CN1472671A - Method for apparatus status monitoring and performance degenerate fore casting based on network - Google Patents
Method for apparatus status monitoring and performance degenerate fore casting based on network Download PDFInfo
- Publication number
- CN1472671A CN1472671A CNA031293867A CN03129386A CN1472671A CN 1472671 A CN1472671 A CN 1472671A CN A031293867 A CNA031293867 A CN A031293867A CN 03129386 A CN03129386 A CN 03129386A CN 1472671 A CN1472671 A CN 1472671A
- Authority
- CN
- China
- Prior art keywords
- equipment
- information
- processing module
- performance
- network
- 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.)
- Granted
Links
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
In the method, the transducer signal processing module carries on time domain and frequency domain analysis for collected transducer signal and to pick up characteristic vector representing equipment state, the watching pre-diagnose module applies cerebellar model nervous network method to make information mixing, equipment performance degenerating predicting to the vector, network information processing module interexchanges the performance and state information of the equipment with decision making commercial, and remote maintaining system through Internet network as the equipment behavior should be adjusted according to information fedback from the systems.
Description
Technical field:
The present invention relates to the Forecasting Methodology that a kind of based on network device status monitoring and equipment performance are degenerated, constitute by sensor signal processing module, intelligent house dog (Watchdog) module and network information processing module, realize the intelligent predicting that the remote monitoring and the equipment performance of equipment state are degenerated.Belong to remote monitoring and maintenance technology field.
Background technology:
The shut-down, the stopping production that cause owing to equipment failure often bring very big economic loss to enterprise.Along with computer technology, the development of mechanics of communication is carried out remote monitoring by network to equipment performance, to the diagnosing malfunction and the prediction of equipment, becomes an important development direction of device status monitoring.
The technology of several keys comprises the collection of sensor signal and processing, the transmission of information, analysis of equipment running status etc. in the long distance control system.At present the long distance control system of reported in literature often adopts the mode of fieldbus, LAN (Local Area Network), telephone wire to be connected into the Internet, and the data of teletransmission are original sensor signal information, the processing of data and to analyze all be that measurement and control center at far-end finishes.This makes data quantity transmitted huger, the data processing more complicated of far-end, and real-time is not high.And in the present long distance control system, the method that equipment state is discerned, such as discerning from traditional statistical model, fuzzy logic expert system and neural network till now, the equipment of needing is in the historical data form the basis of normal condition and various malfunctions, for those novel devices or highly sophisticated device, its complete historical data like this can not be arranged, the method for existing various identifications is just no longer suitable.Present long distance control system, main still to diagnosing malfunction and prediction, the performance of equipment often is divided into normal and malfunction simply, and degenerate state is not had to consider and pay attention to.And equipment often will experience a series of degenerate state before breaking down.If can predict, just can effectively prevent the generation of further decline and fault to various degenerate states.
Summary of the invention:
The objective of the invention is at the deficiencies in the prior art, a kind of based on network device status monitoring and performance degradation Forecasting Methodology are provided, the identification of equipment state and the evaluation of performance are finished at the scene, can not only discern the current state of equipment, and the decline of equipment performance carried out quantitative assessment and prediction, avoid the generation of equipment failure, improve the reliability of equipment.
The present invention adopts sensor signal processing module, the pre-diagnostic module of house dog and network information processing module to realize the monitoring and the prediction of based on network equipment performance.The sensor signal processing module is carried out time and frequency domain analysis with the sensor signal of being gathered, extract signal characteristic, and adopt principal component analytical method that the feature of extracting is carried out correlation analysis, finding out can the indication equipment performance and the main information constitutive characteristic vector of state, as the input of the pre-diagnostic module of house dog.The pre-diagnostic module of house dog adopts cerebellum Model Neural method that the proper vector of input is carried out information fusion, and the state of equipment is discerned, and the degeneration of equipment performance is predicted.Network information processing module is connected the information that predicts the outcome of the pre-diagnostic module mode by wireless data transmission with the Internet, realization and decision system, business system, the information interaction at remote maintenance center etc., and the information according to other system feedback is in time adjusted the behavior of equipment, and the performance of equipment is carried out Active Compensation.
Method of the present invention is undertaken by following concrete steps:
1, the feature of monitored device is analyzed, by the sensor acquisition status information of equipment, for example vibration of the electric current of motor, voltage, main shaft etc.Sensor signal sticks into capable signals collecting through after filtering, amplifying by multi-channel data acquisition, and the signal of collection is sent into the sensor signal processing module and handled.
2, the sensor signal processing module is according to the difference of signal properties, signal is carried out different analyses, carry out the extraction of time domain or frequency domain character, because resulting signal characteristic often has very big correlativity each other, therefore adopt the method for principal component analysis (PCA) to remove correlation of data, keeping under the constant substantially situation of former variable information, reduce the dimension of data, reduce the complicacy of follow-up research, find out the main information of those characteristic features that can represent equipment performance, constitute the proper vector of representing equipment state, as the input of the pre-diagnostic module of house dog.
3, the pre-diagnostic module of house dog adopts based on cerebellum model Node Controller (being called for short the cerebellum Model Neural) method proper vector is carried out information fusion, the performance degradation of equipment is predicted.House dog method based on the cerebellum Model Neural mainly comprises two stages: (1) learning phase: the main information constitutive characteristic vector of sensor signal feature is as the input of cerebellum Model Neural, the feature of the signal when neural network learning equipment is in various state, the behavior of apparatus for establishing and the relation of the Nonlinear Mapping between the equipment performance.(2) application stage: when equipment operation, the cerebellum Model Neural can be according to the variation of actual conditions, state to new appearance, not study is learnt, and readjusts the structure and parameter of neural network, the assessment of the quantification of the output indication equipment performance of neural network.The pre-diagnostic module of house dog as one " digital doctor ", is predicted the degeneration of equipment performance in product and equipment, and contingent fault is reported to the police.The pre-diagnostic module of house dog inputs to network information processing module with the state and the degradation information of equipment performance, carries out information interaction.
4, network information processing module mainly is that performance information and out of Memory with equipment carries out alternately, utilize database technology that the state and the performance information of equipment are write down and be kept at a lane database, can carry out the action learning and the Rule Extraction of expert's maintenance system, in addition also can be as " black box " of preserving facility information, in case equipment breaks down, can be by visit this " black box ", the information of acquisition equipment last a few minutes finds fault rootstock fast.Utilize wireless communication technique information and decision system and business system to be carried out alternately by network, and feedback information according to remote maintenance center and other system, equipment is taked corresponding maintenance measure, perhaps adjust the various parameters of equipment itself and condition of work thereof, equipment performance is compensated and revises.
The present invention adopts and based on the house dog method of cerebellum Model Neural the decline of equipment performance to be predicted, has can merge a plurality of characteristic informations, the input data type is had robustness, calculates advantage such as simple, requirement of real time.The more important thing is, method of the present invention can provide the description of quantification to the recession level of equipment performance, overcome other method can only the simple zones subset deficiency of normal and fault, and method of the present invention is easy to be integrated in other hardware environment, has a very big dirigibility.
Method of the present invention adopts sensor signal processing module, the pre-diagnostic module of house dog and network information processing module, can realize the remote monitoring of based on network equipment state, turns to estimated performance to fail from simple fault diagnosis in the past; The Active Compensation of safeguarding the turning facilities performance that slave unit is passive; Turn to the multiple devices property comparison from the performance diagnogtics of single device, i.e. the lateral comparison of equipment performance; Exchange from the exchanges data steering behaviour; Turn to remote maintenance from field maintemance, can improve the reliability and the production efficiency of equipment greatly, bring higher economic benefit to enterprise.
Description of drawings:
Fig. 1 is the schematic flow sheet of the inventive method by sensor signal processing module, the pre-diagnostic module of house dog and network information processing modules implement.
Among Fig. 1, equipment/system 1, sensor signal processing module 2, the pre-diagnostic module 3 of house dog, network information processing module 4, business system/decision system/remote maintenance center 5.
Embodiment:
Below in conjunction with accompanying drawing technical scheme of the present invention is further described.
As shown in Figure 1, the present invention adopts sensor signal processing module, the pre-diagnostic module of house dog and network information processing module to realize the monitoring and the prediction of based on network equipment state.The status information of equipment/system is by sensor acquisition, send into the sensor signal processing module, the sensor signal processing module is carried out principal component analysis (PCA) to signal characteristic, finds out the main information constitutive characteristic vector that can represent equipment state, as the input of the pre-diagnostic module of house dog.The pre-diagnostic module of house dog adopts neural net method that input information is merged, and the state of equipment is discerned, and the degeneration of equipment performance is predicted the result sends into network information processing module.Network information processing module is connected the information that predicts the outcome of the pre-diagnostic module mode by wireless data transmission with the Internet, realization and decision system, business system, the information interaction at remote maintenance center etc., and in time the behavior of equipment is adjusted according to the information of other system feedback.
Critical component or critical positions in equipment/system 1 are installed different sensors, collection can the consersion unit state various signals (electric current, voltage, vibration signal etc.), multiple signals are through filtering, amplify the back and stick into the row sampling by multi-channel data acquisition, data collecting card is realized the conversion of simulating signal to digital signal simultaneously, and sensor data information enters the sensor signal processing module and analyzes.
Sensor signal processing module 2 is analyzed raw data, extracts the characteristic quantity that can represent equipment state or performance.Sensor signal is had various analytical approachs, and for example the amplitude analysis method has maximin, peak-peak value; Statistical analysis technique has average, variance, mean square deviation; Also can carry out after Fourier transform (FFT) or the wavelet transformation signal frequency-domain feature being analyzed.In addition, also can extract different features according to the speed degree that signal changes, signal for changing at a slow speed extracts the feature relevant with energy, such as amplitude, variance, quadratic sum, for fast-changing signal, extract two kinds of features: the one, can represent the feature of process status instantaneous rate of change, the 2nd, can represent the feature of process status variation tendency at that time, as divergence and average variance.In order to gather information as much as possible, often to extract a plurality of features, this can be avoided the omission of important information certainly, however the angle from adding up may exist very strong correlativity between these characteristic quantities, makes problem analysis increase complicacy.And equipment state also has very big difference to the influence degree between the different characteristic amount, therefore, by principal component analysis (PCA), utilize several incoherent generalized variables to replace original more correlated variables, and these incoherent generalized variables can reflect the most information that former variable provides, not only removed the correlativity between the data, and reduced the dimension of data, made the processing of back simple.Those that obtain after the principal component analysis (PCA) can the characterization device performance the main information of feature constituted the proper vector of equipment, as the input of the pre-diagnostic module 3 of house dog.
The pre-diagnostic module 3 of house dog utilizes the cerebellum Model Neural to finish the identification of equipment current state and the prediction of device performance decay.The cerebellum Model Neural adopts a kind of method of tabling look-up to realize the mapping of the input space to output region, component in each input vector forms the pointer that points to certain address through a series of combinations and hash coding, the output weights of storing in these addresses and that be neural network can make the desired function of cerebellum Model Neural study by adjusting weights.Neural network is output as 1 when supposing that equipment is in normal condition, when being in malfunction, neural network is output as-1 (if the various faults state is arranged, can define respectively, as-2,-3 etc.), the proper vector that obtains when at first utilizing two states is trained the cerebellum Model Neural, makes it set up corresponding relation between input vector and the different conditions.Neural network after the training promptly can be used for the monitor procedure of equipment performance.When the output of all states is bigger when the output of neural network departs from training, just think that new state has appearred in equipment, at this moment new state is defined, and again neural network is trained, upgrade the related parameter that has of neural network.In the monitor procedure of reality, when output during near certain known state of the output of neural network, representative equipment is in this state.If the output of neural network is between 1 and-1, then represent equipment performance decline to occur, and the big or small secondary indication of output valve the decline degree.If novel device or high precision apparatus, the data when lacking equipment failure state, the data in the time of can only adopting normal condition are trained neural network.The output of neural network departs from 1 more, and then devices illustrated departs from normal condition more, determines whether reporting to the police or to adopt suitable maintenance measure according to the certain threshold values of formulations such as experience and handbook.The pre-diagnostic module 3 of house dog produces the performance information of equipment, and sends into network information processing module 4 and handle.
Network information processing module 4 deposits the performance information of equipment in the device performance data storehouse, can check for browsing at any time.And carry out information interaction with remote maintenance center, decision system and business system 5 as required.Compare with the message processing module of mentioning in other supervisory system, its difference be the analysis of data and performance be evaluated at sensor signal processing module 2 and the pre-diagnostic module 3 of house dog is finished, data can reduce to and only keep the useful device health information, what send is the information of equipment performance rather than original sensor data information, thereby when overcoming data transmission, data volume receives the situation of bandwidth constraints greatly.Information transmitted is the state of equipment and the information that equipment performance is degenerated, and mainly comprises: the time that the time that the frequency (4) that the parts (3) that (1) degradation modes (2) is degenerated are degenerated is degenerated and position (5) prevention are used etc.This module is also adjusted the parameter or the duty of equipment itself according to maintenance centre's feedack or order, such as when the main shaft wearing and tearing of discovering device are relatively more serious, can adopt the method for the suitable reduction speed of mainshaft, reduce wear, make equipment work in optimum condition.
Claims (1)
1, a kind of based on network device status monitoring and performance degradation Forecasting Methodology is characterized in that by the sensor signal processing module, and pre-diagnostic module of house dog and network information processing modules implement specifically comprise:
1) by the sensor acquisition status information of equipment, the signal of collection is sent into the sensor signal processing module and is handled;
2) the sensor signal processing module is extracted signal time domain or frequency domain character, adopts principal component analysis (PCA) then, finds out the main information constitutive characteristic vector of representing equipment state, as the input of the pre-diagnostic module of house dog;
3) the pre-diagnostic module of house dog adopts cerebellum Model Neural method that proper vector is carried out information fusion, and the performance degradation of equipment is predicted, the state and the degradation information of equipment performance inputed to network information processing module, carries out information interaction;
4) network information processing module writes down and is kept at lane database with the state and the performance information of equipment, by network information and decision system and business system are carried out alternately, and equipment is carried out parameter adjustment according to the feedback information of remote maintenance center and other system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 03129386 CN1240009C (en) | 2003-06-19 | 2003-06-19 | Method for apparatus status monitoring and performance degenerate fore casting based on network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 03129386 CN1240009C (en) | 2003-06-19 | 2003-06-19 | Method for apparatus status monitoring and performance degenerate fore casting based on network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1472671A true CN1472671A (en) | 2004-02-04 |
| CN1240009C CN1240009C (en) | 2006-02-01 |
Family
ID=34153508
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 03129386 Expired - Fee Related CN1240009C (en) | 2003-06-19 | 2003-06-19 | Method for apparatus status monitoring and performance degenerate fore casting based on network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1240009C (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100347381C (en) * | 2005-04-14 | 2007-11-07 | 上海交通大学 | Remote monitoring and maintaining system of pneumatic paisson unmanned construction equipment |
| CN101110713B (en) * | 2007-09-05 | 2010-05-19 | 中国科学院上海微系统与信息技术研究所 | Information anastomosing system performance test bed based on wireless sensor network system |
| CN101194468B (en) * | 2005-04-14 | 2011-09-14 | 高通股份有限公司 | Apparatus and process for universal diagnostic monitor module on a wireless device |
| CN102521604A (en) * | 2011-11-21 | 2012-06-27 | 上海交通大学 | Device and method for estimating performance degradation of equipment based on inspection system |
| CN104615122A (en) * | 2014-12-11 | 2015-05-13 | 深圳市永达电子股份有限公司 | Industrial control signal detection system and detection method |
| CN106154992A (en) * | 2015-03-31 | 2016-11-23 | 西门子公司 | Production system and the control method of production system |
| CN109460846A (en) * | 2018-06-19 | 2019-03-12 | 国网浙江省电力有限公司湖州供电公司 | A kind of Condition Prediction of Equipment analysis method based on data mining |
| WO2019148960A1 (en) * | 2018-02-05 | 2019-08-08 | 华为技术有限公司 | Data analysis device, system, and method |
| CN110815224A (en) * | 2019-11-14 | 2020-02-21 | 华南智能机器人创新研究院 | Remote fault diagnosis pushing method and device for robot |
| CN110861123A (en) * | 2019-11-14 | 2020-03-06 | 华南智能机器人创新研究院 | Method and device for visually monitoring and evaluating running state of robot |
| CN112348078A (en) * | 2020-11-09 | 2021-02-09 | 南京工程学院 | Gate machine controller with sub-health pre-diagnosis and fault type clustering functions |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109086804B (en) * | 2018-07-12 | 2022-04-22 | 中石化石油机械股份有限公司 | Hydraulic equipment early failure prediction method based on fusion of multi-source state monitoring information and reliability characteristics |
-
2003
- 2003-06-19 CN CN 03129386 patent/CN1240009C/en not_active Expired - Fee Related
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100347381C (en) * | 2005-04-14 | 2007-11-07 | 上海交通大学 | Remote monitoring and maintaining system of pneumatic paisson unmanned construction equipment |
| CN101194468B (en) * | 2005-04-14 | 2011-09-14 | 高通股份有限公司 | Apparatus and process for universal diagnostic monitor module on a wireless device |
| CN101110713B (en) * | 2007-09-05 | 2010-05-19 | 中国科学院上海微系统与信息技术研究所 | Information anastomosing system performance test bed based on wireless sensor network system |
| CN102521604A (en) * | 2011-11-21 | 2012-06-27 | 上海交通大学 | Device and method for estimating performance degradation of equipment based on inspection system |
| CN102521604B (en) * | 2011-11-21 | 2014-11-19 | 上海交通大学 | Device and method for estimating performance degradation of equipment based on inspection system |
| CN104615122A (en) * | 2014-12-11 | 2015-05-13 | 深圳市永达电子股份有限公司 | Industrial control signal detection system and detection method |
| CN106154992A (en) * | 2015-03-31 | 2016-11-23 | 西门子公司 | Production system and the control method of production system |
| WO2019148960A1 (en) * | 2018-02-05 | 2019-08-08 | 华为技术有限公司 | Data analysis device, system, and method |
| CN110121180A (en) * | 2018-02-05 | 2019-08-13 | 华为技术有限公司 | A kind of data analysis set-up, system and method |
| US11523288B2 (en) | 2018-02-05 | 2022-12-06 | Huawei Technologies Co., Ltd. | Data analysis apparatus, system, and method |
| CN109460846A (en) * | 2018-06-19 | 2019-03-12 | 国网浙江省电力有限公司湖州供电公司 | A kind of Condition Prediction of Equipment analysis method based on data mining |
| CN109460846B (en) * | 2018-06-19 | 2022-04-01 | 国网浙江省电力有限公司湖州供电公司 | Equipment state prediction analysis method based on data mining |
| CN110815224A (en) * | 2019-11-14 | 2020-02-21 | 华南智能机器人创新研究院 | Remote fault diagnosis pushing method and device for robot |
| CN110861123A (en) * | 2019-11-14 | 2020-03-06 | 华南智能机器人创新研究院 | Method and device for visually monitoring and evaluating running state of robot |
| CN112348078A (en) * | 2020-11-09 | 2021-02-09 | 南京工程学院 | Gate machine controller with sub-health pre-diagnosis and fault type clustering functions |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1240009C (en) | 2006-02-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108418841B (en) | Next-generation key message infrastructure network Security Situation Awareness Systems based on AI | |
| CN1240009C (en) | Method for apparatus status monitoring and performance degenerate fore casting based on network | |
| EP1543394B1 (en) | Method and device for monitoring a technical installation comprising several systems, in particular an electric power station | |
| KR102428632B1 (en) | Systems and methods for detecting and measuring anomalies in signaling originating from components used in industrial processes | |
| CN112292642B (en) | Control device for controlling a technical system and method for configuring a control device | |
| CN104615122B (en) | A kind of industry control signal detection system and detection method | |
| AU2018251134B2 (en) | Method and apparatus for monitoring the condition of subsystems within a renewable generation plant or microgrid | |
| CN1787116A (en) | System for intelligent maintaince of muclear power paltn based on state monitoring | |
| CN112149877B (en) | Multi-source data driven fault prediction method and system for multi-element complex urban power grid | |
| CN118213936A (en) | Voltage monitoring and voltage protecting system and method | |
| CN118174456B (en) | Self-sensing monitoring intelligent control power distribution system | |
| CN117273402A (en) | Energy-saving management system and method for glass deep processing production line based on Internet of things technology | |
| EP3640477A1 (en) | Development of a superordinate model for controlling and/or monitoring a compressor installation | |
| US11687772B2 (en) | Optimization of cyber-physical systems | |
| KR102360004B1 (en) | Management system of machine based on a vibration | |
| CN111416870B (en) | Industrial Internet of things system | |
| CN118230532B (en) | Real-time monitoring and early warning system of power battery assembly line | |
| CN118612772B (en) | Offshore wind farm communication system based on 5G and Lora networking | |
| KR102683088B1 (en) | Lvdc agricultural and industrial complex microgrid power grid stabilization system and system environment analysis device | |
| CN116146402B (en) | Diagnosis and early warning method for pumping fault of main distributing valve of hydropower station speed regulator | |
| CN118224156B (en) | Hydraulic system abnormality monitoring method and system | |
| CN115604016B (en) | Industrial control abnormal behavior monitoring method and system of behavior feature chain model | |
| Bezerra et al. | RDE with Forgetting: An Approximate Solution for Large Values of with an Application to Fault Detection Problems | |
| AU2021286514B2 (en) | Generating a hybrid sensor to compensate for intrusive sampling | |
| CN116738376B (en) | Signal acquisition and recognition method and system based on vibration or magnetic field awakening |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C19 | Lapse of patent right due to non-payment of the annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |