CN206833239U - A kind of thermal power plant's control system fault detection system based on data-driven - Google Patents
A kind of thermal power plant's control system fault detection system based on data-driven Download PDFInfo
- Publication number
- CN206833239U CN206833239U CN201720527569.2U CN201720527569U CN206833239U CN 206833239 U CN206833239 U CN 206833239U CN 201720527569 U CN201720527569 U CN 201720527569U CN 206833239 U CN206833239 U CN 206833239U
- Authority
- CN
- China
- Prior art keywords
- sensor
- signal
- pls
- control system
- fault detection
- 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.)
- Active
Links
Abstract
The utility model discloses a kind of thermal power plant's control system fault detection system based on data-driven, by carrying out PLS calculating to nonlinear thermal power plant's control signal, realizes the real-time monitoring and detection to the system failure.The system includes：Sensor, it is arranged near the actuator of thermal power plant's control system, including at least one temperature sensor, pressure sensor and flow sensor；Memory module, wired or wireless connection, the signal for storage sensor are carried out with sensor；PLS processing modules, it is electrically connected with memory module, PLS computings is carried out to the signal of sensor for receiving the signal of sensor and establishing PLS models, and based on PLS models；Alarm, it is electrically connected with PLS processing modules, for receiving the PLS operation results of PLS processing modules, concurrently send alarm signal.
Description
Technical field
It the utility model is related to the field of fault detection of control system of power plant, more particularly to a kind of fire based on data-driven
Control system of power plant fault detection system.
Background technology
Thermal power plant's control system is the indispensable important component of Large-scale fire-electricity unit, and its Performance And Reliability is to protect
An important factor for demonstrate,proving fired power generating unit safe operation and economy, can have many potential failures for this large-scale control system
Point, it is necessary to by the process operation data of control system be tracked detection and analyze come the generation position of failure judgement point with
And reason, and then take measures to be repaired in time.The fault detection system of thermal power plant's control system is only to control system at present
Input variable decomposed and analyzed, and ignore the output variable of control system, because output variable is generally become by input
Amount and other immeasurablel factors influence, therefore only investigation input variable can not correctly reflect thermal power plant's control system
Failure situation.
Utility model content
In order to overcome the defects of above-mentioned, the utility model one embodiment provides a kind of thermal power plant based on data-driven
Control system fault detection system, including：Sensor, it is arranged near the actuator of thermal power plant's control system, including at least one
Individual temperature sensor, pressure sensor and flow sensor；Memory module, wired or wireless connection is carried out with sensor, is used
In the signal of storage sensor；PLS processing modules, it is electrically connected with memory module, for receiving signal and the foundation of sensor
PLS models, and PLS computings are carried out to the signal of sensor based on PLS models；Alarm, it is electrically connected with PLS processing modules,
For receiving the PLS operation results of PLS processing modules, alarm signal is concurrently sent.
Preferably, sensor also includes at least one liquid level sensor and/or vapor (steam) velocity sensor.
Preferably, pretreatment module is additionally provided with memory module, pretreatment module is used to carry out in advance the signal of sensor
Processing.
Preferably, thermal power plant's control system includes at least one furnace safeguard supervisory system, coordinated control system, bypass control
System processed, and/or analog control system.
Preferably, the signal of sensor includes fire box temperature signal, vapor temperature signal, steam pressure signal, drum liquid
Position signal, and/or steam flow signal.
Preferably, actuator is nearby provided with 5 monitoring points, and each monitoring point is equiped with a sensor.
Preferably, the actuator of furnace safeguard supervisory system includes steam water-level adjusting actuator, vapor (steam) temperature regulation is held
Row device and furnace pressure adjusting actuator.
Preferably, the actuator of Bypass Control System includes bypass valve.
Preferably, the actuator of analog control system includes flow switch, differential pressure switch and liquid-level switch.
Preferably, actuator is nearby provided with and is no less than 2 monitoring points.
It should be understood that in the scope of the utility model, above-mentioned each technical characteristic of the present utility model and below (such as implementation
Example) in specifically describe each technical characteristic between can be combined with each other, so as to form new or preferable technical scheme.It is limited to
Length, no longer tire out one by one state herein.
Brief description of the drawings
Fig. 1 is the structural representation of the fault detection system in one embodiment of the present utility model.
Fig. 2 is the flow chart of data processing figure of the fault detection system in one embodiment of the present utility model.
Embodiment
Inventor develops a kind of thermal power plant's control system based on data-driven first by in-depth study extensively
Fault detection system.For compared with the prior art, the system is provided with PLS processing modules 103, is determined using PLS
Correlation between the output variable of thermal power plant's control system simultaneously establishes PLS models, and then calculates score matrix according to PLS models
Control with the statistic for the normal operating condition for reflecting thermal power plant's control system limits；Comparison module is also additionally arranged, for comparing
Statistic and limit value during real time execution.
The PLS models that the utility model uses can more be precisely calculated the correlation between each variable, so as to faster
The generation of failure is determined fastly and is more accurately alarmed.
Term
As used herein, term " offset minimum binary (PLS) " refers to a kind of multiple linear regression algorithm, is to common minimum
The improvement of two multiplication algorithms, its basic thought are to think that system (or process) is driven by a small amount of hidden variable, pass through hidden change
The form of amount describes the linear relationship between input variable and output variable, so as to establishing the internal model of system.
A kind of thermal power plant's control system fault detection system based on data-driven is the utility model is related to, is employed partially most
Small square law carries out nonlinear analysis to the signal value of control system.
Typically, the utility model includes sensor, is arranged near the actuator of thermal power plant's control system, including at least
Temperature sensor, pressure sensor and the flow sensor of one；Memory module, wired or wireless connection is carried out with sensor,
Signal for storage sensor；PLS processing modules, it is electrically connected with memory module, for receiving the signal of sensor and building
Vertical PLS models, and PLS computings are carried out to the signal of sensor based on PLS models；Alarm, electrically connect with PLS processing modules
Connect, for receiving the PLS operation results of PLS processing modules, concurrently send alarm signal.
Major advantage of the present utility model includes：
(a) reflection thermal power plant control system normal operating condition can be more accurately calculated using PLS models
Statistic control limit.
(b) correlation between each output variable of thermal power plant's control system is studied, and according to each output variable
Accurately estimate service data when thermal power plant's monoblock breaks down.
With reference to specific embodiment, the utility model is expanded on further.It should be understood that these embodiments are merely to illustrate this
Utility model rather than limitation the scope of the utility model.
It should be noted that in the claim and specification of this patent, such as first and second or the like relation
Term is used merely to make a distinction an entity or operation with another entity or operation, and not necessarily requires or imply
Any this actual relation or order be present between these entities or operation.Moreover, term " comprising ", "comprising" or its
Any other variant is intended to including for nonexcludability so that process, method, article including a series of elements or
Equipment not only includes those key elements, but also the other element including being not expressly set out, or also include for this process,
Method, article or the intrinsic key element of equipment.In the absence of more restrictions, wanted by what sentence " including one " limited
Element, it is not excluded that other identical element in the process including key element, method, article or equipment also be present.
All it is incorporated as referring in this application in all documents that the utility model refers to, just as each document quilt
It is individually recited as with reference to such.In addition, it is to be understood that after above-mentioned instruction content of the present utility model has been read, this area skill
Art personnel can make various changes or modifications to the utility model, and these equivalent form of values equally fall within the application appended claims
Book limited range.
Fig. 1 is the structural representation based on data-driven fault detection system in one embodiment of the present utility model.
As shown in figure 1, the detecting system includes：Sensor 101, it is arranged near the actuator of thermal power plant's control system, including at least
Temperature sensor, pressure sensor and the flow sensor of one；Memory module 102, carried out with sensor 101 wired or wireless
Connection, the signal for storage sensor 101；PLS processing modules 103, it is electrically connected with memory module 102, is passed for receiving
The signal of sensor 101 simultaneously establishes PLS models, and carries out PLS computings to the signal of sensor 101 based on PLS models；Alarm
104, it is electrically connected with PLS processing modules, for receiving the PLS operation results of PLS processing modules, concurrently send alarm signal.
Thermal power plant's control system include at least one furnace safeguard supervisory system, coordinated control system, Bypass Control System,
And/or analog control system.Different control systems are provided with the actuator of difference in functionality.For example, furnace safeguard supervisory system
Actuator includes steam water-level adjusting actuator, vapor (steam) temperature adjusting actuator and furnace pressure adjusting actuator；Bypass Control
The actuator of system includes bypass valve；The actuator of analog control system includes flow switch, differential pressure switch and liquid level
Switch.Accordingly, sensor 101 includes at least one temperature sensor, gas pressure sensor, liquid level sensor, and/or stream
Quantity sensor；The signal of sensor 101 then includes fire box temperature signal, vapor temperature signal, steam pressure signal, liquid level of steam drum
Signal, and/or steam flow signal.1~10 monitoring point is provided near each actuator, each monitoring point can install 1
~5 sensors 101.Each actuator, which is nearby provided with, is no less than 2 monitoring points.
Fig. 2 is the structural representation of the fault detection system based on data-driven in another embodiment of the present utility model
Figure.The system is relative to the fault detection system shown in Fig. 1, and memory module 102 also includes pretreatment module, for sensor
101 signal is pre-processed；The function of PLS processing modules is described in detail below in association with embodiment.
The off-line modeling stage
Sensor 101 is in advance when each control system of thermal power plant is in off-line state, from the number of each control system
Pressure signal, temperature signal, liquid level signal, the flow letter of " thermal power plant's monoblock normal operating condition " are represented according to collection in storehouse
Number and particle concentration signal etc., every kind of signal respectively takes p sample, p 100, and these signals reflect thermal power plant's unit machine
The normal operating condition of group.
Memory module 102 stores the signal of the sensor 101 with a matrix type.If for example, collect steam pressure
Value sg, vapor (steam) temperature value st, fire box temperature value ftEach 100 samples of three kinds of signals, then be configured to three-dimensional matrice 3 × 100；If adopt
Collect liquid level of steam drum value lg, steam stream value qm, vapour pressure force value sg, vapor (steam) temperature value st, fire box temperature value ftFive kinds of signals are given
150 samples, then it is configured to five dimension matrixes 5 × 150.
Memory module 102 is additionally provided with pretreatment module.The core processor (such as CPU) of pretreatment module is provided with standardization
Unit, the Standardisation Cell are standardized to the signal of sensor 101 according to the following equation：
Wherein max { xijBe matrix interior element maximum, min { xijBe matrix interior element minimum value, xijIt is matrix
Element, yijIt is the sample values of standardization, m is the species number of signal, and n is the sample number that each signal is taken.
The signal of sensor 101 after standardization forms offline sample space and is stored in pretreatment module.
PLS processing modules 103 receive offline sample space from memory module 102, offline sample space is trained into
And establish PLS models.Then, PLS processing modules 103 are based on PLS models, and carrying out PLS to offline sample space is calculated
Sub-matrix T, also, offline sample space is subjected to lowering dimension decomposition.In one embodiment, offline sample space can be decomposed
Into the low dimension projective subspace being made up of principal component and a corresponding residual error subspace, principal component here refers to above-mentioned expression
The signal of " thermal power plant's monoblock normal operating condition ", including pressure signal, temperature signal, liquid level signal, flow signal with
And particle concentration signal etc..And the statistic T for reflecting spatial variations characteristic is configured in the two subspaces respectively2
And SPE.
Further, PLS processing modules 103 are according to statistic T2The statistic of offline sample space is calculated with SPE
Control limit (i.e. statistic T2Limitation and statistic SPE limit value), the statistic control limit be stored in PLS processing modules 103
In.
The on-line checking stage
When each control system of thermal power plant is in running order, sensor 101 gathers the real-time of each control system
Service data, including vapour pressure force value, vapor (steam) temperature value, fire box temperature value, liquid level of steam drum value, steam stream value etc., it is every kind of
Value equally takes p sample, and p is 100~200；These values reflect the real-time running state of thermal power plant's monoblock.
Memory module 102 stores the signal of the sensor 101 with a matrix type.If for example, collect steam pressure
Value sg, vapor (steam) temperature value st, fire box temperature value ftEach 100 samples of three kinds of signals, then be configured to three-dimensional matrice 3 × 100；If adopt
Collect liquid level of steam drum value lg, steam stream value qm, vapour pressure force value sg, vapor (steam) temperature value st, fire box temperature value ftFive kinds of signals are given
150 samples, then it is configured to five dimension matrixes 5 × 150.
Memory module 102 is additionally provided with pretreatment module.The core processor (such as CPU) of pretreatment module is provided with standardization
Unit, the Standardisation Cell are standardized to the signal of sensor 101 according to the following equation：
Wherein max { xijBe matrix interior element maximum, min { xijBe matrix interior element minimum value, xijIt is matrix
Element, yijIt is the sample values of standardization, m is the species number of signal, and n is the sample number that each signal is taken.
The signal of sensor 101 after standardization forms real-time sample space and is stored in pretreatment module.
PLS processing modules 103 receive real-time sample space from memory module 102, and are based on PLS models, to real-time sample
Space carries out PLS and score matrix T is calculated, also, real-time sample space is carried out into lowering dimension decomposition.In one embodiment,
Real-time sample space can be broken down into the low dimension projective subspace being made up of principal component and a corresponding residual error subspace, this
In principal component refer to the signal of above-mentioned expression " thermal power plant's monoblock on-line operation state ", including pressure signal, temperature letter
Number, liquid level signal, flow signal and particle concentration signal etc..Moreover, PLS processing modules 103 calculate according to score matrix T
Score vector, and real-time sample is projected to low dimension projective subspace and residual error subspace respectively, real-time system is calculated
Measure T2 ' and SPE '.
Further, PLS processing modules 103 are respectively by statistic T2’, SPE ' and statistic T2With the respective statistics of SPE
Control limit is compared, and obtains comparative result, and the comparative result is output into alarm 104.
Work as statistic T2’And SPE ' is respectively smaller than equal to statistic T2Control with SPE is prescribed a time limit, and alarm 104 is to sensor
101 send the signal for representing " thermal power plant's control system is normal ", so as to which sensor 101 continues to gather each control system of thermal power plant
Real-time running data；As real-time statistics amount T2’And SPE ' is respectively greater than statistic T2Control with SPE is prescribed a time limit, alarm 104
Alarm signal is then sent, represents that thermal power plant's control system breaks down.
Fig. 2 is the flow chart of data processing figure of the fault detection system in one embodiment of the present utility model.Building offline
Mould part, data normalization processing is carried out to the history normal data (training data) of collection first, and establish PLS models, counted
Calculate the statistic T of score matrix T and training data2And SPE, determine statistic T2Control with SPE limits；In on-line checking part,
The online real time data (test data) of system is analyzed, the statistic T of real time data is calculated using score matrix T2
And SPE, statistic are Testing index, then illustrate that failure occurs in system in limited time when the statistic of test data exceedes control,
Otherwise, then illustrate that system is in normal operating condition.
Claims (10)
- A kind of 1. thermal power plant's control system fault detection system based on data-driven, it is characterised in that including：Sensor, it is arranged near the actuator of thermal power plant's control system, including at least one temperature sensor, pressure sensing Device and flow sensor；Memory module, wired or wireless connection is carried out with the sensor, for storing the signal of the sensor；PLS processing modules, it is electrically connected with the memory module, for receiving the signal of the sensor and establishing PLS models, And PLS computings are carried out to the signal of the sensor based on the PLS models；Alarm, it is electrically connected with the PLS processing modules, for receiving the PLS operation results of the PLS processing modules, and Send alarm signal.
- 2. fault detection system as claimed in claim 1, it is characterised in that the sensor also includes at least one liquid level and passed Sensor and/or vapor (steam) velocity sensor.
- 3. fault detection system as claimed in claim 1, it is characterised in that pretreatment mould is additionally provided with the memory module Block, the pretreatment module are used to pre-process the signal of the sensor.
- 4. fault detection system as claimed in claim 1, it is characterised in that thermal power plant's control system includes at least one Furnace safeguard supervisory system, coordinated control system, Bypass Control System, and/or analog control system.
- 5. fault detection system as claimed in claim 1, it is characterised in that the signal of the sensor is believed including fire box temperature Number, vapor temperature signal, steam pressure signal, liquid level of steam drum signal, and/or steam flow signal.
- 6. fault detection system as claimed in claim 1, it is characterised in that the actuator is nearby provided with 5 monitoring points, often The individual monitoring point is equiped with a sensor.
- 7. fault detection system as claimed in claim 4, it is characterised in that the actuator bag of the furnace safeguard supervisory system Include steam water-level adjusting actuator, vapor (steam) temperature adjusting actuator and furnace pressure adjusting actuator.
- 8. fault detection system as claimed in claim 4, it is characterised in that the actuator of the Bypass Control System includes side Road regulating valve.
- 9. fault detection system as claimed in claim 4, it is characterised in that the actuator of the analog control system includes Flow switch, differential pressure switch and liquid-level switch.
- 10. fault detection system as claimed in claim 1, it is characterised in that the actuator is nearby provided with no less than 2 prisons Measuring point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201720527569.2U CN206833239U (en) | 2017-05-12 | 2017-05-12 | A kind of thermal power plant's control system fault detection system based on data-driven |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201720527569.2U CN206833239U (en) | 2017-05-12 | 2017-05-12 | A kind of thermal power plant's control system fault detection system based on data-driven |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN206833239U true CN206833239U (en) | 2018-01-02 |
Family
ID=60771473
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201720527569.2U Active CN206833239U (en) | 2017-05-12 | 2017-05-12 | A kind of thermal power plant's control system fault detection system based on data-driven |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN206833239U (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110794227A (en) * | 2018-08-02 | 2020-02-14 | 阿里巴巴集团控股有限公司 | Fault detection method, system, device and storage medium |
| CN111142501A (en) * | 2019-12-27 | 2020-05-12 | 浙江科技学院 | Fault detection method based on semi-supervised autoregressive dynamic hidden variable model |
| CN114021321A (en) * | 2021-10-27 | 2022-02-08 | 陈义 | Steam turbine control upgrading system of thermal power plant |
-
2017
- 2017-05-12 CN CN201720527569.2U patent/CN206833239U/en active Active
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110794227A (en) * | 2018-08-02 | 2020-02-14 | 阿里巴巴集团控股有限公司 | Fault detection method, system, device and storage medium |
| CN111142501A (en) * | 2019-12-27 | 2020-05-12 | 浙江科技学院 | Fault detection method based on semi-supervised autoregressive dynamic hidden variable model |
| CN114021321A (en) * | 2021-10-27 | 2022-02-08 | 陈义 | Steam turbine control upgrading system of thermal power plant |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110441065B (en) | Gas turbine on-line detection method and device based on LSTM | |
| WO2017191872A1 (en) | Plant abnormality detection method and system | |
| US7275018B2 (en) | Method of condition monitoring | |
| US20160239756A1 (en) | Correlation and annotation of time series data sequences to extracted or existing discrete data | |
| CN104756029B (en) | A kind of system of the parts group of monitoring device | |
| EP3055746B1 (en) | Correlation and annotation of time series data sequences to extracted or existing discrete data | |
| CN206833239U (en) | A kind of thermal power plant's control system fault detection system based on data-driven | |
| CN102880170A (en) | System failure early warning method based on baseline model and Bayesian factor | |
| CN108092623B (en) | A kind of photovoltaic array multisensor fault detecting and positioning method | |
| CN102563362A (en) | Compressed air system and intelligent pipe network leakage detecting method for same | |
| CN103754718A (en) | System and method for monitoring elevator operation safety | |
| CN105894027A (en) | Principal element degree of association sensor fault detection method and apparatus based on density clustering | |
| CN110348150A (en) | A kind of fault detection method based on dependent probability model | |
| CN202648860U (en) | Pressure transmitter of process fluid pressure during measuring industry process and process variable transmitter | |
| CN110187696A (en) | Single order servomechanism sensor fault diagnosis method and system based on dynamic trend | |
| CN111400850A (en) | Equipment fault analysis method, device, equipment and storage medium | |
| CN104615123B (en) | K-nearest neighbor based sensor fault isolation method | |
| JP6674033B2 (en) | Plant abnormality diagnosis device and plant abnormality diagnosis system | |
| CN109990803B (en) | Method and device for detecting system abnormity and method and device for sensor processing | |
| CN104317778A (en) | Massive monitoring data based substation equipment fault diagnosis method | |
| CN110336860A (en) | Key node data guard method based on multidimensional data processing in industrial Internet of Things | |
| CN206115315U (en) | High low temperature test detecting and monitoring system | |
| CN112129893A (en) | Online calibration method for CO sensor of battery thermal runaway monitoring system | |
| CN206656749U (en) | Heat-engine plant sensor fault diagnosis system | |
| CN114936520A (en) | Steam turbine fault early warning analysis method based on improved MSET |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| GR01 | Patent grant | ||
| GR01 | Patent grant |