Rapid Assessment Method and Equipment for Residual Displacement of Contact Network on High-Speed ​​Railway Bridges

Abstract

This application proposes a method and equipment for rapid assessment of residual displacement of the overhead contact system on high-speed railway bridges. First, the residual displacement after the main shock is predicted by using the amplified input features of the bridge deck and the main shock index. Then, the residual state after the main shock is explicitly introduced into the aftershock stage. The weights of each sensitive index in the aftershock stage are adaptively updated by the state constraint coefficient, thereby realizing the progressive intelligent prediction of the residual displacement evolution process under the action path of the main shock and aftershock.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method and equipment for rapid assessment of residual displacement of overhead contact lines on high-speed railway bridges. Background Technology

[0002] High-speed railways generally adopt a bridge-instead-of-road construction model. Under seismic action, the overhead contact system on the bridge is simultaneously affected by ground input, bridge dynamic amplification, and inconsistent spatial responses at multiple support points. Compared with a single mainshock, the action of a mainshock-aftershock sequence causes the contact system to continue to accumulate deformation on top of its residual state after the mainshock, resulting in increased residual displacement and affecting the geometric state and operational safety of the contact system.

[0003] Existing methods for analyzing catenary seismic events on bridges have the following shortcomings: First, the input structures for mainshock and aftershock sequences often employ scaling or random splicing of single peak acceleration, making it difficult to simultaneously consider characteristics such as amplitude, spectrum, energy, and duration that are sensitive to residual displacement of the catenary. Second, the fully coupled bridge-catenary analysis model has high degrees of freedom and high computational cost, making it difficult to support rapid assessment under large-sample mainshock and aftershock scenarios. Third, existing rapid assessment methods often treat the mainshock and aftershock as independent input processes, failing to reflect the impact of the post-mainshock state on the evolution of residual displacement during the aftershock stage. Fourth, existing general-purpose artificial intelligence algorithms typically directly map the overall mainshock and aftershock input characteristics to the final response, without constructing a dedicated prediction mechanism with state progression and adaptive weighting for the mainshock and aftershock problem of catenary on bridges.

[0004] Therefore, it is necessary to propose a rapid assessment method for residual displacement of the overhead contact network on a bridge that can simultaneously consider the amplification effect of the bridge, the multi-index constraint matching of the main shock and aftershock, and the progressive influence of the state after the main shock. Summary of the Invention

[0005] This application proposes a method and equipment for rapid assessment of residual displacement of the overhead contact line on high-speed railway bridges, which can solve one of the problems existing in the background art.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] Firstly, a rapid assessment method for residual displacement of the overhead contact line on a high-speed railway bridge is provided, including:

[0008] Obtain real-time mainshock records;

[0009] Extract the real-time main shock index vector that is sensitive to the residual displacement of the catenary on the high-speed railway bridge from the real-time main shock record;

[0010] Furthermore, the bridge deck magnified input feature vector of the contact wire support points and the real-time main shock index vector are input into the constructed intelligent prediction model to calculate the residual displacement assessment result of the contact wire. The bridge deck magnified input feature vector is defined by the contact wire support point design parameters and the contact wire support point motion peak parameters. The bridge deck magnified input feature vector is obtained through the following methods: obtaining historical main shock records; obtaining historical aftershock sequences matching the historical main shock records; constructing historical main shock and aftershock sequences from the historical main shock records and the historical aftershock sequences; inputting the historical main shock and aftershock sequences into the bridge sub-model to obtain the contact wire corresponding to the historical main shock and aftershock sequences. The intelligent prediction model includes: a main shock stage intelligent prediction module for calculating the real-time main shock residual displacement prediction value of the contact wire support point after the main shock, using the bridge deck magnified input feature vector and the real-time main shock index vector as input; and an aftershock stage intelligent prediction module for calculating the real-time aftershock-induced residual displacement increment of the contact wire support point, using the bridge deck magnified input feature vector and the real-time aftershock index vector matched with the real-time main shock index vector as input. The contact wire residual displacement assessment result consists of the real-time main shock residual displacement prediction value and the real-time aftershock-induced residual displacement increment.

[0011] In one possible design approach of the first aspect, obtaining a historical aftershock sequence that matches historical mainshock index records specifically includes:

[0012] For each historical main shock record, extract the historical main shock index vector that is sensitive to the residual displacement of the catenary on the high-speed railway bridge, and extract the candidate aftershock index vector from the candidate aftershock record library for each candidate aftershock record.

[0013] Furthermore, based on the aftershock target constraint interval and comprehensive matching distance defined by the main shock index vector and aftershock index vector, historical aftershock sequences that match the historical main shock records are selected from the candidate aftershock record library.

[0014] In one possible design approach of the first aspect, the aftershock target constraint interval is:

[0015] in, Let q be the index value of the main shock. Let be the lower limit value of the q-th aftershock index. Let q be the upper limit of the aftershock index. Let q be the target value for the q-th aftershock index. This is the lower limit proportional coefficient. This is the upper limit proportional coefficient. The target proportionality coefficient;

[0016] To satisfy The comprehensive matching distance is calculated from the candidate aftershock index vectors, where q is the index number and Q is the total number of indexes.

[0017] Among them, J j Let j be the candidate aftershock index vector. The overall matching distance, ω q ε is the initial weighting coefficient for the q-th seismic index, and ε is a positive decimal to prevent the denominator from being zero.

[0018] In one possible design approach of the first aspect, a historical mainshock and aftershock sequence is constructed from the historical mainshock record and the historical aftershock sequence, specifically: according to a number of aftershock records, a historical mainshock and aftershock sequence is constructed from the historical mainshock record, the historical aftershock sequence, and a zero-input interval segment.

[0019] In one possible design approach of the first aspect, the rapid assessment method for residual displacement of the overhead contact line on a high-speed railway bridge further includes:

[0020] Using the bridge deck magnified input feature vector, historical main shock index vector, and historical aftershock index vector as inputs, and the predicted value of the main shock residual displacement or the increment of the residual displacement caused by the aftershock as outputs, the intelligent prediction module for the main shock stage and the intelligent prediction module for the aftershock stage are trained.

[0021] In one possible design of the first aspect, the mainshock stage intelligent prediction module includes cascaded mainshock stage feature encoder H1(·) and mainshock stage predictor P1(·), wherein the mainshock stage intelligent prediction module is specifically used for:

[0022] in, B is the input vector for the main shock stage corresponding to the k-th contact wire support point. k I is the magnified input feature vector for the bridge deck. (m) This is the mainshock index vector, where m represents the mainshock label. The implicit characteristic vector of the mainshock stage. Let W be the predicted residual displacement value of the main shock corresponding to the k-th contact wire support point, σ(·) be the nonlinear mapping function, and W be the residual displacement value of the main shock. H1 b H1 W P1 and b P1 These are the parameters to be trained.

[0023] In one possible design of the first aspect, the aftershock stage intelligent prediction module includes cascaded aftershock stage feature encoder H2(·) and aftershock stage predictor P2(·), wherein the aftershock stage intelligent prediction module is specifically used for:

[0024] in, This is the input vector for the aftershock stage corresponding to the k-th contact wire support point. This is the aftershock index vector that matches the mainshock index. This is the adaptive weight vector for the aftershock stage corresponding to the q-th seismic index at the k-th support point. This represents the implicit feature vector during the aftershock phase. χ represents the residual displacement increment caused by aftershocks. k Here, r0 is the post-mains state coefficient corresponding to the k-th contact wire support point, r0 is the residual displacement reference value, and ω is the residual displacement reference value. q γ is the initial weighting coefficient of the q-th aftershock index vector. q W is the initial amplification factor for the q-th aftershock index vector. H2 b H2 W P2 and b P2 These are the parameters to be trained.

[0025] In one possible design approach of the first aspect, the residual displacement assessment result of the contact wire is:

[0026] in, This represents the predicted final residual displacement value corresponding to the k-th contact support point. K represents the predicted maximum residual displacement of the overhead contact system of the high-speed railway bridge, where K is the number of contact support points.

[0027] In one possible design approach of the first aspect, the seismic indices include: peak ground acceleration, peak velocity, control period response spectrum acceleration, Arias intensity, and / or significant duration; the catenary support point design parameters include: the pier height corresponding to the catenary support point and the position parameters of the catenary support point along the bridge direction; the peak motion parameters of the catenary support point include: the peak bridge deck acceleration of the catenary support point, the peak bridge deck displacement of the catenary support point, and the multi-support point input inconsistency index, wherein the multi-support point input inconsistency index is defined by the difference in bridge deck acceleration between adjacent catenary support points.

[0028] In a second aspect, an electronic device is provided, comprising: a processor and a memory coupled to the processor, the memory for storing a computer program; the processor for executing the computer program stored in the memory such that the electronic device performs the rapid assessment method for residual displacement of the overhead contact line on a high-speed railway bridge as described in any possible implementation of the first aspect.

[0029] Beneficial effects:

[0030] Based on the above technical solution, the residual displacement after the main shock is first predicted by using the bridge deck magnification input features and the main shock index. The residual state after the main shock is then explicitly introduced into the aftershock stage. The weights of each sensitive index in the aftershock stage are adaptively updated by the state constraint coefficient, thereby realizing the progressive intelligent prediction of the residual displacement evolution process under the main shock-aftershock action path. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is an overall flowchart of the rapid assessment method for residual displacement of bridge contact network based on multi-index constrained main shock matching and state progression intelligent prediction provided in the embodiments of this application;

[0033] Figure 2 This is a schematic diagram illustrating the separation and analysis of the bridge sub-model and the catenary sub-model provided in the embodiments of this application;

[0034] Figure 3 This is a flowchart of the multi-index constrained mainshock and aftershock matching process provided in the embodiments of this application;

[0035] Figure 4 This is a flowchart of the post-mains shock state progressive intelligent prediction process provided in the embodiments of this application;

[0036] Figure 5 This is a comparison chart of the residual displacement evaluation results of the overhead contact line on the bridge under different methods provided in the embodiments of this application, wherein... Figure 5 (a) shows the evaluation results of the method in this embodiment. Figure 5 (b) shows the evaluation results of the random forest method. Figure 5 (c) shows the evaluation results of support vector regression. Figure 5 (d) shows the linear fit evaluation results. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0038] It should be noted that although functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification and the above-mentioned figures are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0039] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0040] See Figure 1 , Figure 1 This embodiment illustrates a rapid assessment method for residual displacement of overhead contact network on bridges based on multi-index constrained main shock matching and state-progressive intelligent prediction, comprising the following steps:

[0041] Step 1: Obtain bridge structural parameters, catenary structural parameters, and catenary layout parameters along the bridge. For example... Figure 2 As shown, a bridge sub-model M is established. b With the catenary sub-model M c The bridge deck response extraction locations corresponding to each support point of the overhead contact system are determined; the set of all bridge deck response extraction locations P is denoted as Ω. In this embodiment, the overhead contact system sub-model M... c It is used not only to characterize the residual displacement response of the overhead contact line on the bridge under the input of the bridge deck, but also to generate offline positioning points of the contact line as residual displacement samples at key locations, and to compare them with the bridge sub-model M. b The output bridge deck response features together constitute the training sample set for the intelligent prediction module in steps six and seven.

[0042] The bridge sub-model M b The numerical analysis model established based on the finite element method includes at least a set of bridge nodes, a set of bridge elements, material parameters, mass parameters, damping parameters, boundary constraint parameters, and a seismic input interface. Bridge sub-model M b Using the mainshock record or mainshock-aftershock sequence as the seismic motion input, the dynamic response of the bridge structure under seismic action is calculated, and the acceleration time history, displacement time history, and response differences between adjacent support points are output for each contact wire support point corresponding to the bridge deck position.

[0043] The bridge sub-model M b This can be expressed by the following dynamic equilibrium relationship:

[0044] Where g(t) is the input seismic acceleration time history, a b,k (t) and u b,k (t) represents the acceleration time history and displacement time history of the bridge deck position corresponding to the k-th catenary support point, respectively, where K is the total number of catenary support points.

[0045] Contact network sub-model M c Used to receive bridge sub-model M b The output bridge deck multi-support response is calculated, and the residual displacement response at key locations of the overhead contact line is also calculated. The input-output relationship is expressed as follows:

[0046] Where, r k The residual displacement at the critical location corresponding to the k-th contact wire support point is, for example... Figure 2 The key location shown is the positioning point located above the support point, which is the control point for evaluating the post-earthquake state of the overhead contact line.

[0047] Therefore, the bridge sub-model M b Output bridge deck response, used to calculate amplified input features of the bridge deck; contact network sub-model M c The residual displacements at key locations in the overhead contact system are output and used to build the training sample set for the intelligent prediction module offline. After training is complete, the overhead contact sub-model M does not need to be called repeatedly during the online rapid evaluation phase. c Only the bridge sub-model M needs to be called. b With the trained intelligent prediction module, the residual displacement after the main shock and the incremental residual displacement of the aftershock can be obtained.

[0048] After training is complete, the online rapid evaluation phase only requires calling the bridge sub-model M. b Once trained, the intelligent prediction module can output the residual displacement after the main shock and the incremental residual displacement of the aftershock.

[0049] Where K represents the total number of overhead contact line support points.

[0050] In this embodiment, the bridge sub-model is used to calculate the dynamic response of the bridge deck under the input of the main shock and aftershock, while the catenary sub-model is used to characterize the residual displacement response of the catenary on the bridge under the input of the bridge deck. The bridge sub-model and the catenary sub-model are established using a separate modeling approach. The bridge sub-model outputs the response of the bridge deck support points, while the catenary sub-model receives the response characteristics of the bridge deck support points as input, thereby reducing the computational cost of the fully coupled bridge-catenary analysis.

[0051] Step 2: Select 250 fundamental mainshock waves as the mainshock sample library, and for each mainshock record g m (t) Extract seismic intensity indicators that are sensitive to residual displacement of the overhead contact line on the bridge and construct the mainshock index vector, where m represents the mainshock label and t represents the time series of the seismic wave;

[0052] Where Q represents the total number of sensitive earthquake intensity indicators. This represents the q-th mainshock index value; in this embodiment, the sensitive earthquake intensity indices selected are at least peak ground acceleration (PGA), peak velocity (PGV), control period response spectrum acceleration Sa(Tc), Arias intensity IA, and significant duration D. 5−95 Three or more of them, where Tc is the overhead contact line control cycle.

[0053] Step 3: Construct a candidate aftershock matching library, and perform matching on each candidate earthquake record in the candidate aftershock library. Extract the corresponding candidate index vector, where a represents the aftershock label and j represents the sequence number of the j-th aftershock seismic wave that matches the description.

[0054] in, Let j be the candidate index vector set corresponding to the j-th candidate aftershock earthquake record. This is the index value of the qth aftershock of the jth aftershock, corresponding to the main shock.

[0055] like Figure 3 As shown, aftershock target constraint intervals and comprehensive matching distances are established based on the main shock index vector, and aftershock records g matching the main shock are selected from the candidate aftershock database. a (t);

[0056] Establish the aftershock target constraint interval based on the main shock index vector:

[0057] in, Let be the lower limit value of the q-th aftershock index. Let q be the upper limit of the aftershock index. Let q be the target value for the q-th aftershock index. This is the lower limit proportional coefficient. This is the upper limit proportional coefficient. This is the target proportionality coefficient.

[0058] To satisfy The comprehensive matching distance is calculated from the candidate aftershock records, where q is the seismic index number and Q is the total number of seismic indices.

[0059] Among them, J j Let ω be the comprehensive matching distance of the j-th candidate aftershock record. q Let be the initial weighting coefficient for the q-th sensitive earthquake intensity index, and satisfy . Q represents the number of earthquake indicators.

[0060] Where ε is a positive decimal to prevent the denominator from being zero, and J is taken as... j The smallest one is used as the matched aftershock record, denoted as g. a (t).

[0061] Step 4: In one embodiment, 250 fundamental waves of the mainshock are selected as the mainshock sample library, and a candidate aftershock matching library is constructed based on the sensitivity index of residual displacement of the overhead contact line. For each mainshock sample, corresponding mainshock and aftershock sequences are constructed under four site conditions and six distance conditions; the four site conditions are A, B, C, and D, and the six distance conditions are 10km, 15km, 20km, 25km, 35km, and 45km, respectively. A single aftershock sequence is constructed using the following formula:

[0062] The sequence of two aftershocks is constructed using the following formula:

[0063] Among them, g a1 (t) represents the record of the first aftershock, g a2 (t) represents the record of the second aftershock. When the two aftershocks in the two aftershock sequences use the same matching record, g is taken. a2 (t)=g a1 (t). Here, ⊕ represents the time history splicing operator, used to connect multiple acceleration time histories end-to-end in chronological order to form a new input time history, rather than adding each time history segment point by point at the same time; z τ(t) represents a zero-input interval of duration τ, during which the amplitude remains zero. This construction method is consistent with the current time history splicing rule for one maximum aftershock and two intermediate aftershocks.

[0064] In this embodiment, the duration of the interval is set to τ = 40s. A single aftershock sequence is constructed as "mainshock + 40s interval + one aftershock," and a two-aftershock sequence is constructed as "mainshock + 40s interval + first aftershock + 40s interval + second aftershock." The peak ground acceleration (PGA) of the single aftershock is adjusted using the maximum aftershock ratio, i.e., scaling the input PGA of the aftershock seismic record itself. An initial suggested value is empirically set to 0.75. The two aftershocks are repeated twice using a moderate aftershock ratio, with a PGA of 0.5, i.e., let g... a2 (t)=g a1 (t). Wherein, the proportion of maximum aftershock and the proportion of moderate aftershock are preset aftershock peak acceleration proportionality coefficients, denoted as η, respectively. max and η mid And satisfying 0 < η mid <η max ≤1. Aftershock records in a single aftershock sequence are arranged according to η max Peak acceleration adjustment was performed, and the two aftershock records in the two aftershock sequences were adjusted according to η. mid Adjust the peak acceleration.

[0065] Step 5: Input the main shock and aftershock sequence constructed in Step 4 into the bridge sub-model Mb, extract the bridge deck response time history at the set of support points Ω, including the acceleration time history, displacement time history at the support points, and acceleration time history at adjacent support points, and calculate the bridge deck magnified input feature vector B. k ;

[0066] Among them, h k Let s be the pier height corresponding to the k-th support point. k Let A be the positional parameter of the k-th support point along the bridge direction. k U represents the peak acceleration of the bridge deck at that support point. k G represents the peak displacement of the bridge deck at that support point. k Input inconsistency indicators for multiple pivots.

[0067] Among them, the peak acceleration A of the bridge deck k Peak bridge deck displacement U k Inconsistency index G with multi-pivot input k Calculate according to the following formulas:

[0068] in, Let the bridge deck acceleration time history be at the k-th support point. Let N(k) be the time history of the bridge deck displacement at the k-th support point, and N(k) be the set of adjacent support points of the k-th support point. The acceleration time history is at the adjacent support point l. The acceleration time history and displacement time history can be obtained by establishing a finite element model of a high-speed railway bridge and inputting seismic waves for calculation.

[0069] In one embodiment, based on multiple mainshock sequences and mainshock-aftershock sequences, the bridge sub-model M is used respectively. b Extract the magnified input features of the bridge deck, and use the contact network sub-model M c Calculate the residual displacement at the corresponding key locations of the overhead contact system and construct a training sample set. Using the bridge deck magnified input features, the main shock index vector, and the aftershock index vector as inputs, and the residual displacement at the key locations of the overhead contact system or the increment of the residual displacement of the aftershock as outputs, train the intelligent prediction modules for the main shock stage and the intelligent prediction modules for the aftershock stage. When the validation set error meets the preset convergence condition, the trained model is obtained for the rapid evaluation in steps six and seven.

[0070] Step Six: As Figure 4 As shown, based on the bridge deck magnified input feature vector and the main shock index vector, the residual displacement of the contact network after the main shock is calculated using the intelligent prediction module for the main shock stage. First, the main shock stage input vector is constructed:

[0071] in, This is the input vector for the main shock stage corresponding to the k-th support point. The main shock stage intelligent prediction module in step six consists of a main shock stage feature encoder H1(·) and a main shock stage predictor P1(·), and the calculation process is as follows:

[0072] in, The implicit characteristic vector of the mainshock stage. W represents the predicted residual displacement at the key location of the overhead contact line corresponding to the k-th support point after the main shock. σ(·) is a nonlinear mapping function used to nonlinearly characterize the amplified input features of the bridge deck, seismic sensitivity indicators, and state constraint information to improve the adaptability of residual displacement prediction. H1 b H1 W P1 and b P1 These are the parameters to be trained.

[0073] In this embodiment, the main shock stage feature encoder H1(·) is used to extract the nonlinear coupling features between the bridge deck magnification input features and the main shock index, and the main shock stage predictor P1(·) is used to output the predicted value of the residual displacement during the main shock stage.

[0074] Step 7: Intelligent prediction of the aftershock stage following the mainshock. After obtaining the predicted residual displacement values ​​during the mainshock stage, calculate the post-mainshock state coefficients:

[0075] Where, χ k is the post-mains state coefficient corresponding to the k-th support point, and r0 is the residual displacement reference value.

[0076] Then update the weights of each sensitive indicator during the aftershock phase using the following formula:

[0077] in, ω represents the adaptive weight for the aftershock stage of the q-th sensitive earthquake intensity index corresponding to the k-th support point. q Let γ be the initial weighting coefficient for the q-th aftershock index. q is the initial amplification factor for the q-th aftershock index, obtained by comparing the initial seismic intensity index of the support point with the intensity index of the bridge deck acceleration response.

[0078] Further construct the input vector for the aftershock stage:

[0079] in, To match the aftershock index vector, This is the adaptive weight vector for the aftershock phase of the k-th support point.

[0080] Furthermore, the intelligent prediction module for the aftershock phase calculates using the following formula:

[0081] in, This represents the implicit feature vector during the aftershock phase. This represents the residual displacement increment caused by aftershocks.

[0082] σ(·) is a nonlinear mapping function used to nonlinearly characterize the bridge deck magnification input features, seismic sensitivity indicators, and state constraint information, thereby improving the adaptability of residual displacement prediction. H2 b H2 W P2 and b P2 These are the parameters to be trained.

[0083] In this embodiment, the intelligent prediction during the aftershock stage does not treat aftershocks as independent inputs from the main shock for a one-time prediction. Instead, it uses the post-main shock state coefficient χ. k Explicitly introduce the aftershock phase, and according to χ k The weights of various sensitive indicators are adaptively updated during the aftershock phase, thus forming a state progression prediction process under the main shock-aftershock action path.

[0084] Step 8: Calculate the final residual displacement. The final residual displacement of the overhead contact line on the bridge after the main shock is calculated using the following formula:

[0085] in, This represents the predicted final residual displacement value at the critical location of the overhead contact line corresponding to the k-th support point. This represents the predicted maximum residual displacement at key locations of the entire bridge's overhead contact system.

[0086] Based on the above output results, the residual displacement assessment of the overhead contact line on the bridge and the comparison of results from different methods can be completed.

[0087] The output should include at least:

[0088] Residual displacement of each support point after the main shock ;

[0089] Residual displacement increments caused by aftershocks at each support point ;

[0090] Final residual displacement of each support point ;

[0091] Maximum residual displacement of the entire bridge .

[0092] In one embodiment, a single peak acceleration matching method, a mainshock and aftershock independent prediction method, and the multi-index constrained mainshock and aftershock matching and state-progressive intelligent prediction method proposed in this embodiment can be selected for comparison. The prediction results and calculation efficiency of the maximum residual displacement of the overhead contact network under different methods can be compared to verify the advantages of the method in this embodiment in terms of the rationality of the mainshock and aftershock input and the accuracy of rapid prediction. Figure 5The comparison results of the proposed method with random forest, support vector regression, and linear fitting methods in assessing residual displacement of the overhead contact system on bridges are shown. It can be seen that the predicted mean curves of the proposed method are closer to the actual model mean curves in both the mainshock and aftershock phases, and can better reflect the changing trend of residual displacement with increasing ground motion intensity. In contrast, the other comparative methods show varying degrees of deviation in some intensity ranges, with the linear fitting method exhibiting a more significant deviation. These results demonstrate that the multi-index constrained mainshock-aftershock matching and post-mainshock state progressive intelligent prediction mechanism proposed in this embodiment can improve the accuracy and stability of rapid assessment of residual displacement of the overhead contact system on bridges.

[0093] This embodiment discloses a rapid assessment method for residual displacement of overhead contact lines on bridges based on multi-index constraint mainshock and aftershock matching and state-progressive intelligent prediction, belonging to the field of seismic assessment of overhead contact lines on high-speed railway bridges. First, a bridge-network separation analysis framework consisting of a bridge sub-model and an overhead contact line sub-model is established, and the bridge deck response extraction locations corresponding to the overhead contact line support points are determined. Second, multiple seismic intensity indices sensitive to residual displacement of the overhead contact line on the bridge are selected to construct a mainshock index vector. Based on preset upper and lower limits and a comprehensive matching distance, aftershock records matching the mainshock are screened from a candidate ground motion database to form a mainshock and aftershock sequence input. Then, the mainshock and aftershock sequence is input into the bridge sub-model, and the peak bridge deck acceleration, peak bridge deck displacement, and multi-support point input inconsistency features are extracted at each overhead contact line support point to form a bridge deck magnified input feature vector. Furthermore, an adaptive intelligent prediction algorithm for post-mainsh state progression is proposed: first, the residual displacement of the contact network after the mainshock is predicted based on the amplified input features of the bridge deck and the mainshock indices; then, the residual state after the mainshock is used as the state constraint for the aftershock stage, the weights of each sensitive index are adaptively updated, and the residual displacement increment caused by the aftershock is calculated, finally obtaining the residual displacement of the contact network after the mainshock and aftershock. This method integrates the amplified bridge input, multi-index mainshock and aftershock matching, and state progression intelligent prediction into a unified technical process, significantly reducing the computational cost of the fully coupled bridge-contact network while ensuring the rationality of the mainshock and aftershock input. It can be used for post-earthquake state identification, risk analysis, and rapid decision-making of the contact network on the bridge.

[0094] In short, this algorithm first uses the bridge deck magnification input features and mainshock indices to predict the residual displacement after the mainshock, and then explicitly introduces this residual state after the mainshock into the aftershock stage. By adaptively updating the weights of various sensitive indices in the aftershock stage through state constraint coefficients, it achieves progressive intelligent prediction of the residual displacement evolution process under the mainshock-aftershock action path. Compared with existing technologies, this embodiment has the following beneficial effects:

[0095] 1. By constraining the main shock and aftershock matching with multiple indicators, the selection of aftershocks is avoided by using only a single peak acceleration, which improves the ability of the main shock and aftershock input to maintain the sensitivity characteristics of residual displacement of the catenary.

[0096] 2. By extracting the magnified input features of the bridge deck through bridge-over-overhead contact network separation analysis, the dynamic amplification and layout differences of the bridge are introduced into the rapid evaluation of the contact network, taking into account both physical significance and computational efficiency.

[0097] 3. An adaptive intelligent prediction algorithm based on the progressive state after the main shock is proposed, which can dynamically adjust the weights of various sensitive indicators according to the residual state after the main shock during the aftershock stage, and more accurately characterize the residual displacement evolution law under the path dependence of the main shock and aftershock.

[0098] 4. The proposed method can be used for rapid assessment of large-sample residual displacements of the overhead contact line during main and aftershocks on bridges, providing technical support for post-earthquake status identification, risk analysis, and rapid decision-making.

[0099] This application also provides an electronic device, including: a processor, and a memory coupled to the processor, the memory being used to store a computer program; the processor being used to execute the computer program stored in the memory, so that the electronic device performs the method as described in any of the above embodiments.

[0100] Electronic devices can be computing devices such as desktop computers, laptops, handheld computers, and cloud servers. These electronic devices may include, but are not limited to, processors and memory.

[0101] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting various parts of the device via various interfaces and lines.

[0102] The memory can be used to store the computer program, and the processor implements various functions of the electronic device by running or executing the computer program stored in the memory and calling the data stored in the memory.

[0103] The memory may primarily include a program storage area and a data storage area. The program storage area may store the operating system, applications required for at least one function, etc.; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0104] This application also provides a storage medium, which is a computer-readable storage medium. The computer program is stored in the computer-readable storage medium, and when executed by a processor, the computer program can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0105] This application also provides a computer program product, including: a computer program or instructions that, when the computer program or instructions are run on a computer, cause the computer to perform any of the above possible implementation methods.

[0106] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications are also considered to be within the scope of protection of this application.

Claims

1. A method for rapid assessment of residual displacement of the overhead contact line on a high-speed railway bridge, characterized in that, include: Obtain real-time mainshock records; Extract the real-time main shock index vector that is sensitive to the residual displacement of the catenary on the high-speed railway bridge from the real-time main shock record; Furthermore, the bridge deck magnified input feature vector of the contact wire support points and the real-time main shock index vector are input into the constructed intelligent prediction model to calculate the residual displacement assessment result of the contact wire. The bridge deck magnified input feature vector is defined by the contact wire support point design parameters and the contact wire support point motion peak parameters. The bridge deck magnified input feature vector is obtained through the following methods: obtaining historical main shock records; obtaining historical aftershock sequences matching the historical main shock records; constructing historical main shock and aftershock sequences from the historical main shock records and the historical aftershock sequences; inputting the historical main shock and aftershock sequences into the bridge sub-model to obtain the contact wire corresponding to the historical main shock and aftershock sequences. The intelligent prediction model includes: a main shock stage intelligent prediction module for calculating the real-time main shock residual displacement prediction value of the contact wire support point after the main shock, using the bridge deck magnified input feature vector and the real-time main shock index vector as input; and an aftershock stage intelligent prediction module for calculating the real-time aftershock-induced residual displacement increment of the contact wire support point, using the bridge deck magnified input feature vector and the real-time aftershock index vector matched with the real-time main shock index vector as input. The contact wire residual displacement assessment result consists of the real-time main shock residual displacement prediction value and the real-time aftershock-induced residual displacement increment.

2. The method for rapid assessment of residual displacement of the overhead contact line on a high-speed railway bridge as described in claim 1, characterized in that, Obtaining a sequence of historical aftershocks that matches the historical mainshock record specifically includes: For each historical main shock record, extract the historical main shock index vector that is sensitive to the residual displacement of the catenary on the high-speed railway bridge, and extract the candidate aftershock index vector from the candidate aftershock record library for each candidate aftershock record. Furthermore, based on the aftershock target constraint interval and comprehensive matching distance defined by the main shock index vector and aftershock index vector, historical aftershock sequences that match the historical main shock records are selected from the candidate aftershock record library.

3. The rapid assessment method for residual displacement of the overhead contact line on a high-speed railway bridge as described in claim 2, characterized in that, The aftershock target constraint interval is: in, Let q be the index value of the main shock. Let be the lower limit value of the q-th aftershock index. Let q be the upper limit of the aftershock index. Let q be the target value for the q-th aftershock index. This is the lower limit proportional coefficient. This is the upper limit proportional coefficient. The target proportionality coefficient, To satisfy The candidate aftershock index vectors are used to calculate the comprehensive matching distance, where q is the index of the seismic index and Q is the total number of seismic indexes. Among them, J j Let j be the candidate aftershock index vector. The overall matching distance, ω q ε is the initial weighting coefficient for the q-th seismic index, and ε is a positive decimal to prevent the denominator from being zero.

4. The rapid assessment method for residual displacement of the overhead contact line on a high-speed railway bridge as described in claim 1, characterized in that, The historical main shock sequence is constructed from the historical main shock record and the historical aftershock sequence. Specifically, the historical main shock sequence is constructed from the historical main shock record, the historical aftershock sequence, and the zero-input interval segment according to a number of aftershock records.

5. The rapid assessment method for residual displacement of the overhead contact line on a high-speed railway bridge as described in claim 1, characterized in that, The rapid assessment method for residual displacement of the overhead contact line on high-speed railway bridges also includes: Using the bridge deck magnified input feature vector, historical main shock index vector, and historical aftershock index vector as inputs, and the predicted value of the main shock residual displacement or the increment of the residual displacement caused by the aftershock as outputs, the intelligent prediction module for the main shock stage and the intelligent prediction module for the aftershock stage are trained.

6. The rapid assessment method for residual displacement of the overhead contact line on a high-speed railway bridge as described in claim 1, characterized in that, The mainshock stage intelligent prediction module includes a cascaded mainshock stage feature encoder H1(·) and a mainshock stage predictor P1(·). Specifically, the mainshock stage intelligent prediction module is used for: in, B is the input vector for the main shock stage corresponding to the k-th contact wire support point. k I is the magnified input feature vector for the bridge deck. (m) This is the mainshock index vector, where m represents the mainshock label. The implicit characteristic vector of the mainshock stage. Let W be the predicted residual displacement value of the main shock corresponding to the k-th contact wire support point, σ(·) be the nonlinear mapping function, and W be the residual displacement value of the main shock. H1 b H1 W P1 and b P1 These are the parameters to be trained.

7. The rapid assessment method for residual displacement of the overhead contact line on a high-speed railway bridge as described in claim 6, characterized in that, The intelligent prediction module for the aftershock stage includes a cascaded aftershock stage feature encoder H2(·) and an aftershock stage predictor P2(·). Specifically, the intelligent prediction module for the aftershock stage is used for: in, This is the input vector for the aftershock stage corresponding to the k-th contact wire support point. This is the aftershock index vector that matches the mainshock index. This is the adaptive weight vector for the aftershock stage corresponding to the q-th seismic index at the k-th support point. This represents the implicit feature vector during the aftershock phase. χ represents the residual displacement increment caused by aftershocks. k Here, r0 is the post-mains state coefficient corresponding to the k-th contact wire support point, r0 is the residual displacement reference value, and ω is the residual displacement reference value. q γ is the initial weighting coefficient of the q-th aftershock index vector. q W is the initial amplification factor for the q-th aftershock index vector. H2 b H2 W P2 and b P2 These are the parameters to be trained.

8. The method for rapid assessment of residual displacement of the overhead contact line on a high-speed railway bridge as described in claim 7, characterized in that, The assessment result of the residual displacement of the overhead contact line is as follows: in, This represents the predicted final residual displacement value corresponding to the k-th contact support point. K represents the predicted maximum residual displacement of the overhead contact system of the high-speed railway bridge, where K is the number of contact support points.

9. The method for rapid assessment of residual displacement of the overhead contact line on a high-speed railway bridge as described in claim 1, characterized in that, Seismic indices include: peak ground acceleration, peak velocity, control period response spectrum acceleration, Arias intensity and / or significant duration; catenary support point design parameters include: pier height corresponding to the catenary support point and the location parameters of the catenary support point along the bridge direction; catenary support point motion peak parameters include: peak bridge deck acceleration of the catenary support point, peak bridge deck displacement of the catenary support point, and multi-support point input inconsistency index, wherein the multi-support point input inconsistency index is defined by the difference in bridge deck acceleration between adjacent catenary support points.

10. An electronic device, characterized in that, The electronic device includes: a processor, and a memory coupled to the processor. The memory is used to store computer programs; The processor is configured to execute the computer program stored in the memory, so that the electronic device performs the rapid assessment method for residual displacement of the overhead contact line on a high-speed railway bridge as described in any one of claims 1-9.

Images