CN109492777A - A kind of Wind turbines health control method based on machine learning algorithm platform - Google Patents
A kind of Wind turbines health control method based on machine learning algorithm platform Download PDFInfo
- Publication number
- CN109492777A CN109492777A CN201811071601.6A CN201811071601A CN109492777A CN 109492777 A CN109492777 A CN 109492777A CN 201811071601 A CN201811071601 A CN 201811071601A CN 109492777 A CN109492777 A CN 109492777A
- Authority
- CN
- China
- Prior art keywords
- health
- model
- data
- wind turbines
- machine learning
- 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
- 230000036541 health Effects 0.000 title claims abstract description 63
- 238000010801 machine learning Methods 0.000 title claims abstract description 22
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 19
- 238000000034 method Methods 0.000 title claims abstract description 18
- 230000005611 electricity Effects 0.000 claims abstract description 34
- 230000008439 repair process Effects 0.000 claims abstract description 26
- 238000004458 analytical method Methods 0.000 claims abstract description 22
- 238000012423 maintenance Methods 0.000 claims abstract description 13
- 238000003745 diagnosis Methods 0.000 claims abstract description 11
- 238000007689 inspection Methods 0.000 claims abstract description 10
- 238000010248 power generation Methods 0.000 claims abstract description 8
- 238000005516 engineering process Methods 0.000 claims abstract description 7
- 230000003862 health status Effects 0.000 claims abstract description 6
- 230000007704 transition Effects 0.000 claims abstract description 5
- 238000011156 evaluation Methods 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 10
- 230000006870 function Effects 0.000 claims description 8
- 238000007726 management method Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 6
- 238000003066 decision tree Methods 0.000 claims description 5
- 238000013461 design Methods 0.000 claims description 5
- 241000208340 Araliaceae Species 0.000 claims description 4
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims description 4
- 235000003140 Panax quinquefolius Nutrition 0.000 claims description 4
- 230000015556 catabolic process Effects 0.000 claims description 4
- 238000013135 deep learning Methods 0.000 claims description 4
- 238000013210 evaluation model Methods 0.000 claims description 4
- 235000008434 ginseng Nutrition 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 239000000284 extract Substances 0.000 claims description 2
- 230000004927 fusion Effects 0.000 claims description 2
- 238000005259 measurement Methods 0.000 claims description 2
- 230000007246 mechanism Effects 0.000 claims description 2
- 238000005457 optimization Methods 0.000 claims description 2
- 230000003449 preventive effect Effects 0.000 claims description 2
- 238000012827 research and development Methods 0.000 claims description 2
- 238000012360 testing method Methods 0.000 claims description 2
- 238000012800 visualization Methods 0.000 claims description 2
- 230000008901 benefit Effects 0.000 abstract description 3
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000004519 manufacturing process Methods 0.000 description 4
- 238000007418 data mining Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000005684 electric field Effects 0.000 description 2
- 230000008030 elimination Effects 0.000 description 2
- 238000003379 elimination reaction Methods 0.000 description 2
- 238000011002 quantification Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000034303 cell budding Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Educational Administration (AREA)
- Development Economics (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of Wind turbines health control method based on machine learning algorithm platform.The health control method includes: theoretical power generation equilibrium analysis, performance rate assessment, fault diagnosis model and health control modelling, health evaluating and health degree analysis, the realization of health control, that is, repair based on condition of component.According to the deep application of theoretical capacity equilibrium analysis, in conjunction with big data technology, through machine learning in conjunction with artificial intelligence, realize the anticipation and assessment of the health status of blower and its accessory and capital equipment, finally make the maintenance model of Wind turbines from scheduled overhaul, regular inspection periodical repair to repair based on condition of component transition, to reduce Wind turbines scheduled overhaul loss electricity, non-plan repair loss electricity and performance loss electricity, reach the target for improving wind-powered electricity generation station generated energy, improving station economic benefit.
Description
Technical field
The present invention relates to internet big data technology and energy wind power generation fields, and in particular to one kind is based on machine learning
The Wind turbines health control method of algorithm platform.
Background technique
China's wind-powered electricity generation industry is grown rapidly, installed capacity of wind-driven power rapid growth and the special addressing and wind-powered electricity generation of wind power plant
The fluctuation of unit load brings huge challenge to the work of Wind turbines fault alarm.There is alarm in conventional alarm system
It is more, alarm is inaccurate, alarm not in time the problem of, and existing big data system is not fully used, and is lacked there are following
It falls into and insufficient:
1) magnanimity fan operation data, maintenance data and environmental data waste, huge data volume can not but pass through software systems
Reach filtering, processing, analysis within the reasonable time and is formed as the more effective information of wind-powered electricity generation enterprise production and operation management.
2) a large amount of data do not play the role of big data, the low efficiency of data processing and data mining, without accurate
Visual analyzing show.
3) existing alarm causes Threshold Alerts, is not associated associated components, can not achieve the in-advance of alarm
And accuracy.
4) the Wind turbines health control platform based on machine learning algorithm platform is not built, wind power plant Wind turbines are strong
There are management risks for health state unpredictability.
Summary of the invention
The purpose of the present invention is propose that one kind passes through scene for shortcoming existing for current Wind turbines management system
The means such as fault data statistics, expert's assessment, quantification modeling, under the premise of guaranteeing safety and integrity, with the smallest
Repairing shutdown loss and the consumption of the smallest Maintenance Resource is target, provides a kind of Wind turbines based on machine learning algorithm platform
Health control method.
Present general inventive concept are as follows: Wind turbines are health management system arranged, goed deep into according to theoretical capacity equilibrium analysis
Using through machine learning in conjunction with artificial intelligence, realizing the strong of blower and its accessory and capital equipment in conjunction with big data technology
The anticipation and assessment of health state finally make the maintenance model of Wind turbines from scheduled overhaul, regular inspection periodical repair to repair based on condition of component mistake
It crosses, to reduce Wind turbines scheduled overhaul loss electricity, non-plan repair loss electricity and performance loss electricity, reaches raising
Wind-powered electricity generation station generated energy, the target for improving station economic benefit.
The technical scheme is that such:
A kind of Wind turbines health control method based on machine learning algorithm platform.
The health control method includes: theoretical power generation equilibrium analysis, performance rate assessment, fault diagnosis model and strong
The design of health administrative model, health evaluating and health degree analysis, the realization of health control, that is, repair based on condition of component.Each part mentioned above function is such as
Under:
1, theoretical power generation equilibrium analysis.
Calculation formula: theoretical power generation=reality generated energy+scheduled overhaul loses electricity+non-plan repair loses electricity+and rations the power supply
It loses electricity+involvement and loses electricity+performance loss electricity.Scheduled overhaul loses electricity: based on Inspection interval, establishment inspection
The plan of repairing, the electric quantity loss generated during carrying out preventive overhaul to unit are known as scheduled overhaul loss electricity (unit itself
Maintenance);
Non-plan repair loses electricity: the electric quantity loss during unit failure causes shutdown and defect elimination to handle is known as unplanned inspection
It repairs loss electricity and is called breakdown loss electricity (unit faults itself);
It rations the power supply and loses electricity: holding two kinds including rationing the power supply to shut down and ration the power supply to drop, the electric quantity loss during scheduling is rationed the power supply is known as rationing the power supply
Lose electricity;
Involvement loss electricity: involvement is divided into involvement and over-the-counter involvement in field.Involvement is shut down in the field other than referring to because of unit and is set in
It is standby stop transport (such as main transformer, the matched facility failure of collection electric line or scheduled overhaul) cause unit be forced it is out of service during make
At electric quantity loss;Over-the-counter involvement, which is shut down, to be referred to because of over-the-counter power grid reason (such as external power lines road, electric power system fault, maintenance)
Cause unit whithin a period of time (non real-time scheduling) be forced it is out of service in a period of electric quantity loss or natural cause
Caused (sleet, frost, typhoon etc. can not resist weather condition) causes equipment downtime, drop to hold beyond the case where fan design grade
Electric quantity loss.
Performance loss electricity: hold and below standard two kinds of unit performance comprising unit from drop.Unit holds from drop refers to unit certainly
Body reason causes unit drop power output operation, as protected operation after main component (bearing) replacement.Unit performance is below standard to refer to fortune
Row power output be not achieved standard value or wind speed reach specified incision wind speed after unit can not cut operation, the electricity in this period damages
It loses and is known as performance loss electricity.
2, performance rate is assessed.
The assessment of Wind turbines performance rate is exactly to establish evaluation mould on the basis of analyzing Wind turbines performance risk factors
Type estimates Wind turbines performance rate score, and single risk size can be protruded in comprehensive assessment to total evaluation
Influence, more objectively respond actual conditions.Dynamic variable weight fuzzy evaluation model mainly includes two parts:
(1) weight of Wind turbines performance rate evaluation index is determined with the theory that dynamic is weighed surely;
(2) then weight is corrected according to the correlation between index, obtains variable weight model assessment models.
Using variable weight Fuzzy evaluation mode, is graded and scored lower than fan performance risk, reflect unit with data
Performance condition, and equipment performance is divided into A, B, C three grades, unit performance risk situation is calculated, unit performance grade is obtained
Evaluate score, by A grade of Wind turbines it is qualitative for performance it is up to standard, by B, C grades of Wind turbines it is qualitative be performance not
It is up to standard;After the evaluation of passage capacity grade, equipment performance is quantified into the qualitative blower not up to standard for performance, into fault diagnosis mould
Type realizes the anticipation of failure primary to performance blower not up to standard.
3, fault diagnosis model and health control modelling.
Enter big data environment by real time data, overhaul data, video data etc., it is different with multi-source by database interfusion
Structure model, is associated Various types of data, then forms knowledge by computation model, recycles machine learning method and algorithm,
Breakdown judge is carried out by failure of the fault diagnosis to equipment, forms fault type, and pass through the index of setting equipment health degree
Initial value constantly carries out deep learning to health degree initial value to carry out intelligent fault alarm, until index value is completely quasi-
Really, alarm is really accurate.And after completing machine learning and deep learning, accurate health evaluating report and equipment health degree are formed
Real-time tracking and the anticipation of equipment health status are realized in analysis.
Quantification treatment is carried out rather than simple qualitative description to the attribute of each component, main includes representing to disagree
Standard value, acceptable value, the value that transfinites etc. of justice, are described the problem with number, become the numerical chracter that computer can identify, currently
Measured value by with standard value, acceptable value, the value that transfinites according to certain algorithm comparison after by intellectual analysis, provide alarm etc.
Grade analyzes whether it is false alarm, finally provides conclusion, provides decision-making foundation for maintenance staff.
4, health evaluating and health degree analysis, the realization of health control, that is, repair based on condition of component.
Complete machine learning, intelligent platform and it is health management system arranged after, realize the anticipation of equipment health, driving equipment
Inspecting state is from completing overhaul of the equipments and changing from regular inspection periodical repair, scheduled overhaul to repair based on condition of component afterwards into thing with transition in advance.
Comprehensive Evaluation is carried out from multi-level, multi-angle, to realize the accurate, comprehensive of evaluation result.Firstly, more current measured value with
Standard value obtains the degree for deviateing standard value.Secondly, the relatively important carry out fuzzy evaluation to current detection component, provides it
Relative weighting.We are also contemplated that the factor of other aspects, provide evaluation result based on various factors.
The invention has the advantages that:
1) the existing wrong report of wind farm device alarm and failing to report phenomenon are solved, realizes the objective and accurate property of alarm, is fault alarm
Treatment people provides objective accurate alert process foundation;
2) wind power plant Wind turbines health status unpredictability is solved, realizes the higher wind farm device health status of confidence level
Prediction, provides scientific basis for State Maintenance, Failure elimination in budding state, realizes repair based on condition of component;
3) it is predetermined to solve the problems, such as that Wind turbines maintenance program is difficult to, realizes Reliability Maintenance;
4) it solves the problems, such as that the shortage of wind farm device maintenance expert and Maintenance Resource can not be shared, establishes the wind constantly improve
Electric field equipment repairs intelligent expert system;
5) wind farm device is solved because of random stop phenomenon caused by failure and the influence to production, production capacity, profit, realizes wind
Electric field equipment realizes the maximization of wind power plant production capacity, profit maximum because of the Modulatory character shut down caused by failure.
Detailed description of the invention
Fig. 1 is Wind turbines health control flow diagram of the present invention.
Specific embodiment
A specific embodiment of the invention combination attached drawing is illustrated.
As shown in Figure 1, a kind of Wind turbines health control method based on machine learning algorithm platform.The management method packet
Include following procedure:
1, performance rate is evaluated.
Utility grade evaluation model carries out performance rate assessment to blower and its capital equipment and accessory, realizes to wind
The division of motor group performance rate will be defined as performance blower up to standard in A grades of blower, will define in B/C grades of blower
For performance blower not up to standard.
2, fault diagnosis model designs.
By step 1, by fan performance blower not up to standard, into fault diagnosis model, in conjunction with big data platform and machine
The application for learning algorithm platform carries out breakdown judge by failure of the fault diagnosis to equipment, forms fault type, and will be faulty
Blower is alarmed.
3, health evaluating and health degree analysis.
Standby redundancy information, service personnel's information, Weather information are equal to health control models coupling, formed accurately strong
Health assessment report and equipment health degree analysis realize real-time tracking and the anticipation of equipment health status, realize the pre- of equipment health
Sentence, driving equipment inspecting state is from completing overhaul of the equipments from regular inspection periodical repair, scheduled overhaul to shape afterwards into thing with transition in advance
State maintenance transformation.
Specific step is as follows:
1) real time data, overhaul data, video data etc. are written to Hadoop big data platform, realize different types of data
Fusion.We do not start with from sensor, start with from blower structure mechanism, using step analysis means, are split to each
Small component, and the function of each widget, relatively important weight, failure performance, fault message, solution etc. is detailed
Information all enumerates out;For be fitted without sensor part of appliance we intend taking correlation rule, using indirect measurement
Technology analyzes part of appliance that is associated therewith and being mounted with sensor, in conjunction with sensorless technology and some advanced biographies
Sensor obtains;Its bright spot is consideration that each component is associated, rather than traditional independence is analyzed.
2) blower data are analyzed and screened with modeling data, the data that model training needs are imported into machine learning
Platform.
3) to the format of data, legitimacy of data etc. carries out relevant processing.
4) it designs and extracts feature and be used for model training.
5) it chooses GBDT algorithm and carries out model training, export trained machine learning model for subsequent pre-estimation
It calculates.
Algorithms selection: GBDT(Gradient Boosting Decision Tree, Gradient Iteration decision tree)
The applicable scene of GBDT:
Continuous feature is more;
Too many Character adjustment is not needed;
Has certain interpretation after visualization;
Model optimization emphasis-parameter;
GBDT model includes 6 adjustable parameters altogether, is respectively as follows: learning rate, the depth capacity of single tree, the number of tree, most leaflet
Sub- weight, minimum division threshold value, L2 canonical;
Adjusting ginseng work is the process that more preferable result is found in continuous modeling, and project team carries out more wheels to GBDT model and adjusts ginseng, final to select
One group of most suitable parameter configuration is selected, so that model is stablized in training set and assessment concentrated expression;
6) online Prediction service framework is built, data processing, and online Prediction interface and service logic research and development, web services and API are real
It is existing, test.
7) model generated to model training step is assessed using recall rate and logarithm loss function.
Recall rate is a kind of Classification Algorithms in Data Mining evaluation index, the accuracy of decision Tree algorithms, neural network algorithm
It is to be judged according to recall rate come calculated result.Recall rate is the information bar number in the correct information item number and sample extracted
Ratio, (TP: being correctly divided into the number of positive example to recall=TP/ (TP+FN), i.e., practical to be positive example and be classified device stroke
It is divided into the instance number of positive example;FN: mistakenly being divided the number of example of being negative, i.e., practical to be positive example but be classified device division and be negative example
Instance number).
Such as: it realizes alarm and the number that should be alarmed is A, the number that alarm is not implemented but should alarm is B, then alarm is sentenced
Fixed recall rate recall=A/ (A+B).
Logarithm loss function (Log-loss): for estimating the predicted value of model and the inconsistent degree of true value.Log-
Loss is a soft classification accuracy measure, and the confidence level of the classification belonging to it is indicated using probability.(yi refers to
True classification 0 or 1 belonging to i sample, pi indicate that i-th of sample belongs to the probability of classification 1).
In classification output, if output is real number value, that is, belong to the probability of each classification, then Log- can be used
Loss evaluates classification results.Such as: when prediction is exactly matched with concrete class, then two parts of formula are all 0(vacations
Determine 0log0=0), then logarithm loss function value is 0.
4, health evaluating and health degree analysis, the realization of health control, that is, repair based on condition of component.
Complete machine learning, intelligent platform and it is health management system arranged after, realize the anticipation of equipment health, driving equipment
Inspecting state is from completing overhaul of the equipments and changing from regular inspection periodical repair, scheduled overhaul to repair based on condition of component afterwards into thing with transition in advance.
We carry out Comprehensive Evaluation from multi-level, multi-angle, to realize the accurate, comprehensive of evaluation result.Firstly, we
Compare current measured value and standard value, obtains the degree for deviateing standard value.Secondly, we are to the relatively important of current detection component
Fuzzy evaluation is carried out, its relative weighting is provided.We are also contemplated that the factor of other aspects, provide judge based on various factors
As a result.
Claims (3)
1. a kind of Wind turbines health control method based on machine learning algorithm platform, which is characterized in that the health control side
Method includes: theoretical power generation equilibrium analysis, performance rate assessment, fault diagnosis model and health control modelling, health
Assessment and health degree analysis, the realization of health control, that is, repair based on condition of component, each part mentioned above function are as follows:
(1) theoretical power generation equilibrium analysis:
Calculation formula: theoretical power generation=reality generated energy+scheduled overhaul loses electricity+non-plan repair loses electricity+loss of rationing the power supply
Electricity+performance loss electricity is lost in electricity+involvement, and scheduled overhaul loses electricity, based on Inspection interval, establishment maintenance meter
It draws, the electric quantity loss generated during carrying out preventive overhaul to unit is known as scheduled overhaul loss electricity, and (unit itself is examined
It repairs);
(2) performance rate is assessed:
The assessment of Wind turbines performance rate is exactly to establish evaluation model on the basis of analyzing Wind turbines performance risk factors,
Wind turbines performance rate score is estimated, single risk size can be protruded in comprehensive assessment to the shadow of total evaluation
It rings, more objectively responds actual conditions;
(3) fault diagnosis model and health control modelling:
Enter big data environment by real time data, overhaul data, video data etc., passes through database interfusion and multi-source heterogeneous mould
Type is associated Various types of data, then forms knowledge by computation model, recycles machine learning method and algorithm, passes through
Fault diagnosis carries out breakdown judge to the failure of equipment, forms fault type, and initial by the index of setting equipment health degree
Value constantly carries out deep learning to health degree initial value to carry out intelligent fault alarm, until index value entirely accurate, report
It is alert really accurate, and complete machine learning and after deep learning, form accurate health evaluating report and equipment health degree analysis,
Realize real-time tracking and the anticipation of equipment health status;
(4) health evaluating and health degree analysis, the realization of health control, that is, repair based on condition of component:
Complete machine learning, intelligent platform and it is health management system arranged after, realize the anticipation of equipment health, driving equipment is overhauled
State is from completing overhaul of the equipments and changing from regular inspection periodical repair, scheduled overhaul to repair based on condition of component, from more afterwards into thing with transition in advance
Level, multi-angle carry out Comprehensive Evaluation, to realize the accurate, comprehensive of evaluation result.
2. health control method according to claim 1, it is characterised in that: the evaluation that described performance rate assessment is established
Model is dynamic variable weight fuzzy evaluation model, mainly includes two parts: (1) determining Wind turbines with the theory that dynamic is weighed surely
The weight of performance rate evaluation index;(2) then weight is corrected according to the correlation between index, obtains variable weight mould
Type assessment models.
3. health control method according to claim 1, it is characterised in that: described health evaluating and health degree analysis
Specific step is as follows:
1) real time data, overhaul data, video data etc. are written to Hadoop big data platform, realize different types of data
Fusion.We do not start with from sensor, start with from blower structure mechanism, using step analysis means, are split to each
Small component, and the function of each widget, relatively important weight, failure performance, fault message, solution etc. is detailed
Information all enumerates out;For be fitted without sensor part of appliance we intend taking correlation rule, using indirect measurement
Technology analyzes part of appliance that is associated therewith and being mounted with sensor, in conjunction with sensorless technology and some advanced biographies
Sensor obtains;Its bright spot is consideration that each component is associated, rather than traditional independence is analyzed;
2) blower data are analyzed and screened with modeling data, the data that model training needs are imported into machine learning and are put down
Platform.;
3) to the format of data, legitimacy of data etc. carries out relevant processing;
4) it designs and extracts feature and be used for model training;
5) it chooses GBDT algorithm and carries out model training, export trained machine learning model for subsequent and estimate calculating;
Algorithms selection: GBDT(Gradient Boosting Decision Tree, Gradient Iteration decision tree);
The applicable scene of GBDT:
Continuous feature is more;
Too many Character adjustment is not needed;
Has certain interpretation after visualization;
Model optimization emphasis-parameter;
GBDT model includes 6 adjustable parameters altogether, is respectively as follows: learning rate, the depth capacity of single tree, the number of tree, most leaflet
Sub- weight, minimum division threshold value, L2 canonical;
Adjusting ginseng work is the process that more preferable result is found in continuous modeling, and project team carries out more wheels to GBDT model and adjusts ginseng, final to select
One group of most suitable parameter configuration is selected, so that model is stablized in training set and assessment concentrated expression;
6) online Prediction service framework is built, data processing, and online Prediction interface and service logic research and development, web services and API are real
It is existing, test;
7) model generated to model training step is assessed using recall rate and logarithm loss function.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811071601.6A CN109492777A (en) | 2018-09-14 | 2018-09-14 | A kind of Wind turbines health control method based on machine learning algorithm platform |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811071601.6A CN109492777A (en) | 2018-09-14 | 2018-09-14 | A kind of Wind turbines health control method based on machine learning algorithm platform |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109492777A true CN109492777A (en) | 2019-03-19 |
Family
ID=65689239
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811071601.6A Pending CN109492777A (en) | 2018-09-14 | 2018-09-14 | A kind of Wind turbines health control method based on machine learning algorithm platform |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109492777A (en) |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110309981A (en) * | 2019-07-09 | 2019-10-08 | 华能四川水电有限公司 | A kind of power station Decision-making of Condition-based Maintenance system based on industrial big data |
CN110701002A (en) * | 2019-10-18 | 2020-01-17 | 东方电气风电有限公司 | Virtual sensing implementation system for large wind generating set |
CN110750884A (en) * | 2019-10-01 | 2020-02-04 | 深圳市行健自动化股份有限公司 | Method for checking and diagnosing real-time online instrument by self-adaptive support vector machine algorithm |
CN111062508A (en) * | 2019-03-28 | 2020-04-24 | 华北电力大学(保定) | Method for evaluating real-time running state of wind turbine generator based on big data technology |
CN111881617A (en) * | 2020-07-02 | 2020-11-03 | 上海电气风电集团股份有限公司 | Data processing method, and performance evaluation method and system of wind generating set |
CN112215397A (en) * | 2020-09-08 | 2021-01-12 | 中广核风电有限公司 | Wind turbine generator intelligent management method and system based on big data and storage medium |
CN112328590A (en) * | 2020-12-01 | 2021-02-05 | 西安交通大学 | Deep cleaning method for operation data of thermal equipment |
CN112381352A (en) * | 2020-10-19 | 2021-02-19 | 大唐(赤峰)新能源有限公司 | Wind turbine generator running state evaluation method based on multi-source heterogeneous monitoring |
CN112529365A (en) * | 2020-11-19 | 2021-03-19 | 华电电力科学研究院有限公司 | Wind power generation company spare part reserve volume analysis and management equipment and method |
US10954919B1 (en) | 2019-10-02 | 2021-03-23 | General Electric Company | Machine-learning model-based analytic for monitoring wind farm power performance |
CN113095606A (en) * | 2021-06-09 | 2021-07-09 | 北矿智云科技(北京)有限公司 | Equipment maintenance prejudging method, device and system |
CN113360358A (en) * | 2021-06-25 | 2021-09-07 | 杭州优云软件有限公司 | Method and system for adaptively calculating IT intelligent operation and maintenance health index |
CN113435759A (en) * | 2021-07-01 | 2021-09-24 | 贵州电网有限责任公司 | Primary equipment risk intelligent evaluation method based on deep learning |
CN113487182A (en) * | 2021-07-06 | 2021-10-08 | 新智数字科技有限公司 | Equipment health state evaluation method and device, computer equipment and medium |
CN113486868A (en) * | 2021-09-07 | 2021-10-08 | 中南大学 | Motor fault diagnosis method and system |
US11231012B1 (en) | 2020-09-22 | 2022-01-25 | General Electric Renovables Espana, S.L. | Systems and methods for controlling a wind turbine |
CN114021906A (en) * | 2021-10-19 | 2022-02-08 | 广东邦鑫数据科技股份有限公司 | Unattended wind power generation operation and maintenance management method and system |
CN114239870A (en) * | 2021-11-10 | 2022-03-25 | 深圳供电局有限公司 | Health state detection method, system and storage medium |
US11649804B2 (en) | 2021-06-07 | 2023-05-16 | General Electric Renovables Espana, S.L. | Systems and methods for controlling a wind turbine |
CN117192373A (en) * | 2023-08-08 | 2023-12-08 | 浙江凌骁能源科技有限公司 | Power battery fault analysis method, device, computer equipment and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105604807A (en) * | 2015-12-31 | 2016-05-25 | 北京金风科创风电设备有限公司 | Wind turbine generator monitoring method and device |
CN106407589A (en) * | 2016-09-29 | 2017-02-15 | 北京岳能科技股份有限公司 | Wind turbine state evaluation and prediction method and system |
CN107220469A (en) * | 2017-07-14 | 2017-09-29 | 华北电力大学(保定) | A kind of method and system of fan condition estimation |
-
2018
- 2018-09-14 CN CN201811071601.6A patent/CN109492777A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105604807A (en) * | 2015-12-31 | 2016-05-25 | 北京金风科创风电设备有限公司 | Wind turbine generator monitoring method and device |
CN106407589A (en) * | 2016-09-29 | 2017-02-15 | 北京岳能科技股份有限公司 | Wind turbine state evaluation and prediction method and system |
CN107220469A (en) * | 2017-07-14 | 2017-09-29 | 华北电力大学(保定) | A kind of method and system of fan condition estimation |
Cited By (29)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111062508A (en) * | 2019-03-28 | 2020-04-24 | 华北电力大学(保定) | Method for evaluating real-time running state of wind turbine generator based on big data technology |
CN111062508B (en) * | 2019-03-28 | 2023-10-31 | 华北电力大学(保定) | Method for evaluating real-time running state of wind turbine generator based on big data technology |
CN110309981A (en) * | 2019-07-09 | 2019-10-08 | 华能四川水电有限公司 | A kind of power station Decision-making of Condition-based Maintenance system based on industrial big data |
CN110750884A (en) * | 2019-10-01 | 2020-02-04 | 深圳市行健自动化股份有限公司 | Method for checking and diagnosing real-time online instrument by self-adaptive support vector machine algorithm |
CN110750884B (en) * | 2019-10-01 | 2023-06-20 | 深圳市行健自动化股份有限公司 | Instant on-line instrument checksum diagnosis method through self-adaptive support vector machine algorithm |
US10954919B1 (en) | 2019-10-02 | 2021-03-23 | General Electric Company | Machine-learning model-based analytic for monitoring wind farm power performance |
CN110701002A (en) * | 2019-10-18 | 2020-01-17 | 东方电气风电有限公司 | Virtual sensing implementation system for large wind generating set |
CN111881617A (en) * | 2020-07-02 | 2020-11-03 | 上海电气风电集团股份有限公司 | Data processing method, and performance evaluation method and system of wind generating set |
CN111881617B (en) * | 2020-07-02 | 2024-03-26 | 上海电气风电集团股份有限公司 | Data processing method, performance evaluation method and system of wind generating set |
CN112215397A (en) * | 2020-09-08 | 2021-01-12 | 中广核风电有限公司 | Wind turbine generator intelligent management method and system based on big data and storage medium |
US11231012B1 (en) | 2020-09-22 | 2022-01-25 | General Electric Renovables Espana, S.L. | Systems and methods for controlling a wind turbine |
CN112381352A (en) * | 2020-10-19 | 2021-02-19 | 大唐(赤峰)新能源有限公司 | Wind turbine generator running state evaluation method based on multi-source heterogeneous monitoring |
CN112529365A (en) * | 2020-11-19 | 2021-03-19 | 华电电力科学研究院有限公司 | Wind power generation company spare part reserve volume analysis and management equipment and method |
CN112529365B (en) * | 2020-11-19 | 2023-08-18 | 华电电力科学研究院有限公司 | Wind power generation company spare part reserve analysis management device and method |
CN112328590B (en) * | 2020-12-01 | 2023-04-07 | 西安交通大学 | Deep cleaning method for operation data of thermal equipment |
CN112328590A (en) * | 2020-12-01 | 2021-02-05 | 西安交通大学 | Deep cleaning method for operation data of thermal equipment |
US11649804B2 (en) | 2021-06-07 | 2023-05-16 | General Electric Renovables Espana, S.L. | Systems and methods for controlling a wind turbine |
CN113095606A (en) * | 2021-06-09 | 2021-07-09 | 北矿智云科技(北京)有限公司 | Equipment maintenance prejudging method, device and system |
CN113360358B (en) * | 2021-06-25 | 2022-05-27 | 杭州优云软件有限公司 | Method and system for adaptively calculating IT intelligent operation and maintenance health index |
CN113360358A (en) * | 2021-06-25 | 2021-09-07 | 杭州优云软件有限公司 | Method and system for adaptively calculating IT intelligent operation and maintenance health index |
CN113435759A (en) * | 2021-07-01 | 2021-09-24 | 贵州电网有限责任公司 | Primary equipment risk intelligent evaluation method based on deep learning |
CN113435759B (en) * | 2021-07-01 | 2023-07-04 | 贵州电网有限责任公司 | Primary equipment risk intelligent assessment method based on deep learning |
CN113487182A (en) * | 2021-07-06 | 2021-10-08 | 新智数字科技有限公司 | Equipment health state evaluation method and device, computer equipment and medium |
CN113487182B (en) * | 2021-07-06 | 2024-04-26 | 新奥新智科技有限公司 | Device health state evaluation method, device, computer device and medium |
CN113486868A (en) * | 2021-09-07 | 2021-10-08 | 中南大学 | Motor fault diagnosis method and system |
CN114021906A (en) * | 2021-10-19 | 2022-02-08 | 广东邦鑫数据科技股份有限公司 | Unattended wind power generation operation and maintenance management method and system |
CN114239870A (en) * | 2021-11-10 | 2022-03-25 | 深圳供电局有限公司 | Health state detection method, system and storage medium |
CN117192373A (en) * | 2023-08-08 | 2023-12-08 | 浙江凌骁能源科技有限公司 | Power battery fault analysis method, device, computer equipment and storage medium |
CN117192373B (en) * | 2023-08-08 | 2024-05-07 | 浙江凌骁能源科技有限公司 | Power battery fault analysis method, device, computer equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109492777A (en) | A kind of Wind turbines health control method based on machine learning algorithm platform | |
WO2022151890A1 (en) | Method for establishing power generating capacity evaluation and micro-siting model for wind farms | |
CN106054104B (en) | A kind of intelligent electric meter failure real-time predicting method based on decision tree | |
CN107341349A (en) | Method, system, memory and the controller of blower fan health evaluating | |
CN110320892A (en) | The sewage disposal device fault diagnosis system and method returned based on Lasso | |
CN104952000A (en) | Wind turbine operating state fuzzy synthetic evaluation method based on Markov chain | |
CN102411729B (en) | Wind power prediction method based on adaptive linear logic network | |
CN103793859B (en) | A kind of wind power plant operation monitoring and event integrated evaluating method | |
CN106777984A (en) | A kind of method that photovoltaic array Working state analysis and fault diagnosis are realized based on density clustering algorithm | |
CN105846780A (en) | Decision tree model-based photovoltaic assembly fault diagnosis method | |
CN111754020A (en) | Intelligent power plant equipment inspection system and method based on data analysis | |
Bangalore et al. | An approach for self evolving neural network based algorithm for fault prognosis in wind turbine | |
CN109522599A (en) | Transmission line of electricity catastrophic failure method for early warning caused by a kind of typhoon | |
CN110162555A (en) | A kind of fired power generating unit start and stop and drop power output measure of supervision | |
CN115829145A (en) | Photovoltaic power generation capacity prediction system and method | |
CN110968703A (en) | Method and system for constructing abnormal metering point knowledge base based on LSTM end-to-end extraction algorithm | |
Dong et al. | Fault diagnosis and classification in photovoltaic systems using scada data | |
CN110273818A (en) | A kind of fan blade icing fault monitoring method based on the classification of principal axis transformation fineness degree | |
CN116796633A (en) | Power grid monitoring model correction method based on digital twin technology | |
Ren et al. | Research on causes of transmission line fault based on decision tree classification | |
CN108345996B (en) | System and method for reducing wind power assessment electric quantity | |
CN116938676A (en) | Communication risk combined early warning method based on data source analysis | |
CN115936663A (en) | Maintenance method and device for power system | |
CN115310746A (en) | State evaluation method and system for main transmission system of wind generating set | |
Huang et al. | An intelligent operation and maintenance system for power consumption based on deep learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190319 |