CN106529062B - A kind of bridge structure health diagnostic method based on deep learning - Google Patents
A kind of bridge structure health diagnostic method based on deep learning Download PDFInfo
- Publication number
- CN106529062B CN106529062B CN201611029162.3A CN201611029162A CN106529062B CN 106529062 B CN106529062 B CN 106529062B CN 201611029162 A CN201611029162 A CN 201611029162A CN 106529062 B CN106529062 B CN 106529062B
- Authority
- CN
- China
- Prior art keywords
- deep learning
- data
- bridge
- bridge structure
- 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.)
- Expired - Fee Related
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
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- 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
- G06N3/088—Non-supervised learning, e.g. competitive learning
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Geometry (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The present invention provides a kind of bridge structure health diagnostic method based on deep learning, comprising: step 1, data prediction;Step 2 establishes training input of the Deterministic Finite meta-model as deep learning network of bridge structure using the data of design phase and construction stage;Step 3, the deep learning network DN of one five layers of building;Step 4: the layer 5 of fine tuning deep learning network;Another part data input network in the stage that monitors is advanced optimized entire depth learning network by step 5;Step 6, the deep learning network that can be used for structure for health diagnosis formed according to step 5 carries out diagnosing its structural health conditions to any bridge to be normal or is in which rank of deterioration state.Bridge structure health diagnostic method provided by the invention based on deep learning carries out deagnostic test by establishing deep learning network so as to the health status quickly and efficiently to different types of rridges structure.
Description
Technical field
The present invention relates to bridge technology field more particularly to a kind of bridge structure health diagnosis sides based on deep learning
Method.
Background technique
Bridge is the throat of traffic, and the safety problem of bridge structure is related to great property safety even life security, always
All paid much attention to by all circles.With the propulsion of active time, material aging aggravates, in addition the volume of traffic increased increasingly and overload,
It will cause the local damage of bridge structure.The safety of final crisis bridge.However, due to bridge structure complexity, damage influence because
Element is numerous, using traditional diagnostic method often only with the data in a certain stage, it is difficult to obtain effective diagnostic result.Therefore
The Gernral Check-up problem of bridge structure has become generally acknowledged problem.It is equal in design, construction, maintenance, monitoring in view of bridge structure
Mass data is saved, wherein necessarily having contained the various features of bridge structure health state, therefore studies a kind of effective base
It has important practical significance in the bridge structure health diagnostic method of deep learning.
Summary of the invention
The present invention provides a kind of bridge structure health diagnostic method based on deep learning, it is intended to solve due to bridge knot
The problem of structure is complicated, influence factor is numerous, and the health status of bridge structure is difficult to efficient diagnosis.
A kind of bridge structure health diagnostic method based on deep learning, comprising the following steps:
Step 1, data prediction, by more bridge blocks in design, construction, maintenance, the data collection for monitoring four-stage
Afterwards, taxonomic revision is carried out according to data character difference, data fusion is then carried out using big data processing method, form depth
All kinds of input datas needed for practising network;Bridge therein includes the bridge of certain type of different size, different geographical, and contains
There is the case where normal condition and various deterioration states;
Step 2 establishes the Deterministic Finite meta-model of bridge structure using the data of design phase and construction stage, respectively
The case where simulating bridge structure normal condition and various deterioration states, and save training of the related data as deep learning network
Input;
Step 3, the deep learning network DN of one five layers of building, first layer SS use sparse autocoder, second,
Three layers use general autocoder, and the 4th layer of DS uses noise reduction autocoder, and layer 5 SVM uses support vector machines,
Input data is A [i] (i=1,2 ..., n), is formed by input data for step 1, can be Arbitrary Matrix.Output data is S
[j] (j=1,2 ..., m) indicates the specific health status of bridge structure each region and the healthy shape of bridge structure entirety
State can be Arbitrary Matrix;
The data of design, two stages of construction are inputted deep learning network DN, using no prison by step 4, feature learning
Supervise and instruct white silk, obtains the feature of the normal all kinds of parameters of bridge structure;By maintenance phase and monitoring the stage a part of data and
The data of finite element modelling input network, using Training, obtain structure in the spy of all kinds of parameters of various deterioration states
Sign, and finely tune the layer 5 of deep learning network;
Another part data in the stage that monitors are inputted network by step 5, data verification, carry out data verification, and according to
Verification result advanced optimizes entire deep learning network, and different types of rridges structure for health diagnosis can be common to by ultimately forming one
Deep learning network;
Step 6, bridge structure health diagnosis, the deep learning that can be used for structure for health diagnosis formed according to step 5
Network pre-processes any bridge using the method for similar step one to the various data of the bridge, and formation can input number
The deep learning network that the training that succeeded can be input to after is diagnosed, to accurately obtain the structural health shape of the bridge
State is normal or is in which rank of deterioration state.
Further, the bridge structure health diagnostic method based on deep learning as described above, number described in step 1
It include: environment, geology, structure, time according to property.
Further, the bridge structure health diagnostic method based on deep learning as described above, it is bad described in step 2
Change state is divided into slight, slight, intermediate, serious and unacceptable Pyatyi according to bridge structure reliability specification.
Bridge structure health diagnostic method provided by the invention based on deep learning, by establishing deep learning network,
Deagnostic test is carried out so as to the health status quickly and efficiently to different types of rridges structure.
Detailed description of the invention
Fig. 1 is deep learning schematic network structure of the present invention.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, the technical solution below in the present invention carries out clear
Chu is fully described by, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The purpose of the present invention is to provide a kind of bridge structure health diagnostic method based on deep learning, this method include
Following steps:
Step 1, data prediction.By more bridge blocks in design, construction, maintenance, the data collection for monitoring four-stage
Afterwards, taxonomic revision is carried out according to data character (environment, geology, structure, time) is different, then uses big data processing method
Mapreduce carries out data fusion, all kinds of input datas needed for forming following deep learning network.Bridge therein includes certain
The bridge of the different size of seed type, different geographical, and the case where contain normal condition and various deterioration states.
Step 2 establishes the Deterministic Finite meta-model of bridge structure using the data of design phase and construction stage, respectively
The case where simulating bridge structure normal condition and various deterioration states, and save training of the related data as deep learning network
Input.Wherein deterioration state is divided into slight, slight, intermediate, serious and unacceptable Pyatyi according to bridge structure reliability specification.
Step 3, the deep learning network DN of one five layers of building.First layer SS use sparse autocoder, second,
Three layers (respectively GS1, GS2) use general autocoder, and the 4th layer of DS uses noise reduction autocoder, and layer 5 SVM is adopted
With support vector machines (such as Fig. 1).Its input data is A [i] (i=1,2 ..., n), is formed by input data for step 1, can
For Arbitrary Matrix.Output data be S [j] (j=1,2 ..., m), indicate bridge structure each region specific health status and
The health status of bridge structure entirety, can be Arbitrary Matrix.
Step 4, feature learning.The data of design, two stages of construction are inputted into deep learning network, use is unsupervised
Training obtains the feature of the normal all kinds of parameters of bridge structure;By maintenance phase and a part of data in monitoring stage and have
The data of limit member simulation input network, using Training, obtain structure in the feature of all kinds of parameters of various deterioration states,
And finely tune the layer 5 of deep learning network.
Step 5, data verification.Another part data in the stage that monitors are inputted into network, carry out data verification, and according to
Verification result advanced optimizes entire deep learning network.Different types of rridges structure for health diagnosis can be common to by ultimately forming one
Deep learning network.
Step 6, bridge structure health diagnosis.It has passed through step 5 and form the deep learning that can be used for structure for health diagnosis
After network, for any bridge block, when diagnosis, needs to carry out the various data of the bridge using the method for similar step one
Pretreatment, the deep learning network of the training that succeed can be input to after input data by, which being formed, is diagnosed, so that accurate must
Structural health conditions to the bridge are normal or are in which rank of deterioration state.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features;
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (2)
1. a kind of bridge structure health diagnostic method based on deep learning, which comprises the following steps:
Step 1, data prediction, by more bridge blocks design, construction, maintenance, monitor four-stage data collection after, root
Taxonomic revision is carried out according to data character difference, data fusion is then carried out using big data processing method, forms deep learning net
All kinds of input datas needed for network;Bridge therein includes the bridge of certain type of different size, different geographical, and containing just
The case where normal state and various deterioration states;
Step 2 is established the Deterministic Finite meta-model of bridge structure, is simulated respectively using the data of design phase and construction stage
The case where bridge structure normal condition and various deterioration states, and it is defeated as the training of deep learning network to save related data
Enter;
Step 3, deep learning the network DN, first layer SS of one five layers of building use sparse autocoder, second and third layer
Using general autocoder, the 4th layer of DS uses noise reduction autocoder, and layer 5 SVM uses support vector machines, input
Data are A [i] (i=1,2 ..., n), are formed by input data for step 1, can be Arbitrary Matrix;Output data is S [j]
(j=1,2 ..., m) indicates the specific health status of bridge structure each region and the health status of bridge structure entirety, can
To be Arbitrary Matrix;
The data of design, two stages of construction are inputted deep learning network DN, using unsupervised instruction by step 4, feature learning
Practice, obtains the feature of the normal all kinds of parameters of bridge structure;By maintenance phase and monitoring a part of data in stage and limited
The data of member simulation input network, using Training, obtain structure in the feature of all kinds of parameters of various deterioration states, and
Finely tune the layer 5 of deep learning network;
Another part data in the stage that monitors are inputted network, carry out data verification, and according to verifying by step 5, data verification
As a result entire deep learning network is advanced optimized, the depth that can be common to different types of rridges structure for health diagnosis is ultimately formed
Spend learning network;
Step 6, bridge structure health diagnosis, according to step 5 formed the deep learning network that can be used for structure for health diagnosis,
Any bridge pre-processes the various data of the bridge using the method for similar step one, formation can be after input data
The deep learning network that the training that succeeded can be input to is diagnosed, so that the structural health conditions for accurately obtaining the bridge are just
Which rank of deterioration state be often still in;
Data character described in step 1 includes: environment, geology, structure, time.
2. the bridge structure health diagnostic method according to claim 1 based on deep learning, which is characterized in that step 2
Described in deterioration state slight, slight, intermediate, serious and unacceptable Pyatyi is divided into according to bridge structure reliability specification.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611029162.3A CN106529062B (en) | 2016-11-20 | 2016-11-20 | A kind of bridge structure health diagnostic method based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611029162.3A CN106529062B (en) | 2016-11-20 | 2016-11-20 | A kind of bridge structure health diagnostic method based on deep learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106529062A CN106529062A (en) | 2017-03-22 |
CN106529062B true CN106529062B (en) | 2019-10-11 |
Family
ID=58353123
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611029162.3A Expired - Fee Related CN106529062B (en) | 2016-11-20 | 2016-11-20 | A kind of bridge structure health diagnostic method based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106529062B (en) |
Families Citing this family (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110119656A (en) * | 2018-02-07 | 2019-08-13 | 中国石油化工股份有限公司 | Intelligent monitor system and the scene monitoring method violating the regulations of operation field personnel violating the regulations |
CN108511064A (en) * | 2018-02-11 | 2018-09-07 | 河南工程学院 | The system for automatically analyzing healthy data based on deep learning |
CN109102016B (en) * | 2018-08-08 | 2019-05-31 | 北京新桥技术发展有限公司 | A kind of test method for bridge technology situation |
CN109256773B (en) * | 2018-10-19 | 2021-01-26 | 东北大学 | Power system state estimation method of noise reduction self-coding and depth support vector machine |
CN110047145B (en) * | 2019-04-15 | 2023-05-16 | 山东师范大学 | Tissue deformation simulation system and method based on deep learning and finite element modeling |
CN110619345B (en) * | 2019-07-22 | 2022-12-06 | 重庆交通大学 | Cable-stayed bridge monitoring data validity-oriented label reliability comprehensive verification method |
CN111222189A (en) * | 2019-12-31 | 2020-06-02 | 湖南工程学院 | Efficient bridge structure health early warning control system and method |
CN111610406A (en) * | 2020-04-24 | 2020-09-01 | 国网河北省电力有限公司电力科学研究院 | Grounding grid corrosion prediction method based on deep learning |
CN114612473B (en) * | 2022-05-11 | 2022-08-16 | 深圳联和智慧科技有限公司 | Bridge inspection processing method and system based on unmanned aerial vehicle and cloud platform |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2131258A1 (en) * | 2008-06-06 | 2009-12-09 | Inotera Memories, Inc. | A Method for Prognostic Maintenance in Semiconductor Manufacturing Equipments |
CN102663264A (en) * | 2012-04-28 | 2012-09-12 | 北京工商大学 | Semi-supervised synergistic evaluation method for static parameter of health monitoring of bridge structure |
CN103033387A (en) * | 2012-12-28 | 2013-04-10 | 重庆交通大学 | Method for bridge girder dangerous condition diagnosis based on data drive |
CN106017960A (en) * | 2016-06-01 | 2016-10-12 | 清华大学合肥公共安全研究院 | MIS and GIS combination-based bridge safety monitoring system |
-
2016
- 2016-11-20 CN CN201611029162.3A patent/CN106529062B/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2131258A1 (en) * | 2008-06-06 | 2009-12-09 | Inotera Memories, Inc. | A Method for Prognostic Maintenance in Semiconductor Manufacturing Equipments |
CN102663264A (en) * | 2012-04-28 | 2012-09-12 | 北京工商大学 | Semi-supervised synergistic evaluation method for static parameter of health monitoring of bridge structure |
CN103033387A (en) * | 2012-12-28 | 2013-04-10 | 重庆交通大学 | Method for bridge girder dangerous condition diagnosis based on data drive |
CN106017960A (en) * | 2016-06-01 | 2016-10-12 | 清华大学合肥公共安全研究院 | MIS and GIS combination-based bridge safety monitoring system |
Also Published As
Publication number | Publication date |
---|---|
CN106529062A (en) | 2017-03-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106529062B (en) | A kind of bridge structure health diagnostic method based on deep learning | |
EP3447642B1 (en) | System and method for predicting application performance for large data size on big data cluster | |
US20160004627A1 (en) | Utilizing semantic clusters to Predict Software defects | |
US11341034B2 (en) | Analysis of verification parameters for training reduction | |
CA2783492C (en) | Associative memory visual evaluation tool | |
EP3751787A1 (en) | Techniques to generate network simulation scenarios | |
CN110221975A (en) | Create the method and device of interface use-case automatic test script | |
CN105471647B (en) | A kind of power communication network fault positioning method | |
CN103885767B (en) | System and method used for geographical area correlated websites | |
CN106294174B (en) | The various dimensions measure and device of testing adequacy | |
US20180336494A1 (en) | Translating sensor input into expertise | |
US9959182B2 (en) | Replicating test case data into a cache with non-naturally aligned data boundaries | |
CN110262961A (en) | Test method, device, storage medium and the terminal device of Workflow Management System | |
Gao et al. | On the variability of software engineering needs for deep learning: Stages, trends, and application types | |
CN113127955A (en) | Building anti-seismic performance evaluation method, system, device and storage medium | |
CN110059117A (en) | A kind of analysis and processing method and device of big data of imparting knowledge to students | |
Wang et al. | Improving the efficiency of functional verification based on test prioritization | |
CN106200353B (en) | A method of the clock for testing twin-core electric energy meter | |
CN106445973A (en) | Search engine monitoring method and apparatus | |
EP4285268A1 (en) | Generating learned representations of digital circuit designs | |
CN108875374A (en) | Malice PDF detection method and device based on document node type | |
CN111125005B (en) | Method, system and equipment for optimizing IO performance of HDFS distributed file system | |
US11461665B2 (en) | Systems and methods of a Boolean network development environment | |
US6829572B2 (en) | Method and system for efficiently overriding array net values in a logic simulator machine | |
US20200159872A1 (en) | Intelligent Fail Recognition |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | 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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20191011 Termination date: 20201120 |
|
CF01 | Termination of patent right due to non-payment of annual fee |