CN117454232A - Production network construction fault diagnosis, prediction and health management system and method - Google Patents
Production network construction fault diagnosis, prediction and health management system and method Download PDFInfo
- Publication number
- CN117454232A CN117454232A CN202311773490.4A CN202311773490A CN117454232A CN 117454232 A CN117454232 A CN 117454232A CN 202311773490 A CN202311773490 A CN 202311773490A CN 117454232 A CN117454232 A CN 117454232A
- Authority
- CN
- China
- Prior art keywords
- data
- model
- algorithm
- fault
- prediction
- 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
- 238000003745 diagnosis Methods 0.000 title claims abstract description 50
- 230000036541 health Effects 0.000 title claims abstract description 44
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 26
- 238000010276 construction Methods 0.000 title claims abstract description 21
- 238000000034 method Methods 0.000 title claims abstract description 10
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 80
- 238000012549 training Methods 0.000 claims abstract description 28
- 238000000605 extraction Methods 0.000 claims abstract description 4
- 238000005457 optimization Methods 0.000 claims abstract description 4
- 230000007246 mechanism Effects 0.000 claims description 29
- 238000007726 management method Methods 0.000 claims description 25
- 238000012544 monitoring process Methods 0.000 claims description 24
- 238000001514 detection method Methods 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 10
- 238000012423 maintenance Methods 0.000 claims description 10
- 230000005856 abnormality Effects 0.000 claims description 8
- 238000013210 evaluation model Methods 0.000 claims description 6
- 238000002485 combustion reaction Methods 0.000 claims description 5
- 238000013135 deep learning Methods 0.000 claims description 5
- 238000011161 development Methods 0.000 claims description 5
- 238000010801 machine learning Methods 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 3
- 238000013145 classification model Methods 0.000 claims description 3
- 230000006835 compression Effects 0.000 claims description 3
- 238000007906 compression Methods 0.000 claims description 3
- 230000007774 longterm Effects 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000002360 preparation method Methods 0.000 claims description 3
- 230000003449 preventive effect Effects 0.000 claims description 3
- 230000000737 periodic effect Effects 0.000 claims description 2
- 238000004393 prognosis Methods 0.000 claims 2
- 238000011156 evaluation Methods 0.000 claims 1
- 238000013527 convolutional neural network Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 239000000446 fuel Substances 0.000 description 3
- 238000007619 statistical method Methods 0.000 description 3
- 238000007635 classification algorithm Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 102100029469 WD repeat and HMG-box DNA-binding protein 1 Human genes 0.000 description 1
- 101710097421 WD repeat and HMG-box DNA-binding protein 1 Proteins 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000009545 invasion Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Classifications
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/18—File system types
- G06F16/182—Distributed file systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/27—Regression, e.g. linear or logistic regression
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- 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
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
-
- 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
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Databases & Information Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Educational Administration (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Agronomy & Crop Science (AREA)
- Animal Husbandry (AREA)
- Marine Sciences & Fisheries (AREA)
- Mining & Mineral Resources (AREA)
Abstract
The invention relates to the technical field of production network construction, in particular to a production network construction fault diagnosis, prediction and health management system and method, and data uploading: the edge computing system uploads the data acquired by the data source system to the data center system; training a model: the uploaded data is subjected to feature extraction through a data center system, the data is analyzed and calculated through an algorithm library in the data center system, a real-time fault prediction model is established through a model library, training is carried out, continuous optimization is carried out, and finally the fault prediction model is issued to a gas turbine compressor; client side: the decision maker can directly access the running condition of the system, and can make value-added decisions to manage the gas turbine compressor if necessary; edge computing system: judging whether the data source system needs to upload the cloud, if a matching model exists in a model library of the data center system, directly analyzing, calculating and storing data, and executing assignment decision through a default set command for emergency of the gas turbine compressor.
Description
Technical Field
The invention relates to the technical field of production network construction, in particular to a production network construction fault diagnosis, prediction and health management system and method.
Background
In the field of equipment maintenance and assurance, fault Prediction and Health Management (PHM) technology focuses on utilizing advanced sensor integration and predicting, diagnosing, monitoring and managing the status of equipment by means of various algorithms and intelligent models. In recent years, the rapid development of fault diagnosis and health management has led to the transition of maintenance and guarantee modes from state monitoring to state management, which is an innovative scheme for testing and maintenance diagnosis and is a comprehensive fault detection, isolation and prediction and health management technology. PHM technology was introduced not to directly eliminate system failures, but to know and predict when a failure might occur; or triggering a series of maintenance activities when an initial failure occurs, thereby realizing autonomous guarantee and reducing the targets of equipment use and guarantee cost.
In recent years, along with the gradual development of modernization and the improvement of economic level in China, the oil and gas production scale is increasingly enlarged, but due to the lack of reasonable safety management and emergency measures, oil and gas disaster accidents are frequently caused. Large construction projects are constructed near gathering pipelines, and invasion to the gathering pipelines exists. Natural disasters such as landslide or earthquakes cause the gathering lines to break or deform with the risk of line leakage. At present, the overhaul of the gas turbine compressor in an oil-gas pipe network is manual inspection, the protection means is behind, disaster accidents cannot be shut off in time, disaster positions cannot be judged in time, and the accident control is lagged.
Disclosure of Invention
The invention provides a production network construction fault diagnosis, prediction and health management system and method.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
the production network construction fault diagnosis, prediction and health management system comprises a data source system, an edge computing system, a data center platform system and a client, wherein the data source system is an input end of the whole system, the data source system uploads collected data to the edge computing system, the edge computing system is used for abnormality detection and fault diagnosis, and the edge computing system performs data stream compression processing on the received data to remove redundant data; the data center system comprises a data layer, an interface layer, an algorithm library, a model library and service program data, wherein the data layer, the interface layer, the algorithm library, the model library and the service program data are uploaded to the data center system through an edge computing system, and then are collected and processed and stored or released into streaming data; unstructured data comprises voice, images and text, and is stored and managed by an HDFS file system database; the distributed message queue issues monitoring data to be processed in real time, so that real-time state monitoring and fault diagnosis are realized, and the consumption of the diagnosis processing stream program is facilitated; the interface layer is an external interface of the whole system and comprises an interface with a database and an interface with a computing platform; the data read-write operation of the diagnostic software is realized by connecting the data with a database through a standardized and configurable interface program; by means of job submission or API call, the calculation force of the data center system is utilized to realize model training and efficient calculation of real-time diagnosis service; the algorithm library is a customized algorithm set based on mechanism and data driven mixture, and comprises a mechanism algorithm, a data driven algorithm and a rule and knowledge based algorithm.
Further, the mechanism algorithm comprises a fuel engine thermodynamic failure mechanism algorithm, a motor electromagnetic failure mechanism algorithm, a motor electric failure mechanism algorithm and a rotating machinery failure mechanism algorithm; the data driving algorithm comprises a signal processing algorithm, a data preprocessing algorithm, a statistics and machine learning mechanism algorithm and a deep learning algorithm; the algorithm based on the rules and the knowledge comprises an algorithm based on the rules and the knowledge and an inference algorithm based on a fault diagnosis tree.
Further, the statistical and machine learning mechanism algorithm comprises a probability statistical analysis, a clustering algorithm, a classification algorithm and a regression algorithm.
The deep learning algorithm comprises a convolutional neural network CNN and time sequence regression.
Furthermore, the model library is used for various model summaries of unit fault diagnosis and health management, and the model library is used for converting data acquired by the data source system into diagnosis and prediction results.
Further, corresponding models are built aiming at a core device gas turbine, a motor, a body and an accessory system of the unit, and the model mainly comprises a health baseline model, an anomaly detection model, a fault identification classification model and a performance prediction regression model under multiple working conditions. Aiming at a unit system, a layered health evaluation model is constructed, wherein the layered health evaluation model comprises a multi-dimensional health index calculation model, a one-dimensional health index and a discrete health index model. The comprehensive fault reasoning mainly realizes the fusion reasoning judgment of multi-source data, multi-model and multi-equipment diagnosis results, and improves the diagnosis reliability.
Furthermore, the service program is an application layer for realizing unit fault diagnosis and health management, and mainly comprises a model training and testing program, an abnormality detection and diagnosis program, a performance index prediction program and a closed-loop auxiliary decision-making program.
Furthermore, the model training and testing program is mainly offline training, and model training is carried out by adopting a single machine or a cluster according to the data volume, and the cluster is carried out by a distributed model training framework. Considering that the system is in an initial stage of use, the fault samples of the real scene are fewer, and the online training program can be run in a trial mode to perform data stream processing and model training.
Furthermore, the abnormality detection and diagnosis program is realized through a distributed stream processing program, and a large number of data streams of the unit are calculated based on an intelligent model and knowledge rules, so that the state and fault condition of the unit are rapidly identified.
Further, performance index prediction mainly aims at performing periodic long-term tracking and predictive analysis on performance indexes of a combustion engine, a motor and a body, and based on safety threshold setting, preventive maintenance development time is output.
Further, the closed-loop auxiliary decision-making program is mainly based on diagnosis and prediction results, and is used for rapidly linking production pre-operation preparation, post-judgment, fault handling measures, station monitoring instructions and maintenance strategies, so that the station monitoring personnel can be conveniently and remotely instructed to carry out targeted operations.
The production network construction fault diagnosis, prediction and health management method comprises the following steps:
uploading data: the edge computing system uploads the data acquired by the data source system to the data center system;
training a model: the uploaded data is subjected to feature extraction through a data center system, the data is analyzed and calculated through an algorithm library in the data center system, a real-time fault prediction model is established through a model library, training is carried out, continuous optimization is carried out, and finally the fault prediction model is issued to a gas turbine compressor;
client side: the decision maker can directly access the running condition of the system, and can make value-added decisions to manage the gas turbine compressor if necessary;
edge computing system: judging whether the data source system needs to upload the cloud, if a matching model exists in a model library of the data center system, directly analyzing, calculating and storing data, and executing assignment decision through a default set command for emergency of the gas turbine compressor.
The beneficial effects are that: the invention covers all fault diagnosis and prediction services under the technical condition of the current equipment by developing the distributed production network construction fault diagnosis, prediction and health management system and the core module thereof. Firstly, deploying an edge computing system based on a station server, and running a safety key fault detection program; secondly, a high-availability, high-stability, low-delay and telescopic monitoring system is built based on the data center; then developing a system algorithm and model with state monitoring data as a drive, and simultaneously fusing expert experience and a mechanism and knowledge model of equipment, so that the algorithm and the module gradually tend to be mature through continuous actual measurement and iteration; and finally, embedding a mature model and algorithm into the big data distributed stream processing program to perform real-time monitoring, fault diagnosis and early warning on the compressor state monitoring data stream. The distributed production network construction fault diagnosis, prediction and health management system for the oil-gas pipe network equipment is formed by taking a service program, an algorithm library and a model library as cores and through a station server and other components including a message queue, a database and a configuration display interface which are configured in a data center.
Drawings
FIG. 1 is a schematic flow chart of the present invention;
FIG. 2 is a schematic diagram of the structure of the present invention;
FIG. 3 is a schematic diagram of the algorithm library according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments.
In the description of the present invention, it should be understood that the terms "upper," "lower," "front," "rear," "left," "right," "top," "bottom," "inner," "outer," and the like are used in an orientation or positional relationship based on that shown in the drawings, merely to facilitate description of the invention and to simplify the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus should not be construed as limiting the invention.
Referring to fig. 1-3, the production network construction fault diagnosis, prediction and health management system comprises a data source system, an edge computing system, a data center platform system and a client, wherein the data source system is an input end of the whole system, the data source system uploads collected data to the edge computing system, the edge computing system is used for abnormality detection and fault diagnosis, and the edge computing system performs data stream compression processing on the received data to remove redundant data: the data center system comprises a data layer, an interface layer, an algorithm library, a model library and service program data, wherein the data layer, the interface layer, the algorithm library, the model library and the service program data are uploaded to the data center system through an edge computing system, and then are collected and processed and stored or released into streaming data; unstructured data comprises voice, images and text, and is stored and managed by an HDFS file system database; the distributed message queue issues monitoring data to be processed in real time, so that real-time state monitoring and fault diagnosis are realized, and the consumption of the diagnosis processing stream program is facilitated; the interface layer is an external interface of the whole system and comprises an interface with a database and an interface with a computing platform; the data read-write operation of the diagnostic software is realized by connecting the data with a database through a standardized and configurable interface program; by means of job submission or API call, the calculation force of the data center system is utilized to realize model training and efficient calculation of real-time diagnosis service.
The algorithm library is a customized algorithm set based on mechanism and data driven mixture, and comprises a mechanism algorithm, a data driven algorithm and a rule and knowledge based algorithm. The mechanism algorithm comprises a fuel engine thermodynamic failure mechanism algorithm, a motor electromagnetic failure mechanism algorithm, a motor electric failure mechanism algorithm and a rotating machinery failure mechanism algorithm; the data driving algorithm comprises a signal processing algorithm, a data preprocessing algorithm, a statistics and machine learning mechanism algorithm and a deep learning algorithm; the algorithm based on the rules and the knowledge comprises an algorithm based on the rules and the knowledge and an inference algorithm based on a fault diagnosis tree. The statistical and machine learning mechanism algorithm comprises probability statistical analysis, a clustering algorithm, a classification algorithm and a regression algorithm. The deep learning algorithm comprises a convolutional neural network CNN and time sequence regression.
An electrical failure mechanism algorithm of a motor, comprising the following steps:
step 1, determining m fault modes and n fault characteristics of a motor; determining a fault mode and fault characteristics of the motor according to a motor fault mechanism; for example, the total number of failure modes shown in table 1 is m=23; the total number of fault signatures shown in table 2 is n=8;
TABLE 1
Step 2, establishing a relation matrix with sufficient conditions between fault characteristics and fault modes, and firstly giving the following initial values: CF (i, j) =0, i=1 to m, j=1 to n;
traversing i=1 to m, j=1 to n: when meeting the j-th fault signature necessarily results in the i-th fault pattern occurring, then:
CF(i，j)＝1；
for example: since the 2 nd failure feature "the unbalance degree of the voltage is too high" necessarily causes the 14 th failure mode "the three-phase unbalance of the voltage", there is CF (14, 2) =1.
TABLE 2
Step 3, establishing a relation matrix with necessary conditions between the fault characteristics and the fault modes, and giving the initial values as follows: BY (i, j) =0, i=1 to m, j=1 to n;
traversing i=1 to m, j=1 to n: when meeting the ith failure mode occurrence necessarily results in the occurrence of the jth failure feature, then:
BY(i，j)＝1；
for example:
because the 2 nd failure mode "lug loose" necessarily causes the 5 th failure feature "unbalance degree of current is too high", the 7 th failure feature "vibration effective value is too large", BY (2, 5) =1; BY (2, 7) =1
The relationship of all the necessary conditions is integrated, and finally the BY (i, j) array is formed as follows:
TABLE 3 Table 3
Step 4, obtaining signals through monitoring of the motor, comparing the signals with a preset fault characteristic requirement and a fault characteristic parameter to obtain a fault characteristic membership array TZ (j), wherein j=1-n, TZ (j) is a real number between 0 and 1 or a real number between-1, a real number between 0 and 1 represents membership of a j fault characteristic, and-1 represents that the fault characteristic is not monitored;
step 5, calculating to obtain a sufficient condition relation matrix CF1 (i, j) under all known fault characteristic conditions:
traversing i=1 to m, j=1 to n: CF1 (i, j) =0 if TZ (j) = -1, otherwise CF1 (i, j) =cf (i, j). TZ (j);
step 6, calculating to obtain a necessary condition relation matrix BY1 (i, j) under known fault characteristic conditions:
traversing i=1 to m, j=1 to n: if TZ (j) = -1, then BY1 (i, j) = 1; otherwise: if BY (i, j) =1, then BY1 (i, j) =tz (j); if BY (i, j) =0, BY1 (i, j) =1-TZ (j);
step 7, the membership degree array of each fault mode, which is formed by integrating sufficient conditions and does not generate the fault mode, is MS1 (i):
step 8, the membership degree array of each fault mode, which is formed by integrating the necessary conditions and does not generate the fault mode, is MS2 (i):
step 9, the integrated membership array of the fault characteristics corresponding to each fault mode is MS (i): MS (i) =1-MS 1 (i). MS2 (i), i=1-m;
and step 10, taking the fault mode number as an abscissa and the fault mode membership degree as an ordinate, and making a histogram of the MS (i) array.
In step 4, the method for calculating the fault feature membership array TZ (j) includes the following steps:
step 4.1, the time sequence of the motor monitoring parameters corresponding to the jth fault characteristics is as follows: d (j, k), k=1 to k1, k1=p.k2, p is a positive integer, and the time interval of the monitored parameter samples is t1;
forming a group of data according to each p points continuously for statistical analysis to obtain an average value sequence Dav (j, r) and a standard deviation sequence s (j, r) of each group, wherein r=1-k 2;
step 4.2, when the motor normally operates, acquiring a time sequence D (j, k) of each monitoring parameter corresponding to the motor fault characteristic in the monitoring time of k1.t1 according to the method of step 4.1, and calculating to obtain Dav (j, r) and s (j, r), wherein k=1-k 1 and r=1-k 2;
finding a minimum value Dav (j, r 1) and a maximum value Dav (j, r 2) which satisfy |s (j, r)/Dav (j, r) | < e in a Dav (j, r) sequence, wherein e is a given positive number, recording marks r1 and r2, and obtaining s (j, r 1) and s (j, r 2);
let d1=dav (j, r 1), d2=dav (j, r 2), s1=s (j, r 1), s2=s (j, r 2);
step 4.3, when the motor is in normal or fault operation and is monitored, a sequence D (j, k) of p current nearest time points of each monitoring parameter corresponding to the fault characteristic of the motor is obtained according to the method of step 4.1, k=1-p, dav (j, 1) and s (j, 1) are calculated, and the following steps are carried out:
D＝Dav(j，1)
the membership degree of the current monitoring parameter to the normal or fault state is obtained according to the following formula:
the fault feature membership degree array TZ (j) of the fault feature adopts the following 3 different calculation methods according to different conditions of the fault:
membership higher than normal features:
membership below normal characteristics:
membership to failure feature: TZ (j) =v (j)
Step 4.4. Let TZ (j) = -1 for fault signature not monitored.
The model library is used for various model summaries of unit fault diagnosis and health management, and the model library is used for converting data acquired by the data source system into diagnosis and prediction results. Corresponding models are built aiming at a core device fuel engine, a motor, a body and an accessory system of the unit, and the model mainly comprises a health baseline model, an abnormality detection model, a fault identification classification model and a performance prediction regression model under multiple working conditions. Aiming at a unit system, a layered health evaluation model is constructed, wherein the layered health evaluation model comprises a multi-dimensional health index calculation model, a one-dimensional health index and a discrete health index model. The comprehensive fault reasoning mainly realizes the fusion reasoning judgment of multi-source data, multi-model and multi-equipment diagnosis results, and improves the diagnosis reliability.
The business program is an application layer for realizing unit fault diagnosis and health management, and mainly comprises a model training and testing program, an abnormality detection and diagnosis program, a performance index prediction program and a closed-loop auxiliary decision-making program.
The model training and testing program is mainly offline training, and model training is carried out by adopting a single machine or a cluster according to the data volume condition, and the cluster is carried out by a distributed model training framework. Considering that the system is in an initial stage of use, the fault samples of the real scene are fewer, and the online training program can be run in a trial mode to perform data stream processing and model training. The anomaly detection and diagnosis program is realized through a distributed stream processing program, and a large number of data streams of the unit are calculated based on an intelligent model and knowledge rules, so that the state and fault condition of the unit are rapidly identified.
The performance index prediction mainly aims at the performance indexes of the combustion engine, the motor and the body to carry out periodical long-term tracking and prediction analysis, and based on safety threshold setting, preventive maintenance development time is output.
The closed-loop auxiliary decision-making program is mainly based on diagnosis and prediction results, and is used for rapidly linking production pre-operation preparation, post-judgment, fault handling measures, station monitoring instructions and maintenance strategies, so that the station monitoring personnel can be conveniently and remotely instructed to carry out targeted operations.
A method for diagnosing, predicting and managing production network construction faults and health,
uploading data: the edge computing system uploads the data acquired by the data source system to the data center system;
training a model: the uploaded data is subjected to feature extraction through a data center system, the data is analyzed and calculated through an algorithm library in the data center system, a real-time fault prediction model is established through a model library, training is carried out, continuous optimization is carried out, and finally the fault prediction model is issued to a gas turbine compressor;
client side: the decision maker can directly access the running condition of the system, and can make value-added decisions to manage the gas turbine compressor if necessary;
edge computing system: judging whether the data source system needs to upload the cloud, if a matching model exists in a model library of the data center system, directly analyzing, calculating and storing data, and executing assignment decision through a default set command for emergency of the gas turbine compressor.
The client side realizes the interaction and operation requirements of different users on the system through the B/S architecture, and comprises a 2D/3D configuration monitoring interface program, a mobile terminal display program, a domain expert workstation, a modeling expert workstation and other external system accesses.
The configuration monitoring program is developed, and the configuration monitoring program can be operated at a PC end and a mobile end to display the current state and fault information of the compressor unit in real time. The field expert and modeling expert workstation is mainly convenient for updating and maintaining the algorithm library and the model library after the system is put into operation, and improves the safety of the system and data through safety management and level authorization. The unlicensed model cannot be used for online operation by performing sufficient tests and authentication on the model deployed in the business operation environment.
The foregoing is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art, who is within the scope of the present invention, shall cover the same or different embodiments according to the technical solution and the inventive concept of the present invention.
Claims (10)
1. The production network construction fault diagnosis, prediction and health management system is characterized in that: the system comprises a data source system, an edge computing system, a data center platform system and a client, wherein the data source system is an input end of the whole system, the data source system uploads collected data to the edge computing system, the edge computing system is used for anomaly detection and fault diagnosis, and the edge computing system performs data stream compression processing on the received data to remove redundant data; the data center system comprises a data layer, an interface layer, an algorithm library, a model library and service program data, wherein the data layer, the interface layer, the algorithm library, the model library and the service program data are uploaded to the data center system through an edge computing system, and then are collected and processed and stored or released into streaming data; unstructured data comprises voice, images and text, and is stored and managed by an HDFS file system database; the distributed message queue issues monitoring data to be processed in real time, so that real-time state monitoring and fault diagnosis are realized, and the consumption of the diagnosis processing stream program is facilitated; the interface layer is an external interface of the whole system and comprises an interface with a database and an interface with a computing platform; the data read-write operation of the diagnostic software is realized by connecting the data with a database through a standardized and configurable interface program; by means of job submission or API call, the calculation force of the data center system is utilized to realize model training and efficient calculation of real-time diagnosis service; the algorithm library is a customized algorithm set based on mechanism and data driven mixture, and comprises a mechanism algorithm, a data driven algorithm and a rule and knowledge based algorithm.
2. The production grid construction fault diagnosis, prediction and health management system according to claim 1, wherein the mechanism algorithm comprises a combustion engine thermodynamic fault mechanism algorithm, a motor electromagnetic fault mechanism algorithm, a motor electrical fault mechanism algorithm, a rotating machinery fault mechanism algorithm; the data driving algorithm comprises a signal processing algorithm, a data preprocessing algorithm, a statistics and machine learning mechanism algorithm and a deep learning algorithm; the algorithm based on the rules and the knowledge comprises an algorithm based on the rules and the knowledge and an inference algorithm based on a fault diagnosis tree.
3. The system for diagnosing, predicting and managing faults in production network of claim 1, wherein the model library is a collection of various models for diagnosing faults and managing health of the machine set, and the model library is used for converting data collected by the data source system into diagnosis and prediction results.
4. The production network construction fault diagnosis, prediction and health management system according to claim 1, wherein corresponding models are constructed aiming at a core device combustion engine, a motor, a body and an auxiliary system of a unit, and mainly comprise a health baseline model, an abnormality detection model, a fault identification classification model and a performance prediction regression model under multiple working conditions; aiming at a unit system, a layered health evaluation model is constructed, wherein the layered health evaluation model comprises a multi-dimensional health index calculation model, a one-dimensional health index and a discrete health index model.
5. The system of claim 1, wherein the business program is an application layer for implementing unit fault diagnosis and health management, and mainly comprises a model training and testing program, an anomaly detection and diagnosis program, a performance index prediction program, and a closed-loop auxiliary decision program.
6. The system for diagnosing, predicting and managing the construction faults of the production network according to claim 1, wherein the model training and testing program is mainly offline training, model training is carried out by adopting a single machine or a cluster according to the condition of data quantity, and the cluster is carried out by a distributed model training framework.
7. The system for diagnosing, predicting and managing the faults of the production network according to claim 1, wherein the abnormality detection and diagnosis program is realized by a distributed flow processing program, and a large number of data flows of the machine set are calculated based on an intelligent model and knowledge rules to identify the state and the fault condition of the machine set.
8. The production network construction failure diagnosis, prediction and health management system according to claim 1, wherein the performance index prediction mainly performs periodic long-term tracking and prediction analysis on performance indexes of the combustion engine, the motor and the body, and outputs preventive maintenance development timing based on safety threshold setting.
9. The production grid construction fault diagnosis, prognosis and health management system according to claim 1, wherein the closed loop aid decision making procedure is based mainly on diagnosis and prognosis results, fast linking production pre-run preparation, post-evaluation, fault handling measures, site monitoring instructions and maintenance repair strategies.
10. The production network construction fault diagnosis, prediction and health management method is characterized in that: comprising the following steps:
uploading data: the edge computing system uploads the data acquired by the data source system to the data center system;
training a model: the uploaded data is subjected to feature extraction through a data center system, the data is analyzed and calculated through an algorithm library in the data center system, a real-time fault prediction model is established through a model library, training is carried out, continuous optimization is carried out, and finally the fault prediction model is issued to a gas turbine compressor;
client side: the decision maker can directly access the running condition of the system, and make value-added decisions to manage the gas turbine compressor;
edge computing system: judging whether the data source system needs to upload the cloud, if a matching model exists in a model library of the data center system, directly analyzing, calculating and storing data, and executing assignment decision through a default set command for emergency of the gas turbine compressor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311773490.4A CN117454232A (en) | 2023-12-22 | 2023-12-22 | Production network construction fault diagnosis, prediction and health management system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311773490.4A CN117454232A (en) | 2023-12-22 | 2023-12-22 | Production network construction fault diagnosis, prediction and health management system and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117454232A true CN117454232A (en) | 2024-01-26 |
Family
ID=89583991
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311773490.4A Pending CN117454232A (en) | 2023-12-22 | 2023-12-22 | Production network construction fault diagnosis, prediction and health management system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117454232A (en) |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107908864A (en) * | 2017-11-14 | 2018-04-13 | 中国科学院空间应用工程与技术中心 | A kind of complex device remaining life Forecasting Methodology of feature based fusion |
| CN109543210A (en) * | 2018-09-28 | 2019-03-29 | 国电电力宁夏新能源开发有限公司 | A kind of Wind turbines failure prediction system based on machine learning algorithm platform |
| CN109948169A (en) * | 2017-12-20 | 2019-06-28 | 中国中车股份有限公司 | A kind of railway freight-car prognostic and health management system |
| CN110672332A (en) * | 2019-09-10 | 2020-01-10 | 上海电力大学 | Gas turbine fault early warning system based on SARIMA model |
| CN111507489A (en) * | 2020-04-20 | 2020-08-07 | 电子科技大学中山学院 | Cloud-edge-coordinated amusement equipment fault prediction and health management system and method |
| CN114139949A (en) * | 2021-12-01 | 2022-03-04 | 成都西交轨道交通技术服务有限公司 | Intelligent operation and maintenance system and method for rail transit based on edge calculation and machine learning |
| CN114282434A (en) * | 2021-12-16 | 2022-04-05 | 成都航天科工大数据研究院有限公司 | Industrial equipment health management system and method |
| CN114417501A (en) * | 2021-12-30 | 2022-04-29 | 中国航空工业集团公司西安航空计算技术研究所 | Airborne deployment-oriented health management predictive modeling method |
| CN115982123A (en) * | 2022-12-12 | 2023-04-18 | 华中光电技术研究所(中国船舶集团有限公司第七一七研究所) | Photoelectric equipment health management system based on digital twins |
| CN116368355A (en) * | 2021-09-05 | 2023-06-30 | 汉熵通信有限公司 | Internet of things system |
| CN116542651A (en) * | 2023-05-16 | 2023-08-04 | 中国科学院空间应用工程与技术中心 | On-orbit autonomous health management system and method for payload of spacecraft |
| CN116824859A (en) * | 2023-07-21 | 2023-09-29 | 佛山市新基建科技有限公司 | Intelligent traffic big data analysis system based on Internet of things |
| CN116862442A (en) * | 2023-07-21 | 2023-10-10 | 河南盘古科技发展有限公司 | Intelligent equipment management method and system based on big data analysis |
| CN117092918A (en) * | 2023-10-20 | 2023-11-21 | 北京明略政睿信息技术有限公司 | Cloud edge cooperation-based intelligent gas sensing control method |
-
2023
- 2023-12-22 CN CN202311773490.4A patent/CN117454232A/en active Pending
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107908864A (en) * | 2017-11-14 | 2018-04-13 | 中国科学院空间应用工程与技术中心 | A kind of complex device remaining life Forecasting Methodology of feature based fusion |
| CN109948169A (en) * | 2017-12-20 | 2019-06-28 | 中国中车股份有限公司 | A kind of railway freight-car prognostic and health management system |
| CN109543210A (en) * | 2018-09-28 | 2019-03-29 | 国电电力宁夏新能源开发有限公司 | A kind of Wind turbines failure prediction system based on machine learning algorithm platform |
| CN110672332A (en) * | 2019-09-10 | 2020-01-10 | 上海电力大学 | Gas turbine fault early warning system based on SARIMA model |
| CN111507489A (en) * | 2020-04-20 | 2020-08-07 | 电子科技大学中山学院 | Cloud-edge-coordinated amusement equipment fault prediction and health management system and method |
| CN116368355A (en) * | 2021-09-05 | 2023-06-30 | 汉熵通信有限公司 | Internet of things system |
| CN114139949A (en) * | 2021-12-01 | 2022-03-04 | 成都西交轨道交通技术服务有限公司 | Intelligent operation and maintenance system and method for rail transit based on edge calculation and machine learning |
| CN114282434A (en) * | 2021-12-16 | 2022-04-05 | 成都航天科工大数据研究院有限公司 | Industrial equipment health management system and method |
| CN114417501A (en) * | 2021-12-30 | 2022-04-29 | 中国航空工业集团公司西安航空计算技术研究所 | Airborne deployment-oriented health management predictive modeling method |
| CN115982123A (en) * | 2022-12-12 | 2023-04-18 | 华中光电技术研究所(中国船舶集团有限公司第七一七研究所) | Photoelectric equipment health management system based on digital twins |
| CN116542651A (en) * | 2023-05-16 | 2023-08-04 | 中国科学院空间应用工程与技术中心 | On-orbit autonomous health management system and method for payload of spacecraft |
| CN116824859A (en) * | 2023-07-21 | 2023-09-29 | 佛山市新基建科技有限公司 | Intelligent traffic big data analysis system based on Internet of things |
| CN116862442A (en) * | 2023-07-21 | 2023-10-10 | 河南盘古科技发展有限公司 | Intelligent equipment management method and system based on big data analysis |
| CN117092918A (en) * | 2023-10-20 | 2023-11-21 | 北京明略政睿信息技术有限公司 | Cloud edge cooperation-based intelligent gas sensing control method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111445106B (en) | Safety control method and system for fault processing operation site of electricity acquisition equipment | |
| CN110929918B (en) | 10kV feeder fault prediction method based on CNN and LightGBM | |
| CN110766277B (en) | Health assessment and diagnosis system and mobile terminal for nuclear industry field | |
| CN107358366B (en) | Distribution transformer fault risk monitoring method and system | |
| CN109800127A (en) | A kind of system fault diagnosis intelligence O&M method and system based on machine learning | |
| CN111412579B (en) | Air conditioning unit fault type diagnosis method and system based on big data | |
| CN112817280A (en) | Implementation method for intelligent monitoring alarm system of thermal power plant | |
| US20240142063A1 (en) | Method for troubleshooting potential safety hazards of compressor in smart gas pipeline network and internet of things system thereof | |
| CN111444169A (en) | Transformer substation electrical equipment state monitoring and diagnosis system and method | |
| CN113962299A (en) | Intelligent operation monitoring and fault diagnosis general model for nuclear power equipment | |
| CN116228186A (en) | Ship cabin intelligent operation and maintenance system based on human engineering | |
| CN110349048B (en) | Substation multidimensional data operation interactive control platform and fault handling method | |
| CN106570567A (en) | Main network maintenance multi-constraint multi-target evaluation expert system and optimization method | |
| CN112990656A (en) | Health evaluation system and health evaluation method for IT equipment monitoring data | |
| Li et al. | Framework and case study of cognitive maintenance in Industry 4.0 | |
| CN110636066A (en) | Network security threat situation assessment method based on unsupervised generative reasoning | |
| CN110033102A (en) | A kind of huge hydroelectric power plant has the intelligent diagnosing method and expert system of learning functionality | |
| CN116010886B (en) | Security monitoring method, device, electronic equipment and storage medium | |
| CN115358155A (en) | Power big data abnormity early warning method, device, equipment and readable storage medium | |
| CN117041312A (en) | Enterprise-level information technology monitoring system based on Internet of things | |
| CN117032165A (en) | Industrial equipment fault diagnosis method | |
| Mosallam | Remaining useful life estimation of critical components based on Bayesian Approaches. | |
| CN102929241B (en) | Safe operation guide system of purified terephthalic acid device and application of safe operation guide system | |
| CN114118222A (en) | Coal mill state early warning system and fault identification and diagnosis method | |
| CN117408162A (en) | Power grid fault control method based on digital twin |
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 |