A bridge structure performance degradation evaluation method and system based on multi-source data

Abstract

The present application relates to the technical field of bridge engineering, and in particular to a bridge structure performance degradation evaluation method and system based on multi-source data. The method comprises the following steps: extracting damage-sensitive dynamic features based on dynamic response time series data; introducing a structure time-varying stiffness matrix to correct the state equation, and quantifying the deformation trend caused by permanent degradation of the structure; fusing design benchmark deviation features, damage-sensitive dynamic features and deformation trends to obtain overall mechanical performance features; analyzing material-environment coupling degradation features through material performance degradation features and environment-load influence features; and comprehensively evaluating performance degradation conditions in combination with overall mechanical performance features, material-environment coupling degradation features and performance degradation history dependence features. The present application is suitable for quantitative evaluation of bridge structure performance degradation based on multi-source monitoring data, non-destructive testing data, design completion data and historical maintenance data, and can provide precise technical support for bridge operation and maintenance decision-making.

Description

Technical Field

[0001] This invention relates to the field of bridge engineering technology, specifically to a method and system for evaluating the performance degradation of bridge structures based on multi-source data. Background Technology

[0002] As a core component of transportation infrastructure, bridges are subject to the combined effects of multiple factors during their service life, including environmental erosion (temperature and humidity), traffic loads (fatigue effects and overloading), material aging (concrete carbonization and steel corrosion), and maintenance interventions, leading to a gradual degradation of their structural performance. Accurately assessing this performance degradation is crucial for ensuring the safe operation of bridges; however, current technologies have the following significant shortcomings:

[0003] Low accuracy of dynamic feature extraction: Traditional dynamic response data processing uses fixed threshold wavelet denoising, ignoring the influence of the bridge's natural frequency on noise suppression; modal parameter identification does not consider operational load fluctuations, which easily introduces false poles, resulting in a large deviation between damage-sensitive features and the actual stiffness degradation trend.

[0004] Static deformation analysis is inaccurate: Existing methods directly use static deformation as a degradation indicator, failing to distinguish between elastic deformation caused by load and residual deformation caused by permanent structural degradation. They cannot isolate environmental and load interference, making it difficult to quantify the true degradation trend.

[0005] Multi-feature fusion lacks correlation: dynamic and static feature fusion often adopts a simple weighting method without considering the deviation between the design benchmark and the actual structure; material performance evaluation only performs linear superposition of non-destructive testing data, ignoring the error differences and data conflicts of different testing methods; environmental and load effects are analyzed separately without quantifying the synergistic driving effect of the two on degradation.

[0006] Historical dependence is not reflected: Traditional hidden Markov models assume that structural degradation is irreversible and do not incorporate the reversal effect of maintenance measures on performance into the model, resulting in historical maintenance and disease data being unable to effectively support the current degradation assessment.

[0007] The comprehensive assessment lacks dynamism: existing comprehensive assessments are mostly based on static feature combinations and have not established a correlation mechanism of "dynamic mechanical properties-material environment coupling-historical dependence", which makes it impossible to dynamically predict the remaining life and difficult to meet the operation and maintenance needs of bridges throughout their entire life cycle.

[0008] In summary, existing technologies suffer from low accuracy and poor practicality in assessment results because they ignore the correlation, scenario adaptability, and dynamic evolution characteristics of multi-source data. There is an urgent need for a bridge structure performance degradation assessment method that can integrate multi-source data, quantify the coupling effect of multiple factors, and dynamically adapt to actual scenarios. Summary of the Invention

[0009] To address the shortcomings of existing methods and the needs of practical applications, and in order to solve the aforementioned problems, this invention provides a method for evaluating the performance degradation of bridge structures based on multi-source data, comprising the following steps:

[0010] Based on dynamic response time-series data, damage-sensitive dynamic features are extracted; the state equation is corrected by introducing a time-varying stiffness matrix of the structure to quantify the deformation trend caused by permanent structural degradation; the overall mechanical performance characteristics are obtained by integrating the design benchmark deviation characteristics, the damage-sensitive dynamic features, and the deformation trend; the material-environment coupled degradation characteristics are analyzed through material performance degradation characteristics and environment-load influence characteristics; and the performance degradation is comprehensively evaluated by combining the overall mechanical performance characteristics, the material-environment coupled degradation characteristics, and the performance degradation history dependence characteristics.

[0011] This invention extracts damage-sensitive dynamic features based on dynamic response time-series data, which can capture structural stiffness degradation information, providing quantitative basis for early damage identification and preventing hidden damage development. It introduces a time-varying stiffness matrix to correct the state equation, which can isolate load-induced elastic deformation, accurately quantify the deformation trend caused by permanent degradation, and eliminate interference from non-degradation factors. By integrating design benchmark deviations with the first two types of features, it can correct the impact of design and actual deviations, improve the authenticity of overall mechanical performance assessment, and avoid misleading static design benchmarks. By combining material properties with environmental-load characteristic analysis to couple degradation, it can quantify the synergistic effect of both accelerating degradation, compensating for the limitations of separate analysis. By comprehensively assessing degradation using these three types of features, it can integrate multi-dimensional information, comprehensively judge the structural performance status, and provide scientific support for bridge operation and maintenance decisions.

[0012] Optionally, the step of extracting damage-sensitive dynamic features based on dynamic response time-series data includes the following steps:

[0013] This invention employs adaptive wavelet denoising to process dynamic response time-series data; based on the processed data, modal parameters are identified, and damage-sensitive features are then calculated. By using adaptive wavelet denoising (combined with bridge natural frequency-optimized thresholds), this invention addresses the problems of traditional fixed thresholds neglecting frequency characteristics and resulting in poor denoising effects, effectively purifying dynamic response time-series data. Furthermore, based on the processed data, stable poles are screened using load strength to identify modal parameters, avoiding interference from spurious poles. The final calculated damage-sensitive features accurately reflect the structural stiffness degradation trend, providing a reliable dynamic foundation for subsequently integrating design benchmark deviations and deformation trends to construct an overall mechanical performance evaluation model. This directly improves the accuracy of bridge performance degradation assessment and contributes to the scientific rigor of subsequent comprehensive evaluations.

[0014] Optionally, the step of introducing a time-varying stiffness matrix to correct the state equation and quantify the deformation trend caused by permanent structural degradation includes the following steps:

[0015] This invention introduces a time-varying stiffness matrix to correct the state equation; employs a time-varying Kalman filter algorithm to dynamically estimate the bridge structural state; and quantifies the deformation trend caused by permanent structural degradation based on the estimation results. By correcting the state equation with a time-varying stiffness matrix, this invention clearly distinguishes between load-induced elastic deformation and residual deformation caused by permanent structural degradation, effectively eliminating environmental and load interference, avoiding misjudging elastic deformation as degradation, and ensuring the authenticity of the degradation trend quantification. The time-varying Kalman filter breaks through the traditional filter's "constant parameters" assumption, dynamically adapting to the stiffness degradation process, improving the accuracy of residual deformation estimation, and providing reliable data for capturing slowly accumulating permanent degradation. The quantified permanent degradation deformation trend, in conjunction with damage-sensitive dynamic characteristics and design benchmark deviation characteristics, provides accurate static degradation basis for subsequent overall mechanical performance evaluation, avoiding evaluation bias caused by deformation analysis distortion, and supporting the scientific nature of bridge operation and maintenance decisions.

[0016] Optionally, obtaining the overall mechanical performance characteristics by integrating the design baseline deviation characteristics, the damage-sensitive dynamic characteristics, and the deformation trend includes the following steps:

[0017] Based on the design baseline deviation characteristics, the damage-sensitive dynamic characteristics, and the deformation trend, a dynamic weight allocation mechanism is constructed. A neural network fusion evaluation model is then built based on this mechanism to obtain the overall mechanical performance characteristics. This invention dynamically adjusts the weights of the damage-sensitive dynamic characteristics and deformation trends according to the design baseline deviation; the larger the deviation, the stronger the effect of the measured characteristics, adapting to actual structural differences. By using a neural network to handle nonlinear correlations between features, the accuracy loss of linear fusion is avoided, improving the accuracy of the overall mechanical performance evaluation. The output overall mechanical performance characteristics provide core mechanical basis for subsequent comprehensive degradation evaluation combining material-environment coupling and historical dependence characteristics, ensuring that the evaluation is upgraded from a single feature to multi-source correlation, supporting maintenance decisions that are more closely aligned with the actual condition of the bridge.

[0018] Optionally, extracting the design baseline deviation characteristics includes the following steps:

[0019] This invention quantifies the impact of design parameters on structural performance by systematically subjecting them to local perturbations. Bayesian theory is then introduced to dynamically correct the design parameters, thereby calculating the baseline deviation characteristics. This invention quantifies the impact of design parameters on structural performance through systematic local perturbations, accurately identifying key parameters and avoiding interference from irrelevant parameters. The introduction of Bayesian theory dynamically corrects the design parameters, ensuring the baseline aligns with actual operational and completion data. The resulting baseline deviation characteristics provide a basis for subsequent dynamic weight allocation when integrating damage-sensitive dynamic characteristics and deformation trends, effectively eliminating evaluation errors caused by deviations between design and actual structures, and improving the accuracy and scenario adaptability of overall mechanical performance evaluation.

[0020] Optionally, extracting the material property degradation characteristics includes the following steps:

[0021] This invention introduces an evidence theory-based weighting method to quantify the credibility of each testing method. Based on the credibility obtained from the evidence theory weighting, fuzzy clustering is performed to fuse data from different testing methods to obtain material degradation characteristics. This invention quantifies the credibility of each testing method through evidence theory weighting, avoiding interference from methods with large errors. Then, based on the credibility, fuzzy clustering is used to fuse the data, eliminating data conflicts, improving data consistency, and ultimately obtaining accurate material performance degradation characteristics. This provides a reliable foundation for subsequent material-environment coupled degradation characteristic analysis, directly ensuring the overall accuracy of bridge structural performance degradation assessment, and solving the key problem of effectively fusing data from multiple testing methods.

[0022] Optionally, extracting the environment-load influence features includes the following steps:

[0023] This invention first completes environmental data to construct a global distribution function, addressing the issue of incomplete analysis due to missing data and ensuring a comprehensive environmental impact assessment. Then, it converts vehicle axle loads into equivalent uniformly distributed loads, simplifying complex load forms to suit engineering calculations. Finally, it uses grey relational analysis to identify the synergistic effects of load and environment, overcoming the limitations of single-factor analysis and quantifying their combined driving force on structural degradation. This provides accurate input for subsequent material-environment coupled degradation characteristic analysis, avoids biases in single-factor analysis, lays a scientific data foundation for overall performance degradation assessment, and improves assessment accuracy.

[0024] Optionally, the comprehensive assessment of performance degradation, combining the overall mechanical property characteristics, the material-environment coupling degradation characteristics, and the performance degradation history dependence characteristics, includes the following steps:

[0025] The overall mechanical performance characteristics, material-environment coupled degradation characteristics, and historical dependence characteristics of performance degradation are normalized. Based on the processing results, a Bayesian dynamic fusion mechanism is constructed to predict the remaining life of the bridge structure. This invention eliminates the dimensional differences between historical dependence characteristics, real-time load, and environmental characteristics through normalization, solving the problem of incomparability of multi-source features and laying the foundation for fair fusion. Relying on the Bayesian dynamic fusion mechanism, combined with prior information from historical data and real-time data update capabilities, the overall mechanical performance and material-environment coupled degradation characteristics are dynamically integrated to accurately characterize the current performance degradation state. The final output of the remaining life prediction provides a quantitative basis for bridge operation and maintenance, avoiding over- or under-maintenance, and improving the economic efficiency of operation and maintenance while ensuring operational safety. It is a key link connecting feature extraction and actual operation and maintenance decision-making.

[0026] Optionally, analyzing the historical dependency characteristics of the performance degradation includes the following steps:

[0027] Based on the degree of performance degradation of bridge structures, a condition assessment system is constructed; a maintenance intensity quantification model is established, and the state transition probability matrix is ​​improved based on the maintenance intensity quantification model; state prediction entropy is introduced to analyze the historical dependence characteristics of performance degradation. This invention provides a unified standard for measuring the degree of performance degradation by constructing a condition assessment system, solving the problem of a lack of clear reference for historical data; the improved state transition probability matrix, combined with the maintenance intensity quantification model, makes the model more closely reflect the reversal effect of maintenance on performance in actual bridge operation and maintenance; and the introduction of state prediction entropy quantifies the predictive reliability of historical data. Overall, this allows the historical dependence characteristics of performance degradation to more accurately reflect the actual situation, providing reliable historical evidence for subsequent comprehensive assessments, improving the scientific nature of bridge performance assessment and remaining life prediction, and strongly supporting operation and maintenance decisions.

[0028] Secondly, to efficiently execute the bridge structural performance degradation assessment method based on multi-source data provided by this invention, this invention also provides a bridge structural performance degradation assessment system based on multi-source data, including a processor, an input device, an output device, and a memory. The processor, input device, output device, and memory are interconnected. The memory stores a computer program containing program instructions. The processor is configured to call the program instructions to execute the bridge structural performance degradation assessment method based on multi-source data as described in the first aspect of this invention. The bridge structural performance degradation assessment system based on multi-source data of this invention has a compact structure and stable performance, and can stably execute the bridge structural performance degradation assessment method based on multi-source data provided by this invention, further improving the overall applicability and practical application capability of this invention. Attached Figure Description

[0029] Figure 1 A flowchart of a bridge structure performance degradation assessment method based on multi-source data is provided for an embodiment of the present invention.

[0030] Figure 2 This is a framework diagram of a bridge structural performance degradation assessment system based on multi-source data, provided for an embodiment of the present invention. Detailed Implementation

[0031] Specific embodiments of the present invention will now be described in detail. It should be noted that the embodiments described herein are for illustrative purposes only and are not intended to limit the invention. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that these specific details are not necessary to practice the invention. In other instances, well-known circuits, software, or methods have not been specifically described to avoid obscuring the invention.

[0032] Throughout this specification, references to "an embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in connection with that embodiment or example is included in at least one embodiment of the invention. Therefore, the phrases "in an embodiment," "in an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily refer to the same embodiment or example. Furthermore, specific features, structures, or characteristics can be combined in one or more embodiments or examples in any suitable combination and / or sub-combination. Moreover, those skilled in the art will understand that the illustrations provided herein are for illustrative purposes and are not necessarily drawn to scale.

[0033] Please see Figure 1 To address the aforementioned problems, this invention provides a method for evaluating the performance degradation of bridge structures based on multi-source data, such as... Figure 1 As shown, in one embodiment, the method includes the following steps:

[0034] S1. Extract damage-sensitive dynamic features based on dynamic response time-series data.

[0035] By introducing prior constraints based on the inherent modes of the bridge structure to optimize the wavelet threshold function, the problem of excessive smoothing of damage features in traditional wavelet denoising is solved. The stability graph identification rules of the algorithm are implemented by combining the dynamic interval correction feature system of operating load, thereby improving the accuracy of modal parameter identification. In this embodiment, the extraction of damage-sensitive dynamic features based on dynamic response time-series data includes the following steps:

[0036] S11. Dynamic response time series data is processed by adaptive wavelet denoising.

[0037] The dynamic response time-series data collected by the structural health monitoring (SHM) system includes acceleration a(t), strain ε(t), and displacement d(t), with a sampling frequency of 100Hz and a time span of one month.

[0038] Raw monitoring signals are often affected by environmental interference and sensor errors, resulting in outliers and DC components, requiring preprocessing. The 3σ criterion is used to identify and remove outliers. This criterion, based on the characteristics of a normal distribution, considers data deviating from the mean by more than three standard deviations as outliers. Subsequently, mean normalization is performed to eliminate the DC component in the signal, causing it to fluctuate around the zero axis, resulting in a purified signal. This lays the foundation for subsequent analysis. This step effectively improves data quality and reduces the interference of noise on subsequent analysis.

[0039] Based on the characteristics of bridge structural vibration signals, the db8 wavelet basis is used for signal decomposition. This wavelet basis has good time-frequency localization capabilities in capturing structural vibration characteristics. According to the natural frequency range of the bridge (1-5Hz), the number of decomposition layers is set to 8 to achieve fine decomposition of different frequency components of the signal.

[0040] Furthermore, the threshold function is designed to satisfy:

[0041]

[0042] in, This represents the threshold value of the k-th wavelet coefficient in the j-th layer. This represents an empirical correction factor used to adjust the threshold value. This represents the standard deviation of the noise at the j-th layer. Indicates signal length. The frequency correlation coefficient represents the degree to which frequency differences affect the threshold. This represents the center frequency corresponding to the j-th wavelet basis. It represents the fundamental frequency of bridge design and the inherent vibration characteristics of the bridge structure.

[0043] The threshold function designed in this invention can adaptively retain the effective components related to structural vibration based on the natural frequency of the bridge structure, effectively suppressing noise interference and ultimately obtaining a noise-reduced signal. This significantly improves the signal-to-noise ratio and provides high-quality data for modal parameter identification.

[0044] S12. Based on the processed data, identify modal parameters and then calculate damage-sensitive features.

[0045] Based on the noise-reduced signal Constructing the Hankel matrix is ​​a crucial step in ERA (Electronic Recognition Algorithm) for identifying modal parameters. This matrix construction transforms the time-domain signal into a form suitable for parameter identification. During stability graph identification, a load intensity threshold is introduced. This threshold is set by comprehensively considering the normal operating load range of the bridge and the structural response characteristics. In actual bridge operation, the structural response can only more accurately reflect its dynamic characteristics when the load reaches a certain intensity. Therefore, only stable poles during periods where the load is ≥ the load intensity threshold are retained. This effectively removes spurious poles caused by environmental noise and other factors under low-load conditions, improving the accuracy of modal parameter identification and thus obtaining reliable modal frequencies. Damping ratio , where i=1,2,...,n are the modal orders.

[0046] Furthermore, the weighted standard deviation of the modal frequency change rate is defined as the core damage-sensitive feature, satisfying:

[0047]

[0048] in, Indicates damage sensitivity characteristics, Indicates the modal order. This represents the weighting coefficient; the smaller the damping ratio, the larger the corresponding weight. Indicates the number of time steps, reflecting the time span of the monitoring data. The initial health state modal frequency is determined using health monitoring data from the early stages of bridge construction and serves as a benchmark reference.

[0049] This invention comprehensively considers the frequency changes of each modal and their sensitivity to damage, which can more comprehensively and accurately reflect the damage state of bridge structures and provide effective quantitative indicators for the assessment of bridge structural performance degradation.

[0050] S2. Introduce a time-varying stiffness matrix to modify the state equation and quantify the deformation trend caused by permanent structural degradation.

[0051] In the assessment of bridge structural performance degradation, the original monitoring data often contains missing values ​​due to sensor failures, communication interruptions, and other reasons, affecting the accuracy and completeness of subsequent analyses. Cubic spline interpolation is used to complete the missing data. By constructing a piecewise smooth polynomial function, cubic spline interpolation ensures that not only are the function values ​​continuous within the missing data intervals, but also their first and second derivatives are continuous, enabling a more accurate fit to the data change trends.

[0052] In this embodiment, the step of introducing a time-varying stiffness matrix to correct the state equation and quantify the deformation trend caused by permanent structural degradation includes the following steps:

[0053] S21. Introduce the time-varying stiffness matrix of the structure to correct the state equation.

[0054] Deformation of bridges during service From elastic deformation and residual deformation It consists of two parts, with residual deformation directly reflecting the permanent degradation of the bridge structure. Based on the characteristic that bridge stiffness decreases exponentially with time, the state equation is constructed as follows:

[0055]

[0056] in, This represents the stiffness matrix at time t. Indicates the initial stiffness of the bridge. This represents the stiffness degradation coefficient. This represents the combined force of dead load and live load, which is the main external force causing bridge deformation. This represents the residual deformation increment, which follows a normal distribution. This reflects the randomness and uncertainty of the residual deformation increment. This refers to measurement noise, which encompasses the impact of factors such as sensor accuracy errors and environmental interference.

[0057] S22. The time-varying Kalman filter algorithm is used to dynamically estimate the structural state of the bridge.

[0058] Based on the bridge structural state equation, a time-varying Kalman filter algorithm is used to dynamically estimate the bridge structural state, and the filter gain is updated iteratively. and state estimation It can effectively handle noise interference in the data, track changes in the bridge structural status in real time, and meet the following requirements:

[0059]

[0060] Among them, the filter gain The system is dynamically adjusted based on the statistical characteristics of system noise and measurement noise, so that each estimation can make the most of the latest observation data, thereby improving the accuracy of residual deformation estimation.

[0061] S23. Based on the estimation results, quantify the deformation trend caused by permanent structural degradation.

[0062] Specifically, based on the estimation results, the deformation trend caused by permanent structural degradation is quantified, satisfying:

[0063]

[0064] in, This value represents the deformation trend characteristics, i.e., the average cumulative degree of permanent deformation per unit time. The larger the value, the faster the residual deformation of the bridge structure accumulates and the more severe the performance degradation. It can provide an intuitive and effective quantitative basis for bridge maintenance and management. Indicates the total observation duration. Indicates the initial time.

[0065] S3. Integrate the design benchmark deviation characteristics, the damage-sensitive dynamic characteristics, and the deformation trend to obtain the overall mechanical performance characteristics.

[0066] In this embodiment, extracting the design baseline deviation features includes the following steps:

[0067] S311. Systematically perform local disturbances on the design parameters to quantify the degree of influence of each parameter on the structural performance.

[0068] The Morris screening method is used to quantify the impact of each parameter on structural performance by systematically subjecting the design parameters to local perturbations, satisfying the following:

[0069]

[0070] in, Indicates the degree of impact. Indicates the number of samples. This represents the magnitude of the k-th perturbation of parameter i. It represents the change in structural performance indicators (such as bearing capacity, deformation, etc.) under the corresponding disturbance.

[0071] Furthermore, set As screening criteria, key parameters (such as the standard value of concrete axial compressive strength and bridge span deviation) are identified. These key parameters will serve as the core variables for subsequent analysis, and even small changes in them may cause significant fluctuations in structural performance, thus requiring close attention.

[0072] S312. Bayesian theory is introduced to dynamically correct the design parameters, and then the baseline deviation characteristics are calculated.

[0073] To more accurately describe the actual state of bridge structural parameters, Bayesian theory is introduced to dynamically correct the design parameters, and then the baseline deviation characteristics are calculated.

[0074] Using the design value as the prior distribution ,in, Indicates the nominal value of the design parameter. Reflecting the uncertainty of parameters during the design phase; using bridge completion data and long-term monitoring data during operation Constructing the likelihood function ,in, Indicates based on parameters Performance prediction model, This indicates the noise level of the monitoring data.

[0075] According to Bayes' theorem, the posterior distribution can be expressed as: Furthermore, a Markov chain Monte Carlo (MCMC) algorithm is used for numerical simulation. The posterior mean is calculated iteratively by constructing a convergent Markov chain. This value integrates design information and measured data, and can more realistically reflect the actual benchmark parameters of the bridge structure, providing a reliable basis for performance evaluation.

[0076] Furthermore, the baseline deviation characteristic is calculated, satisfying:

[0077]

[0078] in, Indicates the design reference deviation. This indicates the number of key parameters selected through sensitivity analysis. This represents the actual estimated value of the i-th parameter. This represents the design value of the i-th parameter. The parameter deviation is then compared with the sensitivity index. This combination achieves dual weighting: the larger the absolute value of the deviation and the higher the parameter sensitivity, the better the effect. The greater the contribution. The larger the value, the greater the deviation between the design benchmark and the actual structural state. It can be used as one of the important indicators for evaluating the performance degradation of bridge structures, and help engineers quickly locate the key links where the design does not match the reality.

[0079] Furthermore, the process of integrating the design baseline deviation characteristics, the damage-sensitive dynamic characteristics, and the deformation trend to obtain the overall mechanical performance characteristics includes the following steps:

[0080] S321. Based on the design baseline deviation characteristics, the damage-sensitive dynamic characteristics, and the deformation trend, a dynamic weight allocation mechanism is constructed.

[0081] The damage-sensitive dynamic feature matrix is ​​obtained by using normalization techniques. Deformation trend vector and benchmark deviation index The values ​​are uniformly mapped to the [0,1] standard value space to eliminate the influence of dimensional differences on subsequent analysis.

[0082] Furthermore, based on the benchmark deviation index A dynamic weight allocation model is constructed, and the damage-sensitive dynamic feature matrix is ​​implemented through the softmax function. Deformation trend vector The dynamic weight allocation satisfies:

[0083]

[0084]

[0085] This mechanism is achieved through The nonlinear amplification effect enables the automatic increase of the weight of measured features in the comprehensive evaluation when the deviation between measured data and design benchmark increases, effectively weakening the static constraint effect of design benchmark.

[0086] S322. Construct a neural network fusion evaluation model based on the dynamic weight allocation mechanism, and use the neural network fusion evaluation model to obtain the overall mechanical performance characteristics.

[0087] Construct a three-layer fully connected neural network architecture, using weighted feature vectors As the input layer, the overall mechanical performance evaluation index of the bridge is output through layer-by-layer nonlinear transformation, satisfying the following:

[0088]

[0089] in, This represents the overall mechanical performance evaluation index of the bridge. This represents the softmax function. , This represents the weight matrix and bias vector of the i-th layer of the network. Output metric The value ranges from [0,1]. The closer the value is to 1, the closer the bridge structure performance is to the ideal design state.

[0090] This invention uses an attention mechanism to dynamically adjust the weights according to design deviations, thereby integrating dynamic stiffness degradation, static permanent deformation and design deviations to quantify the overall mechanical performance of the structure.

[0091] S4. Analyze the material-environment coupling degradation characteristics by examining the material property degradation characteristics and the environmental-load influence characteristics.

[0092] In an optional embodiment, extracting the material property degradation characteristics includes the following steps:

[0093] S411. Introduce the evidence theory weighting method to quantify the credibility of each detection method.

[0094] In the assessment of bridge structural performance degradation, the original multi-source data (such as stress monitoring data, displacement monitoring data, and material strength test data) often have different dimensions and value ranges. Direct use of these data can lead to imbalances between the data, affecting subsequent analysis. Therefore, a normalization method is used to map all types of data to the [0,1] interval. Linear transformation is then used to eliminate the influence of dimensions, making data from different sources and of different types comparable and laying the foundation for subsequent comprehensive analysis.

[0095] Furthermore, in bridge inspection, different testing methods such as ultrasonic testing, rebound testing, and electromagnetic induction testing each have their own advantages and limitations, and the reliability of their test results also varies. To reasonably quantify the reliability of each testing method, an evidence theory-based weighting method is introduced to define the reliability of each testing method. , Corresponding to ultrasonic waves, springback, and electromagnetic induction respectively, satisfying:

[0096]

[0097] in, This represents the standard deviation of the detection error for the i-th method, obtained through calibration experiments. Calibration experiments involve multiple measurements using the corresponding detection method on a bridge component model with known standard parameters. The deviations between the measured values ​​and the true values ​​are statistically analyzed, and the standard deviation is calculated. A smaller standard deviation indicates a more stable and reliable detection method, thus assigning it a greater weight in the comprehensive evaluation, thereby achieving objective weighting of different detection methods.

[0098] S412. Based on the credibility obtained by weighting evidence theory, fuzzy clustering and fusion are performed on data from different detection methods to obtain material degradation characteristics.

[0099] Credibility derived from the weighting of evidence theory Construct weighted feature vectors Data from different detection methods are weighted and fused according to their reliability.

[0100] Based on this, the fuzzy C-means clustering algorithm is used to classify the fused data. The objective function of FCM clustering is:

[0101]

[0102] in, Describe the objective function. Indicates the total number of samples. Indicates the number of cluster centers. This represents the membership degree of sample i to category c, indicating the degree to which sample i belongs to category c. The value ranges from 0 to 1. This represents the fuzzy coefficient, used to adjust the degree of fuzziness in clustering. The cluster centers are used to represent the structural performance degradation levels. By iteratively adjusting the objective function to minimize these centers, the degradation is categorized into three levels: slight, moderate, and severe. This fuzzy clustering fusion method fully considers the uncertainties in the data and more accurately reflects the actual degradation of the bridge structure.

[0103] Furthermore, to more intuitively quantify the degree of degradation of bridge materials, a weighted sum of sample membership degrees is taken as a characteristic of material degradation, satisfying:

[0104]

[0105] in, Indicates the characteristics of material degradation. The weights of the components are determined based on their importance in the structural stress of the bridge. The beams, as the main load-bearing components, bear a more critical load, hence a weight of 0.6; the piers are weighted at 0.4. This indicates the degradation level, with values ​​of 1 (slight degradation), 2 (moderate degradation), and 3 (severe degradation). The value is calculated using this formula. The value directly reflects the overall degree of material degradation. The closer the value is to 1, the lighter the degradation. The closer it is to 3, the more severe the degradation. This enables a quantitative assessment of the degradation characteristics of bridge materials and provides an important basis for the comprehensive evaluation of bridge structural performance.

[0106] This invention resolves conflicts through evidence theory weighting and fuzzy clustering, and the component weights are assigned based on the importance of stress. It integrates multi-source non-destructive testing data and quantifies the physical property degradation of materials (concrete, steel bars).

[0107] In this embodiment, extracting the environmental-load influence characteristics includes the following steps:

[0108] S421. Complete the environmental data and then construct the global distribution function.

[0109] Kriging interpolation is used to complete the data and construct a global distribution function. Based on this, the average environmental parameters of the components are calculated by taking the mean, thereby achieving spatial consistency and integrity of the environmental data.

[0110] S422. Based on the axle load distribution characteristics of the vehicle, it is converted into an equivalent uniformly distributed load.

[0111] Based on the vehicle axle load distribution characteristics, it is converted into an equivalent uniformly distributed load, satisfying:

[0112]

[0113] in, This represents an equivalent uniformly distributed load. Indicates the number of axle load types. This represents the number of vehicles with axle load of type k at time t. Indicates the corresponding axle load. This indicates the calculated span of the bridge. The conversion process achieves the equivalent quantification of dynamic axle loads into static uniformly distributed loads, facilitating subsequent structural analysis.

[0114] S423. Combine grey relational analysis to identify the synergistic effect between load and environment, and obtain the environment-load influence characteristics.

[0115] Using bridge structural performance degradation indices as a reference sequence A comparison sequence was constructed by selecting environmental parameter (temperature, humidity) sequences and equivalent uniformly distributed load sequences. The grey relational model is used to calculate the degree of correlation between each sequence, satisfying the following:

[0116]

[0117] in, Indicates the degree of correlation. This represents the resolution coefficient, used to adjust the sensitivity of the correlation calculation. The correlation is obtained by averaging the time series data. , This indicates the sequence length, and thus quantifies the degree of correlation between different influencing factors and structural performance degradation.

[0118] Furthermore, the environmental-load influence characteristics are calculated to satisfy:

[0119]

[0120] in, This represents the environment-load coupling effect coefficient. This represents the initial state value of the i-th influencing factor.

[0121] This invention identifies key driving factors and quantifies coupling effects through grey relational analysis, and considers bridge span parameters in the equivalent load calculation, quantitatively describing the accelerating effect of the synergistic effect of environment and load on the performance degradation of bridge structures.

[0122] Furthermore, the coupled degradation characteristics of materials and the influence characteristics of the environment and load are analyzed to determine the characteristics of material-environment degradation. ,satisfy:

[0123]

[0124] The numerical value directly reflects the degree of material degradation under the combined effects of environmental factors and loads. The larger the value, the more significant the material degradation and the more severe the structural performance degradation.

[0125] S5. Combining the overall mechanical performance characteristics, the material-environment coupling degradation characteristics, and the performance degradation history dependence characteristics, comprehensively evaluate the performance degradation situation.

[0126] In this embodiment, analyzing the historical dependency characteristics of the performance degradation includes the following steps:

[0127] S511. Based on the degree of degradation of bridge structural performance, construct a condition assessment system.

[0128] Specifically, based on the degree of bridge structural performance degradation, a five-level condition assessment system is constructed (condition levels 1-5, where level 1 represents the optimal state and level 5 represents the critical failure state). The condition sequence is initialized based on the severity index of the defects. ,in This indicates the performance status level at time t. .

[0129] S512. Establish a maintenance intensity quantification model, and improve the state transition probability matrix based on the maintenance intensity quantification model.

[0130] The maintenance intensity quantification model satisfies:

[0131]

[0132] in, This represents the maintenance strength factor, which is the impact of maintenance practices on structural performance. This represents the actual cost of maintaining the system at time k. This indicates the maximum cost of a single maintenance operation. Let represent the maintenance area maintained at time k. This represents the total area of ​​the bridge structure. The maintenance intensity at each point in time is matched with the impact of the most recent maintenance event. For example, if the second maintenance occurs on the 500th day of operation and the third maintenance occurs on the 800th day of operation, then at t=600 (the 600th day of operation), the maintenance intensity of the bridge directly adopts the intensity factor of the second maintenance, reflecting the continued impact of the maintenance event on the bridge performance in subsequent time periods.

[0133] Furthermore, the state transition probability matrix of the hidden Markov model is improved based on the maintenance intensity quantification model. ,satisfy:

[0134]

[0135] in, This represents the state transition probability of a hidden Markov model. This represents the Kronecker function (which takes the value 1 when j = i - 1, and 0 otherwise).

[0136] S513. Introduce state prediction entropy analysis to analyze the historical dependency characteristics of the performance degradation.

[0137] Based on the output probability matrix Obtain the posterior probability, and then predict the entropy through the state. The ability of historical data to predict future states, i.e., the historical dependency feature of performance degradation, satisfies:

[0138]

[0139] in, Denotes the observation vector at time t. This represents the set of observation sequences from time 1 to time t. The smaller the value, the higher the reliability of the historical data sequence in predicting the future state, that is, the stronger the historical dependence.

[0140] This invention corrects the HMM transition probability by using a maintenance intensity factor, reflecting the impact of historical maintenance and disease records on the current performance degradation trend, and quantifying the historical dependence of the degradation process.

[0141] Furthermore, the comprehensive assessment of performance degradation, combining the overall mechanical property characteristics, the material-environment coupling degradation characteristics, and the performance degradation history dependence characteristics, includes the following steps:

[0142] S521. Normalize the overall mechanical property characteristics, the material-environment coupling degradation characteristics, and the performance degradation history dependence characteristics.

[0143] The overall mechanical properties, the material-environment coupling degradation properties, and the performance degradation history dependence properties are normalized and mapped to the [0,1] interval to eliminate the influence of dimensional differences on subsequent analysis.

[0144] S522. Based on the processing results, construct a Bayesian dynamic fusion mechanism and predict the remaining life of the bridge structure through the Bayesian dynamic fusion mechanism.

[0145] Based on historical dependency characteristics of performance degradation As a prior probability distribution, the feature vectors obtained from real-time monitoring will be used. Defined as the likelihood function, a posterior probability distribution model of the bridge structural performance state is constructed based on Bayesian theory:

[0146]

[0147] in, The bridge structural performance level is represented by values ​​ranging from {1,...,5}, corresponding to different degrees of degradation; the expected value of the posterior probability distribution is calculated. To determine the current performance status level of the bridge.

[0148] Furthermore, based on the current performance state estimate Real-time overall mechanical performance characteristics Material-environment coupled degradation characteristics As input variables, a random forest regression (RF) model was used to establish a bridge structural performance degradation prediction model. By supervising the training on historical degradation data, the model was able to predict the time required for the bridge structural performance to degrade to the critical level (Level 5).

[0149] This invention integrates historical dependencies and real-time features through Bayesian updates and combines them with random forests to achieve dynamic prediction, enabling real-time assessment of bridge structural performance degradation and prediction of remaining life, thus providing a quantitative basis for maintenance decisions.

[0150] Please see Figure 2 In this embodiment, to efficiently execute the bridge structural performance degradation assessment method based on multi-source data provided by the present invention, the present invention also provides a bridge structural performance degradation assessment system based on multi-source data, comprising: an input device, an output device, a processor, and a memory, wherein the input device, output device, processor, and memory are interconnected, and the memory contains program instructions for the steps of the bridge structural performance degradation assessment method based on multi-source data. The bridge structural performance degradation assessment system based on multi-source data of the present invention has a compact structure and stable performance, and can stably execute the bridge structural performance degradation assessment method based on multi-source data of the present invention, further improving the overall applicability and practical application capability of the present invention.

[0151] In this embodiment, the processor may be a central processing unit, but it can also be other general-purpose processors, digital signal processors, application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (OPGs), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor. Input devices can be used to acquire data. Output devices can be used to output the results obtained by storing program instructions contained in a computer program in the memory provided by this invention. The memory may include read-only memory and random access memory (RAM), and provides instructions and data to the processor. A portion of the memory may also include non-volatile random access memory (RAM).

[0152] In one possible implementation, the memory may include a stored program area and a stored data area. The stored program area may store the operating system and applications required for at least one function; the stored data area may store data created during use. Furthermore, the memory may include read-only memory and random access memory, and provides instructions and data to the processor. The memory stores the operating system and operating instructions, executable modules, or data structures, or subsets thereof, or extended sets thereof. The operating instructions may include various operation instructions for implementing various operations. The operating system may include various system programs for implementing various basic tasks and handling hardware-based tasks.

[0153] The embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the steps of the bridge structure performance degradation assessment method based on multi-source data.

[0154] The storage medium can include various media that can store program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0155] In summary, this invention extracts damage-sensitive dynamic features based on dynamic response time-series data, which can capture structural stiffness degradation information, providing quantitative basis for early damage identification and preventing hidden damage development. The introduction of a time-varying stiffness matrix to correct the state equation can isolate load-induced elastic deformation, accurately quantify deformation trends caused by permanent degradation, and eliminate interference from non-degradation factors. Integrating design benchmark deviations with the first two types of features can correct the impact of design-actual deviations, improve the authenticity of overall mechanical performance assessment, and avoid misleading static design benchmarks. Combining material performance with environmental-load characteristic analysis to couple degradation quantifies the synergistic effect of both accelerating degradation, compensating for the limitations of individual analysis. By comprehensively assessing degradation using these three types of features, multi-dimensional information can be integrated to comprehensively judge the structural performance status, providing scientific support for bridge operation and maintenance decisions.

[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the present invention.

Claims

1. A method for evaluating the performance degradation of bridge structures based on multi-source data, characterized in that, Includes the following steps: Based on dynamic response time-series data, damage-sensitive dynamic features are extracted; The state equation is modified by introducing a time-varying stiffness matrix of the structure to quantify the deformation trend caused by permanent degradation of the structure; By integrating the design baseline deviation characteristics, the damage-sensitive dynamic characteristics, and the deformation trend, the overall mechanical performance characteristics are obtained. The characteristics of material-environment coupling degradation are analyzed by examining the material property degradation characteristics and the characteristics of environmental-load influence. The overall mechanical performance characteristics, the material-environment coupling degradation characteristics, and the performance degradation history dependence characteristics are combined to comprehensively evaluate the performance degradation. Extracting the material property degradation characteristics includes the following steps: An evidence theory-based weighting method is introduced to quantify the credibility of each detection method; Based on the credibility obtained by weighting evidence theory, fuzzy clustering and fusion of data from different detection methods are performed to obtain material degradation characteristics. Extracting the environmental-load influence characteristics includes the following steps: Complete the environmental data, and then construct the global distribution function; Based on the axle load distribution characteristics of the vehicle, it is converted into an equivalent uniformly distributed load. By combining grey relational analysis to identify the synergistic effect between load and environment, the environmental-load influence characteristics are obtained. The performance degradation history dependency feature satisfies: in, This indicates the predictive power of historical data for future states. Denotes the observation vector at time t. This represents the set of observation sequences from time 1 to time t. This indicates the performance status level at time t+1.

2. The bridge structural performance degradation assessment method based on multi-source data according to claim 1, characterized in that, The extraction of damage-sensitive dynamic features based on dynamic response time-series data includes the following steps: Dynamic response time-series data is processed using adaptive wavelet noise reduction. Modal parameters are identified based on the processed data, and then damage-sensitive features are calculated.

3. The bridge structural performance degradation assessment method based on multi-source data according to claim 1, characterized in that, The method of introducing a time-varying stiffness matrix to correct the state equation and quantify the deformation trend caused by permanent structural degradation includes the following steps: The state equation is modified by introducing a time-varying stiffness matrix of the structure; A time-varying Kalman filter algorithm is used to dynamically estimate the structural state of the bridge. The deformation trend caused by permanent structural degradation is quantified based on the estimation results.

4. The bridge structural performance degradation assessment method based on multi-source data according to claim 1, characterized in that, The method of integrating the design benchmark deviation characteristics, the damage-sensitive dynamic characteristics, and the deformation trend to obtain the overall mechanical performance characteristics includes the following steps: Based on the design baseline deviation characteristics, the damage-sensitive dynamic characteristics, and the deformation trend, a dynamic weight allocation mechanism is constructed. A neural network fusion evaluation model is constructed based on the dynamic weight allocation mechanism, and the overall mechanical performance characteristics are obtained using the neural network fusion evaluation model.

5. The bridge structural performance degradation assessment method based on multi-source data according to claim 4, characterized in that, Extracting the design baseline deviation characteristics includes the following steps: Systematically perform local perturbations on the design parameters to quantify the degree of influence of each parameter on the structural performance; Bayesian theory is introduced to dynamically correct design parameters, and then the baseline deviation characteristics are calculated.

6. The bridge structural performance degradation assessment method based on multi-source data according to claim 1, characterized in that, The comprehensive assessment of performance degradation, combining the overall mechanical performance characteristics, the material-environment coupling degradation characteristics, and the performance degradation history dependence characteristics, includes the following steps: Normalization processing is performed on the overall mechanical property characteristics, the material-environment coupling degradation characteristics, and the performance degradation history dependence characteristics; Based on the processing results, a Bayesian dynamic fusion mechanism is constructed, and the remaining life of the bridge structure is predicted through the Bayesian dynamic fusion mechanism.

7. The bridge structural performance degradation assessment method based on multi-source data according to claim 6, characterized in that, Analyzing the historical dependency characteristics of the performance degradation includes the following steps: A condition assessment system is constructed based on the degree of degradation of bridge structural performance; Establish a maintenance intensity quantification model, and improve the state transition probability matrix based on the maintenance intensity quantification model; The performance degradation history dependency features are analyzed by introducing state prediction entropy.

8. A bridge structural performance degradation assessment system based on multi-source data, characterized in that, The bridge structure performance degradation assessment system based on multi-source data includes: an input device, an output device, a processor, and a memory, wherein the input device, output device, processor, and memory are interconnected, and the memory includes program instructions for executing the bridge structure performance degradation assessment method based on multi-source data according to any one of claims 1-7.

Images