CN111651840B - Track slab arch state detection method based on deep learning technology - Google Patents
Track slab arch state detection method based on deep learning technology Download PDFInfo
- Publication number
- CN111651840B CN111651840B CN202010275388.1A CN202010275388A CN111651840B CN 111651840 B CN111651840 B CN 111651840B CN 202010275388 A CN202010275388 A CN 202010275388A CN 111651840 B CN111651840 B CN 111651840B
- Authority
- CN
- China
- Prior art keywords
- track
- plate
- data
- mortar
- track plate
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000013135 deep learning Methods 0.000 title claims abstract description 17
- 238000005516 engineering process Methods 0.000 title claims abstract description 16
- 238000001514 detection method Methods 0.000 title claims abstract description 13
- 238000000034 method Methods 0.000 claims abstract description 49
- 238000006073 displacement reaction Methods 0.000 claims abstract description 48
- 238000000605 extraction Methods 0.000 claims abstract description 11
- 239000004570 mortar (masonry) Substances 0.000 claims description 58
- 239000011800 void material Substances 0.000 claims description 30
- 229910000831 Steel Inorganic materials 0.000 claims description 26
- 239000010959 steel Substances 0.000 claims description 26
- 230000006870 function Effects 0.000 claims description 24
- 238000013016 damping Methods 0.000 claims description 23
- 238000012549 training Methods 0.000 claims description 13
- 238000013528 artificial neural network Methods 0.000 claims description 12
- 238000004422 calculation algorithm Methods 0.000 claims description 12
- 238000004088 simulation Methods 0.000 claims description 10
- 230000009471 action Effects 0.000 claims description 9
- 230000000694 effects Effects 0.000 claims description 9
- 241001499740 Plantago alpina Species 0.000 claims description 7
- 238000012805 post-processing Methods 0.000 claims description 6
- 201000010099 disease Diseases 0.000 claims description 5
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 5
- 238000005452 bending Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 239000000463 material Substances 0.000 claims 1
- 230000008569 process Effects 0.000 abstract description 8
- 238000004364 calculation method Methods 0.000 abstract description 5
- 239000012634 fragment Substances 0.000 abstract description 2
- 238000007667 floating Methods 0.000 description 12
- 238000012544 monitoring process Methods 0.000 description 9
- 230000008878 coupling Effects 0.000 description 7
- 238000010168 coupling process Methods 0.000 description 7
- 238000005859 coupling reaction Methods 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 239000010426 asphalt Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 239000004568 cement Substances 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000003780 insertion Methods 0.000 description 2
- 230000037431 insertion Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 239000000725 suspension Substances 0.000 description 2
- 238000011179 visual inspection Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000004886 head movement Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61K—AUXILIARY EQUIPMENT SPECIALLY ADAPTED FOR RAILWAYS, NOT OTHERWISE PROVIDED FOR
- B61K9/00—Railway vehicle profile gauges; Detecting or indicating overheating of components; Apparatus on locomotives or cars to indicate bad track sections; General design of track recording vehicles
- B61K9/08—Measuring installations for surveying permanent way
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Machines For Laying And Maintaining Railways (AREA)
Abstract
The invention belongs to the technical field of track slab arch state detection, and particularly relates to a track slab arch state detection method based on a deep learning technology. The method converts the classification problem into a time sequence classification problem, namely inputting the time sequence classification problem into a series of track plate displacement data signal fragments and outputting the time sequence classification problem into corresponding state classification of the track plate; aiming at the characteristics of the track slab arch state recognition task, the recognition process is divided into two stages: a feature extraction stage and a classification stage; considering that the original signal has more noise, particularly the signal containing track plate irregularity is throughout, various types of characteristics are used for representing the electroencephalogram signal to be divided into two stages so as to simplify the calculated amount of the classification process and improve the calculation speed.
Description
Technical Field
The invention belongs to the technical field of track slab arch state detection, and particularly relates to a track slab arch state detection method based on a deep learning technology.
Background
The novel CRTS II type plate-type ballastless track with the high-speed track structure is developed by introducing the German Boggy plate-type ballastless track and then carrying out study innovation in China. The Jingjin high-speed railway is the first high-speed railway with the speed per hour of 300-350 km/h in China, the whole length of the high-speed railway is 1318km, the high-speed railway with the longest mileage, the largest investment and the highest standard is built once after the establishment of new China, the ballastless track is paved on the whole line, and the CRTS II plate type ballastless track technology is adopted. At present, CRTS II plate-type ballastless tracks are used on a plurality of passenger lines such as Jingjin, jinghu, shanghai and the like. The CRTS II type plate-type ballastless track on the roadbed mainly comprises a steel rail, a fastener system, a track plate, a cement asphalt mortar (Cement Asphalt Mortar, CA mortar for short) layer, a concrete supporting layer and other parts. Because CRTS II type plate-type ballastless track is affected by factors such as train impact and environmental temperature, various diseases can be inevitably caused in the ballastless track lower structure, wherein the occurrence of a gap in the plate-type ballastless track and the subsequent camber of a track plate are typical diseases.
At present, the field detection and maintenance of the off-line arch of the plate-type ballastless track mainly adopts the methods of visual inspection, steel rule insertion measurement, field plate uncovering and rail inspection vehicle. The defects of the visual inspection and the steel rule insertion measurement method are that the middle local gap and the accurate distribution condition of the gap cannot be detected; the plate uncovering method has the defects of being only suitable for railway construction, high in cost and low in efficiency; the rail inspection vehicle has the defects of high manufacturing cost, can only detect in the train operation empty window period, and cannot monitor the real-time state.
Patent CN201910620162.8 discloses a track slab arch-up distributed monitoring system and a monitoring method, which are used for detecting an arch-up angle of a track slab in real time, filtering an angle value detected when a train passes through by performing data processing on the arch-up angle, and thus, retaining the angle value measured when the track slab is static; the track slab arch-up distributed monitoring system and the monitoring method have the following defects:
1) The existing algorithm can only realize the measurement of the angle value measured when the road plate is static, so that the application scene is single;
2) When a train passes, the existing algorithm can directly filter out the data at the moment, and the acquired original data is missing, so that the analysis of the later data is not facilitated;
3) The track slab arch distributed monitoring method and system adopt the node to process data, and have the advantages of high circuit power, high power consumption and reduced service life of the node.
Disclosure of Invention
Aiming at the technical problems, the invention provides a track slab arch state detection method based on a deep learning technology. The classification problem is converted into a time sequence classification problem, namely, the time sequence classification problem is input into a series of track plate displacement data signal fragments, and the time sequence classification problem is output into the corresponding state classification of the track plate.
The invention is realized by the following technical scheme:
a track slab arch state detection method based on a deep learning technology comprises the following steps:
(1) The vehicle-plate type orbit dynamics traditional model is improved into a vehicle-plate type orbit dynamics model under the consideration of the action of CA mortar void;
(2) Constructing a database for deep learning of the classified neural network: the vehicle-plate type orbit dynamics model under the consideration of the action of CA mortar void, which is obtained in the step (1), is utilized to output orbit plate displacement simulation data under different void degrees by setting different parameters, after training is carried out for a plurality of times to obtain enough simulation data, the simulation data is added with labels according to disease types to form a database, and the database comprises a training set for inputting a classified neural network and a test set for inputting the classified neural network;
(3) Training a classification network using the database constructed in step (2): the classified network structure comprises four layers of networks, wherein the first layer of network is an input layer, the second layer of network is BiLSTM, the third layer of network is a full-connection layer, and the fourth layer of network is a softmax layer;
(4) And classifying the track slab arch state according to the classification network result: the track plate displacement data are used as the network input quantity of the trained classification network, and the original track plate displacement data are arranged into a characteristic sequence and are input to the BiLSTM network; and after being circulated by the BiLSTM layer, the track slab is input into the softmax layer to judge the type of the arch state of the track slab, and a final result of the arch state judgment of the track slab is obtained.
Further, in the step (4), the original track plate displacement data is organized into a characteristic sequence, and the specific method comprises the following steps:
(1) Preprocessing the track plate displacement signal: using various types of characteristics to represent track slab displacement signals, normalizing time sequence signals to a (0, 1) range, segmenting the original track slab displacement signals and extracting signal characteristics, wherein the signal characteristic extraction is to rearrange data characteristics in small-segment data obtained by segmentation into a group of data after calculating the data characteristics to become characteristic data; after signal feature extraction, the feature data of each section are connected in sequence to form a final feature data set;
the data features in the signal feature extraction process comprise a maximum value, a minimum value, an average value, a peak-peak value, a rectification average value, a variance, a standard deviation, kurtosis, a root mean square, a waveform factor, a peak factor, a kurtosis factor, a pulse factor and a margin factor;
(2) And calculating a corresponding result correlation weight for each feature of the obtained feature data set by utilizing a Relief algorithm, and eliminating the feature with lower weight according to a set weight threshold value to obtain a final trained feature sequence.
Further, step (4) further comprises step (5): and carrying out post-processing according to the obtained final result of the arch state judgment of the track plate, wherein the post-processing is to judge whether the track plate is arched or not by combining other judging programs, and the classification result can correct obvious recognition errors, further optimize the recognition effect and carry out early warning work.
Further, in the step (1), the detailed shape of the vibration mode coordinate differential equation set of the vehicle-plate type orbit dynamics model under the action of taking the CA mortar into consideration is as follows:
wherein: t (T) n (t) introducing a free beam orthonormal function system { X }, after the track slab vertical vibration differential equation adopts the Ritz method n (n=1-NMS), NMS is the modal order of the orbit board, and NMS generalized coordinates T are selected n (t);Is T n A second derivative of (t); e (E) s I s Is the bending rigidity of the track plate; m is m s The mass of the track plate is the mass of the track plate in unit length; beta n Is a constant; m is m 0 Finger m 0 A plurality of discrete units; c (C) sq Is the distributed damping of CA mortar at q; x is X p Is a free beam orthogonal function system { X ] n Meaning of p is similar to n, the value range is the same as n, X n Is a free beam orthogonal function system of a traditional track slab, X p The free beam orthogonal function system of the rear track plate is improved; t (T) p (t) and X p Similarly, T n P of (t) is similar to n, the value range is the same as n, the n represents the generalized coordinates of the traditional track plate, and the p represents the generalized coordinates of the improved track plate; k (K) sq The distributed rigidity of the CA mortar at q; f corresponds to the concrete supporting layer, z f (x, t) is the vibration displacement (m), z of the concrete support layer f (x q T) is the vibration displacement of the concrete supporting layer of the q-th CA mortar discrete unit; />Is z f (x q T) obtaining a derivative, wherein the derivative represents the vibration speed of the concrete supporting layer of the q-th CA mortar discrete unit; l (L) s A monolithic length for the track plate; n is n 0 The number of the buckles of the steel rail on one track plate is the number; subscript s corresponds to the track plate; x is x i The method is characterized in that the method is a coordinate of an original coordinate system, the coordinate system is established based on a steel rail model, (i=1-N), and N is the number of fasteners in a plate-type track length range L; x is x q Is x i Q=i time; c (C) pi Is in the coordinate system x i Damping the lower cushion layer of the rail when in position; NM is the modal order of the steel rail; k (K) pi Is in the coordinate system x i Rigidity of the lower cushion layer of the rail when in position; y is Y p (x (j-1) × n0+i )q p (t) the regular vibration type function of the vertical vibration displacement application simply supported beams of the steel rail is expressed as +.>In the formula, k=p and x is subscripted to be (j-1) n 0 +i, where n 0 Is the number of buckles of the steel rail on one track plate.
Further, in the step (2), when constructing the database, 104 meters, namely 16 track slabs are selected as the length of the vehicle-slab track dynamics model considering the CA mortar void effect, the vehicle speed is 300 km/h, the displacement of the track slabs is selected as the sample data when the train passes through one track slab, and the sampling interval is 10 -4 And when s, obtaining different sample data by setting different void lengths, setting CA mortar void conditions with longitudinal lengths of 0 meter, 0.325 meter, 0.65 meter, 1.3 meter and 1.95 meter to obtain samples with five void conditions, and obtaining enough sample numbers for each void setting condition to obtain 64 groups of samples.
The beneficial technical effects of the invention are as follows:
(1) The track slab arch state detection method based on the deep learning technology converts the classification problem into a time sequence classification problem, namely, the classification problem is input into a series of track slab displacement data signal segments and output into the corresponding state classification of the track slab.
(2) The invention provides a detection method for the arch state of a track plate based on a deep learning technology, which divides the recognition process into two stages according to the characteristics of the arch state recognition task of the track plate: a feature extraction stage and a classification stage. Considering that the original signal has more noise, particularly the signal containing the irregularity of the track plate is throughout, the calculation resources required by directly using the preprocessed signal training model are very large, and the type of the learned information is single, so that the calculation amount of the classification process can be simplified and the calculation speed can be improved by using various types of characteristics to represent the electroencephalogram signals to be divided into two stages.
(3) In the method provided by the invention, the track slab monitoring equipment paved along the railway is adopted for data acquisition, and the track slab monitoring equipment has the problems of large environmental interference, irregular installation and the like, and has more severe requirements on the aspects of data accuracy, stability and the like; in the method provided by the invention, the artificial intelligence technology is utilized, the interference signals such as track irregularity and the like are added during the training of the classification network, and the classification network can be continuously learned according to the updating of the data set, so that the high accuracy and high robustness are achieved in the aspect of carrying out the abnormal identification of the arch state of the track plate in the track plate displacement signals with poor signal quality and large interference, the subsequent labor cost is greatly reduced, and the period for generating the final conclusion is shortened.
Drawings
FIG. 1 is a conventional model of high-speed rail vehicle-slab track dynamics in an embodiment of the present invention;
FIG. 2 is a diagram of the stress relationship of a conventional model rail in an embodiment of the present invention;
FIG. 3 is a diagram showing a conventional model track plate model stress analysis in accordance with an embodiment of the present invention;
FIG. 4 is a high speed railway vehicle-slab track dynamics improvement model (i.e., a vehicle-slab track dynamics model with CA mortar void considered) in an embodiment of the present invention;
FIG. 5 is a graph showing an improved track slab model force analysis in accordance with an embodiment of the present invention;
FIG. 6 is a BiLSTM neural network structure in an embodiment of the invention;
FIG. 7 is a diagram of a categorized neural network structure in an embodiment of the present invention;
FIG. 8 is a characteristic weight arrangement after the Relief algorithm in an embodiment of the present invention;
FIG. 9 is a feature sequence after low weight rejection in an embodiment of the invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
On the contrary, the invention is intended to cover any alternatives, modifications, equivalents, and variations as may be included within the spirit and scope of the invention as defined by the appended claims. Further, in the following detailed description of the present invention, certain specific details are set forth in order to provide a better understanding of the present invention. The present invention will be fully understood by those skilled in the art without the details described herein.
Aiming at the technical problems of a track slab arch distributed monitoring system and a monitoring method in the prior art, the invention provides a track slab arch state detection method based on a deep learning technology, which comprises the following steps:
(1) The vehicle-plate type orbit dynamics traditional model is improved into a vehicle-plate type orbit dynamics model under the consideration of the action of CA mortar void;
the traditional model of vehicle-slab track dynamics of the high-speed railway is proposed by the teachings of southwest traffic university Wanming in vehicle-track coupling dynamics, and is simply referred to as a traditional model. The conventional model is to use the principle of vehicle-track coupling dynamics, and to take a vehicle system and a plate-type track system as a whole to establish a vertical model of the vehicle-plate-type track coupling system shown in fig. 1.
The conventional model has the following assumption:
(1) Because CA mortar under the track slab mainly has a supporting function, and CA mortar void mainly affects vertical vibration of the ballastless track, only vibration in the vertical direction of the vehicle-slab track coupling system is considered.
(2) The vehicle system and the slab track system are symmetrical to the track center line, and only half of structural researches of the vehicle track coupling system are considered in the calculation process.
(3) The vehicle system is a multi-rigid body model with a secondary spring damping system and comprises a vehicle body, two bogies and four wheel pairs. A primary suspension system is arranged between the bogie and the wheel set, and a secondary suspension system is arranged between the vehicle body and the bogie. Considering the sinking and floating and nodding vibration of the vehicle body and the bogie, the sinking and floating vibration of the wheel sets and the vehicle system have ten degrees of freedom.
(4) The rail is a limited long simple supporting beam on the basis of discrete elastic point support, and the elasticity and damping of the rail lower backing plate and the fastener are respectively realized by using the elasticity coefficient K p And damping coefficient C p And (3) representing.
(5) The track plate is a limited length free beam supported on a linear spring and a linear damper which are continuously distributed, and the distribution rigidity and the distribution damping of CA mortar are respectively realized by the elastic coefficient k s And damping coefficient c s And (3) representing.
(6) The coupling relation between the vehicle system and the plate-type track system is realized through wheel-track interaction force, and a classical Hertz nonlinear elastic contact model is adopted.
The physical quantities represented by the symbols in the model are shown in Table 1
TABLE 1 physical quantity represented by each symbol in the conventional model
The model motion equation includes:
(1) Sinking and floating movement of vehicle body
z c Is used for sinking and floating the car body,is z c After seeking, the sinking and floating speed of the car body is indicated, < ->Representing the sinking and floating acceleration of the vehicle body;and z t1 Similarly, where subscript t1 denotes the front truck and t2 denotes the rear truck.
(2) Spot head movement of vehicle body
(3) Front bogie sinking and floating movement
(4) Front bogie nodding motion
(5) Sinking and floating movement of rear bogie
(6) Rear bogie nodding motion
(7) The first wheel pair moves in a sinking and floating manner
(8) The second wheel pair moves in a sinking and floating manner
(9) Third wheel pair sinking and floating movement
(10) Fourth wheel pair sinking and floating movement
Wherein p is i And (t) is the vertical wheel rail force (i=1 to 4) of the single-side wheel.
The steel rail in the traditional model is regarded as a limited-length simple beam on the basis of discrete elastic point support, and the stress analysis model is shown in figure 2. Wherein p is i Is a wheel rail force, and moves forward with the vehicle at a speed v; f (F) rsi (i=1 to N) is the steel rail fulcrum reaction force, N is the number of fasteners in the plate type track length range L; ox is a fixed coordinate system fixedly connected with the steel rail; o 'x' is a moving coordinate system attached to the vehicle. The relationship between these two coordinates is:
x=x′+x 0 +νt (11)
wherein x is 0 The fixed coordinates of the rear wheel at the starting moment; t is a time variable.
As can be seen from fig. 2, the differential equation of the vibration of the rail is:
wherein the method comprises the steps of
The subscript "r" corresponds to the rail, the subscript "s" corresponds to the rail plate, z r (x i T) and z s (x i T) represents the vertical displacement variable (m) of the rail and the rail plate at the fastener, respectively;and->Representing the vertical velocity variation (m/s) of the rail and rail plate at the fastener, respectively.
Coordinates x of each wheel wj (j=1 to 4) are respectively:
coordinates of each fastener
x i =il p (15)
By adopting the Ritz method and applying the regular vibration type function of the simply supported beam, the vertical vibration displacement of the steel rail can be expressed as:
wherein, NM is the modal order of the steel rail, and NM=0.5N is generally taken; rail vibration mode
Substituting formula (16) into formula (12) to obtain:
both sides are multiplied by Y h (x) (h=1, 2,3,., NM), integrate x from 0 to L, and note the orthogonality of the modes:
has the following components
Depending on the nature of the delta function, formula (20) can be sorted:
because of
Therefore, the formula (21) can be simplified as:
the method is a basic form of a second order ordinary differential equation set of rail vibration mode coordinates.
Further, formula (16) is substituted into formula (13)
Then, expression (24) becomes
The method is a detailed form of a steel rail vibration mode coordinate differential equation set.
The slab track model simplifies the CA mortar into springs and damping distributed continuously along the track slab, which is seen as a finite length free beam supported on continuously distributed linear springs and linear damping as shown in FIG. 3.
The track slab vertical vibration differential equation is:
wherein k is s 、c s CA mortar layers below the track boardsDistributed stiffness (N/m/m) and distributed damping (N.s/m/m) along the track direction; n is n 0 Is the number of buckles of the steel rail on one track plate.
Adopts Ritz method to introduce free beam orthogonal function system { X ] n N=1 to NMS, selecting NMS generalized coordinates T n (t) wherein
Wherein C is n 、β n Is constant. C (C) n 、β n L s The values of (2) are shown in Table 2.
TABLE 2 free Beam function coefficient
The vertical displacement of the track plate can be approximated as:
wherein NMS is the modal order of the track board, taking nms=0.5n 0 Substituting formula (29) into formula (27):
the two sides are multiplied by X p (x) (p=1 to NMS) then L over the full length of the track plate s Integrate x and notice orthogonality of modes
Has the following components
Depending on the nature of the delta function, formula (32) can be sorted:
because of
Therefore, equation (33) can be simplified as:
this is the kinetic equation of the orbital plate model.
Further, formula (29) is substituted into formula (25)
Then, the formula (36) becomes:
this is a detailed form of the track slab mode coordinate differential equation set.
Wherein: wherein: t (T) n (t) introducing a free beam orthogonal function system { X } after the track slab vertical vibration differential equation (27) adopts the Ritz method n (n=1-NMS), NMS is the modal order of the orbit board, and NMS generalized coordinates T are selected n (t);Is T n A second derivative of (t); e (E) s I s Is the bending rigidity of the track plate; m is m s The mass of the track plate is the mass of the track plate in unit length; beta n Is a constant; m is m 0 Finger m 0 A plurality of discrete units; c (C) sq Is the distributed damping of CA mortar at q; x is X p Is a free beam orthogonal function system { X ] n Meaning of p is similar to n, the value range is the same as n, X n Is a free beam orthogonal function system of a traditional track slab, X p The free beam orthogonal function system of the rear track plate is improved; t (T) p (t) and X p Similarly, T n P of (t) is similar to n, the value range is the same as n, the n represents the generalized coordinates of the traditional track plate, and the p represents the generalized coordinates of the improved track plate; k (K) sq The distributed rigidity of the CA mortar at q; f corresponds to the concrete supporting layer, z f (x, t) is the vibration displacement (m), z of the concrete support layer f (x q T) is the vibration displacement of the concrete supporting layer of the q-th CA mortar discrete unit; />Is z f (x q T) obtaining a derivative, wherein the derivative represents the vibration speed of the concrete supporting layer of the q-th CA mortar discrete unit; l (L) s A monolithic length for the track plate; n is n 0 The number of the buckles of the steel rail on one track plate is the number; subscript s corresponds to the track plate; x is x i The method is characterized in that the method is a coordinate of an original coordinate system, the coordinate system is established based on a steel rail model, (i=1-N), and N is the number of fasteners in a plate-type track length range L; x is x q Is x i Q=i time; c (C) pi Is in the coordinate system x i Damping the lower cushion layer of the rail when in position; NM is the modal order of the steel rail; k (K) pi Is in the coordinate system x i Rigidity of the lower cushion layer of the rail when in position; y is Y p (x (j-1)×n0+i )q p (t) the regular vibration type function of the vertical vibration displacement application simply supported beams of the steel rail is expressed as +.>In the formula, k=p and x is subscripted to be (j-1) n 0 +i, where n 0 The number of the buckles of the steel rail on one track plate is the number; x is X p (x i )T p (t) is obtained by substituting formula (29) into formula (42), wherein formula (29) is obtained by introducing a free beam orthogonal function system { X } by using the Ritz method n N=1 to NMS, selecting NMS generalized coordinates T n (t) wherein
Wherein C is n 、β n Is constant. C (C) n 、β n L s The values of (2) are shown in Table 2. The vertical displacement of the track plate can be approximated as:
in a ballastless track dynamics equation of a traditional vehicle-slab track dynamics model, CA mortar is simplified into springs and damping which are continuously distributed along a track slab, and acting force on the track slab is embodied in a form of distributed force in the dynamics equation, so that the void of a CA mortar layer of the length of the whole track slab can be simulated only by changing the rigidity damping value of the CA mortar under the whole track slab, and any range can not be simulated. In the improved model provided by the invention, the CA mortar model is discretized in the longitudinal direction, discrete spring damping is adopted to replace distributed spring damping, the supporting and distributing force of the CA mortar on the track plate is converted into concentrated force, and a vehicle-plate type track dynamics model taking the void effect of the CA mortar into consideration is established, as shown in figure 4.
In the vehicle-plate type orbit dynamics model considering the CA mortar void effect provided by the invention, the CA mortar corresponding to one orbit plate is averagely divided into m 0 The CA mortar of each discrete unit is simplified into a spring and a damping which are concentrated to one point, and the spring coefficient K is used for respectively s And damping coefficient C s The longitudinal length of each cell is denoted by l s =L s /m 0 In the state of no void of CA mortar, the rigidity of the supporting surface is:
E CA the modulus of elasticity of the CA mortar; h is a CA The thickness of the CA mortar layer (CRTS II plate-type ballastless track is 0.03 m), the CA sand is in the working condition of no dechuckingVertical stiffness K of each discrete spring and damper of the pulp model s And damping coefficient C s The method comprises the following steps of:
in the method, in the process of the invention,bearing surface stiffness, b 0 Is the overall width of the track plate (2.55 m for CRTS type II track plate); c s For CA mortar distributed damping, l s A longitudinal length for each discrete unit;
after the CA mortar model is discretized, the distributed force on the track slab is converted into a concentrated force in the track slab vibration differential equation, as shown in FIG. 5, in the improved track slab model, the track slab is regarded as a finite length free beam on the basis of the discrete elastic point support, and the differential equation is:
wherein,
wherein, the subscript f corresponds to the concrete supporting layer; z f (x, t) is the vibration displacement (m) of the concrete support layer; f (F) sfq (t)(q=1~m 0 ) Representing the supporting force (N) of the q-th CA mortar discrete unit on the track slab; z s (x q T) is the vertical displacement variable, K, of the rail plate at q sq For the distributed rigidity of CA mortar at q, C sq Is the distributed damping of CA mortar at q;
wherein E is s I s Representing the bending stiffness of the track slab; z s (x, t) a vertical displacement variation of the track plate at the fastener; x and t are variables; m is m s The mass of the track plate in unit length; m is m 0 Is m 0 A plurality of discrete units; q is the q-th CA mortar discrete unit concrete supporting layer; delta (x-x) q ) Is a differential variable; n is n 0 Is the number of buckles of the steel rail on a track slab; f (F) rsj (t) is the supporting force of the concrete supporting layer of the jth CA mortar discrete unit; delta (x-x) i ) Is a differential variable.
Adopts Ritz method to introduce free beam orthogonal function system { X ] n N=1 to NMS, selecting NMS generalized coordinates T n (t),{X n The values of are given in table 2.
Substituting formula (29) into formula (41) and multiplying X on both sides of the formula p (x) (p=1 to NMS) and then over the full length of the track plate L s Inner integral to x:
the properties using modal orthogonality and delta function can be obtained:
substituting formula (34) and formula (35) into formula (44):
this is the kinetic equation of the orbit plate improvement model.
Further, formula (29) is substituted into formula (42):
substituting formula (37) and formula (46) into formula (45):
the method is a detailed form of improving the track plate model vibration mode coordinate differential equation set.
After the CA mortar is discretized, the distribution force of the CA mortar on the track plate is converted into concentrated force, a vehicle track coupling dynamics model considering the action of CA mortar void can be established, and the problem that the existing dynamics model cannot simulate CA mortar void under any working condition can be solved; the modeling is based on matlab software, so that the practical problems of low simulation efficiency, limited line length, short track irregularity wavelength and the like of a finite element model method can be solved.
(2) Constructing a database for deep learning of the classified neural network: the vehicle-plate type orbit dynamics model under the consideration of the action of CA mortar void, which is obtained in the step (1), is utilized to output orbit plate displacement simulation data under different void degrees by setting different parameters, after training is carried out for a plurality of times to obtain enough simulation data, the simulation data is added with labels according to disease types to form a database, and the database comprises a training set for inputting a classified neural network and a test set for inputting the classified neural network;
when constructing the database, in this embodiment, 104 meters, that is, 16 track boards are selected as the length of the vehicle-board type track dynamics model considering the CA mortar void effect, the vehicle speed is 300 km/h, the displacement of the track boards is selected as the sample data when the train passes through one track board, and the sampling interval is 10 -4 s, obtaining data of different samples by setting different void lengths, and respectively setting CA mortar void lengths of 0 meter, 0.325 meter, 0.65 meter, 1.3 meter and 1.95 meter under the track plate to obtain samples under five conditions; a sufficient number of samples were then obtained for each run-out setting, and 64 sets of samples were obtained.
(3) Training a classification network using the database constructed in step (2): the classified network structure comprises four layers of networks, wherein the first layer of network is an input layer, the second layer of network is BiLSTM, the third layer of network is a full-connection layer, and the fourth layer of network is a softmax layer; both the first-tier network and the second-tier network are often used to model context information in natural language processing tasks, where LSTM is known collectively as Long Short-Term Memory, which is one of RNN (Recurrent Neural Network). LSTM is well suited for modeling time series data, such as text data, due to its design characteristics. BiLSTM is an abbreviation of Bi-directional Long Short-Term Memory, and is formed by combining forward LSTM and backward LSTM, and the structure of BiLSTM is shown in figure 6;
(4) And classifying the track slab arch state according to the classification network result: the track plate displacement data are used as the network input quantity of the trained classification network, and the original track plate displacement data are arranged into a characteristic sequence and are input to the BiLSTM network; and after being circulated by the BiLSTM layer, the track slab is input into the softmax layer to judge the type of the arch state of the track slab, and a final result of the arch state judgment of the track slab is obtained.
The structure of the classification network is shown in fig. 7, specifically, in this step, track slab displacement data as the network input of the trained classification network may be collected by a method provided in patent CN201910620162.8, specifically, using tilt sensing nodes installed on the track slab, collecting vibration displacement data of the track slab, and then converting the vibration displacement data into tilt data of the track slab for use in the present invention.
Specifically, in the step (4), the original track plate displacement data is organized into a characteristic sequence, and the specific method is as follows:
(1) Preprocessing the track plate displacement signal: using various types of characteristics to represent track slab displacement signals, normalizing time sequence signals to a (0, 1) range, segmenting the original track slab displacement signals and extracting signal characteristics, wherein the track slab displacement signals can be generally divided into three segments, if the signals are overlong, the number of segments can be increased, and the signal characteristic extraction is to rearrange data characteristics in small segment data obtained by segmentation into a group of data after calculating the data characteristics, so as to obtain characteristic data; after signal feature extraction, the feature data of each section are connected in sequence to form a final feature data set;
the data features in the signal feature extraction process comprise a maximum value, a minimum value, an average value, a peak-peak value, a rectification average value, a variance, a standard deviation, kurtosis, a root mean square, a waveform factor, a peak factor, a kurtosis factor, a pulse factor and a margin factor;
(2) And calculating a corresponding result correlation weight for each feature of the obtained feature data set by utilizing a Relief algorithm, as shown in fig. 8, and eliminating the feature with lower weight according to a set weight threshold to obtain a final trained feature sequence, as shown in fig. 9.
Wherein the Relief algorithm is a feature weighting algorithm (Feature weighting algorithms) that gives features different weights according to their relevance to the category, and features with weights less than a certain threshold are removed. The relevance of features and categories in the Relief algorithm is based on the ability of the features to distinguish between close range samples. Randomly selecting a sample R from the training set D by adopting a Relief algorithm, then searching a nearest neighbor sample H from samples similar to the sample R, namely a Near Hit, and searching a nearest neighbor sample M from samples different from the sample R, namely a Near Miss, wherein the samples Near Hit and the Near Miss are samples in the training set D; the weights for each feature are then updated according to the following rules: if the distance between R and Near Hit on a feature is less than the distance between R and Near Miss, then the feature is said to be beneficial for distinguishing nearest neighbors of the same class from nearest neighbors of different classes, then the weight of the feature is increased; conversely, if the distance between R and Near Hit is greater than the distance between R and Near Miss for a feature, indicating that the feature is negatively affecting the nearest neighbors that distinguish between the same class and different classes, the weight of the feature is reduced. The above process is repeated m times, and finally the average weight of each feature is obtained. The greater the weight of a feature, the more classification capability that the feature is represented, and conversely, the less classification capability that the feature is represented. The runtime of the Relief algorithm increases linearly with the number of samples m and the number of original features N, and thus the running efficiency is very high.
Further, step (4) further comprises step (5): and carrying out post-processing according to the obtained final result of the arch state judgment of the track plate, wherein the post-processing is to judge whether the track plate is arched or not by combining other judging programs, and the classification result can correct obvious recognition errors, further optimize the recognition effect and carry out early warning work. The other judging procedure may be the judging procedure mentioned in CN201910620162.8, and judges whether the track slab is arched or not according to the comparison between the set inclination threshold and the collected inclination data.
Compared with the prior art, the method provided by the invention can be realized:
(1) Converting the event into a time sequence classification problem, and classifying the application scenes by using new technologies to be more various;
(2) The deep learning technology is adopted, so that the data information is prevented from being lost, and the method can be conveniently expanded to other similar tasks;
(3) According to different collected data sets, more result types can be judged, so that the model has wider application range and better ductility;
(4) Other result judgment can be combined, so that the module is more accurate and reliable.
Claims (4)
1. The track slab arch state detection method based on the deep learning technology is characterized by comprising the following steps of:
(1) The vehicle-plate type orbit dynamics traditional model is improved into a vehicle-plate type orbit dynamics model under the consideration of the action of CA mortar void;
(2) Constructing a database for deep learning of the classified neural network: the vehicle-plate type orbit dynamics model under the consideration of the action of CA mortar void, which is obtained in the step (1), is utilized to output orbit plate displacement simulation data under different void degrees by setting different parameters, after training is carried out for a plurality of times to obtain enough simulation data, the simulation data is added with labels according to disease types to form a database, and the database comprises a training set for inputting a classified neural network and a test set for inputting the classified neural network;
(3) Training a classification network using the database constructed in step (2): the classified network structure comprises four layers of networks, wherein the first layer of network is an input layer, the second layer of network is BiLSTM, the third layer of network is a full-connection layer, and the fourth layer of network is a softmax layer;
(4) And classifying the track slab arch state according to the classification network result: the track plate displacement data are used as the network input quantity of the trained classification network, and the original track plate displacement data are arranged into a characteristic sequence and are input to the BiLSTM network; after being circulated by the BiLSTM layer, the material is input into the softmax layer to judge the type of the arch state of the track plate, and a final result of the arch state judgment of the track plate is given;
the detailed shape of the vibration mode coordinate differential equation set of the vehicle-plate type orbit dynamics model under the action of taking CA mortar into consideration is as follows:
wherein: t (T) n (t) introducing a free beam orthonormal function system { X }, after the track slab vertical vibration differential equation adopts the Ritz method n (n=1-NMS), NMS is the modal order of the orbit board, and NMS generalized coordinates T are selected n (t);Is T n A second derivative of (t); e (E) s I s Is the bending rigidity of the track plate; m is m s The mass of the track plate is the mass of the track plate in unit length; beta n Is a constant; m is m 0 Finger m 0 A plurality of discrete units; c (C) sq Is the distributed damping of CA mortar at q; x is X p Is a free beam orthogonal function system { X ] n Meaning of p is similar to n, the value range is the same as n, X n Is a free beam orthogonal function system of a traditional track slab, X p The free beam orthogonal function system of the rear track plate is improved; t (T) p (t) and X p Similarly, T n P of (t) is similar to n, the value range is the same as n, the n represents the generalized coordinates of the traditional track plate, and the p represents the generalized coordinates of the improved track plate; k (K) sq The distributed rigidity of the CA mortar at q; f corresponds to the concrete supporting layer, z f (x, t) is the vibration displacement (m), z of the concrete support layer f (x q T) is the vibration displacement of the concrete supporting layer of the q-th CA mortar discrete unit;is z f (x q T) obtaining a derivative, wherein the derivative represents the vibration speed of the concrete supporting layer of the q-th CA mortar discrete unit; l (L) s A monolithic length for the track plate; n is n 0 The number of the buckles of the steel rail on one track plate is the number; subscript s corresponds to the track plate; x is x i The method is characterized in that the method is a coordinate of an original coordinate system, the coordinate system is established based on a steel rail model, (i=1-N), and N is the number of fasteners in a plate-type track length range L; x is x q Is x i Q=i time; c (C) pi Is in the coordinate system x i Damping the lower cushion layer of the rail when in position; NM is the modal order of the steel rail; k (K) pi Is in the coordinate system x i Rigidity of the lower cushion layer of the rail when in position; y is Y p (x (j-1) × n0+i )q p (t) the regular vibration type function of the vertical vibration displacement application simply supported beams of the steel rail is expressed as +.>In the formula, k=p and x is subscripted to be (j-1) n 0 +i, where n 0 Is the number of buckles of the steel rail on one track plate.
2. The method for detecting the arch state of the track slab based on the deep learning technology according to claim 1, wherein in the step (4), the original track slab displacement data is organized into a characteristic sequence, and the specific method comprises the following steps:
(1) Preprocessing the track plate displacement signal: using various types of characteristics to represent track slab displacement signals, normalizing time sequence signals to a (0, 1) range, segmenting the original track slab displacement signals and extracting signal characteristics, wherein the signal characteristic extraction is to rearrange data characteristics in small-segment data obtained by segmentation into a group of data after calculating the data characteristics to become characteristic data; after signal feature extraction, the feature data of each section are connected in sequence to form a final feature data set;
the data features in the signal feature extraction process comprise a maximum value, a minimum value, an average value, a peak-peak value, a rectification average value, a variance, a standard deviation, kurtosis, a root mean square, a waveform factor, a peak factor, a kurtosis factor, a pulse factor and a margin factor;
(2) And calculating a corresponding result correlation weight for each feature of the obtained feature data set by utilizing a Relief algorithm, and eliminating the feature with lower weight according to a set weight threshold value to obtain a final trained feature sequence.
3. The method for detecting the arch state of the track slab based on the deep learning technology according to claim 1, wherein the step (4) further comprises the step (5): and carrying out post-processing according to the obtained final result of the arch state judgment of the track plate, wherein the post-processing is to judge whether the track plate is arched or not by combining other judging programs, and the classification result can correct obvious recognition errors, further optimize the recognition effect and carry out early warning work.
4. The method for detecting the arch-up state of a track slab based on the deep learning technique according to claim 1, wherein in the step (2), when constructing the database, 104 meters, namely 16 track slabs are selected as the length of the vehicle-slab type track dynamics model considering the CA mortar void effect, the vehicle speed is 300 km/h, the displacement of the track slabs is selected as the sample data when the train passes through one track slab, and the sampling interval is 10% -4 And when s, obtaining different sample data by setting different void lengths, setting CA mortar void conditions with longitudinal lengths of 0 meter, 0.325 meter, 0.65 meter, 1.3 meter and 1.95 meter to obtain samples with five void conditions, and obtaining enough sample numbers for each void setting condition to obtain 64 groups of samples.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010275388.1A CN111651840B (en) | 2020-04-09 | 2020-04-09 | Track slab arch state detection method based on deep learning technology |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010275388.1A CN111651840B (en) | 2020-04-09 | 2020-04-09 | Track slab arch state detection method based on deep learning technology |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111651840A CN111651840A (en) | 2020-09-11 |
CN111651840B true CN111651840B (en) | 2024-04-02 |
Family
ID=72352449
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010275388.1A Active CN111651840B (en) | 2020-04-09 | 2020-04-09 | Track slab arch state detection method based on deep learning technology |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111651840B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113191051B (en) * | 2021-04-30 | 2022-07-05 | 西南交通大学 | Establishment method and application of nonlinear Hertz spring |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE10111957A1 (en) * | 2001-03-13 | 2002-09-19 | Baumann Theodor | High speed rail track in tubing tube matches tubing width to support spacing appropriate to track scale and support legs splayed out at set angle to bear centrally between tubing ring joins |
CN101719183A (en) * | 2009-12-10 | 2010-06-02 | 北京交通大学 | Test simulation system for rail structures of high-speed railways and urban railway system |
CN103678823A (en) * | 2013-12-25 | 2014-03-26 | 西南交通大学 | Plate-type ballastless track structure evaluation method based on CA mortar damage |
CN105631124A (en) * | 2015-12-28 | 2016-06-01 | 北京交通大学 | Ballastless track damage analysis method performing combined solution based on definite element expansion and fatigue analysis |
CN106650067A (en) * | 2016-12-12 | 2017-05-10 | 贵州理工学院 | Vehicle-track-bridge coupled vibrationfrequency domain model with middle and high frequency response output |
CN107729649A (en) * | 2017-10-16 | 2018-02-23 | 西南交通大学 | A kind of method for calculating stock rail coupled system dynamics |
-
2020
- 2020-04-09 CN CN202010275388.1A patent/CN111651840B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE10111957A1 (en) * | 2001-03-13 | 2002-09-19 | Baumann Theodor | High speed rail track in tubing tube matches tubing width to support spacing appropriate to track scale and support legs splayed out at set angle to bear centrally between tubing ring joins |
CN101719183A (en) * | 2009-12-10 | 2010-06-02 | 北京交通大学 | Test simulation system for rail structures of high-speed railways and urban railway system |
CN103678823A (en) * | 2013-12-25 | 2014-03-26 | 西南交通大学 | Plate-type ballastless track structure evaluation method based on CA mortar damage |
CN105631124A (en) * | 2015-12-28 | 2016-06-01 | 北京交通大学 | Ballastless track damage analysis method performing combined solution based on definite element expansion and fatigue analysis |
CN106650067A (en) * | 2016-12-12 | 2017-05-10 | 贵州理工学院 | Vehicle-track-bridge coupled vibrationfrequency domain model with middle and high frequency response output |
CN107729649A (en) * | 2017-10-16 | 2018-02-23 | 西南交通大学 | A kind of method for calculating stock rail coupled system dynamics |
Non-Patent Citations (3)
Title |
---|
板式无碴轨道CA砂浆的配制和性能;赵东田等;天津大学学报;第41卷(第7期);793-799 * |
横向有限条与无砟轨道板段单元的车轨***竖向振动分析法;向俊;赫丹;曾庆元;;铁道学报(04);全文 * |
高速铁路无砟轨道结构伤损行为及其对动态性能的影响;朱胜阳;中国优秀博士学位论文全文数据库信息科技辑;全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN111651840A (en) | 2020-09-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111144039B (en) | Train dynamic weighing system and weighing method based on deep learning | |
CN110704801B (en) | Bridge cluster structure operation safety intelligent monitoring and rapid detection complete method | |
CN106021789A (en) | Fuzzy-intelligence-based rail car suspension system fault classification method and system | |
JP2010236875A (en) | Structure state change detection system | |
CN101266190A (en) | Normal traffic flow stayd-cable bridge mode parametric measurement device and method | |
Zhang et al. | Subway tunnel damage detection based on in-service train dynamic response, variational mode decomposition, convolutional neural networks and long short-term memory | |
Jiang et al. | Identification of vehicles moving on continuous bridges with rough surface | |
CN110806324A (en) | Wheel polygon abrasion detection method based on rail displacement and data acquisition equipment | |
CN116052435B (en) | Urban road congestion influence range definition and road influence calculation method | |
CN111651840B (en) | Track slab arch state detection method based on deep learning technology | |
CN105923014A (en) | Track longitudinal irregularity amplitude value estimation method based on evidential reasoning rule | |
CN116881819B (en) | Stay cable working state monitoring method based on isolated forest | |
CN114882069A (en) | Taxi track abnormity detection method based on LSTM network and attention mechanism | |
Liu et al. | Deep learning based identification and uncertainty analysis of metro train induced ground-borne vibration | |
CN115388815A (en) | Static measurement method and device for irregularity of track functional part of magnetic suspension system | |
Zhang et al. | Structural Damage Identification System Suitable for Old Arch Bridge in Rural Regions: Random Forest Approach. | |
CN110530589A (en) | A kind of beam bridge impact coefficient detection method considering continuous wagon flow coupling effect | |
Nguyen et al. | Damage detection of structural based on indirect vibration measurement results combined with Artificial Neural Network | |
CN117669389A (en) | Random vibration analysis method for earthquake-vehicle-bridge system based on deep learning | |
CN117390551A (en) | Operation monitoring level high-speed railway bridge influence line tracking identification method | |
CN117093911A (en) | Bridge sliding support abrasion identification method based on longitudinal temperature-induced displacement amplitude monitoring of main beam | |
CN116822024A (en) | Method for determining least favored crossing position of multi-line train on railway bridge | |
Shen et al. | A boom damage prediction framework of wheeled cranes combining hybrid features of acceleration and Gaussian process regression | |
Bahamon-Blanco et al. | Recognition of track defects through measured acceleration using a recurrent neural network | |
Yuan et al. | Unsupervised cross-domain damage detection and localization for vibration isolators in metro floating-slab track |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |