A method for analyzing the seismic time-history response of bridges driven jointly by physics and data

Abstract

The present invention discloses a method for analyzing the seismic time-history response of bridges driven jointly by physics and data, which relates to the technical field of seismic response analysis methods. The present invention includes the following steps: S1: Nonlinear element simulation of support and restraint components; S2: Determination of highway bridge model parameters; S3: Establishment of a parametric finite element representation method for bridges; S4: Establishment of a highway bridge database; S5: Establishment of finite element models for each bridge sample; S6: Design of a deep network for mapping the input and output layers of the bridge; S7: Establishment of a surrogate model for analyzing the seismic time-history response of bridges. The present invention strengthens the physical basis of machine learning methods and makes full use of their powerful nonlinear characterization ability to accurately and efficiently analyze the seismic time-history response of bridges, which has important engineering value for the evaluation methods of seismic performance of new-generation single and regional bridge structures, and also has important significance for the development of basic innovation theories of interdisciplinary intersections.

Description

Technical Field

[0001] The present invention relates to the technical field of seismic response analysis methods, and specifically refers to a physical and data-driven bridge seismic time history response analysis method. Background Art

[0002] The seismic time history response of a bridge structure can comprehensively reflect the effects of various characteristics such as the time domain and frequency domain of ground motion, and is the basis for revealing the detailed damage states of each component and the evolution mechanism of dynamic performance under seismic excitation. With the in-depth study of the seismic performance of structures, bridge seismic analysis methods have gradually developed from methods based on structural capacity to methods based on performance, risk, sustainability, and resilience. Therefore, it is necessary to consider not only a large number of structural and seismic randomnesses, but also to conduct seismic analysis on regional bridges containing a large number of individual structures, which all require large-scale bridge seismic time history analysis. However, the physics-based analysis method conducts static or dynamic time history analysis based on the bridge physical model and structural dynamics principle to obtain its seismic response. For example, in finite element analysis, a macroscopic or microscopic physical model is established according to the mechanical properties of each part of the material or component, and each component or part of the structure is discretized to establish the mass, damping, and stiffness matrices of the structure, and the theoretical or numerical solution of the structural dynamic response is obtained by solving the structural motion equation. Such methods can comprehensively consider the effects of structural and ground motion characteristics and have the advantages of flexibility and strong adaptability. However, the currently commonly used individual structure time history analysis methods are difficult to accurately and efficiently implement or popularize under the current computer and technical conditions.

[0003] In recent years, with the development of computing technology, machine learning methods have been widely applied in the fields of structural health monitoring and signal processing, and have been introduced by relevant scholars into the seismic analysis of civil engineering structures, showing strong high-dimensional non-linear characterization capabilities. However, most of the current machine learning methods (such as BP neural network, convolutional neural network, recurrent neural network) are purely data-driven, and there is no clear physical meaning between the model input and output. In essence, it is a "black box" and cannot effectively guarantee its generalization ability. In addition, there are significant differences in the forms and categories of the experimental and simulation data accumulated during the research process of bridge seismic performance, which contain a large amount of noise and random factors and cannot be directly used for model training. Therefore, there is an urgent need for a physical and data-driven bridge seismic time history response analysis method to solve the above problems. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to overcome the above technical problems and provide a method for analyzing the seismic time-history response of bridges driven jointly by physics and data, strengthening the physical basis of machine learning methods and making full use of their powerful non-linear characterization ability to accurately and efficiently analyze the seismic time-history response of bridges, which has important engineering value for the seismic performance evaluation methods of a new generation of single and regional bridge structures, and is also of great significance for the development of basic innovative theories of interdisciplinary intersection.

[0005] To solve the above technical problems, the technical solution provided by the present invention is: a method for analyzing the seismic time-history response of bridges driven jointly by physics and data, comprising the following steps:

[0006] S1: According to the mechanical characteristics of highway bridge structures, the highway bridge structure is discretized into a distributed mass system connected by elastic beam-column elements, and the piers, bearings and abutments, which are support and restraint components, are simulated by corresponding non-linear elements;

[0007] S2: Summarize the types of each component unit, study the node and element division modeling process of the bridge finite element model, and determine the spatial positions, mass distribution characteristics and detailed model parameters of the nodes and connection units of the highway bridge structure;

[0008] S3: Summarize the types of simulation units and unit parameters of each part of the bridge, identify the hysteretic characteristic identification model of the above-mentioned piers based on this, and conduct a systematic study on the physical mechanisms and laws of the restraint effects of the key components of the abutments, bearings and foundations by using theoretical derivation, experimental research and finite element simulation methods. According to the bridge seismic design method, determine the correlation relationship of the detailed component model parameters, and establish a parametric bridge finite element representation method considering the connection of restraint units;

[0009] S4: Design highway bridge samples with rich features according to the experimental design principle, and calibrate them according to bridge examples and seismic design methods to establish a highway bridge database that conforms to the characteristics of actual bridges;

[0010] S5: Propose an automatic modeling method for each bridge finite element based on the finite element model characterization matrix, and use the OpenSEES finite element software to establish the finite element models of each bridge sample;

[0011] S6: Combine the frequency domain characteristics of ground motion with the high-dimensional characterization method of bridges to enhance the multi-scale characterization ability of the network, design a deep network for the mapping relationship between the input layer and output layer of the bridge, and use cross-validation and early termination techniques to prevent overfitting during the training process;

[0012] S7: A deep network with the finite element model of the bridge and the high-dimensional model of the dynamic response of the pier column as inputs and the dynamic time history response of the bridge components as outputs. Determine the detailed CNN architecture according to the characteristics of the convolutional layer and the pooling layer. Establish a surrogate model for bridge seismic time history response analysis through learning the numerical simulation results, and reveal the specific process of the interaction mechanism between the pier column and the restraint components under seismic excitation.

[0013] As an improvement, the features in S4 include the number of lanes, the span, and the pier height.

[0014] After adopting the above method, the present invention has the following advantages: The model established by the present invention can evaluate the seismic damage status of large-scale bridge structures, and then calculate the seismic risk of regional bridges and the seismic resilience of the transportation network. At the same time, it can strengthen the physical basis of machine learning methods and make full use of their powerful non-linear representation ability to accurately and efficiently analyze the bridge seismic time history response, which has important engineering value for the seismic performance evaluation method of the new generation of single and regional bridge structures, and also has important significance for the development of the basic innovation theory of interdisciplinary intersection, improving the disaster prevention, mitigation and emergency management decision-making ability of relevant departments.

[0015] The above summary is only for the purpose of the specification and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments and features described above, further aspects, embodiments and features of the present invention will be readily apparent by reference to the drawings and the following detailed description. Brief Description of the Drawings

[0016] Figure 1 is a flowchart of a method for analyzing the seismic time history response of a bridge jointly driven by physics and data according to the present invention.

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts. Detailed Embodiments

[0018] In the following, only some exemplary embodiments are briefly described. As those skilled in the art can recognize, the described embodiments can be modified in various different ways without departing from the spirit or scope of the present invention. Therefore, the drawings and the description are considered to be exemplary in nature rather than restrictive.

[0019] The embodiments of the present invention will be described in detail below with reference to the drawings.

[0020] Combined with the attached Figure 1, a method for analyzing the seismic time - history response of bridges driven jointly by physics and data, comprising the following steps:

[0021] S1: According to the mechanical characteristics of highway bridge structures, discretize the highway bridge structure into a distributed mass system connected by elastic beam - column elements, and use corresponding nonlinear elements to simulate the piers, bearings, and abutments as support and restraint components.

[0022] S2: Summarize the types of each component element, study the node and element division modeling process of the bridge finite - element model, and determine the spatial positions, mass distribution characteristics, and detailed model parameters of the nodes and connection elements of the highway bridge structure.

[0023] S3: Summarize the types of simulation elements and element parameters of each part of the bridge, based on the above - mentioned hysteretic characteristic identification model of piers, and use theoretical derivation, experimental research, and finite - element simulation methods to systematically study the physical mechanisms and laws of the restraint effects of key components such as abutments, bearings, and foundations. Determine the correlation relationships of the detailed component model parameters according to the bridge seismic design method, and establish a parametric bridge finite - element representation method considering the connection of restraint elements.

[0024] S4: Design highway bridge samples with rich characteristics according to the experimental design principle, and calibrate them according to bridge examples and seismic design methods to establish a highway bridge database that conforms to the characteristics of actual bridges. The characteristics include the number of lanes, span, and pier height.

[0025] S5: Propose an automatic modeling method for each bridge finite - element based on the finite - element model characterization matrix, and use the OpenSEES finite - element software to establish the finite - element models of each bridge sample.

[0026] S6: Combine the frequency - domain characteristics of ground motion with the high - dimensional characterization method of bridges to enhance the multi - scale characterization ability of the network, design a deep network for the mapping relationship between the input layer and output layer of the bridge, and use cross - validation and early - termination techniques to prevent overfitting during the training process.

[0027] S7: A deep network with the finite - element model of the bridge and the high - dimensional model of the dynamic response of the pier as input and the dynamic time - history response of the bridge components as output, and determine the detailed CNN architecture according to the characteristics of the convolutional layer and pooling layer. Establish a surrogate model for analyzing the seismic time - history response of the bridge by learning the numerical simulation results, and reveal the specific process of the interaction mechanism between the pier and the restraint components under seismic excitation.

[0028] The model established in the present invention can evaluate the seismic damage conditions of large-scale bridge structures, and then calculate the seismic risks of regional bridges and the seismic resilience of transportation networks. At the same time, it can strengthen the physical basis of machine learning methods and make full use of their powerful non-linear representation capabilities to accurately and efficiently analyze the seismic time-history responses of bridges, which has important engineering value for the evaluation methods of the seismic performance of a new generation of single and regional bridge structures. At the same time, it is of great significance to the development of basic innovative theories of interdisciplinary intersections, and improves the disaster prevention, mitigation and emergency management decision-making capabilities of relevant departments.

[0029] The above description of the present invention and its implementation manners is not restrictive. What is shown in the drawings is only one of the implementation manners of the present invention, and the actual embodiments are not limited thereto. Generally speaking, if those of ordinary skill in the art are inspired by it and design similar structural manners and embodiments to this technical solution without creative efforts without departing from the purpose of the present invention, they shall fall within the protection scope of the present invention.

Claims

1. A method for analyzing the seismic time - history response of bridges jointly driven by physics and data, characterized in that: It includes the following steps: S1: According to the mechanical characteristics of highway bridge structures, the highway bridge structure is discretized into a distributed mass system connected by elastic beam-column elements, and corresponding nonlinear elements are used to simulate the piers, bearings and abutments as support and restraint components; S2: Summarize the types of each component unit, study the node and element division modeling process of the bridge finite element model, and determine the spatial position, mass distribution characteristics and detailed model parameters of the nodes and connection units of the highway bridge structure; S3: Summarize the types of simulation units and unit parameters of each part of the bridge, identify the model based on the hysteretic characteristics of the pier columns, and use theoretical derivation, experimental research and finite element simulation methods to systematically study the physical mechanisms and laws of the restraint effects of the key components of the abutments, bearings and foundations. Determine the correlation relationships of the detailed component model parameters according to the bridge seismic design method, and establish a parametric bridge finite element representation method considering the connection of restraint units; S4: Design highway bridge samples with rich features according to the experimental design principle, and calibrate them according to bridge examples and seismic design methods to establish a highway bridge database that conforms to the characteristics of actual bridges; S5: Propose an automatic modeling method for each bridge finite element based on the finite element model characterization matrix, and use the OpenSEES finite element software to establish the finite element models of each bridge sample; S6: Combine the frequency-domain characteristics of ground motion with the high-dimensional characterization method of bridges to enhance the multi-scale characterization ability of the network, design a deep network for the mapping relationship between the input layer and output layer of the bridge, and use cross-validation and early termination techniques to prevent overfitting during the training process; S7: A deep network with the finite element model of the bridge and the high-dimensional model of the dynamic response of the pier column as the input and the dynamic time-history response of the bridge components as the output, and determine the detailed CNN architecture according to the characteristics of the convolutional layer and the pooling layer. Establish a surrogate model for bridge seismic time-history response analysis by learning the numerical simulation results, and reveal the specific process of the interaction mechanism between the pier column and the restraint component under seismic excitation.

2. The method for analyzing the seismic time history response of a bridge jointly driven by physics and data according to claim 1, characterized in that: The features in S4 include the number of lanes, span and pier height.

Images