High-speed railway bridge service state evaluation method and system based on non-contact visual perception

Abstract

The invention belongs to the technical field of intelligent operation and maintenance, and particularly discloses a high-speed railway bridge service state evaluation method and system based on non-contact visual perception, and the method comprises the following steps: carrying out the non-contact monitoring of the displacement of multiple measurement points of a railway bridge in a complex environment; efficiently identifying railway bridge state parameters; and evaluating the service state of the railway bridge. The system monitors the service state of the bridge remotely in a non-contact manner, performs real-time early warning on the safety state of the bridge, further guarantees the driving safety of a high-speed train, and is of great significance in damage diagnosis and disaster prevention and reduction of the bridge in a hard mountainous area.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent operation and maintenance, and particularly to a method and system for evaluating the service status of high-speed railway bridges based on non-contact visual perception. Background Art

[0002] The volume of bridge construction in China ranks first in the world. As an important part of China's transportation, railway bridges have prominent problems such as many potential safety hazards, short service life, serious damage caused by disasters, "aging" appearance, and increasing maintenance and management costs due to factors such as material property degradation, vehicle impact, and natural disasters. Therefore, how to ensure the service safety of large-scale bridges in China is a major national demand and a key concern of the industry. The Outline for the Construction of a Transportation Power clearly requires "strengthening the operation monitoring and detection of infrastructure, and vigorously promoting the deep integration of new technologies such as big data, Internet of Things, and artificial intelligence with the transportation industry". Using structural inspection and monitoring technology to timely discover structural diseases and potential hazards is expected to solve the problem of structural service safety and provide strong support for achieving the long life of structures.

[0003] In recent years, the theory of structural health monitoring using various advanced sensor technologies has developed rapidly. They play an active role in early warning of emergencies such as earthquakes, typhoons, vehicle collisions, and rockfall impacts. However, they are restricted by cost and efficiency in the application of identifying the service states of medium and small-span railway bridges with large quantities and wide areas. Therefore, the structural state identification technology based on wireless sensing technology and long-distance monitoring equipment is the key research and development direction in the future. Domestic and foreign scholars have gone through the research process of "mobile testing - indirect measurement - non-contact monitoring" and have done a lot of research on the rapid monitoring of bridge responses, improving the efficiency of bridge state monitoring. The above rapid response monitoring methods have been widely used in bridge engineering. However, at present, only the identification of basic parameters such as bridge frequency, damping ratio, and displacement vibration mode can be realized. Although the identified parameters are crucial for structural model modification, damage diagnosis, and vibration control, they are difficult to be directly applied to the automatic identification of bridge state parameters and safety early warning. Summary of the Invention

[0004] To solve the problems existing in the prior art, the present invention provides a method and system for evaluating the service status of high-speed railway bridges based on non-contact visual perception, which can overcome the defects of time-consuming and laborious installation of contact sensing technology, simple structural identification results, and large false alarm and missed alarm rates in bridge state early warning, and solve the problems mentioned in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solution: A method for evaluating the service status of high-speed railway bridges based on non-contact visual perception, comprising the following steps:

[0006] Step S1, non-contact monitoring of displacements at multiple measurement points of railway bridges in complex environments;

[0007] Step S2, efficient identification of railway bridge state parameters;

[0008] Step S3, evaluation of the service state of railway bridges.

[0009] Preferably, in step S1, the non-contact monitoring of multi-point displacements of railway bridges in complex environments specifically includes: using a high-definition industrial camera to remotely capture infrared targets distributed at multiple points on the bridge, obtaining a vibration image sequence of the bridge under the running of high-speed trains, building a lightweight deep learning neural network to process images in different environments, and accurately obtaining multi-point dynamic displacement data of the bridge.

[0010] Preferably, in step S2, the efficient identification of railway bridge state parameters specifically includes:

[0011] 1) Identification of time-varying modal parameters of the train-bridge coupling system:

[0012] Based on the multi-point dynamic displacement data of the bridge obtained by non-contact vision monitoring, using the adaptive modal decomposition algorithm to process the monitored displacements, obtaining the short-period dynamic component and long-period static component of the railway bridge under the running of high-speed trains, and then analyzing the dynamic displacement component to identify the time-varying modal parameters of the bridge under train running;

[0013] 2) Identification of the bridge mode scaling factor based on time-varying modal parameters:

[0014] Substitute the identified time-varying modal parameters into the characteristic equation of the high-speed train-bridge coupling system

[0015]

[0016] where: [M w , [M v , [K v , [K B are respectively the diagonal arrangement matrices of the vehicle body mass, wheel mass, suspension system stiffness, and ballast stiffness at the location of n spring-mass units on the bridge, all of which are n×n order; [N c is the sequence of structure shape function matrices at the locations of all spring-mass units on the bridge, that is [M b , [K b are respectively the mass and stiffness matrices of the bridge structure; ω cr is the r-th order frequency of the vehicle-bridge coupling system, is the r-th order displacement mode shape of the bridge structure unit, are respectively the r-th order displacement mode shapes of the wheels and vehicle body mass points in the train spring-mass sequence;

[0017] Expanding the above equation term by term and performing mathematical transformations, the simplified vibration equation can be obtained:

[0018]

[0019] Where: [η r = [I n + [γ r , [γ r = ([K v + [K B - [K v [β r - ω cr 2 [M w ) -1 [K B , [I n is an n×n order unit diagonal matrix;

[0020] Combining the above equation with the bridge vibration equation without train running, we can get:

[0021]

[0022] Where: ωor is the r-th order frequency of the bridge without train running;

[0023] Expressing the parameters in the above vibration equation according to vertical and rotational degrees of freedom, that is:

[0024]

[0025] Where: Ncv and N cθ are the shape function matrices corresponding to vertical and rotational degrees of freedom; and are the r-th order displacement vibration modes corresponding to vertical and rotational degrees of freedom respectively; M bv and M bθ are the bridge mass matrices corresponding to vertical and rotational degrees of freedom respectively;

[0026] Expanding the above formula, we can get:

[0027]

[0028] Where: [T] is the conversion relationship between the rotational displacement vibration mode and the vertical displacement vibration mode;

[0029] Multiplying the left side of by the above expression, and according to the definition of the mass-normalized vibration mode we can get:

[0030]

[0031] Set the unscaled displacement mode shape of the bridge structure and the mass-normalized mode shape The relationship between them is The expression for the modal mode shape scaling factor is obtained as follows:

[0032]

[0033] 3) Identification of bridge modal flexibility:

[0034] Multiply the calculated modal mode shape scaling factor by the mode shape of the railway bridge when there is no train passing through to obtain the mass-normalized mode shape of the bridge Thereby reconstructing the structural frequency response function matrix:

[0035]

[0036] In the formula: m is the order of the identified mode; i represents a complex number, that is, i 2 =-1; ξ r is the damping ratio of the r-th order of the structure;

[0037] The frequency response function at ω = 0 is the modal flexibility of the bridge:

[0038]

[0039] Preferably, in step S3, the evaluation of the service state of the railway bridge specifically includes:

[0040] S31. Based on the long-period static component, considering the correlation between temperature, displacement, and vehicle speed, determine the key characteristic indicators and then establish a multi-level threshold interval state evaluation model to evaluate the state level of the structure according to the threshold interval level;

[0041] S32. Based on the identified bridge state parameters and the monitored displacement data, construct a monitoring information statistical distribution evaluation model to evaluate the state of the heavy-haul railway bridge based on the change amount of the key characteristic indicator monitoring data;

[0042] S33. Integrate the results obtained from the two evaluation models to evaluate the service state of the railway bridge.

[0043] Preferably, in step S31, six key characteristic indicators are selected to evaluate the state level of the structure, including lateral frequency, vertical frequency, mid-span deflection, mid-span lateral displacement, lateral vibration acceleration, and vertical vibration acceleration; at the same time, according to the limited ranges of the thresholds of each index in different railway bridge specifications, four different levels of threshold intervals are constructed: A1, A2, A3, A4; when the monitored values are input into the multi-level evaluation model, the level threshold interval to which each index belongs is obtained, and the state level evaluation of the structural operation performance depends on the threshold interval with the highest evaluation level.

[0044] Preferably, in step S32, statistical distribution processing is performed on six key feature indicators. The monitoring signal is regarded as a population, which approximately satisfies the normal distribution, and each monitoring data is a sample value in the population. Let the total amount of sample X be n, and the sample has a mean and variance The change amount of parameter β is used to reflect the health degree of the structure, which is divided into three health degree levels: B1, B2, and B3. Among them,

[0045]

[0046] The analytic hierarchy process is introduced, and the correlation degree of the monitoring data of each key index is calculated based on the interval extension theory, and then the value ranges of the three bridge health level intervals are defined; the correlation degree calculation formula is:

[0047]

[0048] K t (X i ) represents the degree to which the i-th evaluation index belongs to the state level t; in the formula, ρ is used to represent the distance between the calculation intervals, and the interval <a i ,b i > represents the value range of the index to be evaluated with respect to the i-th feature.

[0049] On the other hand, to achieve the above object, the present invention also provides the following technical solution: A high-speed railway bridge service state evaluation system based on non-contact visual perception, including the following modules:

[0050] Non-contact monitoring module for multi-point displacement of railway bridges: Use a high-definition industrial camera to remotely capture infrared targets distributed at multiple points on the bridge, obtain the vibration image sequence of the bridge under the driving of high-speed trains, and build a lightweight deep learning network to process images in different environments to accurately obtain the dynamic displacement data of multiple points on the bridge;

[0051] Efficient identification module for railway bridge state parameters: Use the adaptive time-varying modal parameter identification algorithm to process the monitored displacement, obtain the time-varying modal parameters of the bridge under the driving of the train, and then establish the mapping relationship between the train spatio-temporal information and the time-varying modal parameters to obtain the identification of the mass-normalized vibration mode, modal mass, and modal flexibility parameters of the bridge;

[0052] Service state evaluation module for railway bridges: Build a multi-level threshold interval state evaluation model and a monitoring information statistical distribution evaluation model, integrate the results obtained from the two evaluation models, and evaluate the service state of the railway bridge.

[0053] The beneficial effects of the present invention are as follows: The present invention uses non-contact visual perception technology to non-contact and remotely monitor the dynamic displacement effects of multiple measuring points on railway bridges under the driving of trains, and uses a lightweight deep neural network to achieve high-precision extraction of bridge displacements in complex environments with changing lighting; then, through the adaptive variational mode decomposition algorithm, the dynamic and static components of the collected responses under the driving of trains are separated, and combined with the train-bridge coupling vibration equation, a mapping relationship between the time-varying frequency of the train-bridge coupling system and the structural mode scaling factor is established, and then the bridge modal flexibility identification and static deflection prediction are realized; finally, a railway bridge structure state evaluation model that integrates multi-level threshold intervals and monitoring information statistical distribution characteristic indicators is proposed to comprehensively evaluate and warn the service state of the bridge. The method of the present invention can non-contact, remotely, and with high precision monitor the dynamic displacements of multiple measuring points on the bridge under the driving of trains, and can overcome the defects of time-consuming and laborious installation of contact sensing technology, simple structural identification results, and large false alarm and missed alarm rates in bridge state early warning, and can effectively ensure the efficient evaluation of the service states of numerous railway bridges on the railway network. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 Schematic diagram of the process of the high-speed railway bridge service state evaluation method based on non-contact visual perception in Embodiment 1;

[0055] Figure 2 Schematic diagram of multi-point displacement monitoring of railway bridges in complex environments in Embodiment 2;

[0056] Figure 3 Schematic diagram of the dynamic displacement of the bridge under the driving of trains in Embodiment 2;

[0057] Figure 4 Schematic diagram of the time-varying frequency of the train-bridge coupling system in Embodiment 2;

[0058] Figure 5 Schematic diagram of the bridge mode scaling factor in Embodiment 2;

[0059] Figure 6 Schematic diagram of the mass-normalized mode of the bridge in Embodiment 2;

[0060] Figure 7 Three-dimensional surface diagram of the bridge modal flexibility matrix in Embodiment 2;

[0061] Figure 8 Schematic diagram of the predicted static deflection of the bridge under virtual load in Embodiment 2;

[0062] Figure 9 Schematic diagram of the comprehensive service state evaluation results of the bridge integrating multi-source monitoring data in Embodiment 2. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0064] Embodiment 1

[0065] The present invention provides a technical solution: a method for evaluating the service status of high-speed railway bridges based on non-contact visual perception. Please refer to Figure 1 , which includes the following steps:

[0066] Step S1, non-contact monitoring of multi-point displacements of railway bridges in complex environments; establish a vibration image database of railway bridges in different environments, construct a lightweight deep neural network suitable for infrared target semantic segmentation and image matching tracking, and train it through a large amount of image databases under different environmental conditions, so as to adjust the real-time collected image data to obtain a sequence of bridge vibration images with obvious image brightness changes and obvious image features; furthermore, realize the high-precision and robust extraction of multi-point dynamic displacements of the bridge structure through an adaptive feature point detection and extraction algorithm.

[0067] Step S2, efficient identification of railway bridge state parameters; realize the separation of dynamic and static components of the collected response under train operation through the adaptive variational mode decomposition algorithm, and then realize the identification of time-varying modal parameters of the bridge when the train is running. Combine the train-bridge coupling vibration equation to establish a mapping relationship between the time-varying frequency of the train-bridge coupling system and the structural vibration mode scaling factor, and realize the identification of bridge modal flexibility and the prediction of static deflection.

[0068] Step S3, evaluation of the service status of railway bridges; establish a multi-level threshold interval state evaluation model to evaluate the state structure level with the threshold interval level; at the same time, construct a monitoring information statistical distribution evaluation model to evaluate the state of heavy-haul railway bridges based on the change amount of key feature index monitoring data. Finally, integrate the results obtained from the two evaluation models to evaluate the operating performance state of the railway bridge.

[0069] Furthermore, in step S1, the non-contact monitoring of multi-point displacements of railway bridges in complex environments specifically includes: using a high-definition industrial camera to remotely capture infrared targets distributed at multiple points on the bridge to obtain a sequence of vibration images of the bridge under the running of high-speed trains, and building a lightweight deep learning neural network to process images in different environments to accurately obtain multi-point dynamic displacement data of the bridge.

[0070] Furthermore, in step S2, the efficient identification of railway bridge state parameters specifically includes:

[0071] 1) Identification of time-varying modal parameters of train-bridge coupling system:

[0072] Based on the dynamic displacement data of multiple measuring points on the bridge obtained by non-contact visual monitoring, the adaptive modal decomposition algorithm is used to process the monitored displacement to obtain the short-period dynamic component and long-period static component of the railway bridge under the running of high-speed trains, and then the dynamic displacement component is analyzed to identify the time-varying modal parameters of the bridge under train running;

[0073] 2) Identification of bridge mode shape scaling factor based on time-varying modal parameters:

[0074] Substitute the identified time-varying modal parameters into the characteristic equation of the high-speed train-bridge coupling system

[0075]

[0076] where: [M w , [M v , [K v , [K B are respectively the diagonal arrangement matrices of the vehicle body mass, wheel mass, suspension system stiffness, and ballast stiffness at the position of n spring-mass units on the bridge, all of which are of order n×n; [N c is the sequence of structure shape function matrices at the positions of all spring-mass units on the bridge, that is [M b , [K b are respectively the mass and stiffness matrices of the bridge structure; ω cr is the r-th order frequency of the vehicle-bridge coupling system, is the r-th order displacement mode shape of the bridge structure unit, are respectively the r-th order displacement mode shapes of the wheel and vehicle body mass points in the train spring-mass sequence;

[0077] Expand the above equation item by item and perform mathematical transformation to obtain the simplified vibration equation:

[0078]

[0079] where: [η r = [I n + [γ r , [γ r = ([K v + [K B - [K v [β r - ω cr 2 [M w ) -1 [K B, [I n is an \(n\times n\) unit diagonal matrix;

[0080] Combining the above equation with the bridge vibration equation without train running, we can get:

[0081]

[0082] where: \(\omega_{or}\) is the \(r\)-th order frequency of the bridge without train running;

[0083] Express the parameters in the above vibration equation according to the vertical and rotational degrees of freedom, that is:

[0084]

[0085] where: \(N_{cv}\) and \(N\) cθ are the shape function matrices corresponding to the vertical and rotational degrees of freedom; and are the \(r\)-th order displacement vibration modes corresponding to the vertical and rotational degrees of freedom respectively; \(M\) bv and \(M\) bθ are the bridge mass matrices corresponding to the vertical and rotational degrees of freedom respectively;

[0086] Expanding the above formula, we can get:

[0087]

[0088] where: \([T]\) is the conversion relationship between the rotational displacement vibration mode and the vertical displacement vibration mode;

[0089] Multiply on the left of the above expression, and according to the definition of the mass-normalized vibration mode we can get:

[0090]

[0091] Assume that the relationship between the unscaled displacement vibration mode of the bridge structure and the mass-normalized vibration mode is to obtain the expression of the modal vibration mode scaling factor:

[0092]

[0093] 3) Bridge modal flexibility identification:

[0094] Multiply the calculated modal vibration mode scaling factor by the vibration mode of the railway bridge without train passing, and we can get the mass-normalized vibration mode of the bridge Thus, the structural frequency response function matrix is reconstructed:

[0095]

[0096] where: m is the order of the identified mode; i represents an imaginary number, i.e., i 2 = -1; ξ r is the damping ratio of the r-th order of the structure;

[0097] The frequency response function at ω = 0 is the modal flexibility of the bridge:

[0098]

[0099] Furthermore, in step S3, the assessment of the service state of the railway bridge specifically includes:

[0100] S31. Based on the long-term static component, considering the correlation between temperature, displacement, and vehicle speed, determine the key characteristic indicators and then establish a multi-level threshold interval state assessment model to evaluate the state level of the structure according to the threshold interval level;

[0101] Six key characteristic indicators are selected to evaluate the state level of the structure, including lateral frequency, vertical frequency, mid-span deflection, mid-span lateral displacement, lateral vibration acceleration, and vertical vibration acceleration; meanwhile, according to the limit ranges of the thresholds of each index in different railway bridge specifications, four different levels of threshold intervals are constructed: A1, A2, A3, A4; when the monitoring values are input into the multi-level assessment model, the level threshold interval to which each index belongs is obtained, and the state level assessment of the structural operation performance depends on the threshold interval with the highest evaluation level.

[0102] S32. Based on the identified bridge state parameters and the monitored displacement data, construct a monitoring information statistical distribution assessment model to evaluate the state of the heavy-haul railway bridge based on the change amount of the key characteristic indicator monitoring data;

[0103] Perform statistical distribution processing on the six key characteristic indicators. Consider the monitoring signal as a population, which approximately satisfies the normal distribution, and each monitoring data is a sample value in the population; assume that the total amount of the sample X is n, and the sample has a mean and variance Use the change amount of the parameter β to reflect the health degree of the structure, which is divided into three health degree levels: B1, B2, B3, where

[0104]

[0105] Introduce the analytic hierarchy process, calculate the correlation degree of the monitoring data of each key index based on the interval extension theory, and then define the value ranges of three bridge health level intervals; the correlation degree calculation formula is:

[0106]

[0107] K t (Xi ) is the degree to which the i-th evaluation index belongs to the state level t; in the formula, ρ is used to represent the distance between calculation intervals, and the interval <a i , b i > represents the value range of the index to be evaluated with respect to the i-th feature.

[0108] S33. Integrate the results obtained from the two evaluation models to evaluate the service status of the railway bridge.

[0109] On the other hand, the present invention also provides the following technical solution: A service status evaluation system for high-speed railway bridges based on non-contact visual perception, including the following modules:

[0110] Non-contact monitoring module for multi-point displacement of railway bridges: Use a high-definition industrial camera to remotely capture infrared targets distributed at multiple points on the bridge to obtain a vibration image sequence of the bridge under the driving of high-speed trains, and build a lightweight deep learning network to process images in different environments to accurately obtain multi-point dynamic displacement data of the bridge;

[0111] Efficient identification module for railway bridge state parameters: Use an adaptive time-varying modal parameter identification algorithm to process the monitored displacement, obtain the time-varying modal parameters of the bridge under the driving of the train, and then establish a mapping relationship between the train spatio-temporal information and the time-varying modal parameters to obtain the identification of the mass-normalized vibration mode, modal mass, and modal flexibility parameters of the bridge;

[0112] Service status evaluation module for railway bridges: Build a multi-level threshold interval state evaluation model and a monitoring information statistical distribution evaluation model, integrate the results obtained from the two evaluation models, evaluate the service status of the railway bridge, and conduct real-time early warning.

[0113] Embodiment 2

[0114] Combined with Figures 2 to 9 , through a typical simply supported railway bridge case, the specific implementation process of a service status evaluation method for high-speed railway bridges based on non-contact visual perception in this embodiment is described. The specific steps are as follows:

[0115] 1) Non-contact monitoring of multi-point displacement of railway bridges in complex environments: According to the size of the shooting field of view, determine the appropriate camera resolution and acquisition frame rate, and control the high-speed camera to capture the vibration image sequence of the simply supported railway bridge under the driving of high-speed trains and under different lighting conditions changes (such as Figure 2As shown); establish a vibration image database of railway bridges under different environmental conditions, construct a lightweight deep neural network suitable for infrared target semantic segmentation and image matching tracking, and train it on a large amount of image databases under different environmental conditions, so as to adjust the real-time collected image data to obtain a bridge vibration image sequence with obvious image brightness changes and obvious image features; furthermore, achieve high-precision and robust extraction of dynamic displacements at multiple measurement points of the bridge structure through an adaptive feature point detection and extraction algorithm.

[0116] 2) Efficient identification of railway bridge state parameters: The total length of the simply supported concrete beam bridge selected in this case is 30 m, the elastic modulus of the material is 34.5 GPa, the weight per unit length is 10832 kg / m, and the sampling frequency is 400 Hz. The high-speed train has 16 carriages, the wheelbase between the front and rear wheels of the same carriage is 18 m, the wheelbase between the front and rear carriages is 7 m, the wheel mass and the car body mass are 5000 kg and 48000 kg respectively, the vehicle system stiffness and damping are 1.5×10 6 N / m and 8.5×10 4 N·s / m, and the driving speed is 100 km / h. The specific steps are as follows:

[0117] ① Identification of time-varying frequencies of the railway-bridge coupling system

[0118] Arrange 20 monitoring points on this simply supported beam structure. The dynamic displacement curve of the bridge monitored by the non-contact vision measurement system during train running is as Figure 3 shown. It can be seen that when the high-speed train passes through the bridge, the bridge displacement decreases rapidly, and then vibrates greatly under the excitation of the train. Subsequently, the displacement gradually increases and slowly returns to 0 mm; the measuring point 10 is the mid-span position of the bridge, and its vibration amplitude is greater than the maximum vibration displacements of the measuring points 5 and 15. Then, use the adaptive mode decomposition algorithm to process the monitored vibration displacement to obtain its short-period static displacement component and long-period dynamic displacement component. Perform Hilbert-Huang transform on the separated dynamic displacement component to identify the time-varying frequency of the train-bridge coupling system. For this simply supported beam bridge case, its first 3-order time-varying modal frequencies are as Figure 4 shown. It can be seen that the time-varying frequency of the train-bridge coupling system changes with the position of the train running. For the first 2-order time-varying frequencies, the coupling system frequency first decreases, then fluctuates, and finally increases with the change of the train position. However, the 3rd-order time-varying frequency first increases, then fluctuates, and finally decreases with the change of the train position.

[0119] ② Identification of bridge mode scaling factors

[0120] Substitute the identified time-varying frequency of the train-bridge coupling system into the mode scaling factor expression of the present invention, and the mass-normalized mode scaling factor can be obtained (as Figure 5As shown in the figure, for this simply supported beam bridge case, the scaling factors of the first three vibration modes are 0.00248, 0.00247, and 0.00251 respectively, and the errors between them and the theoretical scaling coefficients are 0.143%, -0.5144%, and 1.135% respectively, meeting the engineering accuracy requirements.

[0121] ③ Identification of bridge modal flexibility

[0122] Multiplying the identified scaling factors of the first three vibration modes of the bridge by the arbitrarily scaled displacement vibration modes of the bridge can obtain the mass-normalized vibration modes of the bridge. For this embodiment, the identified first three mass-normalized vibration modes are as Figure 6 shown. It can be seen that the mass-normalized vibration modes obtained by the method of the present invention are in good agreement with the theoretical values, verifying the correctness of the method of the present invention. Then, by combining the mass-normalized vibration modes with the basic modal parameters of the bridge, the modal flexibility of the bridge can be reconstructed. The three-dimensional surface of the modal flexibility identified in this embodiment is as Figure 7 shown.

[0123] ④ Prediction of static deflection of bridge under virtual load

[0124] After obtaining the modal flexibility matrix of the bridge, the static deflection of the bridge under any virtual load can be predicted within the elastic range. In this embodiment, a uniformly distributed load with an amplitude of 9.8 kN is applied to 20 monitoring points of the simply supported beam bridge. Multiplying this load vector by the identified modal flexibility matrix can obtain the deformation of the bridge under this load (as Figure 8 shown). It can be seen that the static deformation of the bridge predicted by the method of the present invention is in good agreement with the theoretical values, further verifying the correctness and effectiveness of the method of the present invention.

[0125] 3) Service state assessment and real-time warning of railway bridges: Establish a multi-level threshold interval state assessment model to evaluate the state level of the structure based on the threshold interval level; at the same time, construct a monitoring information statistical distribution assessment model to evaluate the state of heavy-haul railway bridges based on the change amount of the monitoring data of key characteristic indicators. Finally, integrate the results obtained from the two assessment models to evaluate the operating performance state of railway bridges.

[0126] First, determine key characteristic indicators such as the lateral frequency, vertical frequency, mid-span deflection, mid-span lateral displacement, lateral vibration acceleration, and vertical vibration acceleration of the main girder. According to the control ranges of the thresholds of key characteristic indicators in different railway bridge codes, establish four different levels of threshold intervals, namely A1, A2, A3, and A4. Input the six key index values of the lateral frequency, vertical frequency, mid-span deflection, mid-span lateral displacement, lateral vibration acceleration, and vertical vibration acceleration of the main girder into the evaluation model, and the level threshold interval to which each index belongs can be obtained. The four monitoring data values are all A1. Then, establish a statistical distribution model of monitoring information, introduce the "3σ" method of normal distribution to determine the value range of the β health level, and divide the health level into three levels according to relevant standards. Among them, the confidence level of the first-level health level is 0.75; the confidence level of the second-level health level is 0.2; the confidence level of the third-level health level is 0.05. Finally, the following health confidence intervals are obtained:

[0127]

[0128] Then, calculate the correlation degree to obtain which confidence health interval the key characteristic indicators of the bridge structure belong to. The health levels of the lateral and vertical accelerations are B1, and the health levels of the lateral and vertical displacements are B2. Finally, fuse the multi-level threshold interval and the statistical distribution evaluation model of monitoring information to obtain the comprehensive state assessment status of the bridge, as Figure 9 shown. Through the analysis of historical monitoring data, the performance degradation state of the bridge can be tracked, which is convenient for formulating maintenance decisions in a timely manner.

[0129] The present invention uses non-contact visual perception technology to non-contact and remotely monitor the dynamic displacement effects of multiple measurement points of railway bridges under the driving of trains, and uses a lightweight deep neural network to achieve high-precision extraction of bridge displacements in complex environments with changing illumination; then, through an adaptive variational mode decomposition algorithm, the dynamic and static components of the collected responses under the driving of trains are separated. Combining the train-bridge coupling vibration equation, a mapping relationship between the time-varying frequency of the train-bridge coupling system and the structural mode scaling factor is established, and then the modal flexibility identification and static deflection prediction of the bridge are realized; finally, a railway bridge structure state evaluation model integrating multi-level threshold intervals and statistical distribution characteristic indicators of monitoring information is proposed to comprehensively evaluate and warn the service state of the bridge. The method of the present invention can non-contact, remotely, and highly precisely monitor the dynamic displacements of multiple measurement points of the bridge under the driving of trains, and can overcome the defects of time-consuming and laborious installation of contact sensing technology, simple structural identification results, and large false alarm and missed alarm rates of bridge state warnings, and can effectively guarantee the efficient evaluation of the service states of numerous railway bridges on the railway network.

[0130] In several embodiments provided by the embodiments of the present invention, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device and method embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of devices, methods, and computer program products according to multiple embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the part of the module, program segment, or code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0131] In addition, the functional modules in each embodiment of the present invention may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0132] It should be noted that in this article, the terms "including", "comprising", or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or elements inherent to such a process, method, article, or device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of another identical element in the process, method, article, or device including the said element.

[0133] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The singular forms "a", "the", and "said" used in the embodiments of the present invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0134] It should be understood that the term " / and" used herein is only a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after.

[0135] Depending on the context, as used herein, the word "if" can be interpreted as "when" or "while" or "in response to determining" or "in response to detecting". Similarly, depending on the context, the phrase "if determined" or "if detected (stated condition or event)" can be interpreted as "when determined" or "in response to determining" or "when detected (stated condition or event)" or "in response to detecting (stated condition or event)".

[0136] The "first / second" mentioned in the embodiments is only used to distinguish similar objects and does not represent a specific order for the objects. It can be understood that the "first / second" can be interchanged in a specific order or sequence when permitted. It should be understood that the objects distinguished by the "first / second" can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in an order other than those illustrated or described herein.

[0137] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for evaluating the service status of a high-speed railway bridge based on non-contact visual perception, characterized in that: The steps include: Step S1, non-contact monitoring of displacement of multiple measuring points of railway bridges in complex environments; Step S2, efficient identification of railway bridge status parameters; Step S3: evaluating the service status of railway bridges.

2. The method for evaluating the service status of a high-speed railway bridge based on non-contact visual perception according to claim 1 is characterized in that: In step S1, the non-contact monitoring of the displacement of multiple measuring points of a railway bridge in a complex environment specifically includes: using a high-definition industrial camera to shoot infrared targets distributed at multiple points on the bridge at a long distance, obtaining a vibration image sequence of the bridge under the movement of a high-speed train, building a lightweight deep learning neural network to process images in different environments, and obtaining dynamic displacement data of multiple measuring points of the bridge with high precision.

3. The method for evaluating the service status of a high-speed railway bridge based on non-contact visual perception according to claim 1 is characterized in that: In step S2, the efficient identification of railway bridge status parameters specifically includes: 1) Identification of time-varying modal parameters of train-bridge coupling system: Based on the dynamic displacement data of multiple measuring points of the bridge monitored by non-contact visual monitoring, the adaptive modal decomposition algorithm is used to process the monitored displacement to obtain the short-period dynamic component and long-period static component of the railway bridge under the passage of high-speed trains, and then the dynamic displacement component is analyzed to identify the time-varying modal parameters of the bridge under the passage of trains; 2) Identification of bridge vibration mode scaling factors based on time-varying modal parameters: Substitute the identified time-varying modal parameters into the characteristic equation of the high-speed train-bridge coupling system Among them: [M w ]、[M v ]、[K v ]、[K B ] are the diagonally arranged matrices of the vehicle body mass, wheel mass, suspension system stiffness, and ballast stiffness of the n spring mass units located on the bridge, all of which are n×n order; [N c ] is the structural shape function matrix sequence at the locations of all spring mass units on the bridge, that is, [M b ]、[K b ] are the mass and stiffness matrices of the bridge structure respectively; ω cr is the rth order frequency of the vehicle-bridge coupling system, is the r-th displacement mode of the bridge structural unit, are the r-th order displacement vibration modes of the wheels and body mass points in the train spring mass sequence respectively; By expanding the above equation and performing mathematical transformation, we can obtain the simplified vibration equation: Where: [η r ]=[I n ]+[γ r ],[γ r ]=([K v ]+[K B ]-[K v ][β r ]-ω cr 2 [M w ]) -1 [K B ],[I n ] is an n×n unit diagonal matrix; Combining the above equation with the bridge vibration equation when there is no train running, we can get: Where: ωor is the rth order frequency of the bridge without trains; The parameters in the above vibration equation are expressed according to the vertical and rotational degrees of freedom, that is: Where: Ncv and N cθ is the shape function matrix corresponding to the vertical and rotational degrees of freedom; and are the r-th order displacement modes corresponding to the vertical and rotational degrees of freedom respectively; M bv and M bθ are the bridge mass matrices corresponding to the vertical and rotational degrees of freedom, respectively; Expand the above formula to get: Where: [T] is the conversion relationship between the angular displacement mode and the vertical displacement mode; Will Multiply the above expression on the left and according to the definition of mass normalized vibration mode We can get: Assume that the bridge structure has no scaled displacement mode shape Mass normalized vibration mode The relationship between The expression for the mode shape scaling factor is obtained: 3) Bridge modal flexibility identification: The normalized vibration mode of the bridge mass is obtained by multiplying the calculated modal vibration mode scaling factor with the vibration mode of the railway bridge when there is no train passing through. Thus reconstructing the structural frequency response function matrix: Where: m is the identified modal order; i represents a complex number, i.e. 2 =-1;ξ r is the r-th order damping ratio of the structure; The frequency response function at ω = 0 is the modal flexibility of the bridge:

4. The method for evaluating the service status of a high-speed railway bridge based on non-contact visual perception according to claim 1 is characterized in that: In step S3, the railway bridge service status assessment specifically includes: S31. Based on the long-period static component, considering the correlation between temperature, displacement and vehicle speed, a multi-level threshold interval state assessment model is established after determining the key characteristic indicators, and the state structure state level is assessed by the threshold interval level; S32. Based on the identified bridge status parameters and monitoring displacement data, a monitoring information statistical distribution evaluation model is constructed, and the status of the heavy-duty railway bridge is evaluated based on the change in the key characteristic indicator monitoring data; S33. Integrate the results of the two evaluation models to evaluate the service status of railway bridges.

5. The method for evaluating the service status of a high-speed railway bridge based on non-contact visual perception according to claim 4 is characterized in that: In step S31, six key characteristic indicators are selected to evaluate the state level of the structure, including lateral frequency, vertical frequency, mid-span deflection, mid-span lateral displacement, lateral vibration acceleration and vertical vibration acceleration; at the same time, according to the limited range of threshold values ​​of each indicator in different railway bridge specifications, four threshold intervals of different levels are constructed: A1, A2, A3, and A4; when the monitoring values ​​are input into the multi-level evaluation model, the level threshold interval to which each indicator belongs is obtained, and the state level assessment of the structural operating performance depends on the threshold interval with the highest evaluation level.

6. The method for evaluating the service status of a high-speed railway bridge based on non-contact visual perception according to claim 4 is characterized in that: In step S32, the six key characteristic indicators are statistically processed, and the monitoring signal is regarded as a whole, which approximately satisfies the normal distribution. The change in parameter β is used to reflect the health of the structure, and is divided into three health levels: B1, B2, and B3, where The analytic hierarchy process is introduced to calculate the correlation of key indicator monitoring data based on interval extension theory, and then three bridge health level interval values ​​are defined; the correlation calculation formula is:

7. A system for evaluating the service status of a high-speed railway bridge based on non-contact visual perception according to any one of claims 1 to 6, characterized in that: Includes the following modules: Railway bridge multi-point displacement non-contact monitoring module: Use high-definition industrial cameras to shoot infrared targets distributed on the bridge at multiple points at a long distance, obtain vibration image sequences of the bridge under the running of high-speed trains, build a lightweight deep learning network to process images in different environments, and obtain dynamic displacement data of multiple measuring points on the bridge with high precision; Railway bridge state parameter efficient identification module: Use the adaptive time-varying modal parameter identification algorithm to process the monitored displacement, obtain the time-varying modal parameters of the bridge under the train, and then establish the mapping relationship between the train's time-space information and the time-varying modal parameters to obtain the bridge's mass normalized vibration shape, modal mass and modal flexibility parameter identification; Railway bridge service status assessment module: construct a multi-level threshold interval status assessment model and a monitoring information statistical distribution assessment model, integrate the results of the two assessment models, and evaluate the service status of railway bridges.

Images