Railway bridge static load bending test automatic control system

Abstract

The invention discloses an automatic control system for a railway bridge static load bending test, which relates to the technical field of bridge engineering detection and comprises an intelligent sensing layer, a digital twin engine, an intelligent control layer, a trusted data management layer, a green energy supply layer and a man-machine interaction layer. The equipment health prediction model integrates oil temperature, vibration and current data to predict mechanical faults, performs early warning 30 minutes in advance, automatically switches to a slow loading mode when the crack propagation rate exceeds the limit, breaks through a passive protection mode of single threshold alarm, realizes multi-dimensional active protection from structural damage to equipment degradation, and drives finite element model updating through sensor data. The simulation prediction reverse correction loading strategy and the RBF neural network online correction material parameters are simulated and predicted, parameter iteration is triggered when the deviation is larger than a set value, the island problem that a traditional BIM only uploads data in a one-way mode is solved, closed-loop verification of test data and design parameters is achieved, and simulation precision is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of bridge engineering detection, in particular to an automatic control system for static load bending tests of railway bridges. Background Art

[0002] The static load bending test of a bridge is to examine the performance of a bridge structure under static loads, including strength, stiffness, and stability. By applying a static load to the bridge, it simulates the load conditions that may occur in actual use, and observes the deformation, stress distribution, and other responses of the bridge structure to evaluate the working condition and load-bearing capacity of the bridge.

[0003] Existing technologies use traditional fuzzy control algorithms, which are prone to oscillation when sudden loads are applied, resulting in synchronization deviations. They rely solely on threshold alarms and hardware emergency stops, lack multi-dimensional predictions of equipment status, and BIM integration only enables data upload, without establishing closed-loop verification of test data and design parameters. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an automatic control system for static load bending tests of railway bridges.

[0005] In order to achieve the above-mentioned objectives, the present invention adopts the following technical solutions: an automatic control system for static load bending tests of railway bridges, comprising an intelligent perception layer, a digital twin engine, an intelligent control layer, a trusted data management layer, a green energy supply layer and a human-computer interaction layer. The intelligent perception layer constructs a multi-source sensor array, and through the collaborative work of the multi-source sensor array and the edge computing node, all-round monitoring and rapid response of the bridge status are realized. The digital twin engine constructs a virtual-reality interactive closed-loop model based on real-time sensor data, dynamically corrects simulation parameters and generates optimization control instructions. The intelligent control layer realizes dynamic adjustment and multi-objective optimization of the loading process through a hierarchical decision-making mechanism and a composite control algorithm. The trusted data management layer adopts blockchain evidence storage and smart contract audit technology to ensure the non-tamperability of the test data and quantitative assessment of the carbon footprint. The green energy supply layer realizes energy self-sufficiency and low-carbon operation through photovoltaic-supercapacitor collaborative power supply and dynamic scheduling strategy. The human-computer interaction layer provides real-time data interaction and remote collaborative management functions through a three-dimensional visualization platform and augmented reality technology.

[0006] As a further description of the above technical solution:

[0007] The multi-source sensing array of the intelligent perception layer includes distributed fiber grating strain sensors, laser displacement sensors, industrial cameras, dual-function spoke sensors and infrared thermal imagers. The edge computing node preprocesses the data collected by the multi-source sensing array to generate a high-confidence data set.

[0008] As a further description of the above technical solution:

[0009] The digital twin engine imports the three-dimensional design model of the bridge into the finite element analysis software, converts it into a parametric model that can be dynamically simulated, defines the physical properties of the bridge material, sets load boundary conditions, and generates risk prediction and strategy correction instructions.

[0010] As a further description of the above technical solution:

[0011] The intelligent control layer obtains multi-dimensional physical quantities of the bridge structure in real time through a multi-source sensor array, combines edge computing nodes to perform real-time data processing and decision-making, and provides data support for subsequent control and verification.

[0012] As a further description of the above technical solution:

[0013] The trusted data management layer uses blockchain technology and carbon footprint management to achieve tamper-proof storage of test data and quantification of green energy supply benefits, record photovoltaic power generation, energy storage status and system power consumption in real time, and generate a carbon footprint assessment report.

[0014] As a further description of the above technical solution:

[0015] The green energy supply layer achieves energy self-sufficiency and low-carbon operation during the test process through photovoltaic-supercapacitor collaborative power supply technology. The high-efficiency photovoltaic components have high conversion efficiency, and the supercapacitor group prioritizes power supply for core sensors and supports emergency power supply.

[0016] As a further description of the above technical solution:

[0017] The human-computer interaction layer realizes multi-dimensional display and interactive analysis of test data through a three-dimensional visualization platform.

[0018] As a further description of the above technical solution:

[0019] The system process is as follows:

[0020] S1. Digital pre-test stage

[0021] S1.1 Parameter input and model initialization

[0022] Input bridge design parameters, generate a parametric finite element model through the digital twin engine, call the cloud-based historical case library, load historical test data and optimize the initial model parameters;

[0023] S1.2 Virtual Loading Optimization

[0024] Monte Carlo simulations are used to simulate multiple candidate loading paths, generating Pareto front solutions. Finite element analysis is then used to avoid high-risk areas, optimize the distribution of loading points, and select low-risk loading solutions, which are then updated synchronously to the intelligent control layer.

[0025] S2. Physical test stage

[0026] S2.1 Equipment Deployment and Calibration

[0027] Deploy a distributed sensor array, calibrate it to the set accuracy, install a multi-channel servo hydraulic actuator, and complete the pressure and displacement dual closed-loop control calibration;

[0028] S2.2 Gradual Loading and Dynamic Control

[0029] Loads are applied in stages according to the optimal loading scheme, with adjustable holding time for each stage. Edge computing nodes run the MPC-Fuzzy composite algorithm to dynamically adjust the hydraulic output curve. If the deviation between measured data and simulation data exceeds a threshold, the RBF neural network is triggered to correct the loading parameters online. When the AI vision algorithm detects that the crack propagation rate exceeds the limit, it switches to slow loading mode.

[0030] S2.3 Multimodal Data Synchronization

[0031] Sensor data drives the update of the digital twin model in real time, reconstructing the three-dimensional crack expansion path, and temperature field data corrects the material thermal expansion coefficient to improve simulation accuracy;

[0032] S3. Data closed loop stage

[0033] S3.1 Data Compression and Upload

[0034] The edge node compresses non-critical data and uploads it to the cloud via the MQTT protocol to train the LSTM model and optimize subsequent test parameters;

[0035] S3.2 Blockchain Evidence Storage and Compliance Verification

[0036] Key data generates a SHA-256 hash value and is written to the blockchain every 30 seconds. Smart contracts automatically match industry standards to verify test compliance. When data tampering is detected, the report is frozen and an alarm is sent to the supervisor.

[0037] S3.3 Carbon footprint assessment and report generation

[0038] Quantify the contribution of photovoltaic energy supply to emission reduction, generate carbon footprint assessment reports, and automatically generate multi-dimensional analysis charts;

[0039] S4. Green energy supply and emergency guarantee

[0040] PV panels prioritize powering edge nodes and core sensors, while supercapacitors provide emergency power. Even in grid-disconnected or nighttime environments, supercapacitors ensure continuous system operation for at least two hours.

[0041] S5. Human-computer interaction and remote collaboration

[0042] The 3D visualization platform uses Unity3D engine rendering to dynamically generate bridge deformation cloud maps, AR annotation of stress concentration areas, and supports gesture operation to zoom / rotate the model and view the 3D reconstruction results of cracks;

[0043] Remote monitoring: The supervisor accesses the blockchain evidence report in real time through the 5G network, decrypts the private key to verify the data integrity, pushes the test progress and safety warning information to the mobile terminal, and completes the electronic signature approval.

[0044] The present invention has the following beneficial effects:

[0045] 1. In the present invention, a full-life cycle closed-loop system of "intelligent perception → digital twin → intelligent control → trusted data management → green energy supply → human-computer interaction" is first constructed by integrating digital twins, multimodal data fusion, edge computing, blockchain and dynamic hybrid control algorithms. This architecture breaks through the limitations of the functional separation of traditional test systems, realizes the integrated coordination of data acquisition, strategy correction, risk warning and energy optimization, and significantly improves test efficiency and reliability. In terms of synchronous control and anti-disturbance, a composite control algorithm is used to integrate model predictive control and fuzzy PID-RBF neural network to dynamically compensate for nonlinear disturbances of the hydraulic system; high-precision synchronous adjustment is achieved through the dual closed-loop control mechanism of pressure and displacement. Based on the hierarchical decision-making mechanism, the cloud generates a global optimization strategy through the NSGA-III algorithm, the edge adjusts the loading parameters in real time, and the device executes instructions accurately. The above innovation solves the problems of traditional fuzzy control being prone to oscillation and insufficient synchronization accuracy under sudden loads, and ensures dynamic following accuracy and control stability under complex working conditions.

[0046] 2. In the present invention, AI vision and fiber optic sensors jointly reconstruct the three-dimensional expansion path of the crack. The equipment health prediction model integrates oil temperature, vibration, and current data to predict mechanical failures, and issues a 30-minute advance warning. When the crack expansion rate exceeds the limit, it automatically switches to slow loading mode, breaking through the passive protection mode of a single threshold alarm and achieving multi-dimensional active protection from structural damage to equipment degradation. Sensor data drives the update of the finite element model, simulation prediction reversely corrects the loading strategy, and the RBF neural network corrects the material parameters online. When the deviation is greater than the set value, parameter iteration is triggered at all times, solving the island problem of traditional BIM that only uploads data in one direction, realizing closed-loop verification of test data and design parameters, and improving simulation accuracy.

[0047] 3. In the present invention, blockchain evidence storage is integrated with green energy supply, taking into account both data credibility and the low-carbon requirements of the experiment. The photovoltaic-supercapacitor collaborative power supply technology is used to achieve energy self-sufficiency and low-carbon operation in the test process, automatically generate emission reduction quantitative reports, support environmental compliance audits, and use the Unity3D engine to superimpose stress cloud maps on real bridges, mark high-risk areas, trace back to original data from visual reports, associate blockchain hash values, coordinate the division of labor between PC and mobile terminals, process large-scale simulation data on the PC terminal, generate global optimization strategies, configure test parameters and safety thresholds, and focus on real-time monitoring, report approval, and on-site AR annotation on the mobile terminal. It pushes emergency alarm information, realizes multi-terminal collaborative management, breaks the interactive limitations of traditional two-dimensional reports, and realizes multi-dimensional, full-link data transparency management. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a system architecture diagram of the present invention;

[0049] Figure 2 Flowchart of the present invention. DETAILED DESCRIPTION

[0050] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0051] Reference Figure 1-2 , the present invention provides an embodiment: an automatic control system for static load bending test of railway bridges, including an intelligent perception layer, a digital twin engine, an intelligent control layer, a trusted data management layer, a green energy supply layer and a human-computer interaction layer. The intelligent perception layer constructs a multi-source sensor array, and through the collaborative work of the multi-source sensor array and the edge computing node, it realizes all-round monitoring and rapid response of the bridge status. The digital twin engine constructs a virtual-reality interactive closed-loop model based on real-time sensor data, dynamically corrects simulation parameters and generates optimization control instructions. The intelligent control layer realizes dynamic adjustment and multi-objective optimization of the loading process through a hierarchical decision-making mechanism and a composite control algorithm. The trusted data management layer adopts blockchain evidence storage and smart contract audit technology to ensure the non-tamperability of test data and quantitative assessment of carbon footprint. The green energy supply layer realizes energy self-sufficiency and low-carbon operation through photovoltaic-supercapacitor collaborative power supply and dynamic scheduling strategy. The human-computer interaction layer provides real-time data interaction and remote collaborative management functions through a three-dimensional visualization platform and augmented reality technology.

[0052] The multi-source sensor array of the intelligent perception layer includes distributed fiber Bragg grating strain sensors, laser displacement sensors, industrial cameras, dual-function spoke sensors and infrared thermal imagers. The edge computing node pre-processes the data collected by the multi-source sensor array to generate a high-confidence data set. The digital twin engine imports the three-dimensional design model of the bridge into the finite element analysis software and converts it into a parametric model that can be dynamically simulated. It defines the physical properties of the bridge material, sets load boundary conditions, and generates risk prediction and strategy correction instructions. The intelligent control layer obtains the multi-dimensional physical quantities of the bridge structure in real time through the multi-source sensor array, and combines the edge computing node for real-time Data processing and decision-making provide data support for subsequent control and verification. The trusted data management layer uses blockchain technology and carbon footprint management to achieve tamper-proof storage of test data and quantification of green energy supply benefits, record photovoltaic power generation, energy storage status and system power consumption in real time, and generate a carbon footprint assessment report. The green energy supply layer uses photovoltaic-supercapacitor collaborative power supply technology to achieve energy self-sufficiency and low-carbon operation during the test process. High-efficiency photovoltaic modules have high conversion efficiency. Supercapacitor groups prioritize power supply for core sensors and support emergency power supply. The human-computer interaction layer uses a three-dimensional visualization platform to achieve multi-dimensional display and interactive analysis of test data.

[0053] The intelligent perception layer realizes all-round monitoring and rapid response of bridge status through the collaborative work of multi-source sensor arrays and edge computing nodes, deploys various high-precision sensors to form sensor arrays, and builds a distributed sensor network. Based on the high-precision data acquisition of distributed sensors and the real-time decision-making capabilities of edge computing nodes, its workflow follows the "perception-processing-decision-making" closed-loop mechanism to ensure high-precision synchronous control and dynamic response of the test process. At the same time, it supports green energy supply and data trustworthy evidence to ensure the safety of the test process and data reliability, and provides high-precision, low-latency, and highly reliable monitoring and control capabilities for bridge static load tests. Distributed fiber Bragg grating strain sensors are distributed in key sections of the bridge to monitor the strain distribution of key parts of the bridge and capture local Stress concentration phenomenon, providing high-precision micro-strain measurement data, can resist electromagnetic interference, suitable for long-term stable monitoring, laser displacement sensor non-contact measurement of bridge deflection changes, evaluate the overall structural stiffness, support high-resolution displacement feedback, real-time tracking of beam deformation trend, high-frame rate industrial camera captures bridge surface image, supports dynamic crack tracking, combined with AI visual algorithm to identify crack morphology and expansion trend, dual-function spoke sensor real-time feedback hydraulic jack load value and displacement, ensure the synchronization accuracy of loading control, integrated pressure and displacement dual signal output, support closed-loop control, infrared thermal imager monitors bridge temperature field distribution, identifies local temperature difference, corrects the influence of material thermal expansion effect on test results, all sensors pass IEEE1588 protocol It is proposed to synchronize the clock source to achieve millisecond-level synchronous collection of bridge strain, deflection, temperature field, surface deformation and hydraulic load data to ensure the temporal and spatial consistency of the data. The collected raw data is transmitted to the edge computing node via the CAN bus or industrial Ethernet, and the industrial camera image is subjected to denoising, contrast enhancement and distortion correction to improve the accuracy of crack detection. The edge computing node deploys a lightweight YOLOv7 model to perform real-time analysis of the industrial camera image. The model accelerates reasoning through the FPGA chip to detect defects where the crack width is greater than the set value, and classifies the crack type, optimizes the model parameters, adapts to the edge computing power limit, and adopts a composite algorithm of model predictive control and fuzzy logic to integrate strain, deflection and temperature data. The noise is eliminated through Kalman filtering to generate It forms a high-confidence data set, combines binocular vision and AI algorithms to generate a three-dimensional expansion path of the crack, analyzes its impact on structural safety, and thus predicts the load-deflection relationship in the next 5 seconds. Based on the model predictive control and fuzzy logic composite algorithm, it dynamically adjusts the hydraulic system output curve in combination with real-time data, and triggers the loading rate adjustment instruction by the fuzzy rule library. When an abnormal response is triggered, an emergency stop instruction is directly issued to the hydraulic system. The edge node cache pre-loading strategy continues to execute the current loading stage task when the network is interrupted. The supercapacitor group provides at least 30 minutes of emergency power supply to ensure the operation of the core sensors and computing units. The sensor data is input into the digital twin model through the OPCUA protocol, and the boundary conditions of the finite element simulation are updated. When a high-risk working condition is detected,The intelligent control layer triggers graded load reduction protection, and pushes early warnings to the supervision platform via the 5G network. The edge node dynamically switches between high-performance and energy-saving modes of the FPGA chip according to the load. The hydraulic system enters a low-power standby state when idle, and photovoltaic panels are used to power sensors first. Supercapacitors provide emergency power on cloudy days. Sensor status is regularly self-checked, and alarms are triggered and redundant sensors are switched when abnormalities occur.

[0054] The digital twin engine imports the three-dimensional design model of the bridge into the finite element analysis software and converts it into a parametric model that can be dynamically simulated. It defines the physical properties of the bridge material, sets load boundary conditions, installs sensors on the test bridge, collects data such as deformation, temperature, and cracks of the bridge in real time, transmits the sensor data to the virtual model through the communication protocol, updates the simulation boundary conditions, supports dynamic correction of material parameters, ensures the consistency of the model with the physical test, and thus constructs a digital twin model of the experimental bridge, realizes real-time two-way mapping between physical tests and virtual models, and realizes dynamic verification and strategy correction of the test process. Its core mechanism is the "virtual-reality interactive closed loop". Real-time sensor data is input into the digital twin model through the communication protocol, and the boundary conditions of the finite element analysis are dynamically updated. The simulation model predicts the stress distribution and deformation trend of the bridge, generates risk prediction and strategy correction instructions, and triggers model parameter self-optimization and loading path adjustment when the deviation between the measured data and the simulation results exceeds the threshold. Reversely optimize physical test parameters to improve test safety, efficiency, and accuracy. Use AI visual algorithms to analyze industrial camera images, construct a three-dimensional crack expansion path, identify crack morphology and expansion trends, and combine fiber grating sensor data to achieve multimodal fusion analysis of crack evolution. Generate multiple sets of candidate loading paths through Monte Carlo path simulation, screen low-risk solutions through finite element analysis, and optimize the distribution of loading points by combining with historical case libraries. Generate an optimal strategy set based on multiple objectives such as test duration, energy consumption, and loading accuracy through the Pareto front solution set, and dynamically call it from the controller. Sensor data is input into the digital twin model through a protocol, and the boundary conditions of the finite element analysis are updated. Industrial camera and fiber optic sensor data drive the three-dimensional reconstruction of the crack, with analysis accuracy reaching sub-millimeter level. The simulation model predicts future stress distribution. If the deviation from the measured value exceeds the threshold, the neural network is triggered to correct the loading force online, dynamically adjust the control parameters, update the Monte Carlo solution set, and avoid high-risk loading intervals.

[0055] The intelligent control layer realizes dynamic adjustment and multi-objective optimization of the loading process through a hierarchical decision-making mechanism. Its core mechanism is the three-level collaboration of "cloud global optimization-edge real-time control-equipment precise execution", combined with compound control algorithms and health prediction models to ensure high precision, high efficiency and high reliability of the test. The cloud generates long-term test strategies based on historical data and multi-objective optimization algorithms. The edge combines sensor feedback and compound control algorithms to correct loading parameters in real time. The adaptive hydraulic loading system executes loading instructions, and dual closed-loop control ensures synchronization accuracy. The equipment health model warns of mechanical failures in advance and automatically switches redundant channels. The adaptive hydraulic loading system achieves precise loading control through multi-channel servo hydraulic actuators, and dynamically adjusts the loading strategy in combination with compound control algorithms. The channel servo hydraulic actuator independently adjusts the hydraulic output of each loading point, supports asymmetric loading mode, adopts pressure and displacement dual closed-loop control to ensure synchronization accuracy, integrates model predictive control and fuzzy PID-RBF neural network, predicts future loading trends and dynamically compensates for nonlinear disturbances, receives feedback data from edge computing nodes in real time, adjusts the hydraulic output curve, builds an equipment health prediction model, and installs high-precision sensors at key parts of the hydraulic system, including oil temperature sensors to monitor hydraulic oil temperature changes, vibration sensors to detect mechanical component vibration spectra, current sensors to collect servo motor drive current waveforms, and pressure sensors to record hydraulic pipeline pressure values in real time. Time synchronization of multi-sensor data is achieved through CAN bus or industrial Ethernet, and the mean value is extracted. Variance, peak factor and other statistics are extracted through fast Fourier transform to extract the main frequency component of the vibration spectrum, build a multivariate time series data set of oil temperature-pressure-vibration, collect typical fault cases of hydraulic system, mark the fault type and occurrence time, combine maintenance records and sensor data, build a fault feature label library, use long short-term memory network to train equipment health model, input multimodal time series data, output the remaining service life prediction value, configure spare hydraulic channel, automatically switch to the spare system when the main channel fails, use dual power supply and signal redundant transmission to ensure seamless switching process, integrate fault diagnosis and control command interlocking logic to ensure smooth transition of loading force during switching, dynamically adjust control algorithm parameters, compensate for system disturbances caused by switching, collect real-time The fault data is uploaded to the cloud knowledge base, the long short-term memory network is retrained regularly, and a federated learning framework is adopted to optimize the model by combining data from multiple devices under the premise of protecting privacy. Simulated fault data is injected into the digital twin engine to verify the accuracy of warning and switching response time. The robustness of the model under different working conditions is evaluated through Monte Carlo simulation. The weight distribution of key fault features is enhanced through the attention mechanism. When the life prediction value is lower than the set threshold, a maintenance reminder is triggered. When an abnormal vibration mutation or oil temperature exceeds the limit is detected, the redundant channel switching is triggered immediately. The hydraulic system status is monitored in real time, and potential faults are warned in advance. Multimodal data such as oil temperature, vibration, and current are integrated to build a fault feature library. The equipment degradation trend is predicted through the equipment health prediction model. When an abnormality is detected,Automatically switching to redundant hydraulic channels ensures test continuity. The multi-objective optimization controller balances test duration, energy consumption, and loading accuracy to generate a global optimal strategy. The NSGA-III algorithm, based on Pareto frontier theory, selects the multi-objective optimal strategy from the candidate solution set. The cloud-based historical case library is used to optimize initial parameters. The simulation prediction results of the digital twin engine are received, and the loading path is dynamically adjusted to avoid high-risk areas. Combined with real-time sensor feedback, the control parameters are iteratively optimized. The NSGA-III algorithm generates a long-term test strategy and sends it to the edge via the 5G network.

[0056] The trusted data management layer ensures the immutability and traceability of test data through blockchain evidence storage and smart contract auditing technology, and combines carbon footprint management to achieve quantitative evaluation of environmental benefits. It receives original test data from the intelligent perception layer, screens key indicators, and the edge computing nodes screen the key test data in real time and package them into data blocks. The data blocks are doubly encrypted using quantum random number encryption technology to ensure resistance to quantum cracking. Through the Hyperledger Fabric framework, the encrypted data blocks are written to the blockchain every 30 seconds. The smart contract compares the test data with the preset standards to verify the compliance of the test process. If data tampering is detected, the test report is immediately frozen and the supervisor is notified. The photovoltaic power generation, supercapacitor energy storage status and system power consumption are recorded in real time, and a "Carbon Footprint Assessment Report" is automatically generated to quantify the reduction in carbon emissions during system operation.

[0057] The green energy supply layer achieves energy self-sufficiency and low-carbon operation during the test process through photovoltaic-supercapacitor coordinated power supply technology. Its core mechanism is a three-level coordination of "dynamic energy supply scheduling-energy storage buffering-energy consumption optimization". It combines photovoltaic power generation and supercapacitor energy storage to ensure continuous power supply for key equipment and reduce carbon emissions. The photovoltaic power generation module converts solar energy into electricity, providing clean energy for the system. It uses high-efficiency thin-film photovoltaic panels to support power generation in low-light environments. The photovoltaic panels can rotate to track the sun, maximize light reception efficiency, and match energy supply with system demand in real time. They prioritize power supply to edge computing nodes, core sensors, and hydraulic control systems. Dynamically switch power supply modes through load monitoring to ensure the operational stability of key equipment, reduce ineffective system energy consumption, and improve energy utilization. The hydraulic system automatically enters a low-power standby state when idle. The edge computing node adjusts the operating frequency of the FPGA chip according to the real-time task volume to reduce standby power consumption. The emission reduction benefits of photovoltaic energy supply are quantified, and an environmental assessment report is generated. The total energy consumption and equivalent carbon dioxide emissions of the test process are calculated and compared with the traditional mains power supply model to analyze the emission reduction contribution of photovoltaic energy supply. The supercapacitor energy storage module stores surplus energy from photovoltaic power generation and provides response. Emergency power support: A dynamic scheduling algorithm allocates energy storage capacity based on load demand. During grid outages or at night, at least two hours of emergency power is provided to ensure continuous test execution. The emission reduction benefits of PV power supply are quantified, and an environmental assessment report is generated. The total energy consumption and equivalent CO2 emissions of the test process are calculated. The emission reduction contribution of PV power supply is analyzed by comparing it with traditional mains power supply. During daytime hours with ample sunlight, PV panels directly power the system and store excess energy in supercapacitor banks. On cloudy days or at night, the supercapacitor banks release stored energy to prioritize the operation of core equipment. The system monitors the real-time power consumption of the hydraulic system and sensor array, dynamically allocating the output ratio of PV and supercapacitors. When a sudden high load is detected, the supercapacitor bank instantly supplements the peak power demand. PV power generation, supercapacitor energy storage status, and system power consumption data are synchronized in real time with the trusted data management layer. A carbon footprint assessment report is automatically generated, documenting emission reduction results and energy consumption optimization recommendations. The green energy supply layer, through PV-supercapacitor coordinated power supply and dynamic scheduling strategies, addresses the mains-based power dependence and low energy efficiency of traditional test systems. Its modular design not only ensures the continuity of the test process but also promotes the green transformation of bridge inspection technology.

[0058] The human-computer interaction layer realizes multi-dimensional display and interactive analysis of test data through a three-dimensional visualization platform. Its core mechanism is the trinity of "real-time rendering-augmented reality annotation-data penetration". Combining 5G communication and blockchain technology, it provides test personnel and supervisors with an intuitive and reliable interactive experience. The digital twin model of the bridge is displayed through the three-dimensional visualization platform. Based on finite element simulation data and sensor feedback, the three-dimensional visualization platform dynamically generates bridge deformation cloud maps through Unity3D engine rendering, renders bridge structure deformation and stress distribution in real time, marks stress concentration areas with color gradients, and superimposes virtual information such as stress cloud maps and crack extension paths onto real bridge images. It supports real-time AR display, marks high-risk areas, and assists on-site personnel in quickly locating hidden dangers. It has AR interaction functions and can realize model interaction through gestures or mobile device operations. With data penetration query, gesture operations to zoom and rotate the bridge model, view local details, click on specific areas to retrieve related data, trace back from the visual report to the original sensor data, verify data credibility, associate the blockchain evidence number, view the hash value and timestamp information with one click, realize remote monitoring and report management through 5G network, remotely review the test progress, loading curve and safety warning status, receive blockchain evidence report push, support private key decryption to verify data integrity, automatically generate multi-dimensional analysis charts, the supervisor signs the electronic signature through the mobile terminal, completes the test compliance approval, coordinates the division of labor between PC and mobile terminals, the PC terminal processes large-scale simulation data, generates global optimization strategies, configures test parameters and safety thresholds, the mobile terminal focuses on real-time monitoring, report approval and on-site AR annotation, pushes emergency alarm information, and realizes multi-terminal collaborative management.

[0059] The system process is as follows:

[0060] S1. Digital pre-test stage

[0061] S1.1 Parameter input and model initialization

[0062] Input bridge design parameters, generate a parametric finite element model through the digital twin engine, call the cloud-based historical case library, load historical test data and optimize the initial model parameters;

[0063] S1.2 Virtual Loading Optimization

[0064] Monte Carlo simulations are used to simulate multiple candidate loading paths, generating Pareto front solutions. Finite element analysis is then used to avoid high-risk areas, optimize the distribution of loading points, and select low-risk loading solutions, which are then updated synchronously to the intelligent control layer.

[0065] S2. Physical test stage

[0066] S2.1 Equipment Deployment and Calibration

[0067] Deploy a distributed sensor array, calibrate it to the set accuracy, install a multi-channel servo hydraulic actuator, and complete the pressure and displacement dual closed-loop control calibration;

[0068] S2.2 Gradual Loading and Dynamic Control

[0069] Loads are applied in stages according to the optimal loading scheme, with adjustable holding time for each stage. Edge computing nodes run the MPC-Fuzzy composite algorithm to dynamically adjust the hydraulic output curve. If the deviation between measured data and simulation data exceeds a threshold, the RBF neural network is triggered to correct the loading parameters online. When the AI vision algorithm detects that the crack propagation rate exceeds the limit, it switches to slow loading mode.

[0070] S2.3 Multimodal Data Synchronization

[0071] Sensor data drives the update of the digital twin model in real time, reconstructing the three-dimensional crack expansion path, and temperature field data corrects the material thermal expansion coefficient to improve simulation accuracy;

[0072] S3. Data closed loop stage

[0073] S3.1 Data Compression and Upload

[0074] The edge node compresses non-critical data and uploads it to the cloud via the MQTT protocol to train the LSTM model and optimize subsequent test parameters;

[0075] S3.2 Blockchain Evidence Storage and Compliance Verification

[0076] Key data generates a SHA-256 hash value and is written to the blockchain every 30 seconds. Smart contracts automatically match industry standards to verify test compliance. When data tampering is detected, the report is frozen and an alarm is sent to the supervisor.

[0077] S3.3 Carbon footprint assessment and report generation

[0078] Quantify the contribution of photovoltaic energy supply to emission reduction, generate carbon footprint assessment reports, and automatically generate multi-dimensional analysis charts;

[0079] S4. Green energy supply and emergency guarantee

[0080] PV panels prioritize powering edge nodes and core sensors, while supercapacitors provide emergency power. Even in grid-connected environments or at night, supercapacitors ensure continuous system operation for at least two hours.

[0081] S5. Human-computer interaction and remote collaboration

[0082] The 3D visualization platform uses Unity3D engine rendering to dynamically generate bridge deformation cloud maps, AR annotation of stress concentration areas, and supports gesture operation to zoom / rotate the model and view the 3D reconstruction results of cracks;

[0083] Remote monitoring: The supervisor accesses the blockchain evidence report in real time through the 5G network, decrypts the private key to verify the data integrity, pushes the test progress and safety warning information to the mobile terminal, and completes the electronic signature approval.

[0084] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An automatic control system for static load bending test of railway bridges, characterized by: It includes an intelligent perception layer, a digital twin engine, an intelligent control layer, a trusted data management layer, a green energy supply layer and a human-computer interaction layer. The intelligent perception layer constructs a multi-source sensor array, and through the collaborative work of the multi-source sensor array and the edge computing nodes, it realizes all-round monitoring and rapid response of the bridge status. The digital twin engine constructs a virtual-reality interactive closed-loop model based on real-time sensor data, dynamically corrects simulation parameters and generates optimization control instructions. The intelligent control layer realizes dynamic adjustment and multi-objective optimization of the loading process through a hierarchical decision-making mechanism and a composite control algorithm. The trusted data management layer adopts blockchain evidence storage and smart contract audit technology to ensure the non-tamperability of the test data and the quantitative assessment of the carbon footprint. The green energy supply layer realizes energy self-sufficiency and low-carbon operation through photovoltaic-supercapacitor collaborative power supply and dynamic scheduling strategy. The human-computer interaction layer provides real-time data interaction and remote collaborative management functions through a three-dimensional visualization platform and augmented reality technology.

2. The automatic control system for static load bending test of railway bridge according to claim 1, characterized in that: The multi-source sensing array of the intelligent perception layer includes distributed fiber grating strain sensors, laser displacement sensors, industrial cameras, dual-function spoke sensors and infrared thermal imagers. The edge computing node preprocesses the data collected by the multi-source sensing array to generate a high-confidence data set.

3. The automatic control system for static load bending test of railway bridge according to claim 1, characterized in that: The digital twin engine imports the three-dimensional design model of the bridge into the finite element analysis software, converts it into a parametric model that can be dynamically simulated, defines the physical properties of the bridge material, sets load boundary conditions, and generates risk prediction and strategy correction instructions.

4. The automatic control system for static load bending test of railway bridges according to claim 1, characterized in that: The intelligent control layer obtains multi-dimensional physical quantities of the bridge structure in real time through a multi-source sensor array, combines edge computing nodes to perform real-time data processing and decision-making, and provides data support for subsequent control and verification.

5. The automatic control system for static load bending test of railway bridge according to claim 1, characterized in that: The trusted data management layer uses blockchain technology and carbon footprint management to achieve tamper-proof storage of test data and quantification of green energy supply benefits, record photovoltaic power generation, energy storage status and system power consumption in real time, and generate a carbon footprint assessment report.

6. The automatic control system for static load bending test of railway bridge according to claim 1, characterized in that: The green energy supply layer achieves energy self-sufficiency and low-carbon operation during the test process through photovoltaic-supercapacitor collaborative power supply technology. The high-efficiency photovoltaic components have high conversion efficiency, and the supercapacitor group prioritizes power supply for core sensors and supports emergency power supply.

7. The automatic control system for static load bending test of railway bridges according to claim 1, characterized in that: The human-computer interaction layer realizes multi-dimensional display and interactive analysis of test data through a three-dimensional visualization platform.

8. The automatic control system for static load bending test of railway bridge according to claim 1, characterized in that: The system process is as follows: S1. Digital pre-test stage S1.1 Parameter input and model initialization Input bridge design parameters, generate a parametric finite element model through the digital twin engine, call the cloud-based historical case library, load historical test data and optimize the initial model parameters; S1.2 Virtual Loading Optimization Monte Carlo simulations are used to simulate multiple candidate loading paths, generating Pareto front solutions. Finite element analysis is then used to avoid high-risk areas, optimize the distribution of loading points, and select low-risk loading solutions, which are then updated synchronously to the intelligent control layer. S2. Physical test stage S2.1 Equipment Deployment and Calibration Deploy a distributed sensor array, calibrate it to the set accuracy, install a multi-channel servo hydraulic actuator, and complete the pressure and displacement dual closed-loop control calibration; S2.2 Gradual Loading and Dynamic Control Loads are applied in stages according to the optimal loading scheme, with adjustable holding time for each stage. Edge computing nodes run the MPC-Fuzzy composite algorithm to dynamically adjust the hydraulic output curve. If the deviation between measured data and simulation data exceeds a threshold, the RBF neural network is triggered to correct the loading parameters online. When the AI vision algorithm detects that the crack propagation rate exceeds the limit, it switches to slow loading mode. S2.3 Multimodal Data Synchronization Sensor data drives the update of the digital twin model in real time, reconstructing the three-dimensional crack expansion path, and temperature field data corrects the material thermal expansion coefficient to improve simulation accuracy; S3. Data closed loop stage S3.1 Data Compression and Upload The edge node compresses non-critical data and uploads it to the cloud via the MQTT protocol to train the LSTM model and optimize subsequent test parameters; S3.2 Blockchain Evidence Storage and Compliance Verification Key data generates a SHA-256 hash value and is written to the blockchain every 30 seconds. Smart contracts automatically match industry standards to verify test compliance. When data tampering is detected, the report is frozen and an alarm is sent to the supervisor. S3.3 Carbon footprint assessment and report generation Quantify the contribution of photovoltaic energy supply to emission reduction, generate carbon footprint assessment reports, and automatically generate multi-dimensional analysis charts; S4. Green energy supply and emergency guarantee PV panels prioritize powering edge nodes and core sensors, while supercapacitors provide emergency power. Even in grid-disconnected or nighttime environments, supercapacitors ensure continuous system operation for at least two hours. S5. Human-computer interaction and remote collaboration The 3D visualization platform uses Unity3D engine rendering to dynamically generate bridge deformation cloud maps, AR annotation of stress concentration areas, and supports gesture operation to zoom / rotate the model and view the 3D reconstruction results of cracks; Remote monitoring: The supervisor accesses the blockchain evidence report in real time through the 5G network, decrypts the private key to verify the data integrity, pushes the test progress and safety warning information to the mobile terminal, and completes the electronic signature approval.

Images