CN115452282A - Railway bridge structure health monitoring method based on data fusion - Google Patents
Railway bridge structure health monitoring method based on data fusion Download PDFInfo
- Publication number
- CN115452282A CN115452282A CN202211074817.4A CN202211074817A CN115452282A CN 115452282 A CN115452282 A CN 115452282A CN 202211074817 A CN202211074817 A CN 202211074817A CN 115452282 A CN115452282 A CN 115452282A
- Authority
- CN
- China
- Prior art keywords
- strain
- deflection
- bridge
- matrix
- bridge structure
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 47
- 230000004927 fusion Effects 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 title claims abstract description 35
- 230000036541 health Effects 0.000 title claims abstract description 29
- 239000011159 matrix material Substances 0.000 claims abstract description 53
- 239000013598 vector Substances 0.000 claims abstract description 42
- 230000006378 damage Effects 0.000 claims abstract description 35
- 230000004044 response Effects 0.000 claims abstract description 33
- 238000012549 training Methods 0.000 claims abstract description 29
- 238000000611 regression analysis Methods 0.000 claims abstract description 11
- 238000000605 extraction Methods 0.000 claims abstract description 5
- 238000001514 detection method Methods 0.000 claims description 8
- 238000006073 displacement reaction Methods 0.000 claims description 6
- 230000000694 effects Effects 0.000 claims description 6
- 238000005259 measurement Methods 0.000 claims description 5
- 230000009471 action Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000012935 Averaging Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 7
- 238000004422 calculation algorithm Methods 0.000 description 5
- 230000000873 masking effect Effects 0.000 description 5
- 230000007613 environmental effect Effects 0.000 description 4
- 230000002829 reductive effect Effects 0.000 description 4
- 238000004088 simulation Methods 0.000 description 3
- 229910000831 Steel Inorganic materials 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 239000010959 steel Substances 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 239000008186 active pharmaceutical agent Substances 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
Images
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
- G01M5/00—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings
- G01M5/0008—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings of bridges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M5/00—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings
- G01M5/0041—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings by determining deflection or stress
- G01M5/005—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings by determining deflection or stress by means of external apparatus, e.g. test benches or portable test systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a railway bridge structure health monitoring method based on data fusion, which comprises the following steps: acquiring a strain-deflection training data set of a railway bridge structure, and extracting an initial characteristic matrix of strain and deflection; performing stable regression analysis considering pairwise combination of strain-deflection measuring points to obtain a three-dimensional regression coefficient matrix; performing first-time feature fusion to obtain residual error feature vectors corresponding to the strain measuring points; calculating a covariance matrix and a mean vector of the residual error feature vector, performing second feature fusion based on the Mahalanobis distance according to the covariance matrix and the mean vector, and calculating a control limit of the distance feature; acquiring new strain-deflection monitoring data, and obtaining distance characteristics of the strain-deflection monitoring data corresponding to each strain measuring point by adopting the same method through initial characteristic extraction and twice characteristic fusion; and judging and positioning the local damage of the railway bridge structure according to the distance characteristics and the control limit. The method fuses the vehicle-induced vibration response data of the strain-deflection measuring points, and enhances the detectability of the local damage of the railway bridge structure.
Description
Technical Field
The invention belongs to the technical field of structural health monitoring, and relates to a railway bridge structural health monitoring method based on data fusion.
Background
The railway bridge is an important carrier and key nodes for the safe operation of a railway transportation network. As the speed and the transportation volume of the train increase, the operation pressure of the infrastructure increases; meanwhile, during the service period of dozens of years and hundreds of years, the infrastructures are continuously aged, and the damage accumulation and the performance degradation of the bridge structure are caused by the combined action of factors such as natural disasters, fatigue, corrosion and the like, so that the risk of failure and even collapse is increased. In the last two decades, structural health monitoring technology has developed rapidly with the goal of monitoring the operational status of the engineering infrastructure in real time and identifying damage before it has accumulated to a point where performance has degraded significantly to provide a preventative maintenance strategy. These efforts make it possible to compensate for the deficiencies of visual inspection and are particularly useful for the assessment of the condition of bridge structures whose operation is difficult to interrupt.
For railway bridges, especially old bridges in middle and later service periods, providing indexes representing the operation safety and stability of bridges and trains is a key task for implementing health monitoring. Currently, safety indexes based on monitoring information can be divided into two types: global indicators and local indicators. The overall indexes mainly comprise static and dynamic deflection of the bridge, modal frequency and the like; the indexes are suitable for evaluating the running stability of the vehicle-bridge coupling system, but the local damage of the structure is difficult to identify, and the indexes are interfered by the environment and the vehicle load factor, so that the caused masking effect covers the change of the original indexes, and further the practicability of the indexes is influenced. The local indexes mainly comprise stress, fatigue and the like of vulnerable parts which are mainly monitored by the bridge structure, and the indexes are generally obtained by calculating data measured by a stress/strain sensor which is more stable in test; although these indicators are more sensitive to damage to the component, they are still adversely affected by "masking effects" caused by environmental and vehicular loading factors, which can result in localized damage effects that are difficult to detect.
The structural health monitoring system acquires data through various sensors which are arranged on a structure or are not in contact with the structure, and efficient processing and full utilization of heterogeneous monitoring data with discrete sources are always a hotspot problem in the field. At present, the advanced method is to adopt an information processing technology represented by cloud computing to store and manage monitoring data in an integrated and automatic manner, so as to reduce the hardware cost of the monitoring system and improve the efficiency and compatibility of the system. However, the current software platform is weak in analyzing and interpreting the monitoring data. For railway bridges, the overall and local indexes are not enough to realize the detection of the early damage of the structure, and a data analysis algorithm which can fully utilize bridge health monitoring information to enhance the detectability of the early damage of the structure is lacked.
Disclosure of Invention
The invention aims to solve the problems that health monitoring data of a railway bridge structure is easily affected by variable factors such as environment and vehicle load and the like, and early damage detection of the structure is difficult to realize, and a monitoring system software platform lacks a data analysis algorithm for fully utilizing heterogeneous data, and provides a railway bridge structure health monitoring method based on data fusion.
The invention provides a railway bridge structure health monitoring method based on data fusion, which comprises the following steps:
step 1: acquiring a strain-deflection training data set of a railway bridge structure, and extracting a strain initial characteristic matrix and a deflection initial characteristic matrix of the strain-deflection training data set;
and 2, step: performing stable regression analysis considering pairwise combination of the strain-deflection measuring points to obtain a three-dimensional regression coefficient matrix of the strain-deflection training data set;
and 3, step 3: performing first feature fusion based on robust regression analysis to obtain residual error feature vectors of strain-deflection training data sets corresponding to each strain measuring point;
and 4, step 4: calculating a covariance matrix and a mean vector of the residual error feature vector;
and 5: performing secondary feature fusion based on the Mahalanobis distance according to the covariance matrix and the mean vector to obtain distance features of each strain measuring point corresponding to the strain-deflection training data set;
step 6: calculating a control limit of the distance characteristic;
and 7: acquiring new strain-deflection monitoring data, and obtaining distance characteristics of the strain-deflection monitoring data corresponding to each strain measuring point by adopting the same method of the steps 1, 3 and 5 through initial characteristic extraction and twice characteristic fusion; and judging and positioning the local damage of the railway bridge structure according to the distance characteristics and the control limit obtained in the step 6.
The invention discloses a railway bridge structure health monitoring method based on data fusion, which at least has the following beneficial effects:
firstly, the method makes full use of strain and deflection response data with discrete sources and isomerism, and realizes the fusion of the strain and deflection response data excited by the train bridge crossing, wherein the strain is related to the local performance of the structure, and the deflection represents the whole behavior of the structure. The whole process involves two times of feature fusion on the basis of initial feature extraction: the first time of feature fusion is based on that when a train passes a bridge, the load effect of different measuring point positions on the bridge has correlation, the second time of feature fusion is based on distance-based clustering of high-dimensional features, and the two times of fusion have clear physical significance. The method can eliminate the masking effect caused by the environment and the vehicle load, and enhances the detectability of the local small damage of the railway bridge structure.
Secondly, the railway bridge structure health monitoring method based on data fusion can be synchronously executed with data acquisition and control of the structure health monitoring cloud platform. The following advantages are provided to facilitate preventive maintenance of the railroad bridge:
1. the data required by the method is easily acquired response data of strain and deflection excited by the train bridge crossing, the time history data has strong regularity and large amplitude, and time-varying influence factors are few;
2. the proposed method is unsupervised and does not require tag data of a structural damage state;
3. the extracted initial characteristics are integration/summation of dynamic strain and deflection response data, and a subsequent algorithm only processes the initial characteristics, so that the data storage capacity is reduced;
4. the two feature fusions based on the robust regression analysis and the mahalanobis distance are both insensitive to parameters;
5. in both the training stage and the detection stage, the method only needs a small amount of computing resources, and meets the data processing requirement of online early warning of structural health monitoring;
6. the method of the invention does not depend on a physical model of a structure, and has lower case dependence.
Drawings
FIG. 1 is a flow chart of a method for monitoring health of a railroad bridge structure based on data fusion according to the present invention;
FIGS. 2 (a) - (b) are numerical calculations of a steel single span bowstring arch bridge according to an embodiment of the present invention; wherein, 2 (a) is a three-dimensional finite element model diagram of the bridge, and 2 (b) is a simplified constraint schematic diagram of the bridge bearing;
3 (a) - (b) are schematic diagrams of strain and deflection sensor placement and damage location for one embodiment of the present invention; wherein, FIG. 3 (a) is a south side view of the bridge, and FIG. 3 (b) is a north side view of the bridge;
FIG. 4 is a graph of the primary input data for environmental and vehicle load effect simulation in accordance with one embodiment of the present invention;
FIGS. 5 (a) - (d) are typical strain and deflection responses under the influence of environmental and vehicle loading factors for one embodiment of the present invention; wherein, figure 5 (a) the strain response of the strain measuring point s4 of the Suburban type train under the conditions of the speed of passing a bridge of 107km/h and the temperature of 18.2 ℃; FIG. 5 (b) the deflection response of a deflection measuring point d5 of a Suburban type train under the conditions of a bridge passing speed of 107km/h and a temperature of 18.2 ℃; FIG. 5 (c) the strain response of the strain measuring point s6 of the Intercity train at the bridge crossing speed of 50km/h and the temperature of 26 ℃; FIG. 5 (d) deflection response of a deflection measuring point d5 of an Intercity type train under the conditions of a bridge passing speed of 50km/h and a temperature of 26 ℃;
FIG. 6 shows initial feature data corresponding to two strain measurement points s9 and s17 according to an embodiment of the present invention;
FIGS. 7 (a) - (b) are boxplots of initial characteristic data for all strain and deflection points for one embodiment of the present invention; wherein, fig. 7 (a) is a strain initial characteristic, and fig. 7 (b) is a deflection initial characteristic;
FIG. 8 shows distance characteristic data corresponding to strain measurement points s9 and strain measurement points s17 according to an embodiment of the present invention;
FIG. 9 is a box plot of distance signature data for all strain points in one embodiment of the present invention.
Detailed Description
As shown in FIG. 1, the method for monitoring the health of the railroad bridge structure based on data fusion comprises the following steps:
step 1: acquiring a strain-deflection training data set of a railway bridge structure, and extracting a strain initial characteristic matrix and a deflection initial characteristic matrix of the strain-deflection training data set, wherein the step 1 specifically comprises the following steps:
step 1.1: aiming at a railway bridge structure with m strain measuring points and n deflection measuring points, according to the current operation state, the regular detection and the health monitoring history of the bridge, k acquired under the undamaged state of the bridge structure B Combining the strain and deflection response data excited by the secondary train passing through the bridge as a strain-deflection training data set;
step 1.2: the strain response measured by the kth train strain measuring point i is epsilon i,k (t)，i＝1,2,…,m，k＝1,2,…,k B Taking the integral of the strain response time course as the initial strain characteristic SF i,k ：
In the formula: t is k The total time of the action of the load effect of the passing bridge of the kth train;
strain-deflection training data set, total k B The secondary train bridge crossing can obtain a strain initial characteristic matrix SFM of m strain measuring points:
step 1.3: the deflection response measured by the deflection measuring point j is delta when the kth train passes a bridge j,k (t), j =1,2, …, n, and using the deflection response time-course integral as the deflection initial characteristic DF j,k ：
Strain-deflection training data set, total k B The secondary train bridge crossing can obtain a deflection initial characteristic matrix DFM of n deflection measuring points:
step 2: performing robust regression analysis considering pairwise combination of strain-deflection measuring points to obtain a three-dimensional regression coefficient matrix of the strain-deflection training data set, wherein the step 2 specifically comprises the following steps of:
step 2.1: in the strain initial characteristic matrix SFM and the deflection initial characteristic matrix DFM, the ith row of the SFM matrix and the jth row of the DFM matrix corresponding to the strain-deflection measuring point combination (i, j) are picked up, and the two rows of data are subjected to robust regression analysis to obtain a regression coefficient vector CV corresponding to the strain-deflection measuring point combination i,j ∈R 1×2 ；
Step 2.2: considering pairwise combination of m strain measuring points and n deflection measuring points, and obtaining a three-dimensional regression coefficient matrix CM (element for R) of the strain-deflection training data set m×n×2 。
During specific implementation, a strain measuring point and a deflection measuring point for acquiring strain and deflection response data are respectively composed of a strain sensor for measuring the stress/strain of a vulnerable part of a bridge structure and a displacement sensor for measuring the deflection of the bridge; the strain sensor and the displacement sensor are in normal working states before and after the train passes through the bridge.
And step 3: performing first feature fusion based on robust regression analysis to obtain residual error feature vectors corresponding to each strain measuring point of the strain-deflection training data set, wherein the method specifically comprises the following steps:
step 3.1: according to the step 1, the strain initial characteristic vector SFV of m strain measuring points can be obtained by passing the bridge of the kth train k ：
SFV k ＝{SF 1,k SF 2,k … SF m-1,k SF m,k } T (5)
Step 3.2: according to the step 1, the deflection initial characteristic vector DFV of n deflection measuring points can be obtained by passing the bridge of the k-th train k ：
DFV k ＝{DF 1,k DF 2,k … DF n-1,k DF n,k } T (6)
Step 3.3: taking a two-dimensional regression coefficient matrix CM corresponding to a strain measuring point i in the three-dimensional regression coefficient matrix CM i ∈R n ×2 Multiplying it by the matrix [ I ] point by point n ,DFV k ]Summing the obtained matrix rows to obtain a regression estimation vector aiming at the strain measuring point i
I n ∈R n×1 Is a unit column vector, and an upper corner mark e represents an estimated value; computing residual eigenvectors RFV for strain measurement points i i，k ∈R n×1 ：
And sequentially calculating residual error feature vectors corresponding to the m strain measuring points.
And 4, step 4: calculating a covariance matrix and a mean vector of the residual error feature vector, wherein the step 4 specifically comprises:
step 4.1: calculating the total k of a strain-deflection training data set aiming at a strain measuring point i B Residual error feature vectors of the secondary train passing a bridge form a residual error feature matrix
Step 4.2: calculation matrix RFM i Covariance matrix COV of each row vector i ∈R n×n For matrix RFM i Averaging each row vector to obtain an average vector
And sequentially calculating covariance matrixes and mean vectors corresponding to the m strain measuring points.
And 5: according to the covariance matrix and the mean vector, performing second feature fusion based on the Mahalanobis distance to obtain distance features of each strain measuring point corresponding to the strain-deflection training data set, wherein the step 5 specifically comprises the following steps:
step 5.1: calculating the Mahalanobis distance of the residual error characteristic vector aiming at the strain measuring point i by considering the k-th train passing through the bridge to obtain the distance characteristic MD i,k ：
Step 5.2: and sequentially calculating the distance characteristics corresponding to the m strain measuring points.
And 6: calculating a control limit of the distance feature, wherein the step 6 specifically comprises:
calculating the total k of a strain-deflection training data set aiming at a strain measuring point i B And (3) adding the standard deviation of the mean value of the distance characteristics by 3 times to obtain the distance characteristics of the secondary train passing through the bridge as the control limit value CL of the vulnerable part of the bridge structure where the strain measuring point i is located i (ii) a And sequentially calculating control limit values corresponding to the m strain measuring points.
And 7: acquiring new strain-deflection monitoring data, and obtaining distance characteristics of the strain-deflection monitoring data corresponding to each strain measuring point by adopting the same method of the steps 1, 3 and 5 through initial characteristic extraction and twice characteristic fusion; and judging and positioning the local damage of the railway bridge structure according to the distance characteristics and the control limit obtained in the step 6, wherein the step 7 specifically comprises the following steps:
step 7.1: when strain-deflection monitoring data excited by a new train passing bridge in an operating state are obtained, steps 1, 3 and 5 are repeatedly executed, and the distance characteristics corresponding to each strain measuring point are obtained by using the three-dimensional regression coefficient matrix obtained in the step 2 and the covariance matrix and the mean vector obtained in the step 4; wherein for the z-th train passing, z is more than k B Obtaining needleDistance characteristic MD of strain measuring point i i,z ；
Step 7.2: if MD i,z ≤CL i If so, indicating that the vulnerable part of the bridge structure where the strain measuring point i is located is not damaged; if MD i,z ＞CL i Indicating that the vulnerable part of the bridge structure where the strain measuring point i is located is damaged; and obtaining damage judgment and positioning results of the vulnerable parts of the bridge structure where the m strain measuring points are located.
Examples
In the embodiment, the feasibility and the robustness of the proposed method are verified by performing a simulation test on an actual railway bridge. Fig. 2 (a) shows a finite element model of the single-span bowstring arch bridge, and fig. 2 (b) is a simplified schematic diagram of bridge abutment constraints. The bridge main body is a steel structure, the double-track bridge floor is a plane frame bridge floor system consisting of two main beams and 33 cross beams, the bridge floor system is hung on arch ribs at the north and south sides through 32 inclined support rods, and two ends of the bridge are respectively supported by two rubber support seats.
3 (a) - (b) are schematic diagrams of strain gage and deflection sensor placement and damage location for one embodiment of the present invention; FIG. 3 (a) is a south side view of the bridge and FIG. 3 (b) is a north side view of the bridge; as shown in the figure, each diagonal brace is provided with a strain measuring point, and the axial strain response of the diagonal brace is measured through a strain sensor; and (3) arranging deflection measuring points at the joints of the main beam and the diagonal brace on the south side of the bridge, and measuring the vertical displacement response of the joints of the main beam and the diagonal brace through a displacement sensor. Wherein the strain measuring points are numbered as s1-s32 in sequence.
When a real bridge is detected, the bolt connection between each inclined supporting rod of the bridge and the main beam and the arch is found to have construction defects, and then weak parts of the node connection are reinforced. Therefore, the present embodiment considers the local damage of the diagonal strut end node position and the development evolution thereof in the simulation. Three damage conditions (DamageScenario, DS) were considered, representing three damage states of the bridge, respectively: wherein DS1 is that the connecting parts of the diagonal brace 8 and the rod 9 and the main beam are damaged, and the connecting rigidity is reduced by 20%; the DS2 damage position is the same as DS1, but the damage degree, namely the connection rigidity is reduced by 50%; DS3 is that on the basis of DS2, the connecting part of the diagonal brace 17 and the main beam is damaged, and the connecting rigidity is reduced by 50%; further, DS0 represents a bridge undamaged state.
In order to fully verify the robustness of the data fusion algorithm under the influence of time-varying environment and vehicle load, 135 completely random train bridge-crossing events are simulated, as shown in fig. 4. The recorded temperature of each train crossing event is also uniformly distributed between-3 ℃ and 37 ℃ by randomly passing through two European trains of Suburban and Intercity along the direction B at the driving speed uniformly distributed from 50km/h to 200km/h each time, see figure 2 (B). The bridge is in an undamaged state DS0 during the first 45 times of train passing, is in a damaged state DS1 during the 46 th to 75 th train passing, is in a damaged state DS2 during the 76 th to 105 th train passing, and is in a damaged state DS3 during the 106 th to 135 th train passing. Fig. 5 (a) - (d) are typical strain and deflection responses under the influence of environmental and vehicle loading factors for one embodiment of the present invention. Wherein, figure 5 (a) the strain response of the strain measuring point s4 of the Suburban type train under the conditions of the speed of passing a bridge of 107km/h and the temperature of 18.2 ℃; FIG. 5 (b) the deflection response of a deflection measuring point d5 of a Suburban type train under the conditions of a bridge passing speed of 107km/h and a temperature of 18.2 ℃; FIG. 5 (c) the strain response of the strain measuring point s6 of the Intercity train at the bridge crossing speed of 50km/h and the temperature of 26 ℃; FIG. 5 (d) deflection response of deflection measuring point d5 of an Intercity type train at a bridge passing speed of 50km/h and a temperature of 26 ℃.
Fig. 6 lists the initial characteristic data corresponding to the strain measuring points s9 and s17 extracted from 135 train bridge passing events. Some outliers exist in the initial characteristic data, and analysis shows that the outliers are only response characteristics of a high-load interaction train during low-speed bridge crossing and are irrelevant to local damage. The connecting parts of the inclined struts and the main beams where the strain measuring points s9 and s17 are located are damaged, theoretically, response characteristics corresponding to the two strain measuring points after damage are reflected, but response characteristics before and after damage, namely, obvious changes of initial characteristic data are not seen in the graph, and therefore the damage is difficult to identify only by using the initial characteristic data of the single strain measuring point, and the damage effect in the strain initial characteristic data is covered by a masking effect caused by a time-varying environment and a vehicle load.
Fig. 7 (a) - (b) further show the quantitative distribution of the initial characteristic data corresponding to all strain and deflection measuring points in 135 train passing events by using box line diagrams. It can be observed that the vehicle load dominates the amplitude variations in response to the initial signature, resulting in local small lesions, such as DS1 and DS3, that are difficult to detect directly.
Fig. 8 lists distance characteristic data and control limits thereof corresponding to the strain measuring points s9 and s17 in 135 train bridge passing events obtained based on the data fusion algorithm provided by the invention. The control limit can effectively classify the distance characteristics before and after damage, and no damage misjudgment occurs; therefore, the distance characteristics obtained by fusion are effective as the early damage detection indexes of the railway bridge structure.
The distance characteristics corresponding to all the strain measuring points, namely the damage detection result, are shown in the form of a box line diagram in fig. 9. The damage detection result shows that the invention fully utilizes the monitoring data with different measuring points and isomerism, eliminates the masking effect caused by the time-varying environment and the vehicle load in the embodiment and enhances the detectability of local small damage, such as DS1 and DS3.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the scope of the present invention, which is defined by the appended claims.
Claims (9)
1. A railway bridge structure health monitoring method based on data fusion is characterized by comprising the following steps:
step 1: acquiring a strain-deflection training data set of a railway bridge structure, and extracting a strain initial characteristic matrix and a deflection initial characteristic matrix of the strain-deflection training data set;
and 2, step: performing stable regression analysis considering pairwise combination of the strain-deflection measuring points to obtain a three-dimensional regression coefficient matrix of the strain-deflection training data set;
and step 3: performing first feature fusion based on robust regression analysis to obtain residual error feature vectors of strain-deflection training data sets corresponding to each strain measuring point;
and 4, step 4: calculating a covariance matrix and a mean vector of the residual error feature vector;
and 5: performing secondary feature fusion based on the Mahalanobis distance according to the covariance matrix and the mean vector to obtain distance features of each strain measuring point corresponding to the strain-deflection training data set;
step 6: calculating a control limit of the distance characteristic;
and 7: acquiring new strain-deflection monitoring data, and obtaining distance characteristics of the strain-deflection monitoring data corresponding to each strain measuring point by adopting the same method of the steps 1, 3 and 5 through initial characteristic extraction and twice characteristic fusion; and judging and positioning the local damage of the railway bridge structure according to the distance characteristics and the control limit obtained in the step 6.
2. The method for monitoring the structural health of the railroad bridge based on the data fusion according to claim 1, wherein the step 1 specifically comprises:
step 1.1: aiming at a railway bridge structure with m strain measuring points and n deflection measuring points, according to the current operation state, the regular detection and the health monitoring history of the bridge, k acquired under the undamaged state of the bridge structure B Combining the strain and deflection response data excited by the secondary train passing through the bridge as a strain-deflection training data set;
step 1.2: the strain response measured by the kth train strain measuring point i is epsilon i,k (t)，i＝1,2,…,m，k＝1,2,…,k B Taking the integral of the strain response time course as the initial strain characteristic SF i,k ：
In the formula: t is k The total time of the action of the load effect of the passing bridge of the kth train;
strain-deflection training data set, total k B The secondary train bridge crossing can obtain a strain initial characteristic matrix SFM of m strain measuring points:
step 1.3: the deflection response measured by the deflection measuring point j is delta when the kth train passes a bridge j,k (t), j =1,2, …, n, and using the deflection response time-course integral as the deflection initial characteristic DF j,k ：
Strain-deflection training data set, total k B The secondary train bridge crossing can obtain a deflection initial characteristic matrix DFM of n deflection measuring points:
3. the method for monitoring the health of the railroad bridge structure based on the data fusion according to claim 2, wherein the step 2 specifically comprises:
step 2.1: in the strain initial characteristic matrix SFM and the deflection initial characteristic matrix DFM, the ith row of the SFM matrix and the jth row of the DFM matrix corresponding to the strain-deflection measuring point combination (i, j) are picked up, and the two rows of data are subjected to robust regression analysis to obtain a regression coefficient vector CV corresponding to the strain-deflection measuring point combination i,j ∈R 1×2 ；
Step 2.2: considering pairwise combination of m strain measuring points and n deflection measuring points, and obtaining a three-dimensional regression coefficient matrix CM (element for R) of the strain-deflection training data set m×n×2 。
4. The method for monitoring the health of the railway bridge structure based on the data fusion of claim 1, wherein a strain measuring point and a deflection measuring point for obtaining strain and deflection response data are respectively composed of a strain sensor for measuring the stress/strain of a vulnerable part of the bridge structure and a displacement sensor for measuring the deflection of the bridge; the strain sensor and the displacement sensor are in normal working states before and after the train passes through the bridge.
5. The method for monitoring the health of a railroad bridge structure based on data fusion of claim 3, wherein the first feature fusion based on robust regression analysis in step 3 specifically comprises:
step 3.1: according to the step 1, the strain initial characteristic vector SFV of m strain measuring points can be obtained by passing the bridge of the kth train k ：
SFV k ＝{SF 1,k SF 2,k …SF m-1,k SF m,k } T (5)
Step 3.2: according to the step 1, the deflection initial characteristic vector DFV of n deflection measuring points can be obtained by passing the bridge of the k-th train k ：
DFV k ＝{DF 1,k DF 2,k …DF n-1,k DF n,k } T (6)
Step 3.3: taking a two-dimensional regression coefficient matrix CM corresponding to a strain measuring point i in the three-dimensional regression coefficient matrix CM i ∈R n×2 Multiplying it by the matrix [ I ] point by point n ,DFV k ]Summing the obtained matrix rows to obtain a regression estimation vector aiming at the strain measuring point i
Is a unit column vector, and an upper corner mark e represents an estimated value; calculating residual eigenvector RFV for strain measurement point i i,k ∈R n×1 ：
And sequentially calculating residual error feature vectors corresponding to the m strain measuring points.
6. The method for monitoring the health of the railroad bridge structure based on the data fusion of claim 5, wherein the step 4 specifically comprises:
step 4.1: calculating the total k of a strain-deflection training data set aiming at a strain measuring point i B Residual error feature vectors of the secondary train passing a bridge form a residual error feature matrix
Step 4.2: calculation matrix RFM i Covariance matrix COV of each row vector i ∈R n×n For matrix RFM i Averaging each row vector to obtain an average vector
And sequentially calculating covariance matrixes and mean vectors corresponding to the m strain measuring points.
7. The method for monitoring the health of the railroad bridge structure based on the data fusion of claim 6, wherein the step 5 specifically comprises:
step 5.1: calculating the Mahalanobis distance of the residual error characteristic vector aiming at the strain measuring point i by considering the k-th train passing through the bridge to obtain the distance characteristic MD i,k ：
Step 5.2: and sequentially calculating the distance characteristics corresponding to the m strain measuring points.
8. The method for monitoring the health of the railroad bridge structure based on data fusion of claim 7, wherein the step 6 specifically comprises:
calculating the total k of a strain-deflection training data set aiming at a strain measuring point i B Distance characteristics of secondary train passing bridgeThe mean value plus 3 times of standard deviation is used as a control limit value CL of the vulnerable part of the bridge structure where the strain measuring point i is located i (ii) a And sequentially calculating control limit values corresponding to the m strain measuring points.
9. The method for monitoring the health of the railroad bridge structure based on the data fusion according to claim 8, wherein the step 7 is specifically:
step 7.1: when strain-deflection monitoring data excited by a new train passing bridge in an operating state are obtained, steps 1, 3 and 5 are repeatedly executed, and distance features corresponding to each strain measuring point are obtained by using the three-dimensional regression coefficient matrix obtained in the step 2 and the covariance matrix and the mean vector obtained in the step 4; wherein for the z-th train passing, z is more than k B Obtaining a distance characteristic MD for the strain measuring point i i,z ；
Step 7.2: if MD i,z ≤CL i If so, indicating that the vulnerable part of the bridge structure where the strain measuring point i is located is not damaged; if MD i,z ＞CL i Indicating that the vulnerable part of the bridge structure where the strain measuring point i is located is damaged; and obtaining the damage judgment and positioning results of the vulnerable parts of the bridge structure where the m strain measuring points are located.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211074817.4A CN115452282B (en) | 2022-09-02 | Railway bridge structure health monitoring method based on data fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211074817.4A CN115452282B (en) | 2022-09-02 | Railway bridge structure health monitoring method based on data fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115452282A true CN115452282A (en) | 2022-12-09 |
| CN115452282B CN115452282B (en) | 2024-07-02 |
Family
ID=
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116839659A (en) * | 2023-06-21 | 2023-10-03 | 中交三航局第三工程有限公司 | Method and system for identifying damage of girder of bridge girder erection machine |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20120106038A (en) * | 2011-03-17 | 2012-09-26 | 현대건설주식회사 | Measuring method and apparatus of bridge displacement by measure of strain |
| US20150226633A1 (en) * | 2014-01-14 | 2015-08-13 | Iowa State University Research Foundation, Inc. | Method and system for bridge damage detection |
| CN106197910A (en) * | 2016-07-01 | 2016-12-07 | 东南大学 | A kind of Bridges Detection based on Vehicle-bridge Coupling Analysis and detecting system |
| WO2017202139A1 (en) * | 2016-05-26 | 2017-11-30 | 东南大学 | Bridge damage identification method based on long-gauge-length strain influence envelope |
| CN111272366A (en) * | 2020-03-02 | 2020-06-12 | 东南大学 | Bridge displacement high-precision measurement method based on multi-sensor data fusion |
| CN112308826A (en) * | 2020-10-23 | 2021-02-02 | 南京航空航天大学 | Bridge structure surface defect detection method based on convolutional neural network |
| CN113283130A (en) * | 2021-04-21 | 2021-08-20 | 中国铁路设计集团有限公司 | Method for monitoring dynamic deflection of standard beam of 32 meters of high-speed railway based on strain mode |
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20120106038A (en) * | 2011-03-17 | 2012-09-26 | 현대건설주식회사 | Measuring method and apparatus of bridge displacement by measure of strain |
| US20150226633A1 (en) * | 2014-01-14 | 2015-08-13 | Iowa State University Research Foundation, Inc. | Method and system for bridge damage detection |
| WO2017202139A1 (en) * | 2016-05-26 | 2017-11-30 | 东南大学 | Bridge damage identification method based on long-gauge-length strain influence envelope |
| CN106197910A (en) * | 2016-07-01 | 2016-12-07 | 东南大学 | A kind of Bridges Detection based on Vehicle-bridge Coupling Analysis and detecting system |
| CN111272366A (en) * | 2020-03-02 | 2020-06-12 | 东南大学 | Bridge displacement high-precision measurement method based on multi-sensor data fusion |
| CN112308826A (en) * | 2020-10-23 | 2021-02-02 | 南京航空航天大学 | Bridge structure surface defect detection method based on convolutional neural network |
| CN113283130A (en) * | 2021-04-21 | 2021-08-20 | 中国铁路设计集团有限公司 | Method for monitoring dynamic deflection of standard beam of 32 meters of high-speed railway based on strain mode |
Non-Patent Citations (1)
| Title |
|---|
| 劳晓春, 汤立群, 任鹏, 罗立峰: "桥梁内部预应力钢绞线应变的长久监测与分析", 华南理工大学学报(自然科学版), no. 11, 28 November 2003 (2003-11-28) * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116839659A (en) * | 2023-06-21 | 2023-10-03 | 中交三航局第三工程有限公司 | Method and system for identifying damage of girder of bridge girder erection machine |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110704801B (en) | Bridge cluster structure operation safety intelligent monitoring and rapid detection complete method | |
| Cheung et al. | The application of statistical pattern recognition methods for damage detection to field data | |
| KR102158192B1 (en) | Damage Detection Approach for Tendon of Prestressed Concrete PSC Bridge Based on Auto-Encoder | |
| CN105956216A (en) | Finite element model correction method for large-span steel bridge based on uniform temperature response monitoring value | |
| US9969507B2 (en) | Method for performing diagnostics of a structure subject to loads and system for implementing said method | |
| CN117291072B (en) | Bridge damage identification method | |
| JP6458055B2 (en) | Damage estimation device | |
| CN112765881A (en) | Dynamic weighing method and system capable of being expanded to multiple bridges based on neural network | |
| CN110704911A (en) | Bridge damage cross positioning method based on cluster structure similarity | |
| Shahsavari et al. | Damage detection and decreased load-carrying capacity assessment of a vertical-lift steel truss bridge | |
| Shi et al. | Uncertain identification method of structural damage for beam-like structures based on strain modes with noises | |
| CN106446384B (en) | A kind of damnification recognition method of bridging crane main beam structure | |
| Gao et al. | Detection of multi-type data anomaly for structural health monitoring using pattern recognition neural network | |
| Karakostas et al. | Seismic assessment of bridges through structural health monitoring: a state-of-the-art review | |
| CN101487791A (en) | Damage diagnosis method for light grid sandwich materials | |
| CN115452282B (en) | Railway bridge structure health monitoring method based on data fusion | |
| CN115452282A (en) | Railway bridge structure health monitoring method based on data fusion | |
| Meixedo et al. | Structural health monitoring strategy for damage detection in railway bridges using traffic induced dynamic responses | |
| CN109781442B (en) | Detection method for crack fault of bogie of maglev train | |
| Phares et al. | Evolution of a bridge damage-detection algorithm | |
| CN115308054A (en) | Railway bridge local damage diagnosis method using vehicle-induced strain response data | |
| CN115356133A (en) | Bogie operation condition on-line monitoring, real-time classification and identification algorithm | |
| Silva et al. | Traffic monitoring using a structural health monitoring system | |
| CN111259494B (en) | Health monitoring and analyzing method for heavy machine equipment | |
| CN113128053A (en) | Nonlinear system parameter identification method, device, equipment and medium |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20240528 Address after: 110044 Liaoning Province, Shenyang City District wanghuanan Street No. 21 Applicant after: SHENYANG University Country or region after: China Address before: 110186 No. 3, Wenhua street, Hunnan District, Shenyang City, Liaoning Province Applicant before: Ren Peng Country or region before: China |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant |