Railway steel truss bridge bolted joint full life cycle fatigue sensing evaluation method

Abstract

This invention discloses a method for full life-cycle fatigue perception assessment of bolted joints in railway steel truss bridges, belonging to the field of rail transit facility structural monitoring technology. The method includes: S10: acquiring real-time strain data deployed on the physical nodes of the bolted joints; S20: calling the bolted joint feature mapping relationship to generate virtual node real-time strain data; S30: constructing a physical constraint framework based on the basic control equations of elasticity, and solving for the stress distribution that satisfies the known conditions and minimizes the total strain energy under the physical constraint framework, obtaining a dynamic full-field stress cloud map; S40: the fatigue assessment module identifies real-time variable stress data in stress concentration areas based on the dynamic full-field stress cloud map, and performs fatigue damage analysis and life assessment of the bolted joints based on the real-time variable stress data. This significantly improves the accuracy of full life-cycle fatigue assessment of bolted joints in railway steel truss bridges, realizing full-field stress reconstruction and early micro-damage warning based on real loads.

Description

Technical Field

[0001] This invention relates to the field of monitoring technology for rail transit facility structures, specifically to a method for fatigue perception assessment of bolted joints of railway steel truss bridges throughout their entire life cycle. Background Technology

[0002] Railway steel truss bridges are key load-bearing structures in modern railway transportation systems, and their safety is directly related to train operation safety. Bolted joints, as the main connection method in steel truss bridges, are prone to fatigue damage under the long-term action of alternating train loads. Stress concentrations in critical areas such as bolt hole walls and the inside of bolts are high-risk areas for the initiation and propagation of fatigue cracks.

[0003] Currently, fatigue damage detection of bolted joints in railway bridges mainly relies on real-time monitoring using sensors such as displacement gauges and strain gauges deployed at key locations on the bridge. However, steel truss bridges have complex structures, large-scale wiring is costly, and sensors are susceptible to interference from the strong electromagnetic environment of railways, making it difficult to guarantee the stability and reliability of data acquisition. Furthermore, existing technologies for fatigue assessment typically employ either periodic manual inspections or threshold alarm mechanisms based on a limited number of sensors. The former relies on human experience and is difficult to detect early, hidden damage; the latter only reflects the stress state of local measuring points and cannot reconstruct the stress distribution across the entire bolted joint area, lacking the ability to identify deep evolutionary characteristics such as internal plastic accumulation and microcrack initiation. In addition, static simulation models cannot automatically correct themselves as the structure ages and damage accumulates, leading to a decrease in prediction accuracy over service time. Therefore, there is an urgent need for a fatigue sensing and assessment method that can integrate measured surface strain with mechanical theory, eliminate blind spots in internal monitoring, achieve dynamic reconstruction of stress across the entire field, and possess self-evolving capabilities, to support accurate operation and maintenance and risk warning throughout the entire lifecycle of bolted joints in railway steel truss bridges. Summary of the Invention

[0004] To address the problems mentioned in the background section, this invention provides a method for full life-cycle fatigue perception assessment of bolted joints in railway steel truss bridges. The technical solution adopted by this invention is as follows:

[0005] A method for fatigue perception assessment of bolted joints in railway steel truss bridges throughout their entire life cycle includes the following steps:

[0006] S10: Using the data acquisition module deployed on the physical node of the bolted joint, acquire real-time strain data deployed on the physical node of the bolted joint;

[0007] S20: Call a bolt point feature mapping relationship from the virtual-real mapping module to describe the strain transfer association between the physical node and the virtual node inside the bolt point, map the real-time strain data to the virtual node, and generate real-time strain data of the virtual node;

[0008] S30: The field reconstruction module constructs a physical constraint framework based on the basic control equations of elasticity, and inputs the real-time strain data and the real-time strain data of the virtual nodes as known conditions into the physical constraint framework. The stress distribution that satisfies the known conditions and minimizes the total strain energy under the physical constraint framework is solved to obtain a dynamic full-field stress cloud map.

[0009] S40: The fatigue assessment module identifies real-time variable stress data in the stress concentration area based on the dynamic full-field stress cloud map, and performs fatigue damage analysis and life assessment on the bolted joint based on the real-time variable stress data.

[0010] In a preferred embodiment, this application can be further configured such that step S10, which involves acquiring real-time strain data deployed on the physical node of the bolted joint, includes the following steps:

[0011] S101: Based on the bolted joint feature mapping relationship and the real-time strain data of adjacent physical nodes, calculate the theoretical strain value of the target physical node;

[0012] S102: Compare the measured real-time strain data of the target physical node with the theoretical strain value to generate a data reliability index;

[0013] S103: When the data reliability index is lower than the second threshold, trigger the data anomaly verification or data reconstruction process for the physical node.

[0014] In a preferred embodiment, this application can be further configured such that, in step S20, the virtual-real mapping module invokes the bolt point feature mapping relationship, including the following steps:

[0015] S201: Generate the feature mapping relationship of the bolted joint, establish a three-dimensional parametric finite element model of the bolted joint, and set the virtual node at the mechanically sensitive position inside the bolted joint;

[0016] S202: Perform multi-condition mechanical simulation on the three-dimensional parametric finite element model to extract the dynamic response characteristics of the bolted joint under various loads;

[0017] S203: Perform dimensionality reduction processing on the dynamic response features to generate the bolt point feature mapping relationship.

[0018] In a preferred embodiment, this application can be further configured such that, in step S30, obtaining the dynamic full-field stress cloud map includes the following steps:

[0019] S301: Using a three-dimensional parametric finite element model representing the healthy state, simulation calculations are performed under the same load assumptions as the current working conditions to generate a reference stress cloud diagram;

[0020] S302: Compare the dynamic full-field stress cloud map with the reference stress cloud map to generate virtual-real deviation data;

[0021] S303: When the virtual-real deviation data exceeds the first threshold, mark the spatial location corresponding to the virtual-real deviation data as an abnormal interest area, and trace the abnormal state of the abnormal interest area.

[0022] In a preferred embodiment, this application can be further configured such that step S303, the step of tracing the abnormal state of the abnormal interest area, includes the following steps:

[0023] S3031: Extract the dynamic stress field pattern of the abnormal interest region and generate the current load fingerprint;

[0024] S3032: Match the current load fingerprint with a load fingerprint database that stores characteristic stress field patterns corresponding to various known events to obtain a matching result;

[0025] S3033: Based on the matching results, determine the possible root cause of the abnormal state.

[0026] In a preferred embodiment, this application can be further configured such that, in step S40, the fatigue assessment module identifies real-time variable stress data of stress concentration regions based on the dynamic full-field stress cloud map, including the following steps:

[0027] S4011: Extract damage precursor characteristic parameters from the dynamic full-field stress cloud map, wherein the damage precursor characteristic parameters include stress gradient change rate or principal stress direction rotation mode.

[0028] S4012: Generate an early micro-damage warning signal based on the evolution trend of the damage precursor characteristic parameters.

[0029] In a preferred embodiment, this application can be further configured as follows: Step S40, the step of performing fatigue damage analysis and life assessment on the bolted joint based on the real-time alternating stress data, includes the following steps:

[0030] S4021: Perform cyclic counting and statistics on the real-time variable stress data to obtain the stress cycle spectrum;

[0031] S4022: Based on the damage accumulation model and the stress cycle spectrum, calculate the real-time damage increment and the current cumulative damage degree of the bolted joint;

[0032] S4023: Based on the current cumulative damage level and damage evolution trend, predict the remaining fatigue life of the bolted joint and generate a remaining fatigue life prediction result;

[0033] S4024: The early micro-damage warning signal is fused with the remaining fatigue life prediction result to generate a probabilistic life interval prediction.

[0034] In a preferred embodiment, this application may be further configured to include, after step S40, the following step:

[0035] S501: Determine the risk level of the bolted joint based on the current cumulative damage, remaining fatigue life prediction results, and probabilistic life interval prediction;

[0036] S502: When the risk level is the first risk level, a routine inspection record is generated;

[0037] S503: When the risk level is the second risk level, the enhanced monitoring mode is activated;

[0038] S504: When the risk level is the third risk level, generate a precise maintenance work order that includes the coordinates of the abnormal location and the suspected damage mechanism.

[0039] In a preferred embodiment, this application may be further configured to include, after step S3033, the following step:

[0040] S601: When the possible root cause of the abnormal state is determined to be the evolution of structural damage, the parameters of the three-dimensional parametric finite element model representing the healthy state are iteratively optimized using the virtual-real deviation data to generate a corrected three-dimensional parametric finite element model.

[0041] S602: Based on the corrected three-dimensional parametric finite element model, update the bolt joint feature mapping relationship.

[0042] A fatigue perception assessment system for the entire life cycle of bolted joints in railway steel truss bridges, used to execute a fatigue perception assessment method for the entire life cycle of bolted joints in railway steel truss bridges, comprising:

[0043] The data acquisition module is used to acquire real-time strain data deployed on the physical nodes of the bolted joints;

[0044] The virtual-real mapping module is used to call a bolt point feature mapping relationship that describes the strain transfer relationship between the physical node and the virtual node inside the bolt point, and to map the real-time strain data to the virtual node to generate real-time strain data of the virtual node.

[0045] The field reconstruction module is used to construct a physical constraint framework based on the basic control equations of elasticity, and input the real-time strain data and the real-time strain data of the virtual nodes as known conditions into the physical constraint framework to solve the stress distribution that satisfies the known conditions and minimizes the total strain energy under the physical constraint framework, thereby obtaining a dynamic full-field stress cloud map.

[0046] The fatigue assessment module is used to identify real-time variable stress data of stress concentration areas based on the dynamic full-field stress cloud map, and to perform fatigue damage analysis and life assessment of the bolted joint based on the real-time variable stress data.

[0047] The beneficial effects of the present invention, a method for full life-cycle fatigue perception assessment of bolted joints in railway steel truss bridges, are as follows: First, real-time strain data is collected using physical nodes deployed on the surface. A virtual-real mapping module is used to extrapolate surface strain to virtual nodes inside bolts and hole walls, which cannot be directly monitored, eliminating monitoring blind spots. Then, a physical constraint framework is constructed based on the elasticity control equations. Using measured and extrapolated strains as known conditions, the stress distribution that minimizes total strain energy is solved, reconstructing a dynamic full-field stress cloud map. Furthermore, damage precursor characteristic parameters are extracted from the stress cloud map to generate early micro-damage warning signals. Simultaneously, cyclic counting and damage accumulation calculations are performed on the variable stress data in stress concentration areas to obtain the cumulative damage. The system calculates damage degree and remaining fatigue life prediction results, and integrates them with early warning signals to output probabilistic life interval predictions. Then, based on the cumulative damage degree, remaining life, and probability intervals, a risk level is comprehensively assessed. For low-risk situations, routine inspection records are generated; for medium-risk situations, an enhanced monitoring mode is activated; and for high-risk situations, precise maintenance work orders containing abnormal location coordinates and damage mechanisms are generated, achieving risk-driven differentiated operation and maintenance. Furthermore, by comparing the dynamic full-field stress cloud map with the health status benchmark model, the system identifies virtual-real deviations and locates areas of concern with abnormalities. Load fingerprint matching is used to trace the root cause of the anomaly. If structural damage evolution is determined, the parameters of the three-dimensional parametric finite element model are iteratively optimized and the mapping relationship is updated. This significantly improves the accuracy and intelligence level of the full life cycle fatigue assessment of bolted joints in railway steel truss bridges, eliminates the monitoring blind spots of the contact surface between the bolt interior and the hole wall, realizes full-field stress reconstruction and early micro-damage warning based on real loads, and extends remaining life prediction from deterministic to probabilistic intervals to quantify uncertainty. Attached Figure Description

[0048] Figure 1 This is a flowchart of an embodiment of the fatigue perception assessment method for the whole life cycle of bolted joints of railway steel truss bridges according to this application;

[0049] Figure 2 This is a flowchart of step S10 in an embodiment of the fatigue perception assessment method for the whole life cycle of bolted joints of railway steel truss bridges according to this application.

[0050] Figure 3 This is a flowchart illustrating step S30 in an embodiment of the fatigue perception assessment method for the whole life cycle of bolted joints of railway steel truss bridges according to this application.

[0051] Figure 4This is a flowchart illustrating step S402 in an embodiment of the fatigue perception assessment method for the whole life cycle of bolted joints of a railway steel truss bridge according to this application. Detailed Implementation

[0052] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] As attached Figure 1-4 As shown, a method for fatigue perception assessment of bolted joints in railway steel truss bridges throughout their entire life cycle includes the following steps:

[0054] S10: Using the data acquisition module deployed on the physical node of the bolted joint, acquire real-time strain data deployed on the physical node of the bolted joint;

[0055] S20: Call a bolt point feature mapping relationship from the virtual-real mapping module to describe the strain transfer association between the physical node and the virtual node inside the bolt point, map the real-time strain data to the virtual node, and generate real-time strain data of the virtual node;

[0056] S30: The field reconstruction module constructs a physical constraint framework based on the basic control equations of elasticity, and inputs the real-time strain data and the real-time strain data of the virtual nodes as known conditions into the physical constraint framework. The stress distribution that satisfies the known conditions and minimizes the total strain energy under the physical constraint framework is solved to obtain a dynamic full-field stress cloud map.

[0057] S40: The fatigue assessment module identifies real-time variable stress data in the stress concentration area based on the dynamic full-field stress cloud map, and performs fatigue damage analysis and life assessment on the bolted joint based on the real-time variable stress data.

[0058] In this embodiment, the bolted joint of a railway steel truss bridge refers to the bolted connection part used to connect the members in the railway steel truss bridge, which is a key node that bears the alternating load of trains; the full life cycle fatigue perception assessment is to continuously monitor, analyze and predict the fatigue damage state of the bolted joint from its commissioning to its decommissioning; the physical node of the bolted joint refers to the actual sensor measuring point position installed on the exposed surface of the bolted joint, which can obtain real strain data; the data acquisition module is a hardware or software unit responsible for collecting sensor signals; the real-time strain data is the amount of structural deformation measured by the sensor in real time, reflecting the stress level at that point; the virtual-real mapping module is a functional unit that establishes the strain transfer relationship between physical nodes and virtual nodes; the virtual node is a virtual measuring point generated through simulation or mapping in mechanically sensitive locations inside the bolted joint, such as inside the bolt rod or the contact surface of the hole wall, where sensors cannot be directly installed; the bolted joint feature mapping relationship is a model describing the mathematical relationship between the strain of physical nodes and the strain of virtual nodes, which is usually obtained by finite element simulation and dimensionality reduction processing; the real-time strain data of the virtual node is obtained by substituting the measured strain of the physical node into the mapping relationship. The system includes: calculated virtual node strain values; a field reconstruction module, a functional unit for constructing a stress field solution model based on the principles of elasticity; a set of partial differential equations describing the relationship between stress, strain, and displacement of an elastic body; a physical constraint framework, a mathematical framework that uses the mechanical equations as necessary constraints to limit the stress field solution results; a stress distribution with minimum total strain energy, which is the stress field that minimizes the total strain energy of the entire structure under the condition of satisfying known strain constraints, corresponding to the minimum potential energy principle in elasticity; a dynamic full-field stress cloud map, a color map of stress distribution covering the entire area of ​​the bolted joints that is updated over time; a fatigue assessment module, a functional unit for calculating fatigue life and risk classification based on stress results; a stress concentration area, a local location where the stress value is significantly higher than the surrounding average level, which is usually a high-incidence area for fatigue crack initiation; real-time variable stress data, a sequence of stress amplitude and mean values ​​fluctuating over time in the stress concentration area; and fatigue damage analysis and life assessment, which uses variable stress data to calculate the degree of damage that has occurred according to the damage accumulation theory and predict the remaining usable life.

[0059] Specifically, firstly, real-time strain data is collected by data acquisition modules deployed at installable locations on the bolted joint surface, providing a true reflection of the physical world. Secondly, a pre-constructed virtual-real mapping relationship is invoked to extrapolate the strain from surface measuring points to virtual nodes inside the bolt and the hole wall, which cannot be directly monitored, thus obtaining strain information in hidden areas. Then, a physical constraint framework is established based on the governing equations of elasticity. Using the measured and extrapolated strains as known conditions, the stress distribution that satisfies mechanical equilibrium and minimizes total strain energy is solved, thereby reconstructing a dynamic full-field stress cloud map covering the entire bolted joint area. Finally, stress concentration areas are identified from the stress cloud map, and their stress data over time is extracted. Based on this, fatigue damage accumulation calculations and remaining life predictions are performed. By utilizing a limited number of surface measuring points to invert the internal full-field stress, the limitation of sensors only being able to be installed on the surface is overcome. Furthermore, the monitoring blind spots for core vulnerable areas such as the bolt interior and contact surfaces are eliminated, achieving accurate fatigue assessment based on real load data and providing a reliable basis for the full life-cycle maintenance of bridges.

[0060] In one embodiment, step S10, the step of acquiring real-time strain data deployed on the physical node of the bolted joint, includes the following steps:

[0061] S101: Based on the bolted joint feature mapping relationship and the real-time strain data of adjacent physical nodes, calculate the theoretical strain value of the target physical node;

[0062] S102: Compare the measured real-time strain data of the target physical node with the theoretical strain value to generate a data reliability index;

[0063] S103: When the data reliability index is lower than the second threshold, trigger the data anomaly verification or data reconstruction process for the physical node.

[0064] In this embodiment, adjacent physical nodes are other sensor measurement points that are spatially close to the target physical node and have a mechanical relationship with it; the target physical node is a sensor installation location whose data reliability needs to be verified; the theoretical strain value is the strain estimate of the target node calculated by mapping relationship and measured data of adjacent nodes; the measured real-time strain data is the strain value actually collected by the sensor of the target node; the data reliability index is a quantitative score of the degree of agreement between the measured value and the theoretical value, reflecting whether the sensor or data link is normal; the second threshold is a preset acceptable reliability lower limit, and if it is lower than this value, the data is judged to be abnormal; the data abnormality verification is the process of repeatedly collecting, self-checking or comparing and analyzing suspicious data; the data reconstruction process is an alternative method to re-estimate the strain of the point by using adjacent nodes and mapping relationship when the measured data is unreliable.

[0065] Specifically, based on the characteristic mapping relationship of the bolted joints and the measured strain of adjacent physical nodes, the theoretical strain value of the target physical node is calculated and compared with the measured value of the node to generate a quantitative data reliability index. When this index is lower than a preset second threshold, the system automatically triggers a data anomaly verification or data reconstruction process for that physical node. Real-time monitoring of sensor data reliability identifies abnormal data caused by sensor failure, electromagnetic interference, or communication errors, thereby improving the system's anti-interference capability and robustness. This avoids erroneous full-field stress reconstruction and fatigue assessment results due to the failure of individual sensors, ensuring the continuity and reliability of long-term monitoring data.

[0066] In one embodiment, step S20, where the virtual-real mapping module invokes the bolt point feature mapping relationship, includes the following steps:

[0067] S201: Generate the feature mapping relationship of the bolted joint, establish a three-dimensional parametric finite element model of the bolted joint, and set the virtual node at the mechanically sensitive position inside the bolted joint;

[0068] S202: Perform multi-condition mechanical simulation on the three-dimensional parametric finite element model to extract the dynamic response characteristics of the bolted joint under various loads;

[0069] S203: Perform dimensionality reduction processing on the dynamic response features to generate the bolt point feature mapping relationship.

[0070] In this embodiment, the three-dimensional parametric finite element model is a finite element model that describes the bolt joint structure using geometric parameters, which facilitates parameter adjustment for batch simulation; the mechanically sensitive location is the internal location with relatively high stress and most likely to fail due to fatigue in the finite element analysis, used to set virtual nodes; multi-condition mechanical simulation involves performing finite element calculations under different load magnitudes, directions, and combinations; dynamic response characteristics are the numerical set of strain relationships between physical nodes and virtual nodes under different loads; dimensionality reduction is the process of compressing high-dimensional simulation results, such as responses under thousands of load conditions, into a low-dimensional, rapidly computable mapping model using mathematical methods.

[0071] Specifically, a three-dimensional parametric finite element model of the bolted joint is first established, and virtual nodes are set at mechanically sensitive locations within it. Then, multi-condition mechanical simulations are performed on this model to extract the dynamic response characteristics of physical and virtual nodes under various loads. Finally, the high-dimensional response characteristics are dimensionality-reduced to generate a concise and real-time callable feature mapping relationship for the bolted joint. By establishing a mathematical bridge from surface strain to internal strain offline, a high-fidelity mapping model can be obtained through simulation and dimensionality reduction without pre-embedding sensors inside the bolt. This significantly reduces monitoring costs and implementation difficulty, while providing a fast and accurate calculation basis for subsequent virtual-real mapping.

[0072] In one embodiment, step S30, obtaining the dynamic full-field stress cloud map, includes the following steps:

[0073] S301: Using a three-dimensional parametric finite element model representing the healthy state, simulation calculations are performed under the same load assumptions as the current working conditions to generate a reference stress cloud diagram;

[0074] S302: Compare the dynamic full-field stress cloud map with the reference stress cloud map to generate virtual-real deviation data;

[0075] S303: When the virtual-real deviation data exceeds the first threshold, mark the spatial location corresponding to the virtual-real deviation data as an abnormal interest area, and trace the abnormal state of the abnormal interest area.

[0076] In this embodiment, the three-dimensional parametric finite element model is the finite element model of the bolted joint in the initial undamaged state, and the parameters are based on the design drawings and material standards; the current working condition is the actual load conditions that occur during real-time monitoring, such as the support reaction force and temperature when a train passes; the reference stress cloud map is the stress distribution obtained by simulating the healthy model under the same load assumption, serving as a comparison baseline; the virtual-real deviation data is the stress difference field between the dynamic full-field stress cloud map and the reference stress cloud map at the same spatial node; the first threshold is a preset upper limit of allowable deviation, exceeding this value indicates that the structural state has significantly deviated from the healthy state; the abnormal concern area is the spatial sub-region where the deviation continues to exceed the limit, which may correspond to local damage or changes in boundary conditions; the abnormal state tracing is the process of analyzing the cause of the abnormality.

[0077] Specifically, a three-dimensional parametric finite element model representing the healthy state is used to simulate the current measured working conditions under the same load assumptions, generating a baseline stress cloud map. This baseline is then compared point-by-point with a dynamic full-field stress cloud map reconstructed from measured data to calculate the virtual-to-real deviation data. When the deviation exceeds a first threshold, the corresponding location is marked as an area of ​​concern for anomalies, and anomaly tracing is initiated. By identifying whether the structure has deviated from its healthy state and locating the abnormal position, an automatic transition from data-driven to anomaly detection is achieved. This enables the capture of abnormal changes in stress distribution at the micro-damage stage, providing quantitative evidence for early warning.

[0078] In one embodiment, step S303, the step of tracing the abnormal state of the abnormal interest area, includes the following steps:

[0079] S3031: Extract the dynamic stress field pattern of the abnormal interest region and generate the current load fingerprint;

[0080] S3032: Match the current load fingerprint with a load fingerprint database that stores characteristic stress field patterns corresponding to various known events to obtain a matching result;

[0081] S3033: Based on the matching results, determine the possible root cause of the abnormal state.

[0082] In this embodiment, the dynamic stress field pattern is the spatiotemporal characteristics of stress distribution in the abnormal region, such as the movement of stress peaks and local gradient anomalies; the current load fingerprint is a set of numerical feature vectors extracted from the stress field pattern that can uniquely identify the current abnormal state; the load fingerprint database is a pre-stored feature library of typical stress field patterns corresponding to various known events, such as bolt loosening, microcracks, and support settlement; the matching result is the sorting or classification label of the similarity between the current fingerprint and each known event in the database; the possible root cause is the abnormal cause obtained by matching, including bolt preload decay or contact plastic deformation of the hole wall.

[0083] Specifically, the dynamic stress field pattern of the anomaly focus area is extracted to generate a load fingerprint that quantitatively describes the current anomaly state. This fingerprint is then matched with a pre-built load fingerprint database (containing typical stress field patterns corresponding to various known events) to obtain similarity ranking or classification results. Finally, the potential root cause of the anomaly state is determined based on the matching results. By linking anomalies with physical causes, intelligent diagnosis is achieved, significantly reducing the workload of manual analysis. This allows maintenance personnel to directly identify the possible causes of anomalies, enabling them to quickly formulate targeted maintenance strategies and improving decision-making efficiency and accuracy.

[0084] In one embodiment, step S40, where the fatigue assessment module identifies real-time variable stress data of stress concentration regions based on the dynamic full-field stress cloud map, includes the following steps:

[0085] S4011: Extract damage precursor characteristic parameters from the dynamic full-field stress cloud map, wherein the damage precursor characteristic parameters include stress gradient change rate or principal stress direction rotation mode.

[0086] S4012: Generate an early micro-damage warning signal based on the evolution trend of the damage precursor characteristic parameters.

[0087] In this embodiment, the damage precursor characteristic parameter is an early indicator that can reflect the evolution of micro-damage in the stress field before the appearance of macro-cracks, such as the stress gradient change rate or principal stress direction rotation mode; the evolution trend is the direction and rate of change of the above parameters over time; the early micro-damage warning signal is a warning message generated based on the evolution trend of the characteristic parameter, indicating that the structure has entered the fatigue damage initiation stage.

[0088] Specifically, damage precursor parameters such as the rate of change of stress gradient and the rotation mode of principal stress directions are extracted from the dynamic full-field stress cloud map. Their evolution over time is monitored, and an early warning signal for micro-damage is generated when the trend indicates that micro-damage is developing. This enables the identification of fatigue failure instead of waiting for cracks to appear before issuing an alarm, thus significantly advancing the warning time. Maintenance personnel can intervene when the damage is still repairable, avoiding sudden fracture accidents and reducing the overall lifecycle maintenance risks.

[0089] In one embodiment, step S40, which involves performing fatigue damage analysis and life assessment on the bolted joint based on the real-time variable stress data, includes the following steps:

[0090] S4021: Perform cyclic counting and statistics on the real-time variable stress data to obtain the stress cycle spectrum;

[0091] S4022: Based on the damage accumulation model and the stress cycle spectrum, calculate the real-time damage increment and the current cumulative damage degree of the bolted joint;

[0092] S4023: Based on the current cumulative damage level and damage evolution trend, predict the remaining fatigue life of the bolted joint and generate a remaining fatigue life prediction result;

[0093] S4024: The early micro-damage warning signal is fused with the remaining fatigue life prediction result to generate a probabilistic life interval prediction.

[0094] In this embodiment, cycle counting statistics decompose a random variable stress time series into a series of complete stress cycles, commonly using the rainflow counting method; the stress cycle spectrum is a distribution table of each stress amplitude and its corresponding cycle number obtained through statistics; the damage accumulation model is usually the Palmgren-Miner linear accumulation theory, which assumes that each stress cycle causes a certain proportion of damage, and the total damage is the sum of the damage of each cycle; the real-time damage increment is the newly occurring fatigue damage value within the current monitoring period; the current cumulative damage degree is the sum of damage from the start of service to the current moment, a normalized value, and a value of 1 indicates that the component's lifespan has been exhausted; the damage evolution trend is the curve of the cumulative damage degree changing over time, used to extrapolate future states; the remaining fatigue life prediction result is the estimated time or number of cycles that the component can still safely serve based on the current cumulative damage degree and the average damage rate; the probabilistic life interval prediction combines deterministic life prediction with the uncertainty of micro-damage early warning signals to output a life range with a confidence level.

[0095] Specifically, the process first involves cyclic counting of real-time variable stress data to obtain a stress cycle spectrum. Then, based on a damage accumulation model, the real-time damage increment and cumulative damage level up to the present time period are calculated. Next, the remaining fatigue life is predicted based on the cumulative damage level and damage evolution trend, generating a remaining fatigue life prediction result. Finally, early micro-damage warning signals are fused with the remaining fatigue life prediction result to output a probabilistic fatigue life interval prediction that considers uncertainties. This transforms stress history into a quantifiable fatigue life indicator and introduces damage precursor characteristics to correct prediction uncertainties, providing a more scientific risk boundary for maintenance planning. Simultaneously, micro-damage signals are used to provide early warnings of situations where fatigue life may be accelerated.

[0096] In one embodiment, after step S40, the following step is further included:

[0097] S501: Determine the risk level of the bolted joint based on the current cumulative damage, remaining fatigue life prediction results, and probabilistic life interval prediction;

[0098] S502: When the risk level is the first risk level, a routine inspection record is generated;

[0099] S503: When the risk level is the second risk level, the enhanced monitoring mode is activated;

[0100] S504: When the risk level is the third risk level, generate a precise maintenance work order that includes the coordinates of the abnormal location and the suspected damage mechanism.

[0101] In this embodiment, the risk level is classified according to the severity of cumulative damage, remaining lifespan, and probability interval. The first risk level is structurally healthy, with a cumulative damage level far less than 1 and ample remaining lifespan. The routine inspection record is an archived file containing the time, measurement point number, and main health indicators. The second risk level is where there is some damage but it has not reached a dangerous level, requiring enhanced monitoring. The enhanced monitoring mode increases the data acquisition frequency, adds analysis dimensions, and focuses on the evolution of precursory damage characteristics. The third risk level is where the damage is already quite severe or the remaining lifespan is short, requiring immediate intervention. The precise maintenance work order contains the bolt connection point number that needs repair, the three-dimensional coordinates of the abnormal location, and the operation instructions for the suspected damage mechanism.

[0102] Specifically, the risk level of the bolted joint is determined by comprehensively considering the current cumulative damage, remaining fatigue life prediction results, and probabilistic life interval prediction. For low-risk situations, routine inspection records are generated and archived for future reference only; for medium-risk situations, enhanced monitoring mode is activated, increasing sampling and analysis frequency; for high-risk situations, precise maintenance work orders are generated, clearly identifying the coordinates of abnormal locations and damage mechanisms, directly guiding maintenance. By matching differentiated operation and maintenance strategies according to risk levels, resource waste caused by uniform processing of all states is avoided. Simultaneously, directly executable maintenance instructions are provided for high-risk situations, significantly improving operation and maintenance efficiency and bridge safety.

[0103] In one embodiment, after step S3033, the following step is further included:

[0104] S601: When the possible root cause of the abnormal state is determined to be the evolution of structural damage, the parameters of the three-dimensional parametric finite element model representing the healthy state are iteratively optimized using the virtual-real deviation data to generate a corrected three-dimensional parametric finite element model.

[0105] S602: Based on the corrected three-dimensional parametric finite element model, update the bolt joint feature mapping relationship.

[0106] In this embodiment, structural damage evolution is defined as the abnormal root cause being determined as irreversible degradation of the material or component itself, such as fatigue crack propagation or plastic accumulation, rather than external load changes or sensor failure; iterative optimization is the process of repeatedly adjusting model parameters such as elastic modulus and boundary stiffness to make the simulation results approximate the measured reconstruction results; the three-dimensional parametric finite element model is a finite element model that better reflects the current structural state after parameter updates.

[0107] Specifically, when the root cause of the abnormal state is determined to be the evolution of structural damage, the parameters of the three-dimensional parametric finite element model representing the healthy state are iteratively optimized using virtual-to-real deviation data to generate a corrected model. The feature mapping relationship of the bolted joints is then updated based on the corrected model. This allows the digital twin model to evolve synchronously with the actual aging process of the structure, avoiding the problem of gradually increasing prediction bias caused by long-term model fixation. It achieves a closed loop of "monitoring-evaluation-updating," ensuring that fatigue assessment maintains high fidelity throughout the entire lifespan and extending the effective service life of the system.

[0108] A fatigue perception assessment system for the entire life cycle of bolted joints in railway steel truss bridges, used to execute a fatigue perception assessment method for the entire life cycle of bolted joints in railway steel truss bridges, comprising:

[0109] The data acquisition module is used to acquire real-time strain data deployed on the physical nodes of the bolted joints;

[0110] The virtual-real mapping module is used to call a bolt point feature mapping relationship that describes the strain transfer relationship between the physical node and the virtual node inside the bolt point, and to map the real-time strain data to the virtual node to generate real-time strain data of the virtual node.

[0111] The field reconstruction module is used to construct a physical constraint framework based on the basic control equations of elasticity, and input the real-time strain data and the real-time strain data of the virtual nodes as known conditions into the physical constraint framework to solve the stress distribution that satisfies the known conditions and minimizes the total strain energy under the physical constraint framework, thereby obtaining a dynamic full-field stress cloud map.

[0112] The fatigue assessment module is used to identify real-time variable stress data of stress concentration areas based on the dynamic full-field stress cloud map, and to perform fatigue damage analysis and life assessment of the bolted joint based on the real-time variable stress data.

[0113] A specific embodiment of a method for assessing the fatigue perception of bolted joints throughout the entire life cycle of railway steel truss bridges is as follows:

[0114] The steps include: S10: using the data acquisition module deployed on the physical node of the bolted joint to acquire real-time strain data deployed on the physical node of the bolted joint;

[0115] S20: Call a bolt point feature mapping relationship from the virtual-real mapping module to describe the strain transfer association between the physical node and the virtual node inside the bolt point, map the real-time strain data to the virtual node, and generate real-time strain data of the virtual node;

[0116] S30: The field reconstruction module constructs a physical constraint framework based on the basic control equations of elasticity, and inputs the real-time strain data and the real-time strain data of the virtual nodes as known conditions into the physical constraint framework. The stress distribution that satisfies the known conditions and minimizes the total strain energy under the physical constraint framework is solved to obtain a dynamic full-field stress cloud map.

[0117] S40: The fatigue assessment module identifies real-time variable stress data in the stress concentration area based on the dynamic full-field stress cloud map, and performs fatigue damage analysis and life assessment on the bolted joint based on the real-time variable stress data.

[0118] Optionally, fatigue damage analysis and life assessment of bolted joints based on real-time variable stress data also include:

[0119] The stress cycle spectrum is obtained by performing cycle counting statistics on real-time variable stress data. Based on the damage accumulation model and the stress cycle spectrum, the real-time damage increment and current cumulative damage degree of the bolted joint are calculated. The Palmgren-Miner linear damage accumulation theory is adopted, and for the i-th stress amplitude in the stress cycle spectrum... The system first calculates the fatigue life of the component under the specified stress amplitude based on a preset material SN curve. The material SN curve characterizes the relationship between a specific stress level and the number of cycles required for material failure. It is pre-set based on fatigue test data of steel used in bridges and represents the real-time damage increment caused by the current load event. It is the sum of damage caused by all cycles, and its calculation formula is:

[0120] ;

[0121] in Stress amplitude Corresponding loop count, current cumulative damage This is obtained by adding the historical cumulative damage to the current real-time damage increment. Based on the current cumulative damage level and damage evolution trend, the system predicts the remaining fatigue life of the bolted joint. When the critical value (taken as 1.0) is reached, fatigue failure of the component is determined. The system calculates the average damage rate of the bolted joint under a unit standard load by analyzing historical data. Remaining fatigue life It is derived from the following formula:

[0122] ;

[0123] in Based on the current cumulative damage level, this lifespan prediction result will serve as the core basis for risk assessment and maintenance decisions;

[0124] Based on the current cumulative damage and damage evolution trend, predict the remaining fatigue life of the bolted joint.

[0125] Optionally, the fatigue assessment module, which identifies real-time alternating stress data, also includes:

[0126] Damage precursor characteristic parameters are extracted from the dynamic full-field stress cloud map. The damage precursor characteristic parameters include the stress gradient change rate or the principal stress direction rotation mode.

[0127] Based on the evolution trend of the damage precursor characteristic parameters, an early warning signal for micro-damage is generated.

[0128] The system fuses early micro-damage warning signals with remaining fatigue life prediction results to generate a probabilistic fatigue life interval prediction. A correction model is then used to adjust the deterministic remaining fatigue life. The correction model is as follows:

[0129] ;

[0130] in The remaining fatigue life calculated in the previous step. This is the comprehensive intensity index of the quantified early micro-damage warning signal. It is a dimensionless normalized value with a range of 0 to 1. This is a preset correction function, whose value varies with... As the value increases, the strength of the warning signal decreases. This strength is also used to define the confidence interval of the prediction result. A stronger warning signal will not only shorten the average lifespan of the prediction, but also expand its uncertainty range. The final output result is no longer a single value, but a result with a specific confidence level, providing a more comprehensive risk reference for operation and maintenance decisions.

[0131] Optionally, obtaining the dynamic full-field stress contour map also includes:

[0132] Using a three-dimensional parametric finite element model representing the healthy state, simulation calculations are performed under the same load assumptions as the current working conditions to generate a benchmark stress cloud map;

[0133] The dynamic full-field stress contour map is compared with the reference stress contour map to generate a virtual-real deviation data field. The system calculates the von Mises equivalent stress difference of this deviation at each node. Its formula is:

[0134] ;

[0135] in This represents the von Mises equivalent stress value at the nodes in the dynamic full-field stress contour map. The von Mises equivalent stress value is the corresponding node in the baseline stress contour diagram;

[0136] When any node in the virtual-real deviation data field When the value continuously exceeds a preset threshold related to the material fatigue performance, the spatial location corresponding to the virtual-real deviation data is marked as an abnormal area of ​​interest, and the abnormal state of the abnormal area of ​​interest is traced.

[0137] Optionally, the abnormal state of the region of interest can be traced, including:

[0138] Extract the dynamic stress field pattern of the abnormal interest region and generate the current load fingerprint;

[0139] The current load fingerprint is matched with a load fingerprint database that stores characteristic stress field patterns corresponding to various known events to obtain the matching results;

[0140] Based on the matching results, determine the possible root causes of the abnormal state.

[0141] Optionally, the step of acquiring real-time strain data deployed on the physical node of the bolted joint further includes:

[0142] Based on the characteristic mapping relationship of the bolted joint and the real-time strain data of adjacent physical nodes, the theoretical strain value of the target physical node is calculated.

[0143] The measured real-time strain data of the target physical node is compared with the theoretical strain value to generate a data reliability index. This index is a quantitative score used to characterize the degree of agreement between the measured value and the theoretical prediction value. Its calculation formula is as follows:

[0144] ;

[0145] in It is a data reliability indicator, and it is a dimensionless value. These are real-time strain data measured at physical nodes. These are the calculated theoretical strain values. It is the reference strain range for this measuring point set according to the design specifications, used for normalization processing. An index that approaches 1 indicates that the data is highly reliable.

[0146] When the data reliability index continues to fall below a preset, relatively strict second preset threshold, the data anomaly verification or data reconstruction process for that physical node is triggered.

[0147] The present invention and its embodiments have been described above. This description is not restrictive. The accompanying drawings are only one embodiment of the present invention. The actual structure is not limited to this. In short, if a person skilled in the art is inspired by this description and designs a similar structure and embodiment without departing from the spirit of the present invention, such design should fall within the protection scope of the present invention.

Claims

1. A method for full life cycle fatigue perception evaluation of bolted joints of railway steel truss bridges, characterized in that: Including the following steps: S10: Using the data acquisition module deployed on the physical node of the bolted joint, acquire real-time strain data deployed on the physical node of the bolted joint; S20: Call a bolt point feature mapping relationship from the virtual-real mapping module to describe the strain transfer association between the physical node and the virtual node inside the bolt point, map the real-time strain data to the virtual node, and generate real-time strain data of the virtual node; S30: The field reconstruction module constructs a physical constraint framework based on the basic control equations of elasticity, and inputs the real-time strain data and the real-time strain data of the virtual nodes as known conditions into the physical constraint framework. The stress distribution that satisfies the known conditions and minimizes the total strain energy under the physical constraint framework is solved to obtain a dynamic full-field stress cloud map. S40: The fatigue assessment module identifies real-time variable stress data in the stress concentration area based on the dynamic full-field stress cloud map, and performs fatigue damage analysis and life assessment on the bolted joint based on the real-time variable stress data.

2. The method according to claim 1, characterized in that: Step S10, the step of acquiring real-time strain data deployed on the physical node of the bolted joint, includes the following steps: S101: Based on the bolted joint feature mapping relationship and the real-time strain data of adjacent physical nodes, calculate the theoretical strain value of the target physical node; S102: Compare the measured real-time strain data of the target physical node with the theoretical strain value to generate a data reliability index; S103: When the data reliability index is lower than the second threshold, trigger the data anomaly verification or data reconstruction process for the physical node.

3. The method for full life-cycle fatigue perception assessment of bolted joints in railway steel truss bridges according to claim 1, characterized in that: In step S20, the step of the virtual-real mapping module calling the bolt point feature mapping relationship includes the following steps: S201: Generate the feature mapping relationship of the bolted joint, establish a three-dimensional parametric finite element model of the bolted joint, and set the virtual node at the mechanically sensitive position inside the bolted joint; S202: Perform multi-condition mechanical simulation on the three-dimensional parametric finite element model to extract the dynamic response characteristics of the bolted joint under various loads; S203: Perform dimensionality reduction processing on the dynamic response features to generate the bolt point feature mapping relationship.

4. The method for full life-cycle fatigue perception assessment of bolted joints in railway steel truss bridges according to claim 1, characterized in that: In step S30, obtaining the dynamic full-field stress cloud map includes the following steps: S301: Using a three-dimensional parametric finite element model representing the healthy state, simulation calculations are performed under the same load assumptions as the current working conditions to generate a reference stress cloud diagram; S302: Compare the dynamic full-field stress cloud map with the reference stress cloud map to generate virtual-real deviation data; S303: When the virtual-real deviation data exceeds the first threshold, mark the spatial location corresponding to the virtual-real deviation data as an abnormal interest area, and trace the abnormal state of the abnormal interest area.

5. The method for full life-cycle fatigue perception assessment of bolted joints in railway steel truss bridges according to claim 4, characterized in that: Step S303, the step of tracing the abnormal state of the abnormal interest area, includes the following steps: S3031: Extract the dynamic stress field pattern of the abnormal interest region and generate the current load fingerprint; S3032: Match the current load fingerprint with a load fingerprint database that stores characteristic stress field patterns corresponding to various known events to obtain a matching result; S3033: Based on the matching results, determine the possible root cause of the abnormal state.

6. The method for full life-cycle fatigue perception assessment of bolted joints in railway steel truss bridges according to claim 1, characterized in that: Step S40, the step of the fatigue assessment module identifying real-time variable stress data of stress concentration areas based on the dynamic full-field stress cloud map, includes the following steps: S4011: Extract damage precursor characteristic parameters from the dynamic full-field stress cloud map, wherein the damage precursor characteristic parameters include stress gradient change rate or principal stress direction rotation mode. S4012: Generate an early micro-damage warning signal based on the evolution trend of the damage precursor characteristic parameters.

7. The method for full life-cycle fatigue perception assessment of bolted joints in railway steel truss bridges according to claim 6, characterized in that: Step S40, which involves performing fatigue damage analysis and life assessment on the bolted joint based on the real-time alternating stress data, includes the following steps: S4021: Perform cyclic counting and statistics on the real-time variable stress data to obtain the stress cycle spectrum; S4022: Based on the damage accumulation model and the stress cycle spectrum, calculate the real-time damage increment and the current cumulative damage degree of the bolted joint; S4023: Based on the current cumulative damage level and damage evolution trend, predict the remaining fatigue life of the bolted joint and generate a remaining fatigue life prediction result; S4024: The early micro-damage warning signal is fused with the remaining fatigue life prediction result to generate a probabilistic life interval prediction.

8. The method for full life-cycle fatigue perception assessment of bolted joints in railway steel truss bridges according to claim 1, characterized in that: Following step S40, the following step is also included: S501: Determine the risk level of the bolted joint based on the current cumulative damage, remaining fatigue life prediction results, and probabilistic life interval prediction; S502: When the risk level is the first risk level, a routine inspection record is generated; S503: When the risk level is the second risk level, the enhanced monitoring mode is activated; S504: When the risk level is the third risk level, generate a precise maintenance work order that includes the coordinates of the abnormal location and the suspected damage mechanism.

9. The method for full life-cycle fatigue perception assessment of bolted joints in railway steel truss bridges according to claim 5, characterized in that: Following step S3033, the following step is also included: S601: When the possible root cause of the abnormal state is determined to be the evolution of structural damage, the parameters of the three-dimensional parametric finite element model representing the healthy state are iteratively optimized using the virtual-real deviation data to generate a corrected three-dimensional parametric finite element model. S602: Based on the corrected three-dimensional parametric finite element model, update the bolt joint feature mapping relationship.

10. A fatigue perception and assessment system for the entire life cycle of bolted joints in railway steel truss bridges, used to execute the fatigue perception and assessment method for the entire life cycle of bolted joints in railway steel truss bridges as described in any one of claims 1-9, characterized in that: include: The data acquisition module is used to acquire real-time strain data deployed on the physical nodes of the bolted joints; The virtual-real mapping module is used to call a bolt point feature mapping relationship that describes the strain transfer relationship between the physical node and the virtual node inside the bolt point, and to map the real-time strain data to the virtual node to generate real-time strain data of the virtual node. The field reconstruction module is used to construct a physical constraint framework based on the basic control equations of elasticity, and input the real-time strain data and the real-time strain data of the virtual nodes as known conditions into the physical constraint framework to solve the stress distribution that satisfies the known conditions and minimizes the total strain energy under the physical constraint framework, thereby obtaining a dynamic full-field stress cloud map. The fatigue assessment module is used to identify real-time variable stress data of stress concentration areas based on the dynamic full-field stress cloud map, and to perform fatigue damage analysis and life assessment of the bolted joint based on the real-time variable stress data.

Images