CN107526784A - A kind of method for diagnosing faults based on matrix fill-in - Google Patents
A kind of method for diagnosing faults based on matrix fill-in Download PDFInfo
- Publication number
- CN107526784A CN107526784A CN201710623702.9A CN201710623702A CN107526784A CN 107526784 A CN107526784 A CN 107526784A CN 201710623702 A CN201710623702 A CN 201710623702A CN 107526784 A CN107526784 A CN 107526784A
- Authority
- CN
- China
- Prior art keywords
- data
- msub
- matrix
- real
- mrow
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2462—Approximate or statistical queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Probability & Statistics with Applications (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of method for diagnosing faults based on matrix fill-in, this method comprises the following steps：(1) real-time status data to be diagnosed is obtained；(2) the N group historical data related to real-time status data is extracted from database, every group of historical data includes historical state data and corresponding malfunction；(3) real-time status data and N groups historical data are formed into fault diagnosis matrix, parameter to be filled is the malfunction corresponding to real-time status data in the fault diagnosis matrix；(4) matrix fill-in is carried out to fault diagnosis matrix and obtains the numerical value of parameter to be filled, and then determine malfunction.Compared with prior art, fault diagnosis order of accuarcy of the present invention is high.
Description
Technical field
The present invention relates to a kind of method for diagnosing faults, more particularly, to a kind of method for diagnosing faults based on matrix fill-in.
Background technology
Occur from artificial intelligence technology, technique is just rapidly promoted, applies to include industry, finance, business etc.
Numerous areas.Simultaneously because the lifting of computer computation ability, using historical data as training set, passes through suitable machine learning
Algorithm, obtaining the higher and higher prediction result of the degree of accuracy becomes reality.Therefore the mankind have witnessed computer in 1997 in Christian era
Deep Blue defeats world champion chess player cassie Paro husband, and has witnessed AlphaGo in weiqi play chess in 2017 Christian eras then
Defeat Lee's world champion generation stone.But so far, it using fuzzy reasoning is main that the modes of most of fault diagnosises, which is still,
Means, lack effective utilization to historical state data.For to a certain extent, a kind of wave to data resource is caused
Take.And the determination of the setting of fuzzy rule, especially parameter can be different because of the equipment of different batches, or even same batch
Equipment need parameter also have certain difference.This just needs artificial continuous adjusting parameter, and certain degree be present
Subjectivity.
The content of the invention
It is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide one kind is based on matrix fill-in
Method for diagnosing faults.
The purpose of the present invention can be achieved through the following technical solutions：
A kind of method for diagnosing faults based on matrix fill-in, this method comprise the following steps：
(1) real-time status data to be diagnosed is obtained；
(2) the N group historical data related to real-time status data is extracted from database, every group of historical data includes going through
History status data and corresponding malfunction；
(3) real-time status data and N groups historical data are formed into fault diagnosis matrix, it is to be filled in the fault diagnosis matrix
Parameter is the malfunction corresponding to real-time status data；
(4) matrix fill-in is carried out to fault diagnosis matrix and obtains the numerical value of parameter to be filled, and then determine malfunction.
N groups historical data is to be more than the data of setting value with real-time status data similarity in step (2).
Step (2) is specially：The history shape of all historical datas in database is calculated using cosine similarity measure
Historical data, is ranked up by state data and the similarity for implementing status data according to similarity is descending, and then N before selection
Group historical data.
Described similarity is obtained by equation below：
Wherein, aiFor the vector of real-time status data composition, ajFor the history in one group of historical data of similarity to be asked for
The vector of status data composition, sim (ai,aj) it is aiWith ajSimilarity.
Step (3) forms fault diagnosis matrix：
(31) malfunction in every group of historical data is represented to form malfunction numerical value using digital form, and then will
Historical state data and malfunction numerical value are arranged in order history of forming data vector；
(32) clooating sequence of each data in historical data vector is used to be ranked up real-time status data, by failure shape
State numerical value corresponding position is set to parameter to be filled, forms real time data vector；
(33) all historical data Vector Groups of real time data vector sum are shaped as fault diagnosis matrix.
Matrix fill-in is carried out to fault diagnosis matrix using singular value thresholding algorithm in step (4).
Compared with prior art, the invention has the advantages that：
(1) present invention from database selection with the bigger data of real-time status data similarity as training data, with
Mitigate the operand of matrix fill-in algorithm, improve operation efficiency, select the bigger data of similarity can also as training data
Incoherent data interference is excluded, improves prediction accuracy.
(2) present invention realizes matrix fill-in using singular value thresholding algorithm, so as to using historical data as reference sample, realize
Fault diagnosis, the artificial subjectivity for formulating fuzzy parameter is avoided, meets the otherness between distinct device.
Brief description of the drawings
Fig. 1 is the FB(flow block) of the method for diagnosing faults of the invention based on matrix fill-in.
Embodiment
The present invention is described in detail with specific embodiment below in conjunction with the accompanying drawings.
Embodiment
As shown in figure 1, a kind of method for diagnosing faults based on matrix fill-in, this method comprise the following steps：
(1) real-time status data to be diagnosed is obtained；
(2) the N group historical data related to real-time status data is extracted from database, every group of historical data includes going through
History status data and corresponding malfunction；
(3) real-time status data and N groups historical data are formed into fault diagnosis matrix, it is to be filled in the fault diagnosis matrix
Parameter is the malfunction corresponding to real-time status data；
(4) matrix fill-in is carried out to fault diagnosis matrix and obtains the numerical value of parameter to be filled, and then determine malfunction.
Real-time status data may be from different platforms in step (1), therefore by based on DBCOM and change notification machine
The power control interface of system realizes the collection of real-time status data, is specifically realized in the ForceControl configuration software based on COM technologies
The change notification mechanism of data object.COM provides a kind of effective data exchange mechanism, referred to as uniform data between application
Transmission mechanism (UDT, Uniform Data Transfer).It is using data object as data entity, and data object then passes through
IDataObject interfaces expose its internal information.As long as CLIENT PROGRAM can be obtained by data object by certain host-host protocol
IDataObject interfaces, later CLIENT PROGRAM can directly accesses data object.
COM provides tie point mechanism to realize both-way communication.In UDT, CLIENT PROGRAM need to only realize IAdviseSink
Interface, wherein OnDataChange member are used for the notification procedure of data object.IDataObject::DAdvise functions are established
Circular connection between CLIENT PROGRAM receiver object and data object, once connection is set up, when data change,
IAdviseSink will be called::OnDataChange functions.
It is concretely comprised the following steps：The configuration engineering created by ForceControl configuration software is first turned on, the newly-built note in database
Volume point temp, as with the data communication point in C++.DbCommOcxEf controls are exactly packaged com component,
Before operation program, we will ensure the registered successes of DbCom.ocx, it is ensured that we can be implemented control, and it is with regard to suitable
It is loaded into a .dll file in process.CDbCommOcxEf object class that same we create also has its 128
CLSID is identified.
CDbCommOcxEf object class：
Static CLSID const clsid=0x3310fa25,0xa027,0x47b3,0x8c, 0x49,0x10,
0x91,0x07,0x73,0x17,0xe9}}
Because CLIENT PROGRAM is only it is to be understood that the IDataObject interface pointer cans of data object directly access data pair
As.Sometimes CLIENT PROGRAM wants to learn the information that track data object changes in time, to obtain latest data.
Therefore we obtain the information of our desired registration points by it, these by stating an object
Information is stored in inside circular connection receiver object by IDataObject interface pointers, realizes data communication.When ours
When registration point information changes, message map mechanism goes to find corresponding event control function, event control function here
It is exactly our IAdviseSink：：OnDataChange, by it, we just have updated letter of the registration point in C++ engineerings
Breath.So we are achieved that the change notification mechanism of our data objects.
N groups historical data is to be more than the data of setting value with real-time status data similarity in step (2).
Step (2) is specially：The history shape of all historical datas in database is calculated using cosine similarity measure
Historical data, is ranked up by state data and the similarity for implementing status data according to similarity is descending, and then N before selection
Historical data is organized, sort method uses bubble sort here.
Described similarity is obtained by equation below：
Wherein, aiFor the vector of real-time status data composition, ajFor the history in one group of historical data of similarity to be asked for
The vector of status data composition, sim (ai,aj) it is aiWith ajSimilarity.
Step (3) forms fault diagnosis matrix：
(31) malfunction in every group of historical data is represented to form malfunction numerical value using digital form, and then will
Historical state data and malfunction numerical value are arranged in order history of forming data vector；
(32) clooating sequence of each data in historical data vector is used to be ranked up real-time status data, by failure shape
State numerical value corresponding position is set to parameter to be filled, forms real time data vector；
(33) all historical data Vector Groups of real time data vector sum are shaped as fault diagnosis matrix.
Matrix fill-in is carried out to fault diagnosis matrix using singular value thresholding algorithm (SVT algorithms) in step (4).
Matrix fill-in (Matrix Completion) purpose is part (the not observable in order to be lacked in estimated matrix
Part), can be regarded as using matrix X approximate matrix M, then by the use of the element in X as matrix M in can not observation portion element
Estimation.After existing training data forms matrix, these known information need to meet to assume that data matrix is low-rank,
It can be inferred that unknown part.Due to having carried out similar Sexual behavior mode to training data in 2), so the hypothesis of low-rank is to set up
's.It is formulated then as follows：
Object function：minrank(X)；
Constraints：Xi,j=Mi,j,(x,j)∈Ω；
If setting：Y=PΩ(X),
Wherein：
Then above-mentioned film table function and constraints are：Object function：minrank(X)；
Constraints：PΩ(X)=PΩ(M)；
M is fault diagnosis matrix, and X is the fault diagnosis matrix after filling.
Claims (6)
1. a kind of method for diagnosing faults based on matrix fill-in, it is characterised in that this method comprises the following steps：
(1) real-time status data to be diagnosed is obtained；
(2) the N group historical data related to real-time status data is extracted from database, every group of historical data includes history shape
State data and corresponding malfunction；
(3) real-time status data and N groups historical data are formed into fault diagnosis matrix, parameter to be filled in the fault diagnosis matrix
For the malfunction corresponding to real-time status data；
(4) matrix fill-in is carried out to fault diagnosis matrix and obtains the numerical value of parameter to be filled, and then determine malfunction.
A kind of 2. method for diagnosing faults based on matrix fill-in according to claim 1, it is characterised in that N in step (2)
Group historical data is to be more than the data of setting value with real-time status data similarity.
3. a kind of method for diagnosing faults based on matrix fill-in according to claim 2, it is characterised in that step (2) has
Body is：The historical state data of all historical datas and implementation status number in database are calculated using cosine similarity measure
According to similarity, historical data is ranked up according to similarity is descending, so choose before N group historical datas.
4. a kind of method for diagnosing faults based on matrix fill-in according to claim 3, it is characterised in that described is similar
Degree is obtained by equation below：
<mrow>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>a</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>a</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mi>c</mi>
<mi>o</mi>
<mi>s</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>a</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>a</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>a</mi>
<mi>i</mi>
</msub>
<mo>&CenterDot;</mo>
<msub>
<mi>a</mi>
<mi>j</mi>
</msub>
</mrow>
<mrow>
<mo>|</mo>
<msub>
<mi>a</mi>
<mi>i</mi>
</msub>
<mo>|</mo>
<mo>&times;</mo>
<mo>|</mo>
<msub>
<mi>a</mi>
<mi>j</mi>
</msub>
<mo>|</mo>
</mrow>
</mfrac>
<mo>,</mo>
</mrow>
Wherein, aiFor the vector of real-time status data composition, ajFor the historic state in one group of historical data of similarity to be asked for
The vector of data composition, sim (ai,aj) it is aiWith ajSimilarity.
A kind of 5. method for diagnosing faults based on matrix fill-in according to claim 1, it is characterised in that step (3) group
It is specially into fault diagnosis matrix：
(31) malfunction in every group of historical data is represented to be formed malfunction numerical value using digital form, and then by history
Status data and malfunction numerical value are arranged in order history of forming data vector；
(32) clooating sequence of each data in historical data vector is used to be ranked up real-time status data, by malfunction number
Value corresponding position is set to parameter to be filled, forms real time data vector；
(33) all historical data Vector Groups of real time data vector sum are shaped as fault diagnosis matrix.
6. a kind of method for diagnosing faults based on matrix fill-in according to claim 1, it is characterised in that in step (4)
Matrix fill-in is carried out to fault diagnosis matrix using singular value thresholding algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710623702.9A CN107526784A (en) | 2017-07-27 | 2017-07-27 | A kind of method for diagnosing faults based on matrix fill-in |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710623702.9A CN107526784A (en) | 2017-07-27 | 2017-07-27 | A kind of method for diagnosing faults based on matrix fill-in |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107526784A true CN107526784A (en) | 2017-12-29 |
Family
ID=60680199
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710623702.9A Pending CN107526784A (en) | 2017-07-27 | 2017-07-27 | A kind of method for diagnosing faults based on matrix fill-in |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107526784A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111896246A (en) * | 2020-07-29 | 2020-11-06 | 北京天地龙跃科技有限公司 | Health management verifies evaluation system |
| CN115576733A (en) * | 2022-11-17 | 2023-01-06 | 广州信诚信息科技有限公司 | Intelligent equipment fault diagnosis system based on deep reinforcement learning |
| WO2023197461A1 (en) * | 2022-04-11 | 2023-10-19 | 西安热工研究院有限公司 | Gearbox fault early warning method and system based on working condition similarity evaluation |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104481508A (en) * | 2014-09-10 | 2015-04-01 | 大庆明达韦尔信息系统服务有限公司 | Oilfield rod-pumping well fault diagnosis method by combining comentropy and gray level incidence matrix |
| CN104504607A (en) * | 2014-09-04 | 2015-04-08 | 国家电网公司 | Method for diagnosing photovoltaic power station faults on the basis of fuzzy clustering algorithm |
| CN105045256A (en) * | 2015-07-08 | 2015-11-11 | 北京泰乐德信息技术有限公司 | Rail traffic real-time fault diagnosis method and system based on data comparative analysis |
| CN105843212A (en) * | 2016-03-29 | 2016-08-10 | 东北大学 | System and method for fault diagnosis of blast furnace |
| CN106503652A (en) * | 2016-10-21 | 2017-03-15 | 南京理工大学 | Based on the accident detection method that low-rank adaptive sparse is rebuild |
| KR20170038348A (en) * | 2015-09-30 | 2017-04-07 | 한국전력공사 | Fault diagnosis system and fault diagnosis method in dynamic equipment |
-
2017
- 2017-07-27 CN CN201710623702.9A patent/CN107526784A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104504607A (en) * | 2014-09-04 | 2015-04-08 | 国家电网公司 | Method for diagnosing photovoltaic power station faults on the basis of fuzzy clustering algorithm |
| CN104481508A (en) * | 2014-09-10 | 2015-04-01 | 大庆明达韦尔信息系统服务有限公司 | Oilfield rod-pumping well fault diagnosis method by combining comentropy and gray level incidence matrix |
| CN105045256A (en) * | 2015-07-08 | 2015-11-11 | 北京泰乐德信息技术有限公司 | Rail traffic real-time fault diagnosis method and system based on data comparative analysis |
| KR20170038348A (en) * | 2015-09-30 | 2017-04-07 | 한국전력공사 | Fault diagnosis system and fault diagnosis method in dynamic equipment |
| CN105843212A (en) * | 2016-03-29 | 2016-08-10 | 东北大学 | System and method for fault diagnosis of blast furnace |
| CN106503652A (en) * | 2016-10-21 | 2017-03-15 | 南京理工大学 | Based on the accident detection method that low-rank adaptive sparse is rebuild |
Non-Patent Citations (1)
| Title |
|---|
| 马晓慧: "《矩阵填充理论方法分析》", 《中国学位论文全文数据库》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111896246A (en) * | 2020-07-29 | 2020-11-06 | 北京天地龙跃科技有限公司 | Health management verifies evaluation system |
| WO2023197461A1 (en) * | 2022-04-11 | 2023-10-19 | 西安热工研究院有限公司 | Gearbox fault early warning method and system based on working condition similarity evaluation |
| CN115576733A (en) * | 2022-11-17 | 2023-01-06 | 广州信诚信息科技有限公司 | Intelligent equipment fault diagnosis system based on deep reinforcement learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ware et al. | Conflict reduction in map generalization using iterative improvement | |
| CN107526784A (en) | A kind of method for diagnosing faults based on matrix fill-in | |
| CN110378488A (en) | Federal training method, device, training terminal and the storage medium of client variation | |
| Carfì et al. | An algorithm for dynamical games with fractal-like trajectories | |
| CN111046018A (en) | Multi-objective optimization-based power generation equipment operation condition library construction method and device | |
| CN109144988A (en) | A kind of detection method and device of abnormal data | |
| CN108496188A (en) | Method, apparatus, computer system and the movable equipment of neural metwork training | |
| CN110489307A (en) | Interface exception call monitoring method and device | |
| CN109388458A (en) | Management method, terminal device and the computer readable storage medium of interface control | |
| CN106121622A (en) | A kind of Multiple faults diagnosis approach of Dlagnosis of Sucker Rod Pumping Well based on indicator card | |
| Wang et al. | An iterative updating heuristic search strategy for groundwater contamination source identification based on an ACPSO–ELM surrogate system | |
| US11436263B2 (en) | Geocoding methods and systems of correcting latitude and longitude of a point of interest | |
| CN112148566A (en) | Monitoring method and device of computing engine, electronic equipment and storage medium | |
| CN112556698A (en) | Ship behavior change and abnormality recognition method, equipment and medium | |
| CN111581586A (en) | Lake and reservoir water quality anisotropic interpolation method and device based on registration model | |
| CN116608859A (en) | Navigation method, storage medium and device of self-adaptive unscented Kalman filtering based on threshold processing | |
| Hu et al. | An adaptive hybrid PSO multi-objective optimization algorithm for constrained optimization problems | |
| CN111107162B (en) | Indoor positioning data processing method, device and system based on Internet of things | |
| Wu et al. | A Vehicle Map-matching Algorithm based on Measure Fuzzy Sorting. | |
| CN114240654A (en) | Performance capability perspective method, system, equipment and readable storage medium based on transaction link | |
| CN106447711A (en) | Multiscale basic geometrical shape feature extraction method | |
| CN108510778A (en) | A kind of implementation method of the vehicle polymerization display based on recorder management platform | |
| Ge et al. | A new improved CMAC neural network | |
| Jiang et al. | Prediction on the remaining useful life of rolling bearings using ensemble DLSTM | |
| CN112862536B (en) | Data processing method, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171229 |