CN106570237A - Method and system for monitoring stator blade thickness of turbine of blast furnace gas waste heat recovery device - Google Patents
Method and system for monitoring stator blade thickness of turbine of blast furnace gas waste heat recovery device Download PDFInfo
- Publication number
- CN106570237A CN106570237A CN201610933790.8A CN201610933790A CN106570237A CN 106570237 A CN106570237 A CN 106570237A CN 201610933790 A CN201610933790 A CN 201610933790A CN 106570237 A CN106570237 A CN 106570237A
- Authority
- CN
- China
- Prior art keywords
- turbine
- stator blade
- blade thickness
- turbine stator
- data
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F01—MACHINES OR ENGINES IN GENERAL; ENGINE PLANTS IN GENERAL; STEAM ENGINES
- F01D—NON-POSITIVE DISPLACEMENT MACHINES OR ENGINES, e.g. STEAM TURBINES
- F01D21/00—Shutting-down of machines or engines, e.g. in emergency; Regulating, controlling, or safety means not otherwise provided for
- F01D21/10—Shutting-down of machines or engines, e.g. in emergency; Regulating, controlling, or safety means not otherwise provided for responsive to unwanted deposits on blades, in working-fluid conduits or the like
-
- 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
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The invention relates to a method and system for monitoring stator blade thickness of a turbine of a blast furnace gas waste heat recovery device. The method comprises the following steps: acquiring the turbine data, wherein the turbine data comprises gas parameters of the entrance and exit of the turbine, operating parameters of the turbine, and the stator blade thickness of the turbine; analyzing the acquired turbine data based on an auto-encoder algorithm, and extracting the ash deposition mode characteristics of the stator blades of the turbine; obtaining a deep learning network model for the stator blade thickness of the turbine according to the ash deposition mode characteristics of the stator blades of the turbine and the stator blade thickness of the turbine, and ensuring that the deep learning network model takes the turbine data as input and the stator blade thickness of the turbine as output; and real-timely monitoring the stator blade thickness of the turbine by taking the deep learning network model as the basis and the turbine data which is acquired in real time as the input. By adopting the method and system disclosed by the invention, the thickness of the stator blades of the turbine can be monitored in real time, the problem of frequent faults of the turbine can be solved, the maintenance cost of the turbine can be reduced, and the generating capacity can be increased.
Description
Technical field
The present invention relates to gas energy recovery technology field, more particularly to turbine technical field, specifically refer to a kind of height
The turbine stator blade thickness monitor method and system of producer gas waste-heat recovery device.
Background technology
Steel and iron industry is the important foundation industry and the pillar industry for realizing new industrialization of Chinese national economy, while and
It is one of maximum industry of world's power consumption.Ironmaking is the maximum operation of energy consumption and resource consumption, its energy consumption in steel manufacture process
The 60% or so of integrated iron and steel works' total energy consumption are accounted for, significantly larger than other steel and iron manufacturing operations.And in ironmaking processes, wherein
39% energy consumption is used on blast furnace.
Blast-furnace gas energy recovery device is primarily referred to as top gas pressure recovery turbine for power generation device (Blast
Furnace Top Gas Recovery Turbine, abbreviation TRT devices), the device is the secondary of current world's most worthy
One of energy source recovery apparatus.It is using dedusting after top gas in pressure energy and heat energy, Jing turbo-expanders do
Work(is driving electrical power generators.The 30% or so of the recyclable blast furnace blower power consumption of blast-furnace gas energy recovery device, therefore quilt
It is widely used in the energy-saving and emission-reduction of blast furnace process operation.TRT device ton ferrum generated energy can reach 40kWh, and such as 4000m3 blast furnaces are matched somebody with somebody
It is after putting TRT devices, annual to generate electricity up to 1.6 hundred million kWh.According to statistics, current China's blast furnace sum has more than 900 seats, therefore TRT devices
For the energy-saving and emission-reduction of steel and iron industry have great meaning.Revision in 2008《Blast furnace iron-making process design specification》
(GB50427-2008) explicitly pointed out in：Blast furnace must be provided with top gas excess pressure power generating device, and waste pressure turbine generates electricity
Device synchronous with blast furnace should be gone into operation.
According to statistics, the TRT devices of China's operation at present have 655 sets, and average ton ferrum generated energy is less than 30 kilowatt hours/ton ferrum,
And the average ton ferrum generated energy of TRT devices of Japan is up to 40 kilowatt hours/ton ferrum.Although China's blast furnace gas working medium reaches design
Value, but TRT device generated energy is generally low, TRT devices all generally existing turbine expansion machine overhaulings frequently problem.
Turbo-expander is crucial portion's machine of TRT devices.TRT devices turbo-expander typically can using one-level stator blade at present
Adjust, such variable working condition wide ranges, improve off design performance, and energy contract fully, it is adjustable that second level stator blade takes off lid.Blast furnace gas is saturating
Carry out expanding externally acting in flat decompressor, therefore whether turbo-expander normally runs and have important meaning for energy regenerating
Justice.Turbo-expander typical fault mainly has rotor unbalance, rotor misalignment, bearing damage, shaft coupling to damage, lubricating system
Failure, turbine being utilized, blade deposits etc., most common of which are caused due to impeller of rotor heavy wear or blade deposits
Rotor unbalance.Although blast furnace gas has bag-type dust collector into before TRT devices, blast furnace gas can not possibly be completely net
Change, therefore turbo-expander is in During Process of Long-term Operation, the phenomenon of dirty and abrasion easily occurs in blade.
Although adopting the technologies such as blade coatings, the spraying of dry type antisludging agent reduce dirty and abrasion with a certain degree of,
It is that turbo-expander blade deposits wear phenomenon is yet suffered from.The fault diagnosis of most of rotor unbalances is the vibration from unit
Signal sets out, and analyzes its rumble spectrum feature to judge whether to break down.As vibration signals spectrograph analysis major part is all
Carry out offline, for blade deposits and abrasion lack the monitoring policy of real-time online, it is impossible to meet TRT device stable operations
Require, and vibration analysis system is expensive, general medium and small TRT devices are all configured without.
Therefore, hard measurement is carried out to turbine stator blade thickness by other parameters for easily measuring, solution is analyzed for us
This problem of turbine stator blade thickness monitor is there is provided a kind of analytical mathematics.
The content of the invention
The purpose of the present invention is the shortcoming for overcoming above-mentioned prior art, there is provided a kind of blast furnace gas waste-heat recovery device
Turbine stator blade thickness monitor method and system, can carry out real-time monitoring to the thickness of turbine stator blade, solve turbine
Frequently the problem of failure, reduces turbine maintenance cost.
To achieve these goals, the present invention has following composition：
The turbine stator blade thickness monitor method of the blast furnace gas waste-heat recovery device, described method include following step
Suddenly：
(1) gather turbine data, the turbine data include the gas parameters of the turbine entrance and exit with
And turbine operational factor and turbine stator blade thickness；
(2) the turbine data for collecting are analyzed using self-encoding encoder algorithm, extract turbine stator blade dust stratification
Pattern feature；
(3) pattern feature and turbine stator blade thickness according to the turbine stator blade dust stratification, obtains turbine stator blade thick
The deep learning network model of degree, the deep learning network model with the turbine data as input, with the turbine
Stator blade thickness is output；
(4) based on the deep learning network model, with the turbine data of Real-time Collection as input, to saturating
Flat machine stator blade thickness carries out real-time monitoring.
It is preferred that the gas parameters of the turbine entrance and exit include porch gas flow, porch coal gas pressure
Power, porch gas temperature, porch gas dust content, exit gas flow, exit gas pressure and exit coal
Temperature degree, the turbine parameter include turbine rotating speed and turbine acc power.
It is preferred that the own coding algorithm comprises the steps：
(2-1) coding is carried out using encoder to the turbine data for collecting and obtains hidden layer vector；
(2-2) the hidden layer vector is decoded using decoder；
(2-3) the decoded result computational minimization reconstructed error according to the decoder.
More preferably, the turbine data for collecting are encoded according to equation below using encoder：
H=f (x)=Sf(Wx+bj)；
Wherein, x is the characteristic vector constituted by the turbine data for collecting, and Wx is the weights of the characteristic vector, bj
For the threshold value of j-th neuron, h is that hidden layer is vectorial, SfFor the activation primitive of the encoder.
Further, the hidden layer vector is decoded according to equation below using decoder：
Y=g (h)=Sg(Wh+bh)；
Wherein, h is hidden layer vector, and Wh is the weights of hidden layer vector, bhFor threshold value, SgFor the decoder
Activation primitive.
Yet further, according to equation below computational minimization reconstructed error J：
Wherein, x is the characteristic vector constituted by the turbine data for collecting, and D is training sample set, and g (f (x)) is institute
The decoding output valve of decoder is stated, L is reconstruct error function.
More preferably, the deep learning network model for obtaining turbine stator blade thickness, comprises the steps：
(3-1) the turbine data are input into, it is unsupervised to train first self-encoding encoder；
(3-2) using the current output for training the self-encoding encoder for obtaining as the input of next self-encoding encoder, train down
One self-encoding encoder；
(3-3) training of the hidden layer for judging whether to have been completed predetermined number, if it is, continue step (3-4),
Otherwise continue step (3-2)；
(3-4) with turbine stator blade thickness as output, increase reverse transmittance nerve network on last hidden layer pre-
Model is surveyed, weight adjustment is carried out to deep learning network model.
It is preferred that described carry out real-time monitoring to turbine stator blade thickness, specially：
Using multiple stage personal computer, by Hadoop open source softwares, based on the deep learning network model, structure
Build with the turbine data of Real-time Collection as input, with turbine stator blade thickness as the on-line monitoring network of output.
It is preferred that methods described also includes：
User side carries out the inquiry of turbine stator blade thickness using SQL interface JDBC or ODBC.
It is preferred that methods described also includes：
When monitoring that turbine stator blade thickness exceedes systemic presupposition threshold value, reported to the police.
The invention further relates to a kind of turbine stator blade thickness monitor system of blast furnace gas waste-heat recovery device, for described
Blast furnace gas waste-heat recovery device turbine stator blade thickness monitor method, the system includes：
Data acquisition module, to gather turbine data, the turbine data include the turbine entrance and go out
The gas parameters and turbine operational factor and turbine stator blade thickness of mouth；
Self-encoding encoder, to be analyzed to the turbine data for collecting using self-encoding encoder algorithm, extracts turbine
The pattern feature of stator blade dust stratification；
Deep learning network model builds module, to pattern feature and turbine according to the turbine stator blade dust stratification
Stator blade thickness, obtains the deep learning network model of turbine stator blade thickness, and the deep learning network model is with the turbine
Machine data are input, with the turbine stator blade thickness as output；
On-line monitoring network builds module, to based on the deep learning network model, with the institute of Real-time Collection
Turbine data are stated to be input into, real-time monitoring is carried out to turbine stator blade thickness.
The turbine stator blade thickness monitor method and system of the blast furnace gas waste-heat recovery device in the invention are employed, is led to
Multiple stage personal computer is crossed, and deep learning system is built using open source softwares such as Hadoop, based on deep learning algorithm to turbine
The thickness of machine stator blade carries out real-time monitoring, the mode for taking big data to drive, and is analyzed using current existing mass data,
Solve the problems, such as the frequent failure of turbine and maintenance down, reduce turbine maintenance cost, so as to effectively improve the dimension of turbine
Shield efficiency, reduces and safeguards the waste for bringing due to turbine；Generated energy is improved, is that the optimization that blast furnace gas retracting device is safeguarded is carried
For instrument, manpower and materials input is reduced, it is adaptable to large-scale promotion application.
Description of the drawings
Fig. 1 is the schematic diagram of the data acquisition module of the present invention.
Fig. 2 is the process schematic of the deep learning algorithm of the present invention.
Fig. 3 is the structural representation of the blast furnace gas excess pressure power generating device turbine stator blade thickness monitor system of the present invention.
Specific embodiment
In order to more clearly describe the technology contents of the present invention, carry out with reference to specific embodiment further
Description.
It is illustrated in figure 1 the schematic diagram of the data acquisition module of the present invention.Blast furnace gas excess pressure power generating device turbine is quiet
The data acquisition of leaf thickness monitoring is complete by the sensor such as Programmable Logic Controller and flow, pressure, temperature, dust solubility instrument
Into.And turbine stator blade thickness is completed by laboratory inspection.By the collection of a large amount of turbine service datas, turbine is built
The training dataset of stator blade thickness monitor method and system.
It is illustrated in figure 2 the process schematic of the deep learning algorithm of the present invention.First, to turbine entrance Gas Flow
Amount, entrance gas pressure, entrance gas temperature, entrance gas dust content, turbine rotating speed, turbine acc power, turbine go out
The data such as mouth gas pressure, temperature of exit gas, turbine blade thickness carry out pretreatment.After the pre-treatment, by volume automatically
Code device algorithm, obtains the feature of patrol pattern.Finally, by reverse transmittance nerve network (BP), the power to deep learning network
It is finely adjusted again, obtains containing with turbine entrance gas flow, entrance gas pressure, entrance gas temperature, entrance gas dust
Amount, turbine rotating speed, turbine acc power, turbine outlet gas pressure, temperature of exit gas data are input, turbine blade
Thickness is the deep learning network of output.
Specifically, the turbine stator blade thickness monitor method of blast furnace gas waste-heat recovery device of the invention includes following step
Suddenly：
(1) gather turbine entrance gas flow, entrance gas pressure, entrance gas temperature, entrance gas dust content,
The data such as turbine rotating speed, turbine acc power, turbine outlet gas pressure, temperature of exit gas, turbine stator blade thickness；
(2) using self-encoding encoder algorithm, the data of above-mentioned collection are analyzed, extract the pattern of turbine stator blade dust stratification
Feature；
(3) to the pattern feature of turbine stator blade dust stratification that extracted and turbine stator blade thickness, using back propagation god
Jing networks, therefrom obtain the deep learning network model of turbine stator blade thickness；
(4) based on the deep learning network set up, with real-time running data as input, to turbine stator blade thickness
Carry out real-time monitoring；
Wherein, the depth autocoder includes：Encoder, decoder and hidden layer；
The encoder is encoded using following relational expression：
H=f (x)=Sf(Wx+bj)；
Wherein, x is that turbine entrance gas flow, entrance gas pressure, entrance gas temperature, entrance gas dust contain
The characteristic vector constituted by amount, the outlet of turbine rotating speed, turbine acc power, turbine gas pressure, temperature of exit gas, W is
The weights of input vector, bjThe threshold value of j-th neuron is represented, or is referred to as biased, h is the hidden layer value for obtaining.
Decoder is decoded using following relational expression：
Y=g (h)=Sg(Wh+bh)；
Wherein, h be hidden layer vector, be here as input into, W be corresponding weights, bhFor threshold value, SgIt is swashing for decoder
Function living；
It is searching parameter W, b on training sample set D to the training process of depth autocoderj, bhConstitute
Reconstructed error is minimized, the expression formula of reconstructed error is:
Wherein, inputs of the x for above-mentioned formula, g (f (x)) are exported for the decoder of above-mentioned formula, and L is reconstructed error function
The present invention utilizes multiple stage personal computer, builds deep learning system using open source softwares such as Hadoop.
The training process to depth autocoder includes：
(1) turbine entrance gas flow of the input as training, entrance gas pressure, entrance gas temperature, entrance coal
Gas dust content, turbine rotating speed, turbine acc power, turbine outlet gas pressure, temperature of exit gas, it is unsupervised to train
First self-encoding encoder；
(2) using the output of first self-encoding encoder as the input of next self-encoding encoder, train second own coding
Device；
(3) repeat step (2), till completing the training of predetermined number hidden layer；
(4) with turbine stator blade thickness as output, increase a reverse transmittance nerve network on last hidden layer
Forecast model, realizes the weight fine setting to the prediction network model.
Heretofore described blast furnace gas waste-heat recovery device characteristic parameter includes：Turbine entrance gas flow, entrance
Gas pressure, entrance gas temperature, entrance gas dust content, turbine rotating speed, turbine acc power, turbine outlet coal gas pressure
Power, temperature of exit gas, model instruction is carried out to the blast furnace gas waste-heat recovery device parameter using the depth network model
Practice.In actual applications, can also be adjusted with condition according to the actual requirements.
The structure for being illustrated in figure 3 the blast furnace gas excess pressure power generating device turbine stator blade thickness monitor system of the present invention is shown
It is intended to.First, turbine entrance gas flow, entrance gas pressure, entrance gas temperature, entrance coal gas powder are gathered by PLC
Dust content, turbine rotating speed, turbine acc power, turbine outlet gas pressure, temperature of exit gas, turbine blade thickness etc.
Data carry out pretreatment.Secondly, Hadoop softwares, and multiple stage computers is recycled to carry out distributed storage.Then, using depth
Learning algorithm, is predicted modeling to turbine stator blade thickness, builds deep learning network.In man machine interface, connect using SQL
Mouth, JDBC/ODBC enter the inquiry of row information, carry out real-time estimate to turbine stator blade thickness, when predictive value exceedes given threshold
Afterwards, reported to the police.
The turbine stator blade thickness monitor method and system of the blast furnace gas waste-heat recovery device in the invention are employed, is led to
Multiple stage personal computer is crossed, and deep learning system is built using open source softwares such as Hadoop, based on deep learning algorithm to turbine
The thickness of machine stator blade carries out real-time monitoring, the mode for taking big data to drive, and is analyzed using current existing mass data,
Solve the problems, such as the frequent failure of turbine and maintenance down, reduce turbine maintenance cost, so as to effectively improve the dimension of turbine
Shield efficiency, reduces and safeguards the waste for bringing due to turbine；Generated energy is improved, is that the optimization that blast furnace gas retracting device is safeguarded is carried
For instrument, manpower and materials input is reduced, it is adaptable to large-scale promotion application.
In this description, the present invention is described with reference to its specific embodiment.But it is clear that can still make
Various modifications and alterations are without departing from the spirit and scope of the present invention.Therefore, specification and drawings are considered as illustrative
And it is nonrestrictive.
Claims (11)
1. a kind of turbine stator blade thickness monitor method of blast furnace gas waste-heat recovery device, it is characterised in that described method
Comprise the following steps：
(1) turbine data are gathered, the turbine data include gas parameters of the turbine entrance and exit and thoroughly
Flat machine operational factor and turbine stator blade thickness；
(2) the turbine data for collecting are analyzed using self-encoding encoder algorithm, extract the pattern of turbine stator blade dust stratification
Feature；
(3) pattern feature and turbine stator blade thickness according to the turbine stator blade dust stratification, obtains turbine stator blade thickness
Deep learning network model, the deep learning network model with the turbine data as input, with the turbine stator blade
Thickness is output；
(4) based on the deep learning network model, with the turbine data of Real-time Collection as input, to turbine
Stator blade thickness carries out real-time monitoring.
2. the turbine stator blade thickness monitor method of blast furnace gas waste-heat recovery device according to claim 1, its feature
It is that the gas parameters of the turbine entrance and exit include porch gas flow, porch gas pressure, porch coal
Temperature degree, porch gas dust content, exit gas flow, exit gas pressure and exit gas temperature, it is described
Turbine parameter includes turbine rotating speed and turbine acc power.
3. the turbine stator blade thickness monitor method of blast furnace gas waste-heat recovery device according to claim 1, its feature
It is that the own coding algorithm comprises the steps：
(2-1) coding is carried out using encoder to the turbine data for collecting and obtains hidden layer vector；
(2-2) the hidden layer vector is decoded using decoder；
(2-3) the decoded result computational minimization reconstructed error according to the decoder.
4. the turbine stator blade thickness monitor method of blast furnace gas waste-heat recovery device according to claim 3, its feature
It is the turbine data for collecting to be encoded according to equation below using encoder：
H=f (x)=Sf(Wx+bj)；
Wherein, x is the characteristic vector constituted by the turbine data for collecting, and Wx is the weights of the characteristic vector, bjFor jth
The threshold value of individual neuron, h are that hidden layer is vectorial, SfFor the activation primitive of the encoder.
5. the turbine stator blade thickness monitor method of blast furnace gas waste-heat recovery device according to claim 4, its feature
It is the hidden layer vector to be decoded according to equation below using decoder：
Y=g (h)=Sg(Wh+bh)；
Wherein, h is hidden layer vector, and Wh is the weights of hidden layer vector, SgFor the activation primitive of the decoder.
6. the turbine stator blade thickness monitor method of blast furnace gas waste-heat recovery device according to claim 5, its feature
It is, according to equation below computational minimization reconstructed error J：
Wherein, x is the characteristic vector constituted by the turbine data for collecting, bhFor threshold value, D is training sample set, g (f (x))
For the decoding output valve of the decoder, L is reconstruct error function.
7. the turbine stator blade thickness monitor method of blast furnace gas waste-heat recovery device according to claim 3, its feature
It is that the deep learning network model for obtaining turbine stator blade thickness comprises the steps：
(3-1) the turbine data are input into, it is unsupervised to train first self-encoding encoder；
(3-2) using the current output for training the self-encoding encoder for obtaining as the input of next self-encoding encoder, train the next one
Self-encoding encoder；
(3-3) training of the hidden layer for judging whether to have been completed predetermined number, if it is, continue step (3-4), otherwise
Continue step (3-2)；
(3-4) with turbine stator blade thickness as output, increase reverse transmittance nerve network prediction mould on last hidden layer
Type, carries out weight adjustment to deep learning network model.
8. the turbine stator blade thickness monitor method of blast furnace gas waste-heat recovery device according to claim 1, its feature
It is, it is described that real-time monitoring is carried out to turbine stator blade thickness, specially：
Using multiple stage personal computer, by Hadoop open source softwares, based on the deep learning network model, build with
The turbine data of Real-time Collection are input, with turbine stator blade thickness as the on-line monitoring network of output.
9. the turbine stator blade thickness monitor method of blast furnace gas waste-heat recovery device according to claim 1, its feature
It is that methods described also includes：
User side carries out the inquiry of turbine stator blade thickness using SQL interface JDBC or ODBC.
10. the turbine stator blade thickness monitor method of blast furnace gas waste-heat recovery device according to claim 1, its feature
It is that methods described also includes：
When monitoring that turbine stator blade thickness exceedes systemic presupposition threshold value, reported to the police.
The turbine stator blade thickness monitor system of 11. a kind of blast furnace gas waste-heat recovery devices, it is characterised in that will for right
Ask the turbine stator blade thickness monitor method of the blast furnace gas waste-heat recovery device any one of 1 to 10, the system bag
Include：
Data acquisition module, to gather turbine data, the turbine data include the turbine entrance and exit
Gas parameters and turbine operational factor and turbine stator blade thickness；
Self-encoding encoder, to be analyzed to the turbine data for collecting using self-encoding encoder algorithm, extracts turbine stator blade
The pattern feature of dust stratification；
Deep learning network model builds module, to pattern feature and turbine stator blade according to the turbine stator blade dust stratification
Thickness, obtains the deep learning network model of turbine stator blade thickness, and the deep learning network model is with the turbine number
According to be input into, with the turbine stator blade thickness as output；
On-line monitoring network builds module, to based on the deep learning network model, with the described of Real-time Collection
Flat machine data are input, carry out real-time monitoring to turbine stator blade thickness.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610933790.8A CN106570237B (en) | 2016-10-25 | 2016-10-25 | Turbine stator blade thickness monitoring method and system of blast furnace gas waste heat recovery device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610933790.8A CN106570237B (en) | 2016-10-25 | 2016-10-25 | Turbine stator blade thickness monitoring method and system of blast furnace gas waste heat recovery device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106570237A true CN106570237A (en) | 2017-04-19 |
| CN106570237B CN106570237B (en) | 2020-03-17 |
Family
ID=58534248
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610933790.8A Active CN106570237B (en) | 2016-10-25 | 2016-10-25 | Turbine stator blade thickness monitoring method and system of blast furnace gas waste heat recovery device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106570237B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108268359A (en) * | 2017-12-30 | 2018-07-10 | 浙江中睿低碳科技有限公司 | The optimization method of air compression station based on deep learning |
| CN108592812A (en) * | 2018-05-10 | 2018-09-28 | 电子科技大学 | Fan blade optical fiber load strain characteristics extract and crack monitoring method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1303990A (en) * | 2000-01-12 | 2001-07-18 | 三菱重工业株式会社 | Moving vane of turbine |
| US20080111543A1 (en) * | 2006-04-26 | 2008-05-15 | Snecma | Measurement of wall thicknesses, particularly of a blade, by eddy currents |
| CN102004460A (en) * | 2010-11-24 | 2011-04-06 | 东北电力大学 | Online monitoring method for fouling degree of flow passage of steam turbine |
| CN104331553A (en) * | 2014-10-29 | 2015-02-04 | 浙江大学 | Optimal design method of large turbo expander impeller blade structure with defect consideration |
-
2016
- 2016-10-25 CN CN201610933790.8A patent/CN106570237B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1303990A (en) * | 2000-01-12 | 2001-07-18 | 三菱重工业株式会社 | Moving vane of turbine |
| US20080111543A1 (en) * | 2006-04-26 | 2008-05-15 | Snecma | Measurement of wall thicknesses, particularly of a blade, by eddy currents |
| CN102004460A (en) * | 2010-11-24 | 2011-04-06 | 东北电力大学 | Online monitoring method for fouling degree of flow passage of steam turbine |
| CN104331553A (en) * | 2014-10-29 | 2015-02-04 | 浙江大学 | Optimal design method of large turbo expander impeller blade structure with defect consideration |
Non-Patent Citations (3)
| Title |
|---|
| 蔡自兴: "《人工智能及其应用》", 31 July 2016 * |
| 高洪涛等: "基于热力计算的神经网络在汽轮机通流部分热参数诊断中的应用", 《第一届全国诊断工程技术学术会议》 * |
| 高洪涛等: "汽轮机叶片结垢在线诊断的一种新方法", 《大连理工大学学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108268359A (en) * | 2017-12-30 | 2018-07-10 | 浙江中睿低碳科技有限公司 | The optimization method of air compression station based on deep learning |
| CN108592812A (en) * | 2018-05-10 | 2018-09-28 | 电子科技大学 | Fan blade optical fiber load strain characteristics extract and crack monitoring method |
| CN108592812B (en) * | 2018-05-10 | 2019-12-31 | 电子科技大学 | Method for extracting load strain characteristics and monitoring cracks of optical fiber of fan blade |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106570237B (en) | 2020-03-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Trizoglou et al. | Fault detection by an ensemble framework of Extreme Gradient Boosting (XGBoost) in the operation of offshore wind turbines | |
| CN103605757B (en) | High-speed rail power quality data sorting method based on SVM (support vector machine) | |
| CN104006908B (en) | fan energy consumption monitoring method and system | |
| Sun et al. | Fault diagnosis of mechanical equipment in high energy consumption industries in China: A review | |
| Yang et al. | Vibration test of single coal gangue particle directly impacting the metal plate and the study of coal gangue recognition based on vibration signal and stacking integration | |
| CN104865061B (en) | A kind of fatigue life real-time predicting method based on accumulated damage of probability | |
| CN104794550A (en) | WT-KPCA-SVR coupling model based gas emission quantity prediction method | |
| Sun et al. | Assessment of CO2 emission reduction potentials in the Chinese oil and gas extraction industry: From a technical and cost-effective perspective | |
| CN104866923A (en) | Steel enterprise blast furnace by-product gas emergence size prediction method | |
| CN106570237A (en) | Method and system for monitoring stator blade thickness of turbine of blast furnace gas waste heat recovery device | |
| Yin et al. | Dynamic real–time abnormal energy consumption detection and energy efficiency optimization analysis considering uncertainty | |
| CN114386679A (en) | Ball mill energy consumption state monitoring method based on ensemble learning | |
| Sun et al. | Data augmentation strategy for power inverter fault diagnosis based on wasserstein distance and auxiliary classification generative adversarial network | |
| Song et al. | Intelligent diagnosis method for machinery by sequential auto-reorganization of histogram | |
| CN105092457A (en) | Injection-production string corrosion evaluation method under combined action of alternating load and corrosive medium | |
| Wang et al. | A correlation-graph-CNN method for fault diagnosis of wind turbine based on state tracking and data driving model | |
| CN114113773A (en) | Non-invasive load monitoring method based on zero sample learning | |
| CN103294863B (en) | A kind of method according to chemical constitution prediction lubricating base oil wear resistance | |
| Yang et al. | A fault identification method for electric submersible pumps based on dae-svm | |
| Xu et al. | A bran-new performance evaluation model of coal mill based on GA-IFCM-IDHGF method | |
| Ren et al. | Research on fault feature extraction of hydropower units based on adaptive stochastic resonance and fourier decomposition method | |
| CN108709426B (en) | Sintering machine air leakage fault online diagnosis method based on frequency spectrum characteristic bilateral detection method | |
| Liu et al. | Real-time comprehensive health status assessment of hydropower units based on multi-source heterogeneous uncertainty information | |
| CN114048767A (en) | Fault monitoring and early warning method for wind power master control system | |
| Wang et al. | Mass laplacian discriminant analysis and its application in gear fault diagnosis |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| EE01 | Entry into force of recordation of patent licensing contract | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20170419 Assignee: CHANGSHAN XINLONG BEARING Co.,Ltd. Assignor: ZHEJIANG SCI-TECH University Contract record no.: X2022330000079 Denomination of invention: Monitoring method and system of turbine stationary blade thickness of blast furnace gas waste heat recovery device Granted publication date: 20200317 License type: Common License Record date: 20220506 Application publication date: 20170419 Assignee: Zhejiang Zhongjing Bearing Co.,Ltd. Assignor: ZHEJIANG SCI-TECH University Contract record no.: X2022330000080 Denomination of invention: Monitoring method and system of turbine stationary blade thickness of blast furnace gas waste heat recovery device Granted publication date: 20200317 License type: Common License Record date: 20220506 |