CN105389475B - A kind of electric power factory equipment fault detection method based on WAVELET PACKET DECOMPOSITION - Google Patents
A kind of electric power factory equipment fault detection method based on WAVELET PACKET DECOMPOSITION Download PDFInfo
- Publication number
- CN105389475B CN105389475B CN201510970441.9A CN201510970441A CN105389475B CN 105389475 B CN105389475 B CN 105389475B CN 201510970441 A CN201510970441 A CN 201510970441A CN 105389475 B CN105389475 B CN 105389475B
- Authority
- CN
- China
- Prior art keywords
- wavelet packet
- packet decomposition
- fault
- electric power
- model
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
- G06F17/148—Wavelet transforms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16Z—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR
- G16Z99/00—Subject matter not provided for in other main groups of this subclass
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The present invention provides a kind of electric power factory equipment fault detection method based on WAVELET PACKET DECOMPOSITION.Obtain assuming the difference being output between virtual condition data first with existing state-detection model；Then difference data is divided into subpattern one by one using sliding window, again WAVELET PACKET DECOMPOSITION is carried out using the subpattern after segmentation as the input of Fault Model, signal after decomposition is analyzed, and energy of the tracer signal in each frequency range accounts for the threshold value bound of the ratio of whole signal energy.Finally according to feature of the energy proportion of each frequency range after WAVELET PACKET DECOMPOSITION as failure judgement state.The present invention has taken into full account the complexity of electric power factory equipment working status, and its factor of embryonic character more difficult discovery in time domain of data fault, can accurately detect the generation of electric power factory equipment failure, accurately identify fault signature.
Description
Technical field
The present invention relates to a kind of electric power factory equipment fault detection method, more particularly to a kind of power plant based on WAVELET PACKET DECOMPOSITION to set
Standby fault detection method.
Background technology
Ensure electric power factory equipment safely and steadly run be electric power factory equipment management in a critically important job.With society
The development of meeting, thermal power generation occupy more and more important ratio and position in national economy, and event once occurs for Power Plant Equipment
Barrier or shutdown can cause huge economic loss and social influence.Into after 21 century, electric power factory equipment is run intelligent, long-range
Change, networking and real-time monitoring and centralized control technique not only have urgent need, and are increasingly becoming possibility.
The research of electric power factory equipment fault detection technique, pays close attention to the identification of the monitoring technology and initial failure of equipment state
Technology, is established based on status monitoring, the device intelligence early warning system supplemented by fault detect.Due to electric power factory equipment working status
Complexity, not all equipment measuring point can find obvious degradation trend by Condition Monitoring Technology, in order to ensure
The operation of equipment normal reliable, fault detection technique are essential.
Fault detect is exactly to distinguish the abnormal number of normal data state in data set by various inspections and test method
According to the process of state.The problem of fault detect, is how by normal data set to establish model, selects rational characteristic parameter
Accurately identify abnormal data.
Fault detection method mainly has following three kinds at present：Fault detect based on analytic modell analytical model, Knowledge based engineering failure
Detection and the fault detect based on signal processing.It is to establish an accurate mathematical modulo according to detection object based on analytic modell analytical model
Type, when it is in normal state, the output of model is in a stable scope, when it enters abnormality, model
Output can exceed normal range (NR).Knowledge based engineering fault detection method need not be that detection object builds accurate mathematical modulo
Type, this method make full use of expert diagnostics information to be detected failure, many by adding the bulk information of detection object
It is used widely in the especially nonlinear system of field.Fault detection method based on signal processing is usually used in detection object shape
State complexity can not establish the mathematical model of system, while fault scenario can not be accurately found in time domain.The advantages of this method
It is the difficult point for having evaded structure system mathematic model, then utilizes signal model detection failure.Common model has higher order statistical
Amount and correlation function, autoregressive moving average process and Wavelet Decomposition Technology etc..
Since the working status of electric power factory equipment is complicated, it is difficult to be characterized with accurate mathematical model, expert diagnostics information is utilized
Also it is difficult to realize, therefore is difficult with Fault Model and Knowledge based engineering Fault Model based on analytic modell analytical model.But
When using the detection model based on signal, electric power factory equipment data more difficult fine feature for finding failure initial stage in time domain, therefore
To accurately judge that electric power factory equipment failure has certain challenge.
In view of the above-mentioned problems, there is an urgent need to invent to provide a kind of electric power factory equipment fault detect side based on WAVELET PACKET DECOMPOSITION
Method.The present invention is directed to the working environment of electric power factory equipment complexity, the fault detection method based on signal processing is selected, using wavelet packet
The signal that data message in time domain is transformed on frequency domain by decomposition technique is analyzed, and constructs the event based on WAVELET PACKET DECOMPOSITION
Hinder detection model, and emulation experiment is carried out using power plant's real time data.
The content of the invention
For electric power factory equipment data in time domain it is more difficult find failure initial stage fine feature the problem of, the present invention develops one
Electric power factory equipment fault detection method of the kind based on WAVELET PACKET DECOMPOSITION.
Present invention be characterized in that comprise the following steps：
(1) training dataset is built from the normal condition data set by data prediction；
(2) assessed value of data set is obtained using state-detection model；
(3) residual error is calculated, and WAVELET PACKET DECOMPOSITION is carried out to it；
(4) each band energy infomation detection failure after analysis is decomposed；
(5) test data set, the input using test data set as Fault Model, according to through WAVELET PACKET DECOMPOSITION are chosen
Feature of the energy proportion of each frequency range as failure judgement state afterwards.
The assessed value is the input using training dataset as status monitoring model, and mould is calculated using status monitoring model
The hypothesis output of type；
The residual computations are to seek the difference for assuming output i.e. between assessed value and real-time status data；
It is after the WAVELET PACKET DECOMPOSITION is the subpattern being divided into residual error data using sliding window one by one, this is a little
Subpattern carries out WAVELET PACKET DECOMPOSITION as the input of Fault Model；
The analysis frequency range is that the signal after WAVELET PACKET DECOMPOSITION is analyzed, energy of the tracer signal in each frequency range
Account for the threshold value bound of the ratio of whole signal energy.
Brief description of the drawings
Fig. 1 is the flow chart of the electric power factory equipment fault detection method based on WAVELET PACKET DECOMPOSITION；
Fig. 2 is the detection of superthreshold abnormal failure；
Fig. 3 is frequency anomaly fault detect；
Fig. 4 is frequency and the detection of superthreshold abnormal failure；
Fig. 5 is to stretch abnormal failure detection.
Embodiment
Illustrate embodiments of the present invention below in conjunction with the accompanying drawings.
Found from the sampling of real-time status data, the measuring point of primary air fan A mainly can there are 4 kinds of malfunctions：
(1) superthreshold is abnormal：The value of abnormality data exceedes the bound threshold value of normal condition data.
(2) frequency anomaly：The status data of original intensive change becomes sparse within certain a period of time.
(3) superthreshold and frequency anomaly：In certain time, data become sparse while data value exceeds normal model
Enclose.
(4) exception is stretched：Within certain a period of time, real-time status data is in alignment, which is due to real-time
When obtaining data, null value is run into, by caused by its value filling with the previous moment.
For the Fault Model based on WAVELET PACKET DECOMPOSITION, energy proportion after decomposition is beyond normal range (NR)
During upper threshold, the value of testing result at this moment is 1；Energy proportion after decomposition is less than the bottom threshold of normal range (NR)
When, the value of testing result at this moment is -1；When the energy proportion after decomposition is in the threshold value of normal range (NR), testing result
Value at this moment is 0.
According to variance, average or slope be less than normal data bottom threshold when, testing result at this moment can value be -1.
For every kind of unusual condition, the testing result displaying of a measuring point in test result is chosen, it is different for superthreshold
For often, testing result of two Fault Models on a measuring point is as shown in Fig. 2, from figure 2 it can be seen that for super
Threshold value is abnormal, and two kinds of Fault Models have a good testing result, but based on sliding window add the method for time window with
It (should not be that warning message is detected as alarm shape that method based on WAVELET PACKET DECOMPOSITION, which is compared and may produce more wrong reports,
State) phenomenon.
For frequency anomaly, two Fault Models for a measuring point of selection fault detect design sketch,
As shown in figure 3, for frequency anomaly model, in a measuring point of selection, the Fault Model energy based on WAVELET PACKET DECOMPOSITION
Enough generations for detecting failure well, and the detection model testing result of sliding window+variance slope is relatively poor, in addition it is right
In the first two measuring point, the latter is almost without detecting the generation of failure.
For frequency and superthreshold exception, fault detect of two Fault Models for a measuring point of selection
Design sketch, as shown in figure 4, abnormal for frequency and superthreshold, two kinds of detection models can preferably detect the generation of failure,
But Fault Model based on WAVELET PACKET DECOMPOSITION it is opposite with add the Fault Model of variance slope based on sliding window come
Say the accuracy rate with higher and more preferable detection result.
For stretching exception, two Fault Models for a measuring point of selection fault detect design sketch,
As shown in figure 5, for stretching exception, two kinds of detection models have the delay of certain time when taking place extremely, this be by
Caused by selection in sliding window, at the end of nonserviceabling, the former can timely detect that state returns to normal shape
State, the latter also have time delay.
Although the illustrative embodiment of the present invention is described above, but it should be clear that the present invention is unlimited
In the scope of embodiment, for those skilled in the art, as long as various change is in appended right
It is required that in the spirit and scope of the present invention for limiting and determining, these changes are it will be apparent that all utilize present inventive concept
Innovation and creation in the row of protection.
Claims (1)
1. a kind of electric power factory equipment fault detection method based on WAVELET PACKET DECOMPOSITION, it is characterised in that the described method includes following step
Suddenly：
(1) training dataset is built from the normal condition data set by data prediction；
(2) assessed value of data set is obtained using state-detection model, wherein, the assessed value is using training dataset as shape
The input of state monitoring model, the hypothesis calculated using status monitoring model are exported；
(3) residual error is calculated, and WAVELET PACKET DECOMPOSITION is carried out to it, wherein, the residual computations are to ask hypothesis output and real-time status
Difference between data；
(4) each band energy infomation detection failure after analysis is decomposed；
(5) test data set, the input using test data set as Fault Model, according to each after WAVELET PACKET DECOMPOSITION are chosen
Feature of the energy proportion of frequency range as failure judgement state, wherein, the WAVELET PACKET DECOMPOSITION is to utilize sliding window by residual error
Data are divided into after subpattern one by one and carry out WAVELET PACKET DECOMPOSITION using this subpattern as the input of Fault Model, point
It is that the signal after WAVELET PACKET DECOMPOSITION is analyzed to analyse frequency range, and energy of the tracer signal in each frequency range accounts for whole signal energy
Proportion threshold value bound.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510970441.9A CN105389475B (en) | 2015-12-22 | 2015-12-22 | A kind of electric power factory equipment fault detection method based on WAVELET PACKET DECOMPOSITION |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510970441.9A CN105389475B (en) | 2015-12-22 | 2015-12-22 | A kind of electric power factory equipment fault detection method based on WAVELET PACKET DECOMPOSITION |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105389475A CN105389475A (en) | 2016-03-09 |
| CN105389475B true CN105389475B (en) | 2018-04-20 |
Family
ID=55421756
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510970441.9A Expired - Fee Related CN105389475B (en) | 2015-12-22 | 2015-12-22 | A kind of electric power factory equipment fault detection method based on WAVELET PACKET DECOMPOSITION |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105389475B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110362048A (en) * | 2019-07-12 | 2019-10-22 | 上海交通大学 | Blower critical component state monitoring method and device, storage medium and terminal |
| CN113591897A (en) * | 2021-05-28 | 2021-11-02 | 济南浪潮数据技术有限公司 | Method, device and equipment for detecting monitoring data abnormity and readable medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101299004A (en) * | 2008-06-24 | 2008-11-05 | 华南理工大学 | Vibrating failure diagnosis method based on determined learning theory |
| CN102680017A (en) * | 2012-05-31 | 2012-09-19 | 潍柴动力股份有限公司 | Fault diagnostic method and diagnostic device of sensor |
| CN104503235A (en) * | 2014-12-09 | 2015-04-08 | 中国石油大学(华东) | Condition monitoring method based on improved BP neural network for power plant equipment |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE112005002292A5 (en) * | 2004-07-19 | 2007-07-12 | Prüftechnik Dieter Busch AG | Device and method for detecting defects on objects or for locating metallic objects |
| FR2999722B1 (en) * | 2012-12-19 | 2022-01-07 | Electricite De France | LOCATION OF ONE OR MORE FAULTS IN AN ELECTROCHEMICAL ASSEMBLY. |
-
2015
- 2015-12-22 CN CN201510970441.9A patent/CN105389475B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101299004A (en) * | 2008-06-24 | 2008-11-05 | 华南理工大学 | Vibrating failure diagnosis method based on determined learning theory |
| CN102680017A (en) * | 2012-05-31 | 2012-09-19 | 潍柴动力股份有限公司 | Fault diagnostic method and diagnostic device of sensor |
| CN104503235A (en) * | 2014-12-09 | 2015-04-08 | 中国石油大学(华东) | Condition monitoring method based on improved BP neural network for power plant equipment |
Non-Patent Citations (1)
| Title |
|---|
| 《小波分析在虚拟检测系统中的应用研究》;刘建成 等;《仪器仪表学报》;20031031;第24卷(第5期);摘要，说明书第2页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105389475A (en) | 2016-03-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Song et al. | Wind turbine health state monitoring based on a Bayesian data-driven approach | |
| Coble et al. | Applying the general path model to estimation of remaining useful life | |
| CN112202736A (en) | Industrial control system communication network abnormity classification method based on statistical learning and deep learning | |
| CN109186813A (en) | A kind of temperature sensor self-checking unit and method | |
| CN105846780A (en) | Decision tree model-based photovoltaic assembly fault diagnosis method | |
| Hu et al. | Framework for a smart data analytics platform towards process monitoring and alarm management | |
| CN112799898B (en) | Interconnection system fault node positioning method and system based on distributed fault detection | |
| CN109613428A (en) | It is a kind of can be as system and its application in motor device fault detection method | |
| CN103776654A (en) | Method for diagnosing faults of multi-sensor information fusion | |
| CN102880170A (en) | System failure early warning method based on baseline model and Bayesian factor | |
| CN106250709A (en) | Gas turbine abnormality detection based on sensors association network and fault diagnosis algorithm | |
| US12039045B2 (en) | Event analysis in an electric power system | |
| CN112632845B (en) | Data-based mini-reactor online fault diagnosis method, medium and equipment | |
| CN104317778A (en) | Massive monitoring data based substation equipment fault diagnosis method | |
| CN108052954A (en) | The method for diagnosing faults of sample space based on multistage high dimensional feature | |
| CN106354125A (en) | Method for utilizing block PCA (Principal Component Analysis) to detect fault of chemical process | |
| CN103033309A (en) | Pressure transmitter with diagnostics | |
| CN103529337B (en) | The recognition methods of nonlinear correlation relation between equipment failure and electric quantity information | |
| CN105389475B (en) | A kind of electric power factory equipment fault detection method based on WAVELET PACKET DECOMPOSITION | |
| Carratù et al. | A novel methodology for unsupervised anomaly detection in industrial electrical systems | |
| CN117560300B (en) | Intelligent internet of things flow prediction and optimization system | |
| KR20230102431A (en) | Oil gas plant equipment failure prediction and diagnosis system based on artificial intelligence | |
| CN206833239U (en) | A kind of thermal power plant's control system fault detection system based on data-driven | |
| CN111306051B (en) | Probe type state monitoring and early warning method, device and system for oil transfer pump unit | |
| CN105741184A (en) | Transformer state evaluation method and apparatus |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180420 Termination date: 20181222 |