CN105677791B - For analyzing the method and system of the operation data of wind power generating set - Google Patents
For analyzing the method and system of the operation data of wind power generating set Download PDFInfo
- Publication number
- CN105677791B CN105677791B CN201511029436.4A CN201511029436A CN105677791B CN 105677791 B CN105677791 B CN 105677791B CN 201511029436 A CN201511029436 A CN 201511029436A CN 105677791 B CN105677791 B CN 105677791B
- Authority
- CN
- China
- Prior art keywords
- data
- wind power
- generating set
- power generating
- operating status
- 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.)
- Active
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
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational 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
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
Abstract
The embodiment of the present invention provides a kind of for analyzing the method and system of the operation data of wind power generating set.Operation data, status data and the unit fault data of the wind power generating set acquired in scheduled statistical time section are obtained as primary data；These data are analyzed, attribute relevant to the variation of wind power generating set operating status is extracted, and be sampled to primary data, obtains the sample data of multiple operating status classifications；It is for statistical analysis according to sample data difference of the attribute of extraction to the multiple operating status classification, to obtain multiple operation characteristic amounts of the wind power generating set for the predetermined condition for meeting the multiple operating status classification, each operation characteristic amount includes the value of single attribute or the value of multiple attributes；The variation diagram of the operation characteristic amount in the designated state period is drawn, according to multiple operation characteristic amounts of acquisition to characterize the variation of the wind power generating set its operation characteristic amount under prescribed conditions.
Description
Technical field
The present invention relates to the information processing technologies of wind-power electricity generation more particularly to a kind of for analyzing the fortune of wind power generating set
The method and system of row data.
Background technique
The continuous improvement of Computing ability in recent years and the fast development of distributed computing technology make it possible to base
Magnanimity, the data of variform, i.e. big data are handled in distributed structure/architecture.
Big data technology based on distributed structure/architecture is related to data mining, processing, storage and subject analysis.Wherein, data
Excavation refers to the process of and is hidden in wherein information by algorithm search from a large amount of data, by statistics, online analysis and processing,
All multi-methods such as information retrieval, machine learning, expert system (rely on the past rule of thumb) and pattern-recognition obtain target
Information.
In technical field of wind power generation, to wind-force therein during wind power generating set operation and in wind power plant
During generating set is monitored, a large amount of operation data and status data, fault data etc. are generated.Based on big data skill
Art is excavated and is analyzed to these data, and the status monitoring and policy control optimization for wind power generating set have great
Meaning.
Summary of the invention
The object of the present invention is to provide a kind of for analyzing the technical solution of the operation data of wind power generating set, with
The association between operation data and operating status is established based on a variety of data acquired from wind power generating set, assists control strategy
Optimization.
According to an aspect of the present invention, a kind of method for analyzing the operation data of wind power generating set is provided, it is described
Method includes: acquisition primary data relevant to the operation of wind power generating set, and extracts from the primary data and send out with wind-force
Electric operating states of the units changes relevant attribute, and is sampled to the primary data, obtains multiple operating status classifications
Sample data；It is for statistical analysis according to sample data difference of the attribute of extraction to the multiple operating status classification, with
Obtain the multiple operation characteristic amounts for meeting the wind power generating set of predetermined condition of the multiple operating status classification；According to acquisition
Multiple operation characteristic amounts draw the designated state period in operation characteristic amount variation diagram, to characterize the wind-driven generator
Organize the variation of its operation characteristic amount under prescribed conditions.
Optionally, the primary data includes the operation number of the wind power generating set acquired in scheduled statistical time section
According to, status data and unit fault data.
Optionally, the primary data further include in the statistical time section with the operation data, status data with
And the relevant statistical data of unit fault data.
Optionally, described extract from the primary data changes from relevant attribute to wind power generating set operating status
Reason includes: that each attribute in the primary data is screened and returned according to principal component analytical method and factor-analysis approach
Choosing；It is analyzed according to genetic algorithm screening and returning the attribute selected, selects and become with the operating status of wind power generating set
Change relevant attribute.
Optionally, the method also includes: use the sample data of the multiple operating status classification and described initial
Data, according to Wilcoxen signed rank test algorithm respectively to the predetermined condition for meeting the multiple operating status classification
Multiple operation characteristic amounts of wind power generating set test, do not meet normal distribution to reject Wilcoxen statistic
Operation characteristic amount.
Optionally, the method also includes the operation datas of the wind power generating set of the acquisition of scheduled duration before acquisition
And status data；Category relevant to the variation of the operating status of wind power generating set is calculated according to the operation data and status data
The value of property；According to calculating and the operating status of wind power generating set change the value of relevant attribute respectively with the designated state
The variation diagram of operation characteristic amount in period determines the current operating status of the wind power generating set.
Optionally, described that the primary data is sampled, obtain the place of the sample data of multiple operating status classifications
Reason includes: to carry out stratified sampling to the primary data according to the multiple operating status classification, to obtain the multiple operation
The sample data of status categories.
Preferably, the step in the method is executed by the data platform based on HADOOP distributed structure/architecture, wherein with
HIVE form stores the primary data and sample data, and the attribute information of extraction is stored in relational database.
According to another aspect of the present invention, a kind of system for analyzing the operation data of wind power generating set is also provided,
The system comprises primary data acquisition device, for obtaining the relevant primary data of operation to wind power generating set；Attribute
Extraction element, for extracting attribute relevant to the variation of wind power generating set operating status from the primary data；Sampling apparatus,
For being sampled to the primary data, the sample data of multiple operating status classifications is obtained；Feature establishes device, is used for root
It is for statistical analysis to the sample data difference of the multiple operating status classification according to the attribute of extraction, met with acquisition described more
Multiple operation characteristic amounts of the wind power generating set of the predetermined condition of a operating status classification；Model foundation device is used for basis
The multiple operation characteristic amounts obtained draw the variation diagram of the operation characteristic amount in the designated state period, to characterize the wind-force hair
The variation of motor group its operation characteristic amount under prescribed conditions.
Optionally, the primary data includes the operation number of the wind power generating set acquired in scheduled statistical time section
According to, status data and unit fault data.
Optionally, the primary data further include in the statistical time section with the operation data, status data with
And the relevant statistical data of unit fault data.
Optionally, the attributes extraction device includes: attribute primary election unit, for according to principal component analytical method and the factor
Analysis method carries out screening and Gui Xuan to each attribute in the primary data；The selected unit of attribute, for being calculated according to heredity
Method is analyzed screening and returning the attribute selected, and attribute relevant to the variation of the operating status of wind power generating set is selected.
Optionally, the system also includes: characteristic quantities to verify device, for using the sample of the multiple operating status classification
Notebook data and the primary data meet the multiple operation to described respectively according to Wilcoxen signed rank test algorithm
Multiple operation characteristic amounts of the wind power generating set of the predetermined condition of status categories are tested, to reject Wilcoxen statistics
Amount does not meet the operation characteristic amount of normal distribution.
Optionally, the system also includes state determination device, the wind-force for the acquisition of scheduled duration before obtaining is sent out
The operation data and status data of motor group calculate the operation with wind power generating set according to the operation data and status data
The value of the relevant attribute of state change, and relevant attribute is changed to the operating status of wind power generating set according to calculating
Variation diagram of the value respectively with the operation characteristic amount in the designated state period determines the current fortune of the wind power generating set
Row state.
Optionally, the sampling apparatus is for being layered the primary data according to the multiple operating status classification
Sampling, to obtain the sample data of the multiple operating status classification.
Optionally, the system is to pass through the data platform based on HADOOP distributed structure/architecture, wherein the primary data
It is stored in the form of HIVE with sample data, the attribute information that the attributes extraction device extracts is stored in relational database
In.
The method and system of the operation data for analyzing wind power generating set provided according to embodiments of the present invention passes through
Big data digging technology, from the operation data of wind power generating set, status data, unit fault data and statistical Data Mining
The variation diagram of operation characteristic amount corresponding with one or more operating states of the units out, so as to these variation diagrams as state
Judgment models judge its operating status according to the current operation data of wind power generating set.
Detailed description of the invention
Fig. 1 is the frame stream for showing the operation data for analyzing wind power generating set of general plotting according to the present invention
Cheng Tu；
Fig. 2 is the illustrative diagram for showing the data of data warehouse storage of wind-powered electricity generation big data platform；
Fig. 3 be show according to embodiments of the present invention one for analyze wind power generating set operation data method stream
Cheng Tu；
Fig. 4 is that the method for showing the operation data according to an embodiment of the present invention for analyzing wind power generating set is drawn
The example of the variation diagram of operation characteristic amount；
Fig. 5 is that the system for the operation data for analyzing wind power generating set for showing according to embodiments of the present invention two is patrolled
Collect block diagram.
Specific embodiment
Basic conception of the invention is to pass through all kinds of running of wind generating set data, status data, the unit failure to acquisition
Data, based on the Various types of data mining algorithm such as big data technology and classification, cluster, the characteristic value detection of creation wind-powered electricity generation operation data
Algorithm obtains the operation characteristic amount for capableing of the variation of accurate description set state, establishes between set state and operation characteristic amount
Correlation model, and then judged using these features operating status current to wind power generating set.
For this purpose, can be carried out based on the big data platform of HADOOP distributed structure/architecture proposed by the present invention for analyzing wind-force hair
The technical solution of the operation data of motor group.
Every wind power generating set in the process of running, many operation datas of acquisition/generation and status data；In addition,
The control centre of wind power plant, can be from each wind during the wind power generating set in wind power plant is monitored and is controlled
Power generator group gets a large amount of operation data and status data, and can be collected into the unit event about wind power generating set
Hinder data etc..These data all can be used as the object of technical solution processing proposed by the present invention.In addition, to the wind in wind power plant
Power generator group also generates statistical data relevant to operation data and status data, these statistical numbers during being monitored
According to the input data that also can be used as technical solution proposed by the present invention.
Specifically, the operation data of wind power generating set may include, but be not limited to, the number of operating parameter and environmental parameter
According to, such as the revolving speed of generator, environment temperature, the pressure for yawing clamp, wind speed, the generated energy of wind power generating set, power, electricity
Pressure, electric current etc..
The overall operation state of the status data instruction unit of wind power generating set, such as, but not limited to, be switched on, shut down,
Restart, standby mode and its time of origin etc..
The unit fault data of wind power generating set include, but are not limited to the time that unit breaks down, fault type,
Troubleshooting time, the relevant warning message of failure etc..
Statistical data above-mentioned includes to all kinds of of the operating parameter and environmental parameter collected whithin a period of time from unit
Data carry out the statistical values such as maximum value, minimum value, mean value, the frequency of Various types of data of statistics acquisition.
The operation for being used to analyze wind power generating set of general plotting according to the present invention is discussed in detail hereinafter with reference to Fig. 1
The processing of data.Fig. 1 is the illustrative diagram for showing the data warehouse storage data of wind-powered electricity generation big data platform.
Referring to Fig.1, firstly, acquiring Various types of data relevant to the operation of wind power generating set, and these data are deposited
Storage is in the data warehouse of big data platform (10).Various types of data mentioned here is the operation number of wind power generating set above-mentioned
According to, status data, unit fault data and all kinds of statistical data.
Specifically, can be by technologies such as ETL process (extracting, conversion, load), obtaining from multiple data sources includes (transporting
Row data and status data), unit fault data, statistical data etc..
In general, all had differences from the enriched data that multiple data sources obtain from data format, existing way etc.,
Therefore in the data warehouse for storing that data in big data platform before, need to carry out these data to include that data are led
The processing such as enter, standardize, cleaning.
Specifically, firstly, data source can be the real-time data base of particular memory format, relevant database or not apposition
The file etc. of formula needs to import these data from data source.Secondly, according to the data management standard of enterprise, to Various types of data
Name, format are standardized, and transfer and use these data to facilitate.Again, it is set according to the characteristics of Various types of data
Reasonable data area and rule, reject the abnormal data not in reasonable value range.For example, can be counted according to following rule
According to checking and cleaning: whether allow null value, type whether to match, whether format matches, whether data unique, data explain whether
There are ambiguity, whether data are with uniformity, whether meet table incidence relation, whether meet tentation data range, meet
Service logic etc..
After carrying out aforementioned processing, it can be stored in Hadoop data warehouse with the formal distribution formula of Hive data.This
Sample can either accelerate reading and writing data speed, and can be ready for the distributed computing in later period.
Fig. 2 is the illustrative diagram for showing the data of data warehouse storage of wind-powered electricity generation big data platform.As shown in Fig. 2,
It will acquire and treated datum number storage is according under data classification corresponding in warehouse.Specifically, the data of storage are organized
For wind power plant information table, fault data table, machine group type information table, statistics table, real time data table, fan condition tables of data
And with the associated blower information table of these tables.
The processing of aforementioned data acquisition and storage can be cumulatively executed, the processing can also be disposably carried out.Hereafter,
When needing to carry out analysis proposed by the present invention processing, subsequent processing is executed.
Firstly, being extracted based on the related data (" primary data " i.e. described below) stored in data warehouse and wind
Power generator group operating status changes relevant attribute (20) and carries out standard sample (30).
On the one hand, the primary data is analyzed, extracts category relevant to the variation of wind power generating set operating status
Property, that is to say, that changed by force to run to select in relevant attribute from unit with set state according to scheduled selection algorithm
Relevant attribute.
Firstly, carrying out screening and Gui Xuan to each attribute in the primary data.It is, for example, possible to use following methods pair
The attribute to be analyzed carries out screening and Gui Xuan:
I. multiple variables principal component analysis: are selected to a kind of polynary system of less number significant variable by linear transformation
Count analysis method.Can be for all properties that originally proposed, it is extra that duplicate attribute (attribute of close relation) is left out, and establishes
New attribute as few as possible, so that these new attributes are incoherent two-by-two, and these new attributes are in the information of reflection project
Aspect keeps original information as far as possible.
Ii. factorial analysis: the statistical technique of the general character factor is extracted from variable group, the variable (attribute) of homogeneity is included into one
The factor can reduce the number of attribute, can also examine the hypothesis of relationship between attribute.The method of factorial analysis is there are about more than 10 kinds, such as weight
Heart method, image analysis, maximum likelihood solution, least squares method, A Erfa take out because of method, draw typical pumping difficult to understand because of method etc..
Then, the Feature Selection Algorithm based on genetic algorithm can be used, screening is related to the unit operation of choosing is returned from passing through
Attribute in select and set state changes obvious relevant several attributes.In the genetic algorithm used, it can be increased with information
Benefit carries out its individual evaluation, Selecting operation, crossing operation and the index of mutation operator as measure.Specifically, may be used
By the difference of comentropy and conditional entropy come the information gain of computation attribute:
InfoGains [k]=entropy [k]-condEntropy [k]
Wherein, InfoGains [k] indicates the information gain of k-th of attribute, and entropy [k] indicates the letter of k-th of attribute
Entropy is ceased, condEntropy [k] indicates the conditional entropy of k-th of attribute.
Using the attribute selected by the Feature Selection Algorithm based on genetic algorithm as the attribute of strong correlation.
The processing analyzed by aforementioned attributes filters out some pairs of meaningless attributes of analysis, such as when accumulative energization
Between, system normal time, blower self-starting count etc., and extract can be used as analysis input attribute or combinations of attributes, such as
Power, wind speed-revolving speed, temperature-voltage, blade angle-generated energy etc., the combination of these attributes shape by way of linear combination
The attribute of Cheng Xin.Since itself just has certain business association, subsequent analysis can be made more accurate.
On the other hand, the primary data is sampled, obtains the sample data of multiple operating status classifications, preferably
It takes the mode of stratified sampling to be sampled: primary data being first divided into several classifications according to target call, to each layer
(class) carries out such as random sampling, and the sample extracted to every layer is combined and constitutes sample set, establishes for subsequent feature
Processing and model (variation diagram i.e. described below) are established.Different sample sets can be obtained according to different set states, such as
For unit failure, unit restarts equal different conditions event, extracts different sample sets, for subsequent calculating use.
Sample data and attribute data can be stored using different data storage formats.Due to the data volume of sample data
Greatly, therefore sample data can be continued to be stored in Hive data warehouse；And the combinations of attributes that is obtained by attributive analysis and point
Class has determining format and field, and can store directly may have access in the relational database transferred.
After completing attributes extraction and standard sample, the processing (40) that feature shown in Fig. 1 is established is executed.
Specifically, statistical is carried out according to sample data of the attribute of extraction to the multiple operating status classification respectively
Analysis, to obtain multiple operation characteristic amounts of the wind power generating set for the predetermined condition for meeting the multiple operating status classification, often
A operation characteristic amount includes the value of single attribute or the value of multiple attributes.
Various packets can be established for example, by the various data statistical approach such as regression analysis, variance analysis, correlation analysis
The operation characteristic amount for including the value of attribute or combinations of attributes, by the sample data for the classification that standard sample before obtains, using point
Class and algorithm for pattern recognition are judged, the operation characteristic amount for meeting judging result is preserved, the knot established as feature
Fruit.
For example, setting related coefficient during using correlation analysis and being sentenced as scheduled value as threshold value
It is disconnected, it is compared two-by-two by the value to attribute or combinations of attributes, finds combination and the blade angle-generated energy of wind speed spin rates
Between combination, the related coefficient of most of the time has been more than threshold value 0.8, then illustrates that the value of the attribute or combinations of attributes is that have one
Variation tendency is determined, using the value of the attribute or combinations of attributes as selected operation characteristic amount.
On this basis, it is preferable that these selected operation characteristic amounts can be verified by scheduled check algorithm,
To ensure the reasonability and accuracy of these operation characteristic amounts.For example, wilcoxon method (Wilcoxen signed rank can be used
The method of inspection) verifying is compared with the primary data of totality in the sample data set that extracts before use, with the spy of Double Data collection
Sign point formula is as follows:
Wherein, z indicates the wilcoxon statistic of this feature point.If the statistic meets the normal distribution of (0,1) N, then
Assert that the operation characteristic amount is effective.WxFor the operation characteristic amount sum of ranks and all values at a distance from median the sum of absolute value.μ
For the mean value of the operation characteristic amount.σ is the standard deviation of the operation characteristic amount.n1The primary data used for the operation characteristic amount
The number of (totality), n2For the number for second data set (some sample data set) that the operation characteristic amount uses.τ is operation
Knot value in characteristic quantity, τjFor the number of wherein j-th of knot value.
If illustrating that the operation characteristic figureofmerit shows apparent variation at the sample data set by verifying,
Further screening meets multiple operation characteristic amounts of the wind power generating set of the predetermined condition of the multiple operating status classification.
For example, the multiple operation characteristic amounts for meeting these conditions can be established respectively for normal operation, rebooting status, any malfunction.
Hereafter, the processing (50) of model foundation shown in Fig. 1 is carried out.Specifically, according to multiple operation characteristics of acquisition
Amount draws the variation diagram of the operation characteristic amount in the designated state period, to characterize the wind power generating set under prescribed conditions
The variation of its operation characteristic amount.
In the processing that feature is established, the wind power generating set for meeting the predetermined condition of multiple operating status classifications is got
Multiple operation characteristic amounts.In the processing of model foundation, it is first determined the operating status classification and the operation of concrete analysis
Then the detection period of state, the predetermined amount of time before occurring such as failure are directed to the detection period corresponding fortune of the operating status
Row characteristic quantity carries out the drafting of dot matrix variation diagram, thus Observable wind power generating set its operation characteristic amount under the operating status
Variation.
Fig. 4 is shown when the variation diagram of the operation characteristic amount detected when certain malfunction occurs occurs for unit, and expression exists
Certain set generator break down before a period of time, wind speed-temperature this feature excavated according to aforementioned processing
It is worth changed figure, it can thus be seen that the operation characteristic amount produces apparent reaction for the state change of unit.
Operation that can be current according to unit as state model using the variation diagram of the operation characteristic amount corresponding with operating status of drafting
Data further predict the operating status of unit.
On this basis, it is preferable that all same class events can be also calculated by the statistical data of these operation characteristic amounts
Before barrier point occurs, a possibility that this feature transportation load is abnormal variation.If the possibility reaches certain ratio value (such as
75% or more), then can be directed to the curvilinear characteristic of the variation diagram, establish for the mean value of the operation characteristic amount, variance, maximum value,
The calibration model of normal distribution etc. detects the mean value of the operation characteristic amount within continuous a period of time, is set according to preset
It sets the threshold decision operation characteristic amount and whether exceeds normal level range, to be further modified to the variation diagram.
According to data acquisition above-mentioned, attributes extraction, standard sample, feature foundation and the processing of model foundation, from wind
Operation data, status data, unit fault data and the statistical Data Mining of power generator group go out and one or more units
The variation diagram of the corresponding operation characteristic amount of operating status, so as to these variation diagrams as state judgment models, according to wind-force
The current operation data of generating set judges its operating status, and carries out the prediction of succeeding state.
3~Fig. 5 detailed description of the present invention exemplary embodiment with reference to the accompanying drawing.
Embodiment one
Fig. 3 be show according to embodiments of the present invention one for analyze wind power generating set operation data method stream
Cheng Tu.
Referring to Fig. 3, in step S310, primary data (S310A) relevant to the operation of wind power generating set is obtained, and from
The primary data extracts attribute (S310B) relevant to the variation of wind power generating set operating status.
Primary data may include, but be not limited to, the operation of the wind power generating set acquired in scheduled statistical time section
Data, status data and unit fault data.Preferably, primary data further include in the statistical time section with the fortune
Row data, status data and the relevant statistical data of unit fault data.
As previously mentioned, the input that these data are excavated as follow-up data can be obtained from HADOOP data warehouse.
Specifically, as previously mentioned, the processing of step S310BA includes: according to principal component analytical method and factor-analysis approach
Screening and Gui Xuan are carried out to each attribute in the primary data, and according to genetic algorithm to screening and return the attribute selected
It is analyzed, selects attribute relevant to the variation of the operating status of wind power generating set.
In step S320, the primary data is sampled, obtains the sample data of multiple operating status classifications.
Specifically, stratified sampling is carried out to the primary data according to the multiple operating status classification, described in obtaining
The sample data of multiple operating status classifications.
In step S330, according to the attribute extracted in step S310 to the sample data point of the multiple operating status classification
It is not for statistical analysis, to obtain multiple fortune of the wind power generating set for the predetermined condition for meeting the multiple operating status classification
Row characteristic quantity, each operation characteristic amount include the value of single attribute or the value of multiple attributes.
Specifically, as previously mentioned, can unite for example, by the various data such as regression analysis, variance analysis, correlation analysis
Meter method, the operation for establishing the various values including attribute or combinations of attributes for the predetermined condition for meeting multiple operating status classifications are special
Sign amount.
In addition, the correlation in order to further ensure that multiple operation characteristic amounts that aforementioned processing obtains, passes through standard before
The sample data for the classification that sampling obtains is judged using classification and algorithm for pattern recognition, retains the fortune for meeting judging result
Row characteristic quantity.
Specifically, using the sample data and the primary data of the multiple operating status classification, according to Weir section
Gram gloomy signed rank test algorithm is respectively to the wind power generating set of the predetermined condition for meeting the multiple operating status classification
Multiple operation characteristic amounts test, to reject the operation characteristic amount that Wilcoxen statistic does not meet normal distribution.
In step S340, the operation characteristic amount in the designated state period is drawn according to multiple operation characteristic amounts of acquisition
Variation diagram, to characterize the variation of the wind power generating set its operation characteristic amount under prescribed conditions.
By the processing of abovementioned steps, can from the operation data of wind power generating set, status data, unit fault data with
And statistical Data Mining goes out the variation diagram of operation characteristic amount corresponding with one or more operating states of the units, so as to these
Variation diagram judges its operating status according to the current operation data of wind power generating set as state judgment models.
Correspondingly, the method for the operation data for analyzing wind power generating set that exemplary embodiment of the present proposes is also
It include: the operation data and status data of the wind power generating set of the acquisition of scheduled duration before obtaining；According to the operation number
According to the value for changing relevant attribute to the operating status of wind power generating set with status data calculating；According to being sent out with wind-force for calculating
The operating status of motor group changes the value change with the operation characteristic amount in the designated state period respectively of relevant attribute
Change figure and determines the current operating status of the wind power generating set.
On this basis, it can also establish based on Time-series Techniques, unit behavior prediction model further be established, to one section
Unit behavior in time is predicted that auxiliary monitoring management personnel formulate or adjustment unit operation reserve.
Embodiment two
Fig. 5 is that the system for the operation data for analyzing wind power generating set for showing according to embodiments of the present invention two is patrolled
Collect block diagram.
Referring to Fig. 5, the system for the operation data for analyzing wind power generating set that embodiment two provides includes: initial number
Device 540 and model foundation device 550 are established according to acquisition device 510, attributes extraction device 520, sampling apparatus 530, feature.
Primary data acquisition device 510 is used to obtain the relevant primary data of operation to wind power generating set, at the beginning of described
Beginning data include operation data, status data and the unit event of the wind power generating set acquired in scheduled statistical time section
Hinder data.
Optionally, the primary data further include in the statistical time section with the operation data, status data with
And the relevant statistical data of unit fault data.
Attributes extraction device 520 extracts and wind power generating set operating status for analyzing the primary data
Change relevant attribute.
Specifically, attributes extraction device 520 includes: attribute primary election unit, for according to principal component analytical method and the factor
Analysis method carries out screening and Gui Xuan to each attribute in the primary data；The selected unit of attribute, for being calculated according to heredity
Method is analyzed screening and returning the attribute selected, and attribute relevant to the variation of the operating status of wind power generating set is selected.
Sampling apparatus 530 obtains the sample data of multiple operating status classifications for being sampled to the primary data.
Specifically, sampling apparatus 530 is for being layered the primary data according to the multiple operating status classification
Sampling, to obtain the sample data of the multiple operating status classification.
Feature establishes device 540 for distinguishing according to the attribute of extraction the sample data of the multiple operating status classification
It is for statistical analysis, to obtain multiple operations of the wind power generating set for the predetermined condition for meeting the multiple operating status classification
Characteristic quantity, each operation characteristic amount include the value of single attribute or the value of multiple attributes.
Optionally, system further include: characteristic quantity verifies device, for using the sample of the multiple operating status classification
Data and the primary data meet the multiple operation shape to described respectively according to Wilcoxen signed rank test algorithm
Multiple operation characteristic amounts of the wind power generating set of the predetermined condition of state classification are tested, to reject Wilcoxen statistic
The operation characteristic amount of normal distribution is not met.
Model foundation device 550 is used to draw the operation in the designated state period according to multiple operation characteristic amounts of acquisition
The variation diagram of characteristic quantity, to characterize the variation of the wind power generating set its operation characteristic amount under prescribed conditions.
Optionally, system further include: state determination device 560, the wind-force for the acquisition of scheduled duration before obtaining
The operation data and status data of generating set calculate the fortune with wind power generating set according to the operation data and status data
The value of the relevant attribute of row state change, and according to the attribute relevant to the variation of the operating status of wind power generating set of calculating
Value determine that the wind power generating set is current with the variation diagram of the operation characteristic amount in the designated state period respectively
Operating status.
Optionally, the system is to pass through the data platform based on HADOOP distributed structure/architecture, wherein the primary data
It is stored in the form of HIVE with sample data, the attribute information that the attributes extraction device extracts is stored in relational database
In.
It may be noted that all parts/step described in this application can be split as more multi-section according to the needs of implementation
The part operation of two or more components/steps or components/steps can also be combined into new components/steps by part/step,
To achieve the object of the present invention.
It is above-mentioned to be realized in hardware, firmware according to the method for the present invention, or be implemented as being storable in recording medium
Software or computer code in (such as CD ROM, RAM, floppy disk, hard disk or magneto-optic disk), or it is implemented through network downloading
Original storage in long-range recording medium or nonvolatile machine readable media and the meter that will be stored in local recording medium
Calculation machine code, so that method described herein can be stored in using general purpose computer, application specific processor or programmable or specially
It is handled with such software in the recording medium of hardware (such as ASIC or FPGA).It is appreciated that computer, processor, micro-
Processor controller or programmable hardware include can store or receive software or computer code storage assembly (for example, RAM,
ROM, flash memory etc.), when the software or computer code are by computer, processor or hardware access and execute, realize herein
The processing method of description.In addition, when general purpose computer accesses the code for realizing the processing being shown here, the execution of code
General purpose computer is converted to the special purpose computer for being used for executing the processing being shown here.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.Therefore, protection scope of the present invention should be based on the protection scope of the described claims.
Claims (16)
1. a kind of method for analyzing the operation data of wind power generating set, which is characterized in that the described method includes:
Primary data relevant to the operation of wind power generating set is obtained,
And attribute relevant to the variation of wind power generating set operating status is extracted from the primary data；
The primary data is sampled, the sample data of multiple operating status classifications is obtained；
It is for statistical analysis according to sample data difference of the attribute of extraction to the multiple operating status classification, include to establish
The operation characteristic amount of the value of attribute or combinations of attributes is obtained from the operation characteristic amount of foundation using classification and algorithm for pattern recognition
The multiple operation characteristic amounts for meeting the predetermined condition of the multiple operating status classification are drawn according to multiple operation characteristic amounts of acquisition
The variation diagram of operation characteristic amount in the designated state period processed, to characterize the wind power generating set its fortune under prescribed conditions
The variation of row characteristic quantity.
2. the method according to claim 1, wherein the primary data includes in scheduled statistical time section
Operation data, status data and the unit fault data of the wind power generating set of acquisition.
3. according to the method described in claim 2, it is characterized in that, the primary data further includes in the statistical time section
Statistical data relevant to the operation data, status data and unit fault data.
4. according to the method described in claim 3, it is characterized in that, described from primary data extraction and wind power generating set
The processing that operating status changes relevant attribute includes:
Screening and Gui Xuan are carried out to each attribute in the primary data according to principal component analytical method and factor-analysis approach；
It is analyzed according to genetic algorithm screening and returning the attribute selected, selects and become with the operating status of wind power generating set
Change relevant attribute.
5. according to the method described in claim 4, it is characterized in that, the method also includes:
Using the sample data and the primary data of the multiple operating status classification, examined according to Wilcoxen signed rank
Checking method is special to multiple operations of the wind power generating set of the predetermined condition for meeting the multiple operating status classification respectively
Sign amount is tested, to reject the operation characteristic amount that Wilcoxen statistic does not meet normal distribution.
6. the method according to any one of claim 2~5, which is characterized in that the method also includes:
The operation data and status data of the wind power generating set of the acquisition of scheduled duration before acquisition,
The value for changing relevant attribute to the operating status of wind power generating set is calculated according to the operation data and status data,
According to calculating and the operating status of wind power generating set change the value of relevant attribute respectively with the designated state when
Between the variation diagram of operation characteristic amount in section determine the current operating status of the wind power generating set.
7. the method according to any one of claim 3~5, which is characterized in that described to be taken out to the primary data
Sample, the processing for obtaining the sample data of multiple operating status classifications include:
Stratified sampling is carried out to the primary data according to the multiple operating status classification, to obtain the multiple operating status
The sample data of classification.
8. method according to any one of claims 1 to 5, which is characterized in that by being based on HADOOP distributed structure/architecture
Data platform execute the step in the method, wherein store the primary data and sample data in the form of HIVE, will mention
The attribute information taken is stored in relational database.
9. a kind of system for analyzing the operation data of wind power generating set, which is characterized in that the system comprises:
Primary data acquisition device, for obtaining the relevant primary data of operation to wind power generating set；
Attributes extraction device, for extracting attribute relevant to the variation of wind power generating set operating status from the primary data；
Sampling apparatus obtains the sample data of multiple operating status classifications for being sampled to the primary data；
Feature establishes device, for being united respectively according to the attribute of extraction to the sample data of the multiple operating status classification
Meter analysis, with establish include attribute or combinations of attributes value operation characteristic amount, using classification and algorithm for pattern recognition from foundation
Operation characteristic amount in obtain meet the multiple operating status classification predetermined condition wind power generating set multiple operations
Characteristic quantity；
Model foundation device, for drawing the operation characteristic amount in the designated state period according to multiple operation characteristic amounts of acquisition
Variation diagram, to characterize the variation of the wind power generating set its operation characteristic amount under prescribed conditions.
10. system according to claim 9, which is characterized in that the primary data includes in scheduled statistical time section
Operation data, status data and the unit fault data of the wind power generating set of interior acquisition.
11. system according to claim 10, which is characterized in that the primary data further includes in the statistical time section
Interior statistical data relevant to the operation data, status data and unit fault data.
12. system according to claim 11, which is characterized in that the attributes extraction device includes:
Attribute primary election unit, for according to principal component analytical method and factor-analysis approach to each kind in the primary data
Property carry out screening and Gui Xuan；
The selected unit of attribute is selected and is sent out with wind-force for being analyzed according to genetic algorithm screening and returning the attribute selected
The operating status of motor group changes relevant attribute.
13. system according to claim 12, which is characterized in that the system also includes:
Characteristic quantity verifies device, for using the sample data and the primary data of the multiple operating status classification, root
According to Wilcoxen signed rank test algorithm respectively to the wind-force of the predetermined condition for meeting the multiple operating status classification
Multiple operation characteristic amounts of generating set are tested, to reject the operation spy that Wilcoxen statistic does not meet normal distribution
Sign amount.
14. system described in any one of 0~13 according to claim 1, which is characterized in that the system also includes:
State determination device, operation data and status number for the wind power generating set of the acquisition of scheduled duration before obtaining
According to, the value for changing relevant attribute to the operating status of wind power generating set is calculated according to the operation data and status data,
And according to calculating and the operating status of wind power generating set change the value of relevant attribute respectively with the designated state when
Between the variation diagram of operation characteristic amount in section determine the current operating status of the wind power generating set.
15. system described in any one of 1~13 according to claim 1, which is characterized in that the sampling apparatus is used for according to institute
It states multiple operating status classifications and stratified sampling is carried out to the primary data, to obtain the sample of the multiple operating status classification
Data.
16. the system according to any one of claim 9~13, which is characterized in that the system is by being based on
The data platform of HADOOP distributed structure/architecture, wherein the primary data and sample data are stored in the form of HIVE, the category
Property extraction element extract attribute information be stored in relational database.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201511029436.4A CN105677791B (en) | 2015-12-31 | 2015-12-31 | For analyzing the method and system of the operation data of wind power generating set |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201511029436.4A CN105677791B (en) | 2015-12-31 | 2015-12-31 | For analyzing the method and system of the operation data of wind power generating set |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105677791A CN105677791A (en) | 2016-06-15 |
| CN105677791B true CN105677791B (en) | 2019-03-08 |
Family
ID=56298429
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201511029436.4A Active CN105677791B (en) | 2015-12-31 | 2015-12-31 | For analyzing the method and system of the operation data of wind power generating set |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105677791B (en) |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106640547B (en) * | 2016-11-24 | 2020-08-18 | 东北电力大学 | Method and system for monitoring state of wind turbine generator |
| CN109961374B (en) * | 2017-12-22 | 2021-08-06 | 北京金风科创风电设备有限公司 | Method and device for processing operating data of wind generating set |
| CN110649842B (en) * | 2018-06-27 | 2021-08-13 | 北京金风科创风电设备有限公司 | Control method, device and system of wind generating set and storage medium |
| CN109558982B (en) * | 2018-12-03 | 2022-06-10 | 中国水利水电科学研究院 | Method and device for predicting water intake of thermal power plant |
| CN109299834A (en) * | 2018-12-03 | 2019-02-01 | 中国水利水电科学研究院 | A kind of the water process amount prediction technique and device of water treatment plant |
| CN109657943A (en) * | 2018-12-06 | 2019-04-19 | 中国科学院深圳先进技术研究院 | Dynamic assessment method, device and the electronic equipment of wind power plant operating states of the units |
| CN110196855B (en) * | 2019-05-07 | 2023-03-10 | 中国人民解放军海军航空大学岸防兵学院 | Consistency checking method of performance degradation data and fault data based on rank sum |
| CN113027678B (en) * | 2019-12-25 | 2022-07-12 | 北京金风科创风电设备有限公司 | Wind driven generator group data acquisition system, method and device |
| CN111102139B (en) * | 2019-12-26 | 2021-02-02 | 河北振创电子科技有限公司 | Fan yaw caliper alarm method and system |
| CN111160865A (en) * | 2019-12-31 | 2020-05-15 | 新奥数能科技有限公司 | Workflow management method and device |
| CN111426892B (en) * | 2020-03-10 | 2022-04-22 | 贵州电网有限责任公司 | Unit online state statistical method under complex condition |
| CN111695608B (en) * | 2020-05-27 | 2022-07-29 | 电子科技大学 | Data expansion method for preserving original sample distribution characteristics |
| CN115640285B (en) * | 2022-10-24 | 2023-10-27 | 北京国电通网络技术有限公司 | Power abnormality information transmission method, device, electronic equipment and medium |
| CN117235480B (en) * | 2023-11-16 | 2024-02-13 | 深圳市吾股大数据科技有限公司 | Screening method and system based on big data under data processing |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012149901A1 (en) * | 2011-05-03 | 2012-11-08 | 北京中瑞泰科技有限公司 | Similarity curve-based device malfunction early-warning and optimization method and system |
| CN103234767A (en) * | 2013-04-21 | 2013-08-07 | 蒋全胜 | Nonlinear fault detection method based on semi-supervised manifold learning |
| CN103323772A (en) * | 2012-03-21 | 2013-09-25 | 北京光耀能源技术股份有限公司 | Wind driven generator operation state analyzing method based on neural network model |
| CN103674234A (en) * | 2013-12-23 | 2014-03-26 | 北京天源科创风电技术有限责任公司 | State early warning method and system for abnormal vibration of wind generating set |
-
2015
- 2015-12-31 CN CN201511029436.4A patent/CN105677791B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012149901A1 (en) * | 2011-05-03 | 2012-11-08 | 北京中瑞泰科技有限公司 | Similarity curve-based device malfunction early-warning and optimization method and system |
| CN103323772A (en) * | 2012-03-21 | 2013-09-25 | 北京光耀能源技术股份有限公司 | Wind driven generator operation state analyzing method based on neural network model |
| CN103234767A (en) * | 2013-04-21 | 2013-08-07 | 蒋全胜 | Nonlinear fault detection method based on semi-supervised manifold learning |
| CN103674234A (en) * | 2013-12-23 | 2014-03-26 | 北京天源科创风电技术有限责任公司 | State early warning method and system for abnormal vibration of wind generating set |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105677791A (en) | 2016-06-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105677791B (en) | For analyzing the method and system of the operation data of wind power generating set | |
| CN110097297B (en) | Multi-dimensional electricity stealing situation intelligent sensing method, system, equipment and medium | |
| CN111782472B (en) | System abnormality detection method, device, equipment and storage medium | |
| CN108897954A (en) | Wind turbines temperature pre-warning method and its system based on BootStrap confidence calculations | |
| CN110750524A (en) | Method and system for determining fault characteristics of active power distribution network | |
| CN111259947A (en) | Power system fault early warning method and system based on multi-mode learning | |
| CN109409444B (en) | Multivariate power grid fault type discrimination method based on prior probability | |
| CN114723285B (en) | Power grid equipment safety evaluation prediction method | |
| CN112818604A (en) | Wind turbine generator risk degree assessment method based on wind power prediction | |
| CN112464995A (en) | Power grid distribution transformer fault diagnosis method and system based on decision tree algorithm | |
| CN111126820A (en) | Electricity stealing prevention method and system | |
| CN115660262B (en) | Engineering intelligent quality inspection method, system and medium based on database application | |
| CN116450399A (en) | Fault diagnosis and root cause positioning method for micro service system | |
| CN110968703B (en) | Method and system for constructing abnormal metering point knowledge base based on LSTM end-to-end extraction algorithm | |
| CN115204536A (en) | Building equipment fault prediction method, device, equipment and storage medium | |
| CN115526258A (en) | Power system transient stability evaluation method based on Spearman correlation coefficient feature extraction | |
| CN111881159A (en) | Fault detection method and device based on cost-sensitive extreme random forest | |
| CN111612149A (en) | Main network line state detection method, system and medium based on decision tree | |
| Fu et al. | Trimming outliers using trees: winning solution of the large-scale energy anomaly detection (LEAD) competition | |
| CN115879915A (en) | Cross-platform standardized overhauling method for power plant | |
| CN116150692A (en) | Power grid fault diagnosis method and system based on cloud platform and improved isolated forest | |
| CN114626433A (en) | Fault prediction and classification method, device and system for intelligent electric energy meter | |
| CN114139408A (en) | Power transformer health state assessment method | |
| CN113726558A (en) | Network equipment flow prediction system based on random forest algorithm | |
| CN115598459A (en) | Power failure prediction method for 10kV feeder line fault of power distribution network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 830026 No. 107, Shanghai Road, Urumqi economic and Technological Development Zone, the Xinjiang Uygur Autonomous Region Patentee after: Jinfeng Technology Co.,Ltd. Address before: 830026 No. 107, Shanghai Road, Urumqi economic and Technological Development Zone, the Xinjiang Uygur Autonomous Region Patentee before: XINJIANG GOLDWIND SCIENCE & TECHNOLOGY Co.,Ltd. |