1. Introduction
Since the beginning of the 21st century, high-speed railway (HSR) construction worldwide has entered a period of rapid development, with operational mileage and its construction scale continuously setting historical records [1]. As of 2021, China’s operational mileage has reached 40,000 km. In addition, many railways pass through regions where earthquake zones are widely distributed and active. HSR bridges are highly susceptible to structural damage and performance deterioration under earthquakes, posing a threat to train operational safety [2]. Therefore, rapid earthquake response detection and operational performance assessment for HSR bridges are crucial to ensure the safety and stability of train operations.
Due to HSR trains’ high operating speeds and large traffic densities, HSR bridges have stricter technical requirements than ordinary bridges [3]. To ensure long-term safety, the service performance detection of HSR bridges has become an integral component of HSR safety operation [4]. At present, research on the dynamic response and damage detection of HSR bridges mainly focuses on two aspects: one is structural damage detection based on vibration measurement, and the other is visual detection based on artificial intelligence. Mao et al. [5] utilized a structured light sensor to acquire the structural point cloud of high-speed railway fasteners to detect damage and loosening of the fasteners. Kodai et al. [6] employed a novel technique involving tracking irregular differences in two-track sensors mounted at various positions on the vehicle to investigate the detection of resonance phenomena in high-speed railway bridges. Yang et al. [7] utilized an iterative variational mode decomposition method to separate free and forced vibration modes, achieving automatic detection of the free vibration data of high-speed railway bridges. However, due to the uncertainty and non-uniformly distributed environmental factors, the vibration data collected by sensors often contain “noise”. This creates challenges for engineers in extracting typical damage characteristic data from the vast amount of vibration data acquired, affecting detection efficiency [8,9].
Artificial intelligence technology is gradually being applied in civil engineering to achieve automation and intelligence in HSR bridge structural detection, further reducing labor costs and improving detection efficiency [10,11]. Jing et al. [12] provided an overview of the latest research and applications in various types of railway detection robots, analyzing the feasibility of railway detection robots in efficiently completing detection tasks in the future. Wang et al. [13] employed deep convolutional neural networks to establish a distributed optical fiber acoustic sensing technology for track detection in high-speed railway bridges. Chen et al. [14] established a deep learning system based on a convolutional neural network for detecting and identifying multiple types of damage in high-speed railway bridges. Han et al. [15] proposed a visual detection method for the fracture of high-speed railway catenary U-clamps based on deep learning and utilizing multiple local features and the RSF model. Liang et al. [16] used VGG-16 as a transfer learning pre-trained model for high-speed railway pier damage detection and verified the model’s robustness using Bayesian optimization methods. However, the progress of these studies requires extensive image datasets and engineering application environments that match real-world conditions to ensure the effectiveness of the detection systems. In recent years, with the continuous development of virtual reality technology, the use of synthetic images and virtual scene models to augment datasets has further propelled the application of artificial intelligence technology in civil engineering. Unreal Engine, developed by EPIC Corporation, is a professional game development engine that features a complete animation editing system and various physical collision simulation systems. It allows developers to achieve realistic physical effects on objects that closely resemble the natural world and simulate various motion behaviors and interaction settings for characters [17,18,19]. The Unreal Engine software not only demonstrates powerful capabilities in game development but also shows potential in intelligent civil engineering structures. With its powerful graphics and rendering capabilities, Unreal Engine can generate highly realistic structural components and damage images, enriching the training dataset and thereby enhancing the accuracy of structural damage recognition algorithms [20,21]. Furthermore, developers can create various building scenes and set up obstacle collisions to test the intelligent obstacle avoidance capabilities of drones and the shelter path planning algorithms for construction personnel [22,23]. These studies collectively demonstrate the immense potential of Unreal Engine for the health monitoring of civil engineering structures.
As large-span and super-tall civil engineering structures become increasingly prevalent, the health monitoring of civil engineering structures based on artificial intelligence technology not only requires high detection accuracy but also aims to achieve low cost and high participation rates in the monitoring objectives. Nowadays, smartphones are not only communication tools but also essential links in the “Internet of Everything” in smart cities. The combination of “People + Smartphones” is driving innovation across many industries. For example, O2O platforms like Uber have rapidly expanded in recent years, offering the public smarter and more convenient services for shopping, travel, education, and smart homes. In addition, smartphones for structural health monitoring and damage identification are convenient and cost-effective [24]. Zhao et al. [25,26,27,28,29] utilized smartphones to detect various aspects, including cable tension of bridge cables, inter-story displacement of building structures, structural vibration and strain, and surface damage in concrete. In addition, Ekin et al. [30] installed multiple smartphones at critical locations on the Golden Gate Bridge in the United States, collected vibration data from the bridge, and extracted the bridge’s modal characteristics. Zhang et al. [31] utilized time-frequency fusion features to categorize building seismic response data collected from smartphones automatically. Na et al. [32] proposed a method for assessing building seismic response by utilizing accelerometer data from multiple smartphones to calculate inter-story displacement ratios and compare them with damage limits. Shrestha et al. [33] deployed smartphones and intelligent network units on bridges and conducted long-term monitoring of seismic responses. These studies demonstrate that smartphones can promote large-scale, long-term structural health monitoring.
Earthquakes are among the most sudden and destructive natural disasters, often leading to irreversible structural damage to HSR bridges [34]. Therefore, seismic response assessment of HSR bridges has become a necessary part of their seismic hazard management and emergency response for disaster prevention and mitigation. Building upon the established HSR bridge and track interaction model, Cui et al. [35] utilized the vulnerability analysis to assess the operational risk of the HSR track-bridge interaction system using numerical convolution methods. Chen et al. [36] established a nonlinear model for the coupling of high-speed railway track-train-bridge, studied the influence of bridge-end-induced track deformation under seismic conditions on train operational safety, and proposed using the track deformation rate to assess the safety of high-speed railway bridge-train operation under seismic conditions. Yu et al. [37] focused on the high-speed railway-supported beam-bridge track system and proposed two indices for assessing peak seismic response and residual seismic response. Kang et al. [38] conducted the seismic damage assessment of crucial components of high-speed railway bridges under varying seismic intensities using vibration table experiments and finite element numerical simulation methods.
While there has been improvement in detecting HSR bridge seismic response and operational performance evaluation due to advances in artificial intelligence technology, it still requires professional engineers with specialized knowledge to carry out the corresponding detection and assessment tasks. This study proposes a research framework for rapid detection and assessment of HSR bridge seismic responses based on smartphones in a public participation approach. The research framework primarily assigns the task of HSR bridge seismic response detection to the public through the Smart Bridge Brain (SBB). The public utilizes smartphone-based structural seismic response detection technology to complete the corresponding structural detection tasks. Simultaneously, the SBB collects and analyzes the data, assessing bridges’ operational safety. It also researches the public’s work efficiency index and analyzes the impact of different factors on the operation of the system. Finally, the feasibility of this research framework was validated through virtual experiments created using Unreal Engine 5 (UE5) software. The research framework proposed in this study fully leverages the advantages of crowd detection by the public. This approach is relevant in achieving the intelligence and efficiency of HSR bridge seismic response detection and evaluation.
2. Research Methodology
2.1. Overview of Research Framework
The research framework for rapid detection and assessment of HSR bridge seismic response based on smartphone public participation is shown in Figure 1. This research framework mainly comprises the following three components: (1) the Smart Bridge Brain, (2) bridge seismic response detection, and (3) structural safety and system operational performance evaluation. The seismic response detection of high-speed rail bridges mainly engages trained members of the public who use smartphones to carry out tasks assigned by the intelligent bridge management system. These tasks include measuring seismic response parameters of regional high-speed rail bridges, such as strain detection, surface damage assessment, and displacement monitoring. Importantly, citizens conduct these detection tasks within a consistently safe environment.
2.2. The Smart Bridge Brain (SBB)
To achieve rapid detection and assessment of HSR bridge seismic response through public participation, it is necessary to closely integrate multi-source data and achieve a high degree of integration among different systems. Therefore, based on cluster-based comprehensive data fusion and intelligent analysis technology, this study integrates the HSR bridge comprehensive management terminal (“Smart Bridge Brain”). The Smart Bridge Brain terminal manages the fundamental structural information of regional HSR bridges. It continuously updates the bridge’s health status based on various detection data. At the same time, when an earthquake occurs, the Smart Bridge Brain allocates different detection tasks to the public based on their real-time locations and collects detection data. This is done to achieve rapid detection and assessment of bridge seismic response (Figure 2). The Smart Bridge Brain primarily includes several functions: (1) Storage of basic information for regional HSR bridges: the essential location coordinates, original structural design information, and daily operational and maintenance data for regional HSR bridges will be stored in the intelligent bridge brain information terminal. This allows real-time updates and on-demand access to fundamental information about HSR bridges. (2) Allocation and retrieval of public inspection task data: using the structural inspection requirements for the routine maintenance and comprehensive disaster management of HSR bridges, the Smart Bridge Brain allocates bridge inspection tasks (including routine inspections and emergency detection in response to earthquakes) to the respective inspection public based on real-time location information. This allocation is facilitated through a smartphone APP. Concurrently, the structural detection data acquired by the inspection public is promptly retrieved and stored in the data storage terminal. (3) Intelligent analysis and assessment of bridge seismic response detection data: the SBB conducts intelligent analysis and calculations on various inspection data it acquires.
2.3. Seismic Response Detection Method for HSR Bridge Based on Smartphones
As mentioned in the previous section, smartphones’ high prevalence and numerous built-in sensors make them powerful detection tools for public participation in structural health monitoring. Therefore, in this research, the detection projects for HSR bridge seismic response using smartphones primarily include the following.
(1) Strain detection. Strain detection is one of the necessary inspection projects in civil engineering to ensure the continued safe operation of structures under various disaster conditions. Therefore, in this study, a novel microscopic image strain sensing (MISS) method is employed, which combines smartphones with mobile microscopes for cooperative measurements. This method utilizes the centers of three circles of different sizes to form a triangular region (Figure 3). An arc-supported circular detection algorithm identifies the distance change between the center of the movable circle in the triangular region, denoted as L1. The center of the reference circle is denoted as L. This, combined with the classical cosine theorem, calculates the average strain on the structural surface, denoted as D (as shown in Formula (1)). In the formula, “k” represents the actual distance for each pixel, “M” represents the actual length of the sensor, and “D” represents the detected strain value (positive for tensile strain, negative for compressive strain) [39]. The piston-type sensor (MISS) designed based on this method exhibits high adaptability and can achieve widespread and cost-effective structural strain measurements (Figure 4). The MISS sensor enables the measurement of surface strain on different structures using various smartphones at different times.
(1)
(2) Surface damage detection. Structures are susceptible to surface damage under seismic forces, and such surface damage can sometimes affect the overall structural integrity. Therefore, in the HSR bridge seismic response detection project, it is equally essential to conduct detailed investigations and inspections of the surface damage to the primary structures of HSR bridges. Detecting surface damage on HSR bridge structures can still utilize smartphones to capture images of structural surface damage. Through surface damage recognition and localization algorithms, it is possible to achieve the detection of surface damage on HSR bridge structures. The structural surface damage localization and recognition detection algorithm based on smartphones can achieve pixel-level detection of structural surface damage images. During the detection process, the damaged images are categorized, and the areas of structural damage are entirely segmented from the images. Finally, structural surface damage features are extracted based on the segmentation results, aiming to accurately identify surface damage (Figure 5) [40].
2.4. Structural Safety and System Operational Performance Evaluation
Daily detection and earthquake response emergency detection tasks will generate a large amount of detection data, and the system intelligently analyzes a variety of detection data. According to the relevant inspection specifications, the structural safety and operational performance of HSR bridges will be reassessed as well as the maintenance and repair operation recommendations for their safe and stable operation, so as to ensure that HSR bridges will continue to serve safely. Simultaneously, the system comprehensively reviews and analyzes the execution process of inspection tasks. This includes assessing the sensitivity of the public executing inspection tasks, the required time, the safety importance of inspection indicators, the efficiency of inspection data, and various internal and external factors that impact inspection tasks. The optimization of inspection task workflows is carried out to guarantee the smooth and efficient completion of the system.
3. The Creation of Synthetic Virtual Simulation Experiment Environment
To achieve rapid detection and assessment of seismic responses in high-speed rail bridges, integrating essential regional information such as terrain, structural data, and public contributions is crucial. Additionally, developing a virtual experimental environment can mitigate issues related to time, cost, and safety in detection work, effectively addressing these limitations. Unreal Engine, renowned for its advanced capabilities like graphics rendering, rigid body collision, and physical simulation, empowers developers to craft lifelike scenarios and immersive user interactions effortlessly. In the virtual experimental environment constructed in this study, to achieve a high level of fidelity between each virtual entity in the virtual environment and their real-world physical counterparts, HTML visualization technology was employed alongside the Unreal Engine 5.0 software signaling and web server modules. This facilitated real-time three-dimensional rendering for each virtual entity, resulting in a notably high rendering quality of the virtual models [41,42]. Meanwhile, the virtual experimental environment includes numerous intelligent AI systems and AI data acquisition behavior designs. Unreal Engine and ROS systems utilize YAML-based communication, ROS bridging, and HTTP protocols for interfacing, enabling communication between data streams [43]. Therefore, this research was based on the Unreal Engine 5 professional development platform to establish a comprehensive virtual simulation experiment environment for rapidly detecting and assessing seismic responses in the regional HSR bridge.
3.1. Real Engineering Background
In this research, we selected a newly constructed HSR route in North China. Plains predominantly characterize the geological environment of the project area. The length of the railway route was 10 km, and the entire railway line adopted a construction method known as “bridge in place of road”. This included the construction of three three-span prestressed concrete continuous box girder bridges, named as follows: Bk60 Bridge (48 + 60 + 48 m), Bk64 Bridge (40 + 64 + 40 m), and Bk48 Bridge (32 + 48 + 32 m). (Figure 6). All the continuous girder bridges in this study utilized variable cross-section prestressed concrete box girders. The bridge piers were uniformly designed as circular solid piers, and the bridge bearings employed for all these structures were railway bridge spherical bearings. (Figure 7). The HSR line was designed for a maximum operational train speed of 350 km/h. The track system employed was a ballastless track, and the primary operating vehicle type was the CR400AF model electric multiple unit (EMU) of the Fuxing.
3.2. General Information of the Virtual Scene
In this study, the construction of the virtual scene environment was based on modeling the geographical surroundings of the chosen HSR route. This included parameters such as the elevation of the terrain, vegetation coverage, and road layout in the vicinity of the mountainous areas. However, considering the complexity of the terrain along the HSR route, appropriate simplifications would also be made during the scene construction process. For example, simplifications might be applied to factors such as the types and quantity of vegetation, intricate mountainous structures, and structures along the railway line. This study’s final constructed overall scene model was a rectangle with dimensions of 15 km × 8 km, covering a total area of 1.2 × 108 m2. Simultaneously, the geographical terrain characteristics along the HSR created 25 mountain models of varying heights, three rivers, and ten types of vegetation within the scene. Furthermore, to enhance the content of the scene facilities, we not only constructed 3D models of HSR bridges and but also established operational adjunct facilities associated with HSR bridges (Figure 8). These adjunct facilities encompassed roads, buildings, stations, and so forth. (Table 1).
3.3. Unreal Engine 3D Models
To achieve a higher degree of simulation of various types of structures and facilities in the virtual experimental environment, this study used both 3D modeling software and Unreal Engine 5.0 to establish their simulation models. First, it was necessary to create basic three-dimensional construction models for each structure based on the measurement data of various structures of the HSR bridge. Subsequently, a static mesh model of the bridge was created in the virtual simulation experiment environment based on the three-dimensional construction models. Simultaneously, depending on the different primary structural materials used in the bridge, the corresponding material properties were assigned to the structural models (such as concrete, steel, wood, etc.), thereby completing the creation of the foundational model of the bridge based on measurement data (Figure 9). In addition, we downloaded the corresponding HSR EMU (Electric Multiple Unit) blueprint model components from the Unreal Engine online marketplace. We also made appropriate modifications to these components to align them with the specific structural characteristics of the established HSR bridge model. These modifications encompassed aspects such as the number and length of train carriages, station configurations for train arrivals, track and sleeper structures, materials, etc. (Figure 10). Finally, the train operation was controlled by designing the high-speed train operation blueprint behavior tree to meet the experimental requirements.
3.4. Intelligent AI Public Behavior Creation
In this study, based on the use of smartphones for public participation in HSR bridge seismic response detection and assessment, it was essential to have a public body capable of engaging in a series of intelligent behavioral activities. Therefore, this study was based on the third-person character model within the Unreal Engine beginner’s package. Different AI controllers and task blueprints were designed in a comprehensive virtual experimental environment according to the tasks different user groups would perform in inspection missions. This was done to achieve various intelligent detection activities of AI, as illustrated in Figure 11. At the same time, navigation grids were added to the scene to ensure that the created intelligent AI entities could autonomously navigate and accurately reach the inspection target points under the control of blueprints. They finally integrated the AI controllers of various intelligent AI entities into the Smart Bridge Brain Terminal, achieving controllability and configurability of intelligent AI entities. The specific behavioral activities of intelligent AI public during inspection tasks will be detailed in Section 4.3.
3.5. Smart Bridge Brain GUI
The Smart Bridge Brain (SBB) is an intelligent integrated terminal that combines functions such as storing and displaying primary information data of HSR bridge structures, task assignment and retrieval for inspections, storage, and analysis of detection data. The graphical user interface of the SBB is illustrated in Figure 12. The graphical user interface of the SBB primarily displays the following modules:
(1) Regional HSR Bridge Overview Module. This feature module primarily displays the main bridge types in the regional HSR bridge network, the total lengths of each bridge type, and the proportion of each bridge type’s length to the total length of the HSR line.
(2) Bridge Structural Health Information Module. This feature module displays the regional bridges’ overall structural health information, including the bridge system’s overall health score, the cumulative duration of bridge structure monitoring, the total number of sensors, emergency response occurrences, and the efficiency of emergency response and disposal.
(3) Detection Task Statistics Module. This feature module first displays the total number of bridge detection tasks and the total completed detection tasks. The total number of detection tasks and the number of completed detection tasks will dynamically adjust based on the progress of task execution. Furthermore, this feature module dynamically updates and displays specific information about completed detection tasks, including the type of detection task (routine inspection or emergency inspection), structural condition (normal or in warning state), detection personnel ID, and detection time.
(4) Bridge Video Surveillance Module. This module primarily utilizes surveillance cameras deployed at key locations on bridges. Using image data from these cameras, it is possible to observe whether the bridges have experienced significant displacement, deformation, and noticeable surface damage due to earthquakes, thereby providing an initial assessment of the bridges’ operational status.
(5) High-Speed Train Set Operations Information Module. This feature module primarily displays basic train schedule information for high-speed train sets traveling on the regional HSR line, including train numbers, departure and destination stations, and arrival times of high-speed train sets at the regional HSR bridges.
(6) Data Visualization Module. This feature module primarily visualizes structural strain data obtained from public inspections through curve graphs on the user management interface. Moreover, administrators can access real-time strain data for different regional bridge structures by utilizing a dropdown table menu.
(7) Emergency Event Management Information Module. This feature module primarily focuses on the real-time and dynamic updates of emergency events that may threaten the safe operation of bridges. These events encompass issues related to monitoring cameras, incidents involving train passages, and detecting structural damage through public smartphones.
4. Seismic Response Detection and Evaluation of HSR Bridge
4.1. Earthquake Event Simulation
The geological and environmental conditions established in this study’s virtual comprehensive experimental simulation environment were ideal, with the geological strata at the bottom of the entire virtual simulation environment exhibiting continuous, uniform, and isotropic characteristics. This study, taking into account the site conditions of the actual operating area of high-speed railway bridges, selected a natural earthquake with a magnitude of 7.0 from the PEER Strong Motion Database as the primary seismic input in the virtual experimental environment. The acceleration time-history curve of the earthquake motion is shown in Figure 13a. In order to investigate the seismic response detection and assessment of regional HSR bridges, the epicenter of the seismic motion was randomly positioned along the HSR operating route, as shown in Figure 13b.
4.2. Theoretical Calculation of Seismic Response in HSR Bridge
The theoretical calculations of seismic structural responses for HSR bridges assist inspection personnel in determining the distribution of vulnerable structural positions under seismic conditions and optimizing sensor deployment. Meanwhile, the results of the theoretical analysis can provide input data for virtual experiments on the seismic response of bridge structures.
4.2.1. Finite Element Analysis Model of HSR Bridges
HSR continuous beam bridges have large spans and complex structural forms, and they experience significant loads and vibrations during earthquakes [44]. Therefore, this study selected three continuous beam bridges along the HSR operating route as the subjects for theoretical seismic response analysis of regional HSR bridges. The piers of all three continuous beam bridges were constructed using C35 ordinary reinforced concrete, with longitudinal and lateral reinforcement materials of HRB400. The superstructures are composed of C50 prestressed concrete box girders. Furthermore, actual bridge engineering damage data indicates that piers are prone to structural damage under seismic effects, whereas the upper bridge deck typically remains relatively undamaged. Therefore, this paper did not consider damage to the bridge deck and adopted an elastic material model for the bridge deck material [45]. Concrete and steel reinforcements in the piers were modeled using the ideal elastic-plastic damage model. The finite element model of the continuous beam bridge (using the Bk60 Bridge as an example) is illustrated in Figure 14. Furthermore, this study, considering the actual site conditions of the bridge location (Type II site, Seismic Design Category VII, 0.4 s) [46], inputted the adjusted seismic acceleration time-history curve along the bridge piers’ bottom in the direction of the bridge to obtain the seismic response of the bridge.
4.2.2. Theoretical Calculation Results of Seismic Response in HSR Bridge
Through an analysis of previous investigations into bridge damage during earthquakes, it has been observed that the primary form of damage to bridges is pier destruction [47]. Therefore, the analysis of the response of bridges under seismic conditions primarily focused on the mechanical characteristics of the piers during seismic activity. Figure 15 shows the distribution map of tensile damage throughout the continuous beam bridge under seismic excitation input (using the Bk60 Bridge as an example). Figure 15 shows that under the influence of horizontal seismic motion in the X-direction, the tensile damage to the bridge primarily occurs at the bottom of the bridge piers. This study further extracted the concrete strain time history at the bottom of the piers for each continuous beam bridge to obtain their maximum strain values, as shown in Table 2. As shown in Table 2, the maximum tensile strain values of the bridge piers for various continuous beam bridges under seismic action are 2279.8 με (Bk60 Bridge), 2182.6 με (Bk64 Bridge), and 2302.7 με (Bk48 Bridge), all of which exceed 100 με. Numerous statistical experiments have indicated that when the tensile strain of concrete approaches 100 με–300 με, concrete cracking occurs in bridge piers [44]. Therefore, the bridge piers of various continuous beam bridges may crack under seismic action, leading to concrete cracking and structural damage. In conclusion, for continuous beam bridges subjected to seismic input, vulnerable locations were found at the bottom of the bridge piers, accompanied by damage characterized by concrete cracking. Therefore, in the subsequent seismic response assessment of regional bridges, particular emphasis should be placed on detecting and evaluating the seismic response of bridge piers to fully understand their response characteristics under seismic actions.
4.3. Seismic Response Detection of HSR Bridge
In the previous section, the mechanical response characteristics of regional HSR bridges under seismic conditions were obtained through numerical simulation calculations. The SBB conducted an intelligent analysis of the theoretical calculation result data to determine the distribution of vulnerable positions, the structural emergency detection parameters under seismic conditions for bridges, and the rational allocation of public detection tasks.
4.3.1. Task Assignment
In a survey report on public participation in social disaster management, 59.02% of respondents expressed their willingness to engage in disaster prevention and mitigation efforts [48]. Therefore, fully mobilizing public involvement is a new approach to achieving rapid detection of high-speed rail bridge structural safety in the region. In daily mode, public members who can perform detection tasks near the geographical location of HSR bridges were dispersed in various locations, engaging in their daily production and life activities as usual. However, when an earthquake occurs in the geographical region where the high-speed rail bridge is located, the bridge requires emergency seismic response detection. The SBB assigns different detection tasks to SHM-trained public volunteers and expert volunteers based on the real-time location information of the public, following the principle of assigning tasks to those who can reach the detection target point the fastest. Furthermore, for the convenience of management personnel in identifying the types of detection tasks performed by each general public member and in compiling detection data, each member was assigned a unique numerical label. For example, the label for individuals detecting strain indicators was “SR”, with numbers ranging from 1 to 100 and no repetitions, while the label for those detecting bridge surface damage indicators was “CR”, with numbers from 1 to 100 and no repetitions. In addition, if the public detection results indicate anomalies, the intelligent bridge management system will immediately dispatch and organize expert technicians to perform targeted inspections of the anomalous structures. Based on the public detection data and evaluations, expert technicians will use advanced equipment for on-site testing and conduct more detailed supplementary assessments to achieve a more accurate evaluation of structural safety. This study utilized AI intelligent pathfinding algorithms to plan the movement paths of AI individuals, finding the optimal route from AI individuals’ coordinates to the detection target point. This enabled AI individuals to rush from their initial positions to the detection target point, as shown in Figure 16 [49].
4.3.2. Strain Detection
Sensor optimization layout. In Section 4.2 of this paper, numerical simulation theory calculation obtained bridges’ vulnerable positions and maximum strain values under earthquake action. Therefore, while the SBB assigned detection tasks to intelligent AI individuals, it also transmitted detailed coordinates of vulnerable location distribution and the points with maximum strain values to the intelligent AI individuals. Intelligent AI individuals optimized sensor placement based on coordinate positions, enabling precise detection of vulnerable locations, locations with maximum strain values, and critical bridge structural points. In this study, the damage locations of bridges under seismic effects were primarily situated at the bottom of the bridge piers, and the maximum strain values occurred at unit 543 (Bk60 Bridge), unit 621 (Bk64 Bridge), and unit 473 (Bk48 Bridge). Therefore, sensors were primarily placed in units where significant strain values were distributed at the bottom of the bridge piers. Additionally, sensors were also positioned at critical structural locations, such as h/2 of the piers and the pier tops for each bridge. Furthermore, to enhance the accuracy of detection data, sensors were symmetrically positioned on the opposite side of the pier in the same direction (Figure 17).
Strain detection. Strain detection for the seismic response of HSR bridges in this study was carried out using MISS sensors, specifically targeting strain detection at vulnerable and structurally critical locations of bridges. After receiving the strain detection tasks assigned by the SBB, intelligent AI individuals (SR1-SR3) conducted strain detection at the corresponding locations based on the theoretically calculated vulnerable positions of bridges under seismic response (Figure 18). Based on the development platform of Unreal Engine 5, this study designed and developed a virtual smartphone within a virtual experimental environment. This virtual smartphone incorporated functions such as taking photos, location tracking, and internet access, aiming to simulate the use of smartphones in real-world engineering applications for participatory strain detection. Furthermore, to facilitate the collaborative measurement of bridge strain by various intelligent AI individuals, a virtual smartphone application called “Virtual MISS 1.0 Software” was designed and developed within the virtual smartphone application interface. This application utilized MISS sensors for strain measurements (Figure 19a). During the strain detection task, intelligent AI individuals used virtual smartphones to capture internal micro-images of virtual MISS strain sensors, acquiring detection image data. Simultaneously, the image data were uploaded to a cloud server using the Virtual MISS 1.0 software The cloud server then invoked the MISS algorithm to process the image data, as shown in Figure 19b. Finally, by comparing and calculating the processed results of the detection images with the initial image processing results of sensor installation, the strain values of the measured structure by the sensors could be computed.
To simplify the simulation model and facilitate structural calculations, it is assumed that the strain parameters detected by the public using smartphones for high-speed rail bridges follow a normal distribution [29]. For example, the strain value detected by the public for the bridge structure is Sm~N(μm,), where μm is the expected value and is the variance. The actual strain response value of the bridge structure is represented as Sm0. The strain value detected by the public, Sm, correlates with the actual strain value of the bridge structure, Sm0. Considering sensor uncertainty and the complexity of on-site conditions, the maximum deviation between Sm and Sm0 is set at ±5%. The standard deviation is determined by the minimum and maximum boundary limits. Under these conditions, the expected value μm = Sm0, and the standard deviation is 0.025 Sm0. The public-detected strain values of high-speed rail bridge structures are compared with the strain values under the normal service limit state to evaluate the safety state under seismic response. When detecting the seismic response strain of high-speed rail bridges in the real world, the public collects actual strain detection results. However, in simulation tests conducted in a virtual experimental environment, the focus is on the application of the detection strategy. Therefore, the assumption regarding the strain values for high-speed rail bridge seismic response is acceptable.
Structural assessment. Each intelligent AI individual, based on the coordinates of the structures assigned by the Smart Bridge Brain, sequentially performed seismic response strain detection at the vulnerable locations and critical structural points of the regional HSR bridges. Subsequently, they promptly provided feedback on the detection data to the SBB through their smartphones. The SBB evaluated the structural seismic response state based on the strain data obtained by intelligent AI individuals at various measurement points. This evaluation was performed by analyzing the ratio (Rs) between the measured strain data values (Sm) and the theoretically calculated peak values (Sc) at each measurement point. The seismic response states of the bridge structures were categorized as “normal” (Rs < 0.2), “warning” (0.2 ≤ Rs < 0.4), and “failure” (Rs ≥ 0.4) [50]. Table 3 presents the strain detection results for the main structures of the regional HSR bridges. Table 3 shows that the ratio of strain detection values to the theoretically calculated strain peak values for all detection units of the Bk64 Bridge did not exceed 0.4 and was less than 0.2. As a result, the structural operational status was assessed as “normal”, and the bridge can continue to provide safe service after seismic events. However, significant strain values were recorded in the strain detection at vulnerable locations and critical structural points of the Bk60 Bridge and Bk48 Bridge, with their ratios to the theoretically calculated strain peak values exceeding 0.2. This indicated a potential risk to the safety of high-speed train operations, potentially affecting the structural serviceability. Therefore, to ensure the safety of the bridge’s serviceability, its operational status was assessed as “warning”. At the same time, the intelligent bridge management system will immediately dispatch and coordinate expert technicians to conduct critical inspections of the structure. Expert technicians will use advanced equipment to perform on-site secondary detection and assessment of the bridge’s seismic performance, identify potential structural safety issues, and take appropriate maintenance and repair measures based on the degree of structural damage. This ensures the safe and stable operation of high-speed rail lines.
4.3.3. Surface Damage Detection
Based on the theoretical seismic response calculations presented in Section 4.2 of this study, it was evident that extensive damage occurs in the lower sections of bridge piers under seismic loading. Additionally, there was the occurrence of concrete cracking in the piers as a result of this structural damage. Therefore, during the regional bridge seismic response detection tasks, it was imperative to carry out the necessary damage detection of surface damage on the bridges. The SBB allocated surface damage detection tasks to the respective intelligent AI individuals (CR1–CR3) based on their real-time locations. The intelligent AI individuals, equipped with virtual smartphones, followed the task-assigned coordinates of the structures to be inspected. They planned the optimal path to reach the target detection location and performed surface damage detection tasks. After reaching the target detection location, the intelligent AI individuals, utilizing virtual smartphones, sequentially captured detailed surface damage image data at the damage detection points, as shown in Figure 20a. After capturing image data, the intelligent AI individuals transmitted the collected HSR area bridge surface damage image data back to the SBB. Simultaneously, utilizing the built-in positioning functionality of the virtual smartphones, the location information of bridge damage was obtained, enhancing the accuracy of surface damage detection. In addition, to obtain seismic response information for the upper supports and beams of the bridge, an unmanned aerial vehicle (UAV) equipped with flying, image capturing, and autonomous navigation capabilities was deployed in a virtual experimental environment [51]. This UAV was used to detect surface damage and displacement of the upper supports and beams of the bridge (as shown in Figure 20b), with the detection data sent to the SBB.
The SBB promptly stored the image data collected by the intelligent AI individuals and the UAV. Simultaneously, the SBB performed computational analysis on the uploaded original image data. It extracted surface damage information related to the seismic response of bridges, including cracks, spalling, and exposed reinforcement steel. Finally, the SBB compiled and summarized the surface damage information of regional bridge structures under the seismic event, as shown in Figure 21. Figure 21 shows that the predominant types of surface damage in regional bridges under seismic events are cracks, spalling, and exposed reinforcement steel. Moreover, there is a higher quantity of surface crack damage distribution. Therefore, during bridges’ post-seismic operational safety monitoring, attention should be paid to developing and distributing surface cracks in the structures to ensure their safe and stable operation.
4.4. Comprehensive Assessment of Seismic Response Detection in HSR Bridge
Through evaluating the safety operation performance of the HSR bridge’s seismic response structure, the structure’s operation safety problems can be found in time, and the necessary maintenance measures can be taken to ensure the safe operation of the HSR bridge when an earthquake occurs. At the same time, this study also focused on the efficiency of public participation in the inspection task, which was essential for the rapid inspection and repair of the HSR bridge.
4.4.1. Comprehensive Assessment of HSR Bridge Structural Safety and Operational Performance
Intelligent AI individuals used smartphones to complete various HSR bridge seismic response detection tasks and transmit the detection data back to the SBB. Through in-depth data mining and analysis of seismic response detection data and routine inspection data, and based on specialized regulations for HSR bridge design, inspection, and safety assessment [46,52,53], it conducted the post-seismic operational performance assessment and health management analysis of HSR bridges, ensuring the safety and stable operation of regional HSR bridges after seismic events. Finally, the SBB presented the comprehensive analysis and assessment results of HSR bridges as a user interface on the “HSR Bridge Integrated Management Platform”, as depicted in Figure 22. Managers could access HSR bridges’ overall structural health assessment after seismic events and critical seismic response data through this user interface. For example, they could access post-seismic overall structural performance ratings of HSR bridges, the earthquake acceleration curve for this event, structural seismic response detection data, the total volume of stored detection data, and data composition. This enhances the utility and confidence in the detection data and assists managers in making informed decisions regarding post-seismic maintenance and repairs of bridges, ensuring their safe and stable operation [54].
4.4.2. Comprehensive Assessment of Public Participation in Efficiently Executing Detection Work
During the execution of detection tasks by the public, they are inevitably influenced by various factors from both internal and external environments. These influences may increase the detection cost and reduce the efficiency of the detection work. In this study, we primarily considered the following three factors: (1) Sensitivity of public task execution. The sensitivity of the public in executing detection tasks, which refers to the willingness of the public to carry out tasks, is influenced by factors such as an individual’s social background, social perception, and organizational initiative. This directly relates to whether the public accepts detection tasks and their attitude towards the work [55]. (2) Data collection completeness. The completeness and accuracy of data collected during the execution of public detection tasks are the most critical factors in assessing the work of the detection public. They directly relate to the effectiveness of task execution. (3) Completion time of detection. To achieve the goal of rapid bridge inspection, the time taken for the entire detection task execution is also an essential evaluative factor. To more accurately assess the efficiency of the detection work, this study employed a stratified weighting method to calculate the work efficiency index (WEI) based on the degree of influence of the three factors and the results of the detection work. The calculation formula is shown below in Equation (2):
(2)
In Formula (2), WEI represents the work efficiency index; represents the score for data collection work conditions, with scores assigned based on the completeness and accuracy of data collection, ranging from 0 to 100; represents the weight value of data influence factors, with a value of 0.7; represents the score for task completion time conditions, with higher scores for shorter completion times, ranging from 1 to 100; represents the time influence factor with a value of 0.3; represents the task execution sensitivity factor, which is assigned values based on the attitude towards task acceptance, with “Very Willing” receiving 0.8, “Somewhat Willing” receiving 0.6, and “Uncertain” receiving 0.3. In this study, based on various parameters recorded during the execution of public detection tasks in the experimental process (such as distance, time, detection data, etc.), we calculated the work efficiency index for each public member, as shown in Table 4 below.
When the SBB assigned detection tasks to the public, the principle of allocating tasks was based on the proximity to the detection target point. Therefore, the distance of each public member to the detection target point was 400 m. Table 4 shows that under the same conditions of completing the total number of measurement points and the time taken for detection, the more valid measurement data obtained, the higher the WEI. In other words, the WEI of SR1 was 9.9% higher than that of SR2. At the same time, when the total number of completed points and valid points is the same, shorter completion times result in higher WEI values. Regarding the public’s sensitivity to accepting detection tasks, higher sensitivity factors lead to a higher total number of completed detection tasks and valid data, resulting in higher WEI assessment values. In summary, during the allocation of detection tasks, the SBB should prioritize the public’s sensitivity to performing detection tasks and assign detection tasks to those members of the public who are more willing to execute the tasks. Meanwhile, the SBB, in combination with the WEI obtained by each member of the public, intelligently analyzes the impact of various factors during the execution of detection tasks, optimizing the task allocation mechanism and process, thereby ensuring the rapid and efficient completion of detection tasks.
5. Discussion
The feasibility of the research framework in this paper is verified by establishing a virtual experimental environment for the seismic response of high-speed railroad bridges, and the WEI that affects the execution of the system framework is analyzed. Specifically, for the input ground shaking, we used seismic waves from actual engineering site conditions. High-speed rail lines are selected for operation with classic bridge types and structures. The intelligent AI public body completes the detection of the seismic response parameters of the predetermined damaged bridge structures on the operational routes based on smartphones, and the SBB completes the analysis of the detection data and the structural safety assessment. In addition, based on the effects of the three influencing factors of task sensitivity, detection time and data detection accuracy on WEI, the inspection work execution strategy is optimized with the orientation of obtaining a higher WEI. It fully leverages the proactiveness of public participation in infrastructure disaster prevention and mitigation, thereby contributing to disaster resilience in cluster-based HSR bridge structures. Moreover, the method employed in this paper, conducting virtual experiments in Unreal Engine 5.0 software, offers a viable alternative for large-scale transportation infrastructure disaster prevention and reduction experiments in the real world.
In the real world, citizens from different cultural backgrounds, professions, and prior knowledge have different subjective energies in carrying out their tasks, which brings about a great deal of uncertainty in the efficient execution of work. Ethics, law, personal emotions, etc. can also affect the public’s willingness to perform tasks. However, based on the existing research framework, the development and establishment of an effective reward mechanism for task completion and the full mobilization of citizens’ motivation can further improve the success rate of task completion. Meanwhile, the training for public structure detection technology and work safety, the optimization of task routes, and the reduction of subjectivity in data reporting will also have a significant impact on the efficient operation of the entire system. The study of these influencing factors will be a focus of future research. In addition, while smartphone-based structural seismic response detection and assessment methods introduce innovative approaches for structural health monitoring, more reliable methods can still be developed and integrated into the research framework to achieve more comprehensive and detailed assessments of structural safety. Finally, the occurrence of actual natural disasters has significant randomness and complexity, and the response characteristics of structures are also variable. The system framework proposed in this study needs to be used in practical engineering applications to continuously improve the core content and technical level, which is also a need for further research work in the future.
6. Conclusions
This study is based on smartphone structural health detection technology and the public participation-based rapid detection strategy for structural seismic response. The rapid detection and evaluation system for seismic response of HSR bridge structures is studied, and a smartphone-based public participation-based rapid detection and evaluation framework for seismic response of HSR bridge beams is proposed. Finally, based on the development software of Unreal Engine 5.0, this study constructed a mixed-reality virtual simulation experimental environment, confirming the system’s feasibility. The main conclusions are as follows:
(1) This study proposes a rapid detection and assessment system for the seismic response of HSR bridge structures based on smartphones and public participation. This systemic framework comprises the Smart Bridge Brain (SBB), public participants, and smartphone-based structural seismic response detection technology. The SBB encompasses the fusion of multiple data sources and the highly integrated operation of various systems. It guides public participants to swiftly carry out HSR bridge structural seismic response detection using smartphones, along with managing and analyzing detection data. Public participants engaged in the detection tasks utilize smartphones to perform emergency seismic response detection for HSR bridges, collecting data on the seismic response of bridge structures. In addition, comparing the detection data information with the bridge structure’s seismic response results also validates the system’s feasibility.
(2) This study utilized Unreal Engine 5 software to create a mixed-reality virtual simulation experiment scene. Experimental tests were conducted to observe the operational process of the smartphone-based public participation HSR bridge seismic response rapid detection and assessment system. The SBB acquires seismic information for the HSR line area and allocates emergency seismic response detection tasks to relevant public participants. The public participants, engaged in executing the detection tasks, utilize smartphones to detect and collect bridge structural seismic response information, including structural strain and surface damage. The results indicate that the SBB can effectively allocate detection tasks to public participants. Public participants can also reach the predefined target detection points, complete the detection of structural seismic response indicators, and achieve a faster detection speed with accurate data collection.
(3) This study analyzed the impact of three factors—sensitivity in public participants executing detection tasks, time taken to complete detection, and accuracy in data collection—on the system’s efficient operation. The results indicate that the public’s sensitivity in executing detection tasks significantly impacts the WEI (work efficiency index) assessment score. As the sensitivity factor increases, the total number of detection tasks completed by the public and the effective count also increase, leading to a higher WEI assessment score. Additionally, shorter detection task completion times result in higher WEI values.
Conceptualization, X.Z., J.L. and W.L.; methodology, X.Z.; software, J.L. and C.Z.; validation, J.L. and Y.J.; formal analysis, J.L., W.L. and C.Y.; investigation, J.L., Y.J. and C.Y.; data curation, J.L., W.L. and C.Z.; writing—original draft preparation, J.L. and W.L.; writing—review and editing, J.L., W.L. and C.Z.; visualization, J.L. and Y.J.; supervision, X.Z. and W.L. All authors have read and agreed to the published version of the manuscript.
Data are contained within the article.
The authors declare no conflicts of interest.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Enlarge this image.Figure 7. Bk48 Bridge pier and box girder cross-section diagram (unit: cm): (a) pier cross-section (b) box girder cross-section.
Enlarge this image.Figure 10. High-speed train and associated facilities 3D rendering model: (a) high-speed train (b) HSR station.
Enlarge this image.Figure 11. Intelligent AI public model: (a) intelligent AI public model; (b) partial AI control blueprint.
Enlarge this image.Figure 13. Seismic wave acceleration time history curve and epicenter location map: (a) seismic wave acceleration time history curve; (b) seismic epicenter location.
Enlarge this image.Figure 16. AI public movement process: (a) initial position; (b) move to the target points.
Enlarge this image.Figure 17. Sensor layout schematic and 3D real scene image: (a) sensor layout schematic; (b) 3D real scene image of sensor layout.
Enlarge this image.Figure 19. Smartphone application and virtual miss software interface: (a) smartphone application interface; (b) virtual miss software interface.
Enlarge this image.Figure 20. Intelligent AI public and UAV image data collection: (a) intelligent AI public image data collection; (b) UAV image data collection.
Summary table of key information for virtual scenes.
| Number | Name | Form | Quantity |
|---|---|---|---|
| 1 | Geographic terrain | Vegetation, Mountain, River | 1.2 × 108 m2 |
| 2 | Road | Third-class highway | 20 km |
| 3 | High-speed railway bridge | Continuous beam bridge, Simply supported beam bridge | 10 km |
| 4 | High-speed train | CR400AF | 1 |
| 5 | Public | AI | 10 |
| 6 | High-speed train station | Small-sized station | 2 |
| 7 | Duty room | High-speed railway bridge integrated management platform | 1 |
Maximum concrete strain in each bridge pier under earthquake effects.
| Number | Bridge | Maximum Strain (με) | Structure | Element | |
|---|---|---|---|---|---|
| + | − | ||||
| 1 | Bk60 | 2279.8 | 1598.3 | Pier | 543 |
| 2 | Bk64 | 2182.6 | 1418.7 | Pier | 621 |
| 3 | Bl48 | 2302.7 | 1542.1 | Pier | 473 |
Strain detection results of key structural components of regional HSR bridges.
| Bridge | Site | Element | Coordinates | Value (Sm) | Ratio (Rs) | State |
|---|---|---|---|---|---|---|
| Bk60 | 13# Pier | 543 | (541, 762, 200) | 574.2 με | 0.25 | Warning |
| 13# Pier | 412 | (542, 762, 200) | 492.4 με | 0.22 | Warning | |
| 12# Pier | 214 | (512, 623, 182) | 152.6 με | 0.15 | Normal | |
| 15# Pier | 753 | (574, 452, 181) | 148.3 με | 0.12 | Normal | |
| Bk64 | 24# Pier | 621 | (617, 871, 175) | 218.3 με | 0.10 | Normal |
| 24# Pier | 661 | (620, 871, 175) | 195.4 με | 0.11 | Normal | |
| 23# Pier | 497 | (592, 863, 170) | 82.6 με | 0.04 | Normal | |
| 26# Pier | 840 | (663, 892, 180) | 48.7 με | 0.03 | Normal | |
| Bk48 | 34# Pier | 473 | (1214, 965, 162) | 736.2 με | 0.32 | Warning |
| 34# Pier | 495 | (1218, 965, 162) | 702.4 με | 0.29 | Warning | |
| 33# Pier | 392 | (1163, 952, 165) | 324.6 με | 0.21 | Warning | |
| 36# Pier | 598 | (1232, 982, 167) | 285.1 με | 0.17 | Normal |
Calculation table of efficiency status index for public detection work.
| Number | Distance | Time | Completed Points | Valid | Assignment of Impact Factor Scores | Sensitivity Factor | WEI | |
|---|---|---|---|---|---|---|---|---|
| Data DWm | Time TCs | |||||||
| SR1 | 400 m | 20.0 | 20.0 | 20.0 | 100 | 60 | 0.8 | 70.4 |
| SR2 | 400 m | 20.0 | 20.0 | 16.0 | 88 | 60 | 0.8 | 63.7 |
| SR3 | 400 m | 12.0 | 20.0 | 20.0 | 100 | 100 | 0.8 | 80.0 |
| CR1 | 400 m | 20.0 | 19.0 | 15.0 | 83 | 60 | 0.8 | 60.9 |
| CR2 | 400 m | 25.0 | 16.0 | 10.0 | 62 | 48 | 0.6 | 34.7 |
| CR3 | 400 m | 30.0 | 12.0 | 5.0 | 39 | 40 | 0.3 | 11.8 |
References
1. Hanley, D.; Li, J.; Wu, M. High-speed railways and collaborative innovation. Reg. Sci. Urban Econ.; 2022; 93, 103717. [DOI: https://dx.doi.org/10.1016/j.regsciurbeco.2021.103717]
2. Liu, X.; Jiang, L.; Xiang, P.; Feng, Y.; Lai, Z.; Zhou, W. Running Safety of High-Speed Railway Train on Bridge During Earthquake Considering Uncertainty Parameters of Bridge. Int. J. Struct. Stab. Dyn.; 2022; 22, 2241008. [DOI: https://dx.doi.org/10.1142/S0219455422410085]
3. Zhou, T.; Jiang, L.; Liu, X.; Xiang, P.; Lai, Z.; Zhang, Y. An efficient simplified model for high-speed railway simply supported bridge under earthquakes. Struct. Infrastruct. Eng.; 2023; 19, pp. 1811-1825. [DOI: https://dx.doi.org/10.1080/15732479.2022.2058562]
4. Lu, C.; Cai, C. Overview on safety management and maintenance of high-speed railway in China. Transp. Geotech.; 2020; 25, 100397. [DOI: https://dx.doi.org/10.1016/j.trgeo.2020.100397]
5. Mao, Q.; Cui, H.; Hu, Q.; Ren, X. A rigorous fastener inspection approach for high-speed railway from structured light sensors. ISPRS J. Photogramm. Remote Sens.; 2018; 143, pp. 249-267. [DOI: https://dx.doi.org/10.1016/j.isprsjprs.2017.11.007]
6. Matsuoka, K.; Tanaka, H.; Kawasaki, K.; Somaschini, C.; Collina, A. Drive-by methodology to identify resonant bridges using track irregularity measured by high-speed trains. Mech. Syst. Signal Process.; 2021; 158, 107667. [DOI: https://dx.doi.org/10.1016/j.ymssp.2021.107667]
7. Yang, X.-M.; Yi, T.-H.; Qu, C.-X.; Li, H.-N.; Liu, H. Modal Identification of High-Speed Railway Bridges through Free-Vibration Detection. J. Eng. Mech.; 2020; 146, 04020107. [DOI: https://dx.doi.org/10.1061/(ASCE)EM.1943-7889.0001847]
8. Dang, H.V.; Raza, M.; Nguyen, T.V.; Bui-Tien, T.; Nguyen, H.X. Deep learning-based detection of structural damage using time-series data. Struct. Infrastruct. Eng.; 2021; 17, pp. 1474-1493. [DOI: https://dx.doi.org/10.1080/15732479.2020.1815225]
9. Lin, Y.; Nie, Z.; Ma, H. Structural Damage Detection with Automatic Feature-Extraction through Deep Learning: Structural damage detection with automatic feature-extraction through deep learning. Comput-Aided Civ. Infrastruct. Eng.; 2017; 32, pp. 1025-1046. [DOI: https://dx.doi.org/10.1111/mice.12313]
10. Zhang, C.; Lu, Y. Study on artificial intelligence: The state of the art and future prospects. J. Ind. Inf. Integr.; 2021; 23, 100224. [DOI: https://dx.doi.org/10.1016/j.jii.2021.100224]
11. Dong, C.-Z.; Catbas, F.N. A review of computer vision-based structural health monitoring at local and global levels. Struct. Health Monit.; 2021; 20, pp. 692-743. [DOI: https://dx.doi.org/10.1177/1475921720935585]
12. Jing, G.; Qin, X.; Wang, H.; Deng, C. Developments, challenges, and perspectives of railway inspection robots. Autom. Constr.; 2022; 138, 104242. [DOI: https://dx.doi.org/10.1016/j.autcon.2022.104242]
13. Wang, S.; Liu, F.; Liu, B. Research on application of deep convolutional network in high-speed railway track inspection based on distributed fiber acoustic sensing. Opt. Commun.; 2021; 492, 126981. [DOI: https://dx.doi.org/10.1016/j.optcom.2021.126981]
14. Chen, L.; Chen, W.; Wang, L.; Zhai, C.; Hu, X.; Sun, L.; Tian, Y.; Huang, X.; Jiang, L. Convolutional neural networks (CNNs)-based multi-category damage detection and recognition of high-speed rail (HSR) reinforced concrete (RC) bridges using test images. Eng. Struct.; 2023; 276, 115306. [DOI: https://dx.doi.org/10.1016/j.engstruct.2022.115306]
15. Han, Y.; Liu, Z.; Lyu, Y.; Liu, K.; Li, C.; Zhang, W. Deep learning-based visual ensemble method for high-speed railway catenary clevis fracture detection. Neurocomputing; 2020; 396, pp. 556-568. [DOI: https://dx.doi.org/10.1016/j.neucom.2018.10.107]
16. Liang, X. Image-based post-disaster inspection of reinforced concrete bridge systems using deep learning with Bayesian optimization. Comput-Aided Civ. Infrastruct. Eng.; 2019; 34, pp. 415-430. [DOI: https://dx.doi.org/10.1111/mice.12425]
17. Chu, E.; Zaman, L. Exploring alternatives with Unreal Engine’s Blueprints Visual Scripting System. Entertain. Comput.; 2021; 36, 100388. [DOI: https://dx.doi.org/10.1016/j.entcom.2020.100388]
18. Nowak, L.; Bak, A.; Czajkowski, T.; Wojciechowski, K. Modeling and Rendering of Volumetric Clouds in Real-Time with Unreal Engine 4. Proceedings of the International Conference on Computer Vision and Graphics; Warsaw, Poland, 17–19 September 2018; Chmielewski, L.J.; Kozera, R.; Orlowski, A.; Wojciechowski, K.; Bruckstein, A.M.; Petkov, N. Springer International Publishing Ag: Cham, Switzerland, 2018; Volume 11114, pp. 68-78.
19. Rosenbeiger, M.; Bak, A.; Czajkowski, T. Real-Time Simulation of Animated Characters Crowd in Unreal Engine 4. Proceedings of the International Conference on Computer Vision and Graphics; Warsaw, Poland, 17–19 September 2018; Chmielewski, L.J.; Kozera, R.; Orlowski, A.; Wojciechowski, K.; Bruckstein, A.M.; Petkov, N. Springer International Publishing Ag: Cham, Switzerland, 2018; Volume 11114, pp. 79-88.
20. Siu, C.; Wang, M.; Cheng, J.C. A framework for synthetic image generation and augmentation for improving automatic sewer pipe defect detection. Autom. Constr.; 2022; 137, 104213. [DOI: https://dx.doi.org/10.1016/j.autcon.2022.104213]
21. Rasmussen, I.; Kvalsvik, S.; Andersen, P.-A.; Aune, T.N.; Hagen, D. Development of a Novel Object Detection System Based on Synthetic Data Generated from Unreal Game Engine. Appl. Sci.; 2022; 12, 8534. [DOI: https://dx.doi.org/10.3390/app12178534]
22. Song, C.; Chen, Z.; Wang, K.; Luo, H.; Cheng, J.C.P. BIM-supported scan and flight planning for fully autonomous LiDAR-carrying UAVs. Autom Constr.; 2022; 142, 104533. [DOI: https://dx.doi.org/10.1016/j.autcon.2022.104533]
23. Narazaki, Y.; Hoskere, V.; Chowdhary, G.; Spencer, B.F. Vision-based navigation planning for autonomous post-earthquake inspection of reinforced concrete railway viaducts using unmanned aerial vehicles. Autom Constr.; 2022; 137, 104214. [DOI: https://dx.doi.org/10.1016/j.autcon.2022.104214]
24. Sony, S.; LaVenture, S.; Sadhu, A. A literature review of next-generation smart sensing technology in structural health monitoring. Struct. Control Health Monit.; 2019; 26, e2321. [DOI: https://dx.doi.org/10.1002/stc.2321]
25. Liu, C.; Zhou, L.; Wang, W.; Zhao, X. Concrete Surface Damage Volume Measurement Based on Three-Dimensional Reconstruction by Smartphones. IEEE Sens. J.; 2021; 21, pp. 11349-11360. [DOI: https://dx.doi.org/10.1109/JSEN.2021.3067739]
26. Li, H.; Chen, X.; Chen, H.; Wang, B.; Li, W.; Liu, S.; Li, P.; Qi, Z.; He, Z.; Zhao, X. Simulation of Smartphone-Based Public Participation in Earthquake Structural Response Emergency Monitoring Using a Virtual Experiment and AI. Buildings; 2022; 12, 492. [DOI: https://dx.doi.org/10.3390/buildings12040492]
27. Zhao, X.; Han, R.; Yu, Y.; Hu, W.; Jiao, D.; Mao, X.; Li, M.; Ou, J. Smartphone-Based Mobile Testing Technique for Quick Bridge Cable–Force Measurement. J. Bridg. Eng.; 2017; 22, 06016012. [DOI: https://dx.doi.org/10.1061/(ASCE)BE.1943-5592.0001011]
28. Li, J.; Xie, B.; Zhao, X. Measuring the interstory drift of buildings by a smartphone using a feature point matching algorithm. Struct. Control Health Monit.; 2020; 27, e2492. [DOI: https://dx.doi.org/10.1002/stc.2492]
29. Xie, B.; Chen, X.; Ding, M.; Zhou, G.; Zhao, X. Design and development of a new strain measuring method based on smartphone and machine vision. Measurement; 2021; 182, 109724. [DOI: https://dx.doi.org/10.1016/j.measurement.2021.109724]
30. Ozer, E.; Purasinghe, R.; Feng, M.Q. Multi-output modal identification of landmark suspension bridges with distributed smartphone data: Golden Gate Bridge. Struct. Control Health Monit.; 2020; 27, e2576. [DOI: https://dx.doi.org/10.1002/stc.2576]
31. Zhang, Y.; Yuen, K.-V. Time-frequency fusion features-based incremental network for smartphone measured structural seismic response classification. Eng. Struct.; 2023; 278, 115575. [DOI: https://dx.doi.org/10.1016/j.engstruct.2022.115575]
32. Na, Y.; El-Tawil, S.; Ibrahim, A.; Eltawil, A. Automated Assessment of Building Damage from Seismic Events Using Smartphones. J. Struct. Eng.; 2020; 146, 04020076. [DOI: https://dx.doi.org/10.1061/(ASCE)ST.1943-541X.0002618]
33. Shrestha, A.; Dang, J.; Wang, X.; Matsunaga, S. Smartphone-Based Bridge Seismic Monitoring System and Long-Term Field Application Tests. J. Struct. Eng.; 2020; 146, 04019208. [DOI: https://dx.doi.org/10.1061/(ASCE)ST.1943-541X.0002513]
34. Meng, D.; Yang, M.; Sun, T.; He, X.; Chouw, N. Seismic-induced poundings in highway and high-speed railway bridges: A state-of-the-art review. Struct. Infrastruct. Eng.; 2023; pp. 1-15. [DOI: https://dx.doi.org/10.1080/15732479.2023.2230554]
35. Cui, S.; Guo, C.; Su, J.; Cui, E.; Liu, P. Seismic fragility and risk assessment of high-speed railway continuous-girder bridge under track constraint effect. Bull. Earthq. Eng.; 2019; 17, pp. 1639-1665. [DOI: https://dx.doi.org/10.1007/s10518-018-0491-9]
36. Chen, Y.; Jiang, L.; Lai, Z.; Li, J. A novel evaluation of train running safety on the HSR bridges under earthquakes based on the observed track deformation rate. Eng. Struct.; 2023; 275, 115260. [DOI: https://dx.doi.org/10.1016/j.engstruct.2022.115260]
37. Yu, J.; Zhou, W.; Jiang, L. Study on the estimate for seismic response of high-speed railway bridge-track system. Eng. Struct.; 2022; 267, 114711. [DOI: https://dx.doi.org/10.1016/j.engstruct.2022.114711]
38. Kang, X.; Jiang, L.; Bai, Y.; Caprani, C.C. Seismic damage evaluation of high-speed railway bridge components under different intensities of earthquake excitations. Eng. Struct.; 2017; 152, pp. 116-128. [DOI: https://dx.doi.org/10.1016/j.engstruct.2017.08.057]
39. Xie, B.; Zhang, L.; Ding, M.; Li, W.; Zhao, X. Strain measurement based on cooperative operation with different smartphones. Comput. Civ. Infrastruct. Eng.; 2023; 38, pp. 241-252. [DOI: https://dx.doi.org/10.1111/mice.12919]
40. Li, S.; Zhao, X.; Zhou, G. Automatic pixel-level multiple damage detection of concrete structure using fully convolutional network. Comput. Aided Civ. Infrastruct. Eng.; 2019; 34, pp. 616-634. [DOI: https://dx.doi.org/10.1111/mice.12433]
41. Scorpio, M.; Laffi, R.; Teimoorzadeh, A.; Ciampi, G.; Masullo, M.; Sibilio, S. A calibration methodology for light sources aimed at using immersive virtual reality game engine as a tool for lighting design in buildings. J. Build. Eng.; 2022; 48, 103998. [DOI: https://dx.doi.org/10.1016/j.jobe.2022.103998]
42. Calabuig-Barbero, E.; Davia-Aracil, M.; Mora-Mora, H.; Herrero-Pérez, F. Computational model for hyper-realistic image generation using uniform shaders in 3D environments. Comput. Ind.; 2020; 123, 103337. [DOI: https://dx.doi.org/10.1016/j.compind.2020.103337]
43. Lin, Y.-C.; Chen, Y.-P.; Yien, H.-W.; Huang, C.-Y.; Su, Y.-C. Integrated BIM, game engine and VR technologies for healthcare design: A case study in cancer hospital. Adv. Eng. Inform.; 2018; 36, pp. 130-145. [DOI: https://dx.doi.org/10.1016/j.aei.2018.03.005]
44. Jiang, L.; Kang, X.; Li, C.; Shao, G. Earthquake response of continuous girder bridge for high-speed railway: A shaking table test study. Eng. Struct.; 2019; 180, pp. 249-263. [DOI: https://dx.doi.org/10.1016/j.engstruct.2018.11.047]
45. Wei, B.; Yang, T.; Jiang, L.; He, X. Effects of uncertain characteristic periods of ground motions on seismic vulnerabilities of a continuous track–bridge system of high-speed railway. Bull. Earthq. Eng.; 2018; 16, pp. 3739-3769. [DOI: https://dx.doi.org/10.1007/s10518-018-0326-8]
46. Ministry of Railways of the People’s Republic of China. Code for Seismic Design of Railway Engineering; China Planning Press: Beijing, China, 2006.
47. Zsarnóczay, A.; Vigh, L.G.; Kollár, L.P. Seismic Performance of Conventional Girder Bridges in Moderate Seismic Regions. J. Bridg. Eng.; 2014; 19, 04014001. [DOI: https://dx.doi.org/10.1061/(ASCE)BE.1943-5592.0000536]
48. Zhou, Y. Investigation and Analysis on Public Participation in Community Disaster Risk Governance in China. J. Catastrophol.; 2022; 37, pp. 179-185.
49. Xie, Y.; Chen, Y.; Wang, Z.; Ou, Y. AI intelligent wayfinding based on Unreal Engine 4 static map. J. Phys. Conf. Ser.; 2022; 2253, 012016. [DOI: https://dx.doi.org/10.1088/1742-6596/2253/1/012016]
50. Chen, X.; Wang, B.; Chen, J.; Zhang, X.; Liu, S.; Zhou, G.; Li, P.; Zhao, X. Innovative Life-Cycle Inspection Strategy of Civil Infrastructure: Smartphone-Based Public Participation. Struct. Control Health Monit.; 2023; 2023, pp. 1-17. [DOI: https://dx.doi.org/10.1155/2023/8715784]
51. Esfahlani, S.S. Mixed reality and remote sensing application of unmanned aerial vehicle in fire and smoke detection. J. Ind. Inf. Integr.; 2019; 15, pp. 42-49. [DOI: https://dx.doi.org/10.1016/j.jii.2019.04.006]
52. Ministry of Railways of the People’s Republic of China. Code for Rating Existing Railway Bridges; China Railway Publishing House: Beijing, China, 2004.
53. China State Railway Group Co., Ltd. High-Speed Railway Bridge Operation Performance Verification Regulations; China Railway Publishing House: Beijing, China, 2014.
54. Montenegro, P.; Carvalho, H.; Ribeiro, D.; Calçada, R.; Tokunaga, M.; Tanabe, M.; Zhai, W. Assessment of train running safety on bridges: A literature review. Eng. Struct.; 2021; 241, 112425. [DOI: https://dx.doi.org/10.1016/j.engstruct.2021.112425]
55. Ma, Y.; Zhu, W.; Zhang, H.; Zhao, P.; Wang, Y.; Zhang, Q. The Factors Affecting Volunteers’ Willingness to Participate in Disaster Preparedness. Int. J. Environ. Res. Public Health; 2021; 18, 4141. [DOI: https://dx.doi.org/10.3390/ijerph18084141]
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
; Zhao, Chenhao 1 ; Jing, Yicheng 1 ; Yin, Chao 1 ; Zhao, Xuefeng 1
1 School of Infrasturcture Engineering, Dalian University of Technology, Dalian 116024, China;