CN109556849A - Simply supported girder bridge structure damage monitoring system based on machine learning - Google Patents
Simply supported girder bridge structure damage monitoring system based on machine learning Download PDFInfo
- Publication number
- CN109556849A CN109556849A CN201811544783.4A CN201811544783A CN109556849A CN 109556849 A CN109556849 A CN 109556849A CN 201811544783 A CN201811544783 A CN 201811544783A CN 109556849 A CN109556849 A CN 109556849A
- Authority
- CN
- China
- Prior art keywords
- data
- dynamic strain
- simply supported
- monitoring
- girder bridge
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Bridges Or Land Bridges (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The present invention relates to bridge structure real-time monitoring systems to be related to a kind of structure damage monitoring system and corresponding structure monitoring sensor arrangement method and Structural Damage Assessment method particular for simply supported girder bridge.Dynamic strain sensor is arranged across, 3/4ths cross-locations across, half by a quarter of the corresponding single plate beam beam bottom of wheel trajectories when each lane loaded vehicle of simply supported girder bridge passes through, dynamic response situation when monitoring loaded vehicle passes through;Smooth, noise reduction process is carried out to data by low-pass filtering;Using K mean cluster algorithm rejecting abnormalities data;Clustering is carried out to monitoring big data using machine learning techniques on this basis, the lateral connection situation between beam slab is assessed according to the separate condition of cluster.The present invention can effectively monitor the lateral connection damage of simply supported girder bridge, the case where issuing early warning in time, avoid single-beam stress.
Description
Technical field
The present invention relates to bridge structure real-time monitoring systems to be related to a kind of based on machine learning particular for simply supported girder bridge
Structure damage monitoring system and corresponding structure monitoring sensor arrangement method and Structural Damage Assessment method.
Background technique
Related data show, the bridge of China's Mid and minor spans accounts for 80% or so of the existing bridge total amount in China, and simply supported girder bridge
It is an important component of Mid and minor spans concrete-bridge again.With the extension of simply supported girder bridge service phase, traffic loading etc.
The raising of grade, the simply supported girder bridge bearing capacity built in early days are no longer satisfied the demand of Modern Traffic.Currently, in-service freely-supported
Beam bridge quite a few had already appeared in various degree and various forms of diseases, and lateral connection failure be simply supported girder bridge the most
One of main disease.The typical feature of lateral connection failure is that junction steel plate surface concrete peels off and junction steel plate weld seam is opened
It splits, prevent bearing beam is from effectively transmitting load to non-stress beam.The failure of simply supported girder bridge lateral connection can reduce bridge
The globality of structure influences the cross direction profiles rule of load, stress beam deflection is caused to increase, and non-stress beam deflection becomes smaller, so that
It cannot cooperate well between bearing beam and non-stress beam, bearing beam is easy to appear destruction.Will appear when serious " single-beam by
The case where power ", buries huge security risk to communications and transportation.Currently, there is no a set of effective, system simply supported beams
Bridge lateral connection failure monitor method.Therefore, the monitoring for carrying out the damage of simply supported girder bridge lateral connection has a very important significance.
Summary of the invention
In order to solve the above technical problem, the present invention provides a kind of dynamic strain to monitor system, in the specific bit of simply supported girder bridge
Installation dynamic strain sensor is set, dynamic response situation when being passed through by monitoring loaded vehicle, and application machine learning techniques are to monitoring
Big data carries out clustering, can monitor simply supported girder bridge dynamic strain lateral transport situation, assesses Bridge Structural Damage situation.
Simple beam structure damage monitoring system based on machine learning, the technical solution adopted is that: it include data acquisition
System and data process subsystem.
The data acquisition subsystem includes dynamic strain sensor and Multi-channel data acquisition equipment, data processing
System includes database, data preprocessing module and data analysis module.The output of the dynamic strain sensor is connected to multi-pass
Track data acquires the input of equipment, and the output of the Multi-channel data acquisition equipment is connected to the input of database, the data
Library is connected with data preprocessing module and data analysis module respectively.
The data acquisition subsystem, the corresponding single plate beam beam of wheel trajectories when each lane loaded vehicle of simply supported girder bridge passes through
The a quarter at bottom arranges dynamic strain sensor across, 3/4ths cross-locations across, half, passes through for monitoring single loaded vehicle
When dynamic strain situation.For each lane, two dynamic strain sensors of each cross sectional arrangement, to the freely-supported for having N number of lane
Beam bridge, the number of each cross section dynamic strain sensor are 2N.Often across there are three monitoring cross sections, thus often across number of probes
For 6N.
The Multi-channel data acquisition equipment carries out at signal condition and low-pass filtering collected dynamic strain signal
Reason, to reduce signal noise.
The database is responsible for data collection and storage, analyzes including initial data, the result of data prediction and data
Result.
The data preprocessing module uses K mean cluster algorithm by data clusters for valid data and invalid data, and
Rejecting abnormalities data, so that it is guaranteed that the validity of data.
The K mean cluster algorithm, by the acquisition data of dynamic strain sensorAs training sample,For cluster label, whereinIndicate valid data,Indicate invalid data.ClusterDispersion situation
For, whereinFor clusterCenter,To belong to clusterTotal sample number.For two clusters,
The sum of dispersion value determines the cluster label belonging to it, i.e. formula when minimumIt is minimum.
The data analysis module carries out clustering, data analysis to monitoring big data using machine learning techniques
It is divided into three phases:
S1: the monitoring data of each dynamic strain sensor when loading to single each time loaded vehicle extract its peak value.For cross section k,
Sensor peak-data is expressed as matrix, wherein every a line corresponds to a lane:
S2: dynamic strain peak-data when each lane loaded vehicle is loaded is divided into M section;
S3: clustering is carried out using machine learning techniques.For loaded vehicle by lane n () the case where, needle
To m () a dynamic strain peak value section, extract the dynamic strain peak-data in all lanes cross section k.For
Lane i() jth () a sensor, data are a set.To the number
Clustering is carried out according to collection, and according to the separate condition of cluster, judges the lateral connection shape between the beam slab of lane i and other beam slabs
Condition.
Preferably, the loaded vehicle selects 20 tons or more of vehicle.
The invention has the advantages that arranging dynamic strain sensor by the key position in simply supported girder bridge, single is monitored
The dynamic strain situation of beam slab under loaded vehicle load action, and monitoring big data is analyzed by the method for machine learning, it can
Lateral connection situation effectively between assessment beam slab, fails for lateral connection and issues early warning in time, avoid the feelings of single-beam stress
Condition.
Detailed description of the invention
Fig. 1 is veneer simply supported beam dynamic strain sensor layout schematic diagram.
Fig. 2 is simply supported girder bridge dynamic strain sensor integral installation arrangement sectional view.
Fig. 3 is simply supported girder bridge structure damage monitoring system principle diagram of the present invention.
Fig. 4 is the acquisition of structure monitoring data, convergence and Analysis of Structural Damage flow chart.
Fig. 5~Fig. 8 is big data cluster analysis result schematic diagram.
Specific embodiment
Below in conjunction with attached drawing, the present invention will be described in detail.
Veneer simply supported beam dynamic strain sensor arrangement in the present invention is as shown in Figure 1, simply supported girder bridge dynamic strain sensor is whole
Mounting arrangements section as shown in Fig. 2, system principle diagram as shown in figure 3, Fig. 4 be structure monitoring data acquisition, convergence and structure
Breakdown diagnosis flow chart.
As shown in figure 3, system includes data acquisition subsystem and data process subsystem.
The data acquisition subsystem includes dynamic strain sensor and Multi-channel data acquisition equipment, data processing
System includes database, data preprocessing module and data analysis module.The output of the dynamic strain sensor is connected to multi-pass
Track data acquires the input of equipment, and the output of the Multi-channel data acquisition equipment is connected to the input of database, the data
Library is connected with data preprocessing module and data analysis module respectively.
The data acquisition subsystem, the corresponding single plate beam 1 of wheel trajectories when each lane loaded vehicle of simply supported girder bridge passes through
The a quarter of beam bottom arranges dynamic strain sensor 2 across, 3/4ths cross-locations across, half, for monitoring single loaded vehicle
By when dynamic strain situation.For each lane 3, two dynamic strain sensors of each cross sectional arrangement, to there is N number of lane
Simply supported girder bridge, the number of each cross section dynamic strain sensor are 2N.Often across there are three monitoring cross sections, thus often across sensor
Number is 6N.The loaded vehicle selects 20 tons or more of vehicle.
The Multi-channel data acquisition equipment carries out at signal condition and low-pass filtering collected dynamic strain signal
Reason, to reduce signal noise.
The database is responsible for data collection and storage, analyzes including initial data, the result of data prediction and data
Result.
The data preprocessing module uses K mean cluster algorithm by data clusters for valid data and invalid data, and
Rejecting abnormalities data, so that it is guaranteed that the validity of data.
The K mean cluster algorithm, by the acquisition data of dynamic strain sensorAs training sample,For cluster label, whereinIndicate valid data,Indicate invalid data.ClusterDispersion situation
For,For clusterCenter,To belong to clusterTotal sample number.For two clusters, dispersion
The sum of value determines the cluster label belonging to it, i.e. formula when minimumIt is minimum.
The data analysis module carries out clustering, data analysis to monitoring big data using machine learning techniques
It is divided into three phases:
S1: the monitoring data of each dynamic strain sensor when loading to single each time loaded vehicle extract its peak value.For cross section k,
Sensor peak-data is expressed as matrix, wherein every a line corresponds to a lane:
S2: dynamic strain peak-data when each lane loaded vehicle is loaded is divided into M section;
S3: clustering is carried out using machine learning techniques.For loaded vehicle by lane n () the case where, needle
To m () a dynamic strain peak value section, extract the dynamic strain peak-data in all lanes cross section k.It is right
In lane i() jth () a sensor, data are a set.To this
Data set carries out clustering, and according to the separate condition of cluster, judges the lateral connection between the beam slab of lane i and other beam slabs
The situation of variation.If cluster result as shown in figure 5, if illustrate that lateral connection is all right.If there is Fig. 6~Fig. 8 institute
The case where showing, then there are lateral connection damages for explanation.Fig. 6 illustrates occur paroxysmal lateral connection at some time point between beam slab
Damage, Fig. 8 shows lateral connection damages slowly to accumulate, gradually deepens.The case where Fig. 7, illustrates lateral connection situation by one
Stable state reaches another stable state through damage accumulation after a period of time.
The present invention is not limited to the above embodiments, made any to above embodiment aobvious of those skilled in the art and
The improvement or change being clear to, all protection scope without departing from design of the invention and appended claims.
Claims (2)
1. the simply supported girder bridge structure damage monitoring system based on machine learning includes data acquisition subsystem and data processing subsystem
System, it is characterised in that:
The data acquisition subsystem includes dynamic strain sensor and Multi-channel data acquisition equipment, the data process subsystem
Including database, data preprocessing module and data analysis module;The output of the dynamic strain sensor is connected to multichannel number
According to the input of acquisition equipment, the output of the Multi-channel data acquisition equipment is connected to the input of database, the database point
It is not connected with data preprocessing module and data analysis module;
The data acquisition subsystem, the corresponding single plate beam beam bottom of wheel trajectories when each lane loaded vehicle of simply supported girder bridge passes through
A quarter arranges dynamic strain sensor across, 3/4ths cross-locations across, half, for monitoring when single loaded vehicle passes through
Dynamic strain situation;
The Multi-channel data acquisition equipment carries out signal condition and low-pass filtering treatment to collected dynamic strain signal, from
And reduce signal noise;
The database is responsible for data collection and storage, the knot including the analysis of initial data, the result of data prediction and data
Fruit;
The data preprocessing module uses K mean cluster algorithm by data clusters for valid data and invalid data, and rejects
Abnormal data, so that it is guaranteed that the validity of data;
The data analysis module carries out clustering to monitoring big data using machine learning techniques, and data are divided into
Three phases:
S1: the monitoring data of each dynamic strain sensor when loading to single each time loaded vehicle extract its peak value, for cross section k,
Sensor peak-data is expressed as matrix, wherein every a line corresponds to a lane:
S2: dynamic strain peak-data when each lane loaded vehicle is loaded is divided into M section;
S3: using machine learning techniques carry out clustering: for loaded vehicle by lane n () the case where, for
M () a dynamic strain peak value section, extract the dynamic strain peak-data in all lanes cross section k;For vehicle
Road i() jth () a sensor, data are a set, to the data set
Clustering is carried out, and according to the separate condition of cluster, judges the lateral connection situation between the beam slab of lane i and other beam slabs.
2. the simply supported girder bridge structure damage monitoring system according to claim 1 based on machine learning, which is characterized in that institute
State the vehicle that loaded vehicle is 20 tons or more.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811544783.4A CN109556849A (en) | 2018-12-17 | 2018-12-17 | Simply supported girder bridge structure damage monitoring system based on machine learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811544783.4A CN109556849A (en) | 2018-12-17 | 2018-12-17 | Simply supported girder bridge structure damage monitoring system based on machine learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109556849A true CN109556849A (en) | 2019-04-02 |
Family
ID=65870293
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811544783.4A Pending CN109556849A (en) | 2018-12-17 | 2018-12-17 | Simply supported girder bridge structure damage monitoring system based on machine learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109556849A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110210149A (en) * | 2019-06-06 | 2019-09-06 | 交通运输部公路科学研究所 | A kind of road internal stress, strain Dynamic Response Information obtain system and method |
CN110274537A (en) * | 2019-07-20 | 2019-09-24 | 交通运输部公路科学研究所 | Can cooperated computing the synchronous dynamic strain sensor of intelligent multi-channel |
CN111749092A (en) * | 2020-06-02 | 2020-10-09 | 交通运输部公路科学研究所 | Porous asphalt pavement flying disease detection method based on noise signal identification |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101051334A (en) * | 2006-04-06 | 2007-10-10 | 香港理工大学 | Structure health monitoring and information managing system and its method |
CN101281117A (en) * | 2008-05-29 | 2008-10-08 | 上海交通大学 | Wide span rail traffic bridge damnification recognition method |
CN101281116A (en) * | 2008-05-29 | 2008-10-08 | 上海交通大学 | Wide span rail traffic bridge damnification detecting system |
CN101763053A (en) * | 2008-12-26 | 2010-06-30 | 上海交技发展股份有限公司 | Movable type bridge security detection and analysis management system |
CN106650113A (en) * | 2016-12-26 | 2017-05-10 | 招商局重庆交通科研设计院有限公司 | Method for recognizing abnormal condition of bridge monitoring data based on fuzzy clustering |
CN108229568A (en) * | 2018-01-09 | 2018-06-29 | 上海海事大学 | Gantry crane metal structure loaded-up condition detection method based on K mean cluster algorithm |
-
2018
- 2018-12-17 CN CN201811544783.4A patent/CN109556849A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101051334A (en) * | 2006-04-06 | 2007-10-10 | 香港理工大学 | Structure health monitoring and information managing system and its method |
CN101281117A (en) * | 2008-05-29 | 2008-10-08 | 上海交通大学 | Wide span rail traffic bridge damnification recognition method |
CN101281116A (en) * | 2008-05-29 | 2008-10-08 | 上海交通大学 | Wide span rail traffic bridge damnification detecting system |
CN101763053A (en) * | 2008-12-26 | 2010-06-30 | 上海交技发展股份有限公司 | Movable type bridge security detection and analysis management system |
CN106650113A (en) * | 2016-12-26 | 2017-05-10 | 招商局重庆交通科研设计院有限公司 | Method for recognizing abnormal condition of bridge monitoring data based on fuzzy clustering |
CN108229568A (en) * | 2018-01-09 | 2018-06-29 | 上海海事大学 | Gantry crane metal structure loaded-up condition detection method based on K mean cluster algorithm |
Non-Patent Citations (2)
Title |
---|
刘纲: ""基于长期静态监测数据的大型桥梁安全状态评估方法研究"", 《中国博士学位论文全文数据库 工程科技II辑》 * |
王潇碧: ""移动荷载作用下桥梁损伤识别模型试验研究"", 《中国优秀硕士学位论文全文数据库库 工程科技II辑》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110210149A (en) * | 2019-06-06 | 2019-09-06 | 交通运输部公路科学研究所 | A kind of road internal stress, strain Dynamic Response Information obtain system and method |
CN110210149B (en) * | 2019-06-06 | 2023-01-31 | 交通运输部公路科学研究所 | System and method for acquiring dynamic response information of stress and strain in road |
CN110274537A (en) * | 2019-07-20 | 2019-09-24 | 交通运输部公路科学研究所 | Can cooperated computing the synchronous dynamic strain sensor of intelligent multi-channel |
CN110274537B (en) * | 2019-07-20 | 2024-04-02 | 交通运输部公路科学研究所 | Intelligent multichannel synchronous dynamic strain sensor capable of being cooperatively calculated |
CN111749092A (en) * | 2020-06-02 | 2020-10-09 | 交通运输部公路科学研究所 | Porous asphalt pavement flying disease detection method based on noise signal identification |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109556849A (en) | Simply supported girder bridge structure damage monitoring system based on machine learning | |
CN111256924B (en) | Intelligent monitoring method for expansion joint of large-span high-speed railway bridge | |
US5529267A (en) | Railway structure hazard predictor | |
CN111710165B (en) | Bridge supervision and early warning method and system based on multi-source monitoring data fusion and sharing | |
CN108021732B (en) | Online damage early warning method for modular expansion joint of cable-supported bridge | |
CN102735320B (en) | Method for identifying weights of cars based on dynamic strain of bridges | |
CN107609989A (en) | A kind of bridge health monitoring intelligence CS architecture systems of road network level | |
CN107391811A (en) | Steel truss bridge member checking method and system | |
Al-Khateeb et al. | Computing continuous load rating factors for bridges using structural health monitoring data | |
CN112818444A (en) | Railway concrete bridge linear real-time control method based on operation and driving safety | |
CN114112103A (en) | Plate-type ballastless track and all-line temperature field monitoring system and health monitoring method thereof | |
CN209495764U (en) | A kind of dynamic vehicle weighting platform structure | |
CN206399672U (en) | Motor train unit bogie unstability detecting system | |
CN109556847A (en) | A kind of novel simply supported girder bridge structure damage monitoring system | |
CN112626944B (en) | Monitoring method and system for beam-end telescopic device of long-span railway bridge | |
CN205748953U (en) | It is applicable to the fault pre-alarming device of point machine | |
McElwain et al. | Experimental verification of horizontally curved I-girder bridge behavior | |
CN207147423U (en) | One kind hinge seam detecting system | |
CN109556848A (en) | A kind of simply supported girder bridge structure damage monitoring system based on Transverse Distribution | |
CN113624313A (en) | Dynamic weighing method, equipment, system and storage medium for parallel vehicles | |
CN110987499A (en) | Bridge dynamic load test method | |
CN216207462U (en) | Bolt looseness detection system for bridge expansion joint | |
Zhu et al. | Investigation on the pattern for train-induced strains of the long-span steel truss railway bridge | |
CN114333331A (en) | Method and system for identifying vehicle passing information and vehicle weight of multi-lane bridge | |
Ieng et al. | Analysis of B-WIM signals acquired in Millau orthotropic viaduct using statistical classification |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190402 |