | 2021-03-15 | 2021-06-29 | 南京邮电大学 | Integrated circuit scanning test vector generation method |
| CN114236374A (en) * | 2021-12-13 | 2022-03-25 | 中国矿业大学 | Real-time diagnosis method for open circuit fault of rectifier |
| CN114236374B (en) * | 2021-12-13 | 2023-11-14 | 中国矿业大学 | Real-time diagnosis method for open-circuit fault of rectifier |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107301243A (en) | Switchgear fault signature extracting method based on big data platform | |
| CN111177095B (en) | Log analysis method, device, computer equipment and storage medium | |
| CN104580519B (en) | A kind of method of rapid deployment openstack cloud computing platforms | |
| US20160292591A1 (en) | Streamlined analytic model training and scoring system | |
| CN102236672A (en) | Method and device for importing data | |
| CN102880990B (en) | Fault processing system | |
| CN106503268B (en) | Data comparison methods, devices and systems | |
| CN103455346B (en) | Application program deployment method, deployment main control computer, deployment client side and cluster | |
| CN106062751A (en) | Managing data profiling operations related to data type | |
| CN109739919A (en) | A kind of front end processor and acquisition system for electric system | |
| CN109522228A (en) | Interface automatic test data configuration method, apparatus, platform and storage medium | |
| CN104050276A (en) | Cache processing method and system of distributed database | |
| CN105760272B (en) | Monitoring backstage business customizing method and its system based on plug-in unit | |
| CN104794095B (en) | Distributed Calculation processing method and processing device | |
| EP3470992A1 (en) | Efficient storage and utilization of a hierarchical data set | |
| JP2023506239A (en) | Systems and methods for autonomous monitoring and recovery in hybrid energy management | |
| US10163060B2 (en) | Hierarchical probability model generation system, hierarchical probability model generation method, and program | |
| CN106371931B (en) | A kind of high-performance geoscience computing service system based on Web frame | |
| CN101996257A (en) | Method for searching reconfigurable route exchange platform component | |
| CN104299170B (en) | Intermittent energy source mass data processing method | |
| CN107679701B (en) | Load reduction parallel computing method and device | |
| CN109684517A (en) | A kind of historical data storage method, reading/writing method, storage device and equipment | |
| US20240004778A1 (en) | Method for processing command, device for processing command, and electronic device | |
| CN101387582A (en) | Failure diagnosis system and method based on PDA | |
| Dai et al. | Cyber physical power system modeling and simulation based on graph computing |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20171027 |