Road and bridge parameter anomaly detection method

Abstract

The invention discloses a road and bridge parameter anomaly detection method, and belongs to the technical field of civil engineering. Comprising the following steps: step 1, establishing space-time relevance and a data mapping relation between monitoring points; 2, self-adaptive updating of the judgment rule is achieved so as to adapt to state evolution of a bridge service stage; step 3, performing intelligent attribution analysis on abnormity in the monitoring data; step 4, carrying out joint identification and comprehensive evaluation on the sudden structural damage and the slow degeneration degradation process; and 5, carrying out quantitative analysis on the deviation degree of the bridge health state, and generating a standardized bridge health index. According to the method, the space-time correlation network and the environment-structure mapping function between the monitoring points are established, and a multi-dimensional feature extraction and dynamic weight distribution mechanism is combined, so that environment interference elimination, abnormal attribution intelligent analysis and joint identification of sudden damage and slow degeneration are realized, and finally, a standardized health index is quantified and generated.

Description

Technical Field

[0001] The present application relates to the field of civil engineering technology, and more specifically, to a method for detecting anomalies in road and bridge parameters. Background Art

[0002] To ensure the safe operation and extend the service life of road and bridge structures, real-time monitoring and analysis of their structural parameters has become a routine task for highway operations and maintenance units. Currently, common structural monitoring parameters include strain, deflection, displacement, inclination, acceleration, and temperature. These parameters reflect the response characteristics of bridge structures under varying loads and environmental conditions. Analysis of these parameters can effectively identify the structural health status and potential safety hazards.

[0003] However, in actual operation, numerous factors influence road and bridge parameters, including natural conditions such as traffic load and ambient temperature, as well as engineering defects such as structural aging, uneven foundation settlement, and loose connections. The combination of these factors often results in complex variations in parameter data, including nonlinearity, sudden changes, and periodic disturbances, making it difficult to identify abnormal conditions. Currently, widely adopted anomaly detection methods include threshold judgment, statistical modeling, and intelligent recognition methods based on machine learning. While traditional threshold methods are simple to implement, they struggle with the joint analysis of complex multi-dimensional parameters and are prone to false positives and false negatives. Statistical modeling methods are sensitive to data distribution assumptions and lack adaptability. While intelligent recognition methods offer some improvements in accuracy, they often struggle to be widely adopted in practical deployments due to algorithmic complexity, insufficient model generalization, and strong sample dependency.

[0004] In addition, existing methods generally have the following problems: first, the parameter data quality control mechanism is not sound, and it is difficult to effectively distinguish between anomalies and sensor noise; second, there is a lack of modeling constraints on the physical characteristics of road and bridge structures, resulting in a certain "black box" nature in the detection results; third, the coupling relationship between spatial distribution and temporal evolution characteristics is not fully considered, making it difficult to identify some potential anomalies in a timely manner.

[0005] In summary, how to construct a high-precision, robust and interpretable anomaly detection method based on the multi-source parameter characteristics of road and bridge structures has become a technical problem that needs to be solved urgently. Summary of the Invention

[0006] In order to overcome the shortcomings of the prior art, the present invention aims to provide a method for detecting abnormalities in road and bridge parameters, including the following steps:

[0007] Step 1: Establish the spatiotemporal correlation and data mapping relationship between monitoring points;

[0008] Step 2: Adaptively update the decision rules to adapt to the state evolution of the bridge during its service life.

[0009] Step 3: Conduct intelligent attribution analysis on anomalies in monitoring data;

[0010] Step 4: Jointly identify and comprehensively evaluate sudden structural damage and slow degradation processes;

[0011] Step 5: Quantitatively analyze the degree of deviation of the bridge health status and generate a standardized bridge health index.

[0012] Furthermore, step 1 includes the following steps:

[0013] Structural response parameters and environmental impact parameters are collected through a distributed sensor network and pre-processed to form a structured multi-dimensional time series data set;

[0014] Based on the physical structural characteristics of the bridge, a spatiotemporal correlation network is constructed to reflect the correlation between the spatial topology and time series of the monitoring points;

[0015] Identify and quantify the impact of environmental factors on structural response and its hysteresis effect, and form an environment-structure mapping function library;

[0016] Based on measured data, the response transfer characteristics between monitoring points are calculated, the signal propagation path and its attenuation law are identified, and a spatial transfer function matrix reflecting the structural dynamic characteristics is constructed;

[0017] The high-dimensional monitoring data is reduced in dimension and features are extracted, and the feature combination that affects the structural health status is identified to form a feature fingerprint set that characterizes the structural health status.

[0018] Furthermore, the construction of the spatiotemporal correlation network includes:

[0019] Based on the physical structure layout of the bridge, the monitoring points are mapped into a three-dimensional coordinate system to construct an initial spatial topology model;

[0020] Calculate the Euclidean distance and structural connectivity between monitoring points to form a spatial adjacency matrix;

[0021] For different types of monitoring parameters, the mutual correlation coefficient, Granger causality coefficient and mutual information entropy of the time series are calculated respectively to quantitatively characterize the strength of the time series correlation;

[0022] The spatial adjacency relationship is integrated with the temporal correlation strength to generate a weighted spatiotemporal correlation network, where the edge weight reflects the degree of physical correlation between monitoring points;

[0023] Identify community structures and key nodes in the network, and calculate global and local characteristic parameters of the network, including average path length, clustering coefficient, centrality and robustness indicators;

[0024] With the continuous accumulation of monitoring data, the edge weights and topological structure of the spatiotemporal correlation network are regularly updated to dynamically reflect the changes in the correlation characteristics of the bridge structure over time.

[0025] Furthermore, the process of establishing the environment-structure mapping function library includes:

[0026] Establish a linear mapping relationship between environmental parameters and structural responses, and calculate the sensitivity coefficient of each environmental factor;

[0027] The response delay parameter τ is introduced to quantify the time lag effect between environmental changes and structural responses. For temperature field changes, the short-term response lag τ1 and the long-term response lag τ2 are set;

[0028] Establish a nonlinear mapping function f(E,τ) between environmental parameters and structural responses, where E represents the environmental parameter vector;

[0029] Based on seasonal analysis, the seasonal adjustment coefficient matrix S of the environment-structure mapping is constructed to reflect the differences in structural responses under different seasonal conditions.

[0030] Furthermore, step 2 includes the following steps:

[0031] Based on the historical monitoring data of the bridge during normal operation, a multi-level dynamic envelope system integrating time evolution and periodic change factors is constructed as the basic threshold for subsequent anomaly identification.

[0032] Using the environment-structure mapping function library, the measured structural response parameters are corrected for environmental influences to obtain the residual response after eliminating environmental interference, eliminating the structural response changes caused by environmental fluctuations;

[0033] According to the sensitivity and importance of each monitoring parameter in the characterization of the structural state, a differentiated weight system of spatial partitioning and parameter classification is established;

[0034] The environmentally corrected structural response parameters are compared with the reference envelope in real time, and the comprehensive deviation is calculated using differentiated weights;

[0035] Regularly update the benchmark dataset and envelope parameters to highlight recent operational characteristics and dynamically adapt to performance changes caused by aging of the bridge;

[0036] In view of the differences in performance characteristics of bridges during their service life, a stage-by-stage judgment standard library is constructed, and the abnormal judgment threshold is dynamically adjusted based on the service life, maintenance records and performance evaluation results, and the judgment rule parameters are optimized.

[0037] Furthermore, the establishment of the differentiated weight system includes:

[0038] Based on the structural function of the bridge, the monitoring area is divided into the main beam area, the bridge tower area, the anchorage area and the auxiliary component area, and the spatial weight coefficient w is assigned to each area according to its importance to the overall structural safety. i ;

[0039] According to the sensitivity analysis results of monitoring parameters to structural damage, the monitoring parameters are divided into displacement, strain, vibration and environment categories, and the parameter categories are assigned weights w j ;

[0040] According to the importance difference of the same category parameters in different spatial positions, a position-parameter joint weight matrix W = {w ij}, where w ij =w i ×w j ×α ij , α ij is the adjustment factor;

[0041] A weight adaptive update mechanism is designed to dynamically adjust the weight coefficient of each monitoring point according to the historical accuracy and false alarm rate of anomaly detection, thereby optimizing the accuracy of anomaly judgment.

[0042] Furthermore, step 3 includes the following steps:

[0043] Establish a conditional probability table that reflects the relationship between environmental conditions, material state, load characteristics and the response of each monitoring point, forming a complete probabilistic reasoning framework;

[0044] Summarize the characteristic patterns of different types of anomalies and record the spatiotemporal distribution characteristics, duration, and recovery characteristics of multi-parameter responses to form an anomaly pattern feature library;

[0045] Decompose abnormal responses in monitoring data into three parts: environmental factors, load factors, and structural performance changes. Quantitatively analyze the contribution of each factor to the abnormality, and identify and evaluate the true abnormality caused by structural performance changes.

[0046] Based on the differences in the time scale of structural responses caused by different factors, the relationship between different time patterns and environmental disturbances, load changes and structural damage is analyzed;

[0047] By analyzing the spatial distribution characteristics of the structural response, the spatial propagation pattern of the abnormal response is identified, and the measured spatial distribution is compared with the expected pattern to evaluate the localization or global nature of the anomaly;

[0048] Calculate the confidence and uncertainty of the cause of the anomaly, and determine the root cause based on the set rules to make anomaly attribution decisions.

[0049] Furthermore, the construction of the abnormal pattern feature library includes:

[0050] Collect and organize historical abnormal event data, including abnormalities caused by environmental factors, abnormalities caused by excessive loads, abnormalities caused by structural damage, and abnormalities caused by sensor failures;

[0051] For each type of abnormal event, its time domain feature vector FT = {duration, rise time, recovery time, peak intensity, root mean square value, kurtosis, skewness, and form factor} is extracted to describe the time evolution characteristics of the abnormality;

[0052] Extract the frequency domain feature vector FF = {spectral center of gravity, frequency band energy distribution, main frequency offset, harmonic ratio, spectrum entropy} of the abnormal event to describe the frequency characteristics of the abnormality;

[0053] Extract the spatial distribution feature vector FS = {influence range, attenuation gradient, propagation speed, spatial correlation, principal component direction} of the abnormal event to describe the spatial diffusion characteristics of the abnormality;

[0054] Summarize the recovery characteristics of different types of anomalies, including fully recoverable anomalies, partially recoverable anomalies and unrecoverable anomalies, and define the recovery index IR∈[0,1] to quantify the degree of recovery;

[0055] Classify historical abnormal events into patterns and form an abnormal pattern set M = {M1, M2, ..., M k}, each pattern Mg contains a feature vector [FT, FF, FS] and a category label;

[0056] Establish a mapping table R(M→C) between abnormal patterns and possible causes to provide attribution reference for newly discovered abnormalities;

[0057] Design a similarity-based anomaly pattern matching mechanism to compare newly detected anomalies with the patterns in the feature library to identify the most likely anomaly type.

[0058] Furthermore, step 4 includes the following steps:

[0059] Based on the time series data of monitoring points, statistical feature analysis, trend analysis and change point detection are carried out to extract time domain features to identify long-term change trends, sudden damage events and dynamic response transmission laws;

[0060] Convert the time domain signal to the frequency domain, extract the dynamic characteristic parameters of the structure, analyze the frequency drift, modal changes and high-order harmonic components, identify stiffness changes and local damage, and construct a spectrum feature vector for structural parameter change identification;

[0061] Establish a quantitative evaluation index system for sudden damage identification, combined with the spatial distribution characteristics of the damage location, to achieve rapid response and accurate positioning of sudden damage;

[0062] Extract the slowly varying characteristics of structural parameters, identify the overall stiffness attenuation and material degradation process, calculate the degradation rate and cumulative effect, and predict the future evolution trend of structural performance;

[0063] Construct a unified feature space that integrates multi-dimensional features to achieve collaborative characterization and effective differentiation of sudden damage and slow degradation;

[0064] Based on the above analysis results, the impact of damage and degradation on structural safety is quantitatively evaluated, the structural risk index is calculated, and a comprehensive assessment report containing qualitative judgments and quantitative indicators is generated.

[0065] Furthermore, step 5 includes the following steps:

[0066] Based on multi-physics coupling theory and structural characteristic parameters, a high-precision finite element reference model is constructed that comprehensively considers geometry, material properties, boundary conditions, and environmental influences to truly reflect the service behavior of the bridge;

[0067] Comparing and analyzing measured data with the reference model predictions, a multi-dimensional deviation index system was established that included frequency domain characteristics, time domain characteristics, and energy distribution characteristics to fully capture abnormal bridge conditions.

[0068] Based on the spatiotemporal distribution of deviation indicators at each monitoring point, a comprehensive scoring mechanism considering structural importance, parameter sensitivity, and anomaly persistence is constructed, and the multidimensional deviation indicators are mapped into a standardized health index in the range of [0-100].

[0069] Combine seasonal changes with long-term deterioration factors to make probabilistic predictions of the future state of the bridge;

[0070] The overall health index is decomposed into key component-level and functional unit-level indicators to achieve health status mapping from macro to micro, accurately locate potential problem areas, and quantitatively evaluate their impact on the overall structural safety.

[0071] Compared with the prior art, this application has the following beneficial effects:

[0072] This application establishes a spatiotemporal correlation network and an environment-structure mapping function between monitoring points, combined with multidimensional feature extraction and a dynamic weight allocation mechanism. This approach eliminates environmental interference, intelligently analyzes anomaly attribution, and jointly identifies sudden damage and gradual degradation, ultimately generating a standardized health index. This method significantly improves the accuracy, adaptability, and interpretability of bridge anomaly identification, demonstrating clear engineering practicality and technological innovation. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 This is a flow chart of a method for detecting anomalies in road and bridge parameters disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0074] The present invention provides a method for detecting anomalies of road and bridge parameters. This method realizes real-time monitoring of the operating status of road and bridge structures and anomaly warning by establishing spatiotemporal correlation of monitoring points, adaptively updating judgment rules, intelligently attributing anomalies, jointly identifying structural damage and degradation, and quantitatively evaluating the health status of bridges. Figure 1 Specific embodiments of the present invention are described in detail.

[0075] like Figure 1 As shown, a method for detecting anomalies of road and bridge parameters includes the following steps:

[0076] Step 1: Establish spatiotemporal correlation and data mapping between monitoring points

[0077] This step collects structural response parameters and environmental impact parameters through a distributed sensor network, performs preprocessing and feature extraction, and constructs a network model and mapping function library that reflects the spatiotemporal correlation characteristics of the monitoring points. The specific implementation is as follows:

[0078] 1) Data collection and preprocessing

[0079] A distributed sensor network deployed at key locations on the bridge collects structural response parameters and environmental impact parameters in real time. These parameters include vibration acceleration, displacement, strain, inclination angle, crack width, and cable tension; environmental impact parameters include temperature distribution, humidity, wind speed and direction, rainfall, vehicle volume, and vehicle type distribution. The collected data undergoes preliminary preprocessing, including data cleaning, denoising, and interpolation, to form a structured, multidimensional time series dataset.

[0080] 2) Construction of spatiotemporal correlation network

[0081] Based on the physical structural characteristics of the bridge, a spatiotemporal correlation network reflecting the correlation between the spatial topology and time series of monitoring points is constructed. The specific steps are as follows:

[0082] Map the monitoring points into a three-dimensional coordinate system and construct an initial spatial topology model based on the physical structure layout of the bridge;

[0083] Calculate the Euclidean distance and structural connectivity between monitoring points to form a spatial adjacency matrix;

[0084] For different types of monitoring parameters, the mutual correlation coefficient, Granger causality coefficient and mutual information entropy of the time series are calculated respectively to quantitatively characterize the strength of the time series correlation;

[0085] The spatial adjacency relationship is integrated with the temporal correlation strength to generate a weighted spatiotemporal correlation network, where the edge weight reflects the degree of physical correlation between monitoring points;

[0086] Identify community structures and key nodes in the network, and calculate global and local characteristic parameters of the network, including average path length, clustering coefficient, centrality, and robustness indicators;

[0087] With the continuous accumulation of monitoring data, the edge weights and topological structure of the spatiotemporal correlation network are regularly updated to dynamically reflect the temporal changes in the correlation characteristics of the bridge structure.

[0088] 3) Establishment of environment-structure mapping function library

[0089] Identify and quantify the impact of environmental factors on structural response and their hysteresis effect, and form an environment-structure mapping function library. The specific steps are as follows: establish a linear mapping relationship between environmental parameters and structural response, and calculate the sensitivity coefficient of each environmental factor; introduce a response delay parameter τ to quantify the time lag effect between environmental changes and structural responses. For temperature field changes, set the short-term response lag τ1 and the long-term response lag τ2; establish a nonlinear mapping function f(E,τ) between environmental parameters and structural responses, where E represents the environmental parameter vector; based on seasonal analysis, construct the seasonal adjustment coefficient matrix S of the environment-structure mapping to reflect the structural response under different seasonal conditions.

[0090] 4) Construction of spatial transfer function matrix

[0091] Based on the measured data, the response transfer characteristics between monitoring points are calculated, the signal propagation path and its attenuation law are identified, and a spatial transfer function matrix reflecting the dynamic characteristics of the structure is constructed.

[0092] 5) Feature extraction and fingerprint set generation

[0093] The high-dimensional monitoring data is subjected to dimensionality reduction and feature extraction, and methods such as principal component analysis (PCA) and independent component analysis (ICA) are used to identify the feature combinations that affect the structural health status and form a feature fingerprint set that characterizes the structural health status.

[0094] Step 2: Adaptively update the decision rules to adapt to the state evolution of the bridge during its service life.

[0095] This step achieves adaptive updating of the judgment rules by constructing a dynamic envelope system and a differentiated weight system to adapt to the state evolution of the bridge during its service period. The specific implementation is as follows:

[0096] 1) Construction of multi-level dynamic envelope system

[0097] Based on the historical monitoring data of the bridge during normal operation, a multi-level dynamic envelope system integrating time evolution and periodic change factors is constructed. The specific steps are as follows:

[0098] Decompose each monitoring parameter sequence in the time domain and extract four components: long-term trend T(t), seasonal cycle S(t), daily variation cycle D(t) and random fluctuation R(t);

[0099] For T(t), calculate the dynamic trend envelope [T u (t),T I (t)], reflecting the long-term performance change characteristics of the structure;

[0100] For S(t), a seasonal pattern library is constructed and the seasonal variation envelope is calculated [S u (t),S I (t)], and its envelope width increases near the seasonal transition point;

[0101] For D(t), considering the difference between working days and rest days, the daily variation envelope [D u (t),D I (t)];

[0102] For R(t), set the fluctuation tolerance range [R u (t),R I (t)], the width of the allowable interval is proportional to the historical fluctuation range of the monitoring parameter;

[0103] The envelopes of each level are synthesized to form the comprehensive judgment envelope of monitoring parameters [U u (t),U I (t)], for anomaly detection;

[0104] A dynamic update mechanism for envelope band parameters is designed to adjust the width and baseline of each level of envelope band according to structural aging characteristics, maintenance activities and changes in the use environment, reflecting the natural evolution of structural performance over time.

[0105] 2) Environmental impact correction

[0106] The environment-structure mapping function library is used to correct the environmental impact of the measured structural response parameters, eliminate the structural response changes caused by environmental fluctuations, and obtain the residual response after eliminating the environmental interference.

[0107] 3) Establishment of a differentiated weighting system

[0108] According to the sensitivity and importance of each monitoring parameter in the characterization of the structural state, a differentiated weighting system of spatial partitioning and parameter classification is established. The specific steps are as follows:

[0109] Based on the structural function of the bridge, the monitoring area is divided into the main beam area, the bridge tower area, the anchorage area and the auxiliary component area, and the spatial weight coefficient w is assigned to each area according to its importance to the overall structural safety. i ,where,i=1,2,...,N, where N represents the number of monitoring areas, i.e., the number of areas under the functional division of the bridge structure;

[0110] According to the sensitivity analysis results of monitoring parameters to structural damage, the monitoring parameters are divided into displacement, strain, vibration and environment categories, and the parameter categories are assigned weights w j , where,j=1,2,...,M, M represents the number of monitoring parameter categories, i.e. the parameter types divided based on the sensitivity of structural damage;

[0111] According to the importance difference of the same category parameters in different spatial positions, a position-parameter joint weight matrix W = {w ij}, where w ij =w i ×w j ×α ij , α ij is the adjustment factor;

[0112] A weight adaptive update mechanism is designed to dynamically adjust the weight coefficient of each monitoring point based on the historical accuracy and false alarm rate of anomaly detection.

[0113] 4) Comprehensive deviation calculation and judgment rule optimization

[0114] The environmentally corrected structural response parameters are compared with the reference envelope in real time, and the comprehensive deviation is calculated using the differentiated weights. The specific steps are as follows:

[0115] Corrected response And synchronously read the corresponding upper and lower limits of the reference envelope

[0116] For each structural response parameter, the standardized deviation is calculated according to the following formula:

[0117]

[0118] Among them, δ r (t) represents the standardized deviation of the monitoring parameter r at time t;

[0119] According to the sensitivity of the structural response parameter r to the overall state perception and the importance of the structure, the corresponding weight coefficient is set and dynamically updated according to the structural stage characteristics;

[0120] The deviation of each response parameter is weighted and summarized to calculate the overall deviation index D of the structure total (t).

[0121] The benchmark data set and envelope parameters are updated regularly to highlight recent operational characteristics. Based on the differences in performance characteristics during the service life of bridges, a phased judgment standard library is constructed to dynamically adjust the abnormal judgment threshold and optimize the judgment rule parameters.

[0122] Step 3: Perform intelligent attribution analysis on anomalies in monitoring data

[0123] This step implements intelligent attribution analysis of anomalies in monitoring data by establishing a probabilistic reasoning framework and anomaly pattern feature library. The specific implementation is as follows:

[0124] 1) Establishment of probabilistic reasoning framework

[0125] A conditional probability table reflecting the relationship between environmental conditions, material status, load characteristics and the response of each monitoring point is established to form a complete probabilistic reasoning framework.

[0126] 2) Construction of abnormal pattern feature library

[0127] Summarize the characteristic patterns of different types of anomalies and build an abnormal pattern feature library: collect and organize historical abnormal event data, including anomalies caused by environmental factors, load exceeding, structural damage and sensor failure; for each type of abnormal event, extract its time domain feature vector FT, frequency domain feature vector FF and spatial distribution feature vector FS; summarize the recovery characteristics of different types of anomalies, and define the recovery index IR to quantify the degree of recovery; classify historical abnormal events into patterns to form an abnormal pattern set M, and establish a mapping relationship table R(M→C) between abnormal patterns and possible causes; design an abnormal pattern matching mechanism based on similarity to compare newly detected abnormal events with the patterns in the feature library.

[0128] 3) Abnormal contribution decomposition and analysis

[0129] The abnormal response in the monitoring data is decomposed into three parts: environmental factors, load factors and structural performance changes. The contribution of each factor to the abnormality is quantitatively analyzed, and the real abnormality caused by the structural performance change is identified and evaluated.

[0130] 4) Temporal and spatial feature analysis

[0131] Based on the differences in the time scale of structural responses caused by different factors, the relationship between different time patterns and environmental disturbances, load changes and structural damage is analyzed; by analyzing the spatial distribution characteristics of structural responses, the spatial propagation laws of abnormal responses are identified and compared with the expected patterns to evaluate the locality or globality of the anomaly.

[0132] 5) Root cause determination and decision-making

[0133] Calculate the confidence and uncertainty of the cause of the anomaly, and determine the root cause based on the set rules to make anomaly attribution decisions.

[0134] Step 4: Jointly identify and comprehensively evaluate sudden structural damage and slow degradation processes

[0135] This step extracts the time domain and frequency domain features of the structural response to achieve joint identification and comprehensive evaluation of sudden structural damage and slow degradation processes. The specific implementation method is as follows:

[0136] 1) Time Domain Feature Analysis and Change Point Detection

[0137] Based on the time series data of monitoring points, statistical feature analysis, trend analysis and change point detection are carried out, and time domain features are extracted to identify long-term change trends, sudden damage events and dynamic response transmission laws.

[0138] 2) Frequency domain feature extraction and analysis

[0139] The time domain signal is converted to the frequency domain, the dynamic characteristic parameters of the structure are extracted, the frequency drift, modal changes and high-order harmonic components are analyzed, the stiffness changes and local damage are identified, and the spectrum feature vector is constructed for the identification of structural parameter changes.

[0140] 3) Identification and location of sudden injuries

[0141] A quantitative evaluation index system for sudden damage identification is established, combined with the spatial distribution characteristics of the damage location, to achieve rapid response and precise positioning of sudden damage.

[0142] 4) Identification and prediction of slow degradation

[0143] Extract the slowly varying characteristics of structural parameters, identify the overall stiffness attenuation and material degradation process, calculate the degradation rate and cumulative effect, and predict the future evolution trend of structural performance.

[0144] 5) Collaborative representation and differentiation

[0145] Construct a unified feature space that integrates multi-dimensional features to achieve collaborative characterization and effective distinction between sudden damage and slow degradation.

[0146] 6) Comprehensive evaluation and report generation

[0147] Based on the above analysis results, the impact of damage and degradation on structural safety is quantitatively evaluated, the structural risk index is calculated, and a comprehensive assessment report containing qualitative judgments and quantitative indicators is generated.

[0148] Step 5: Quantitatively analyze the degree of deviation from the bridge health status and generate a standardized bridge health index

[0149] This step achieves quantitative assessment and standardized representation of bridge health status by constructing a finite element reference model and a multi-dimensional deviation index system. The specific implementation is as follows:

[0150] 1) Finite element reference model construction

[0151] Based on multi-physics field coupling theory and structural characteristic parameters, a high-precision finite element reference model is constructed that comprehensively considers the geometric shape, material properties, boundary conditions and environmental influences to truly reflect the service behavior of the bridge.

[0152] 2) Establishment of a multidimensional deviation indicator system

[0153] The measured data are compared and analyzed with the predicted values ​​of the reference model, and a multi-dimensional deviation index system including frequency domain characteristics, time domain characteristics and energy distribution characteristics is established to fully capture the abnormal status of the bridge.

[0154] 3) Health index calculation and mapping

[0155] Based on the spatiotemporal distribution of deviation indicators at each monitoring point, a comprehensive scoring mechanism considering structural importance, parameter sensitivity, and anomaly persistence is constructed, and the multidimensional deviation indicators are mapped into a standardized health index in the range of [0-100].

[0156] 4) Future state probability prediction

[0157] Combining seasonal changes with long-term deterioration factors, time series analysis, machine learning and other methods are used to make probabilistic predictions of the future state of the bridge.

[0158] 5) Health Index Decomposition and Positioning

[0159] The overall health index is decomposed into key component-level and functional unit-level indicators to achieve health status mapping from macro to micro, accurately locate potential problem areas, and quantitatively evaluate their impact on the overall structural safety.

[0160] Through the above steps, the present invention realizes comprehensive monitoring of road and bridge parameters and intelligent abnormality detection, can capture structural state changes in real time, accurately identify abnormal causes, and provide a scientific basis for bridge maintenance decisions.

[0161] Finally, it should be noted that the above embodiments are intended only to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art will appreciate that modifications may be made to the technical solutions described in the above embodiments, or that some of the technical features may be replaced with equivalents; such modifications or replacements do not deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for detecting anomalies of road and bridge parameters, characterized in that: The following steps are involved: Step 1: Establish the spatiotemporal correlation and data mapping relationship between monitoring points; Step 2: Adaptively update the decision rules to adapt to the state evolution of the bridge during its service life. Step 3: Conduct intelligent attribution analysis on anomalies in monitoring data; Step 4: Jointly identify and comprehensively evaluate sudden structural damage and slow degradation processes; Step 5: Quantitatively analyze the degree of deviation of the bridge health status and generate a standardized bridge health index.

2. The method for detecting anomalies of road and bridge parameters according to claim 1, characterized in that: Step 1 includes the following steps: Structural response parameters and environmental impact parameters are collected through a distributed sensor network and pre-processed to form a structured multi-dimensional time series data set; Based on the physical structural characteristics of the bridge, a spatiotemporal correlation network is constructed to reflect the correlation between the spatial topology and time series of the monitoring points; Identify and quantify the impact of environmental factors on structural response and its hysteresis effect, and form an environment-structure mapping function library; Based on measured data, the response transfer characteristics between monitoring points are calculated, the signal propagation path and its attenuation law are identified, and a spatial transfer function matrix reflecting the structural dynamic characteristics is constructed; The high-dimensional monitoring data is reduced in dimension and features are extracted, and the feature combination that affects the structural health status is identified to form a feature fingerprint set that characterizes the structural health status.

3. The method for detecting anomalies of road and bridge parameters according to claim 2, characterized in that: The construction of the spatiotemporal correlation network includes: Based on the physical structure layout of the bridge, the monitoring points are mapped into a three-dimensional coordinate system to construct an initial spatial topology model; Calculate the Euclidean distance and structural connectivity between monitoring points to form a spatial adjacency matrix; For different types of monitoring parameters, the mutual correlation coefficient, Granger causality coefficient and mutual information entropy of the time series are calculated respectively to quantitatively characterize the strength of the time series correlation; The spatial adjacency relationship is integrated with the temporal correlation strength to generate a weighted spatiotemporal correlation network, where the edge weight reflects the degree of physical correlation between monitoring points; Identify community structures and key nodes in the network, and calculate global and local characteristic parameters of the network, including average path length, clustering coefficient, centrality and robustness indicators; With the continuous accumulation of monitoring data, the edge weights and topological structure of the spatiotemporal correlation network are regularly updated to dynamically reflect the changes in the correlation characteristics of the bridge structure over time.

4. The method for detecting anomalies of road and bridge parameters according to claim 2, characterized in that: The process of establishing the environment-structure mapping function library includes: Establish a linear mapping relationship between environmental parameters and structural responses, and calculate the sensitivity coefficient of each environmental factor; The response delay parameter τ is introduced to quantify the time lag effect between environmental changes and structural responses. For temperature field changes, the short-term response lag τ1 and the long-term response lag τ2 are set; Establish a nonlinear mapping function f(E,τ) between environmental parameters and structural responses, where E represents the environmental parameter vector; Based on seasonal analysis, the seasonal adjustment coefficient matrix S of the environment-structure mapping is constructed to reflect the differences in structural responses under different seasonal conditions.

5. The method for detecting anomalies of road and bridge parameters according to claim 1, characterized in that: Step 2 includes the following steps: Based on the historical monitoring data of the bridge during normal operation, a multi-level dynamic envelope system integrating time evolution and periodic change factors is constructed as the basic threshold for subsequent anomaly identification. Using the environment-structure mapping function library, the measured structural response parameters are corrected for environmental influences to obtain the residual response after eliminating environmental interference, eliminating the structural response changes caused by environmental fluctuations; According to the sensitivity and importance of each monitoring parameter in the characterization of the structural state, a differentiated weighting system of spatial partitioning and parameter classification is established; The environmentally corrected structural response parameters are compared with the reference envelope in real time, and the comprehensive deviation is calculated using differentiated weights; Regularly update the benchmark dataset and envelope parameters to highlight recent operational characteristics and dynamically adapt to performance changes caused by aging of the bridge; In view of the differences in performance characteristics of bridges during their service life, a stage-by-stage judgment standard library is constructed, and the abnormal judgment threshold is dynamically adjusted based on the service life, maintenance records and performance evaluation results, and the judgment rule parameters are optimized.

6. The method for detecting anomalies of road and bridge parameters according to claim 1, characterized in that: The establishment of the differentiated weight system includes: Based on the structural function of the bridge, the monitoring area is divided into the main beam area, the bridge tower area, the anchorage area and the auxiliary component area, and the spatial weight coefficient w is assigned to each area according to its importance to the overall structural safety. i ; According to the sensitivity analysis results of monitoring parameters to structural damage, the monitoring parameters are divided into displacement, strain, vibration and environment categories, and the parameter categories are assigned weights w j ; According to the importance difference of the same category parameters in different spatial positions, a position-parameter joint weight matrix W = {w ij }, where w ij =w i ×w j ×α ij , α ij is the adjustment factor; A weight adaptive update mechanism is designed to dynamically adjust the weight coefficient of each monitoring point according to the historical accuracy and false alarm rate of anomaly detection, thereby optimizing the accuracy of anomaly judgment.

7. The method for detecting anomalies of road and bridge parameters according to claim 1, characterized in that: Step 3 includes the following steps: Establish a conditional probability table that reflects the relationship between environmental conditions, material state, load characteristics and the response of each monitoring point, forming a complete probabilistic reasoning framework; Summarize the characteristic patterns of different types of anomalies and record the spatiotemporal distribution characteristics, duration, and recovery characteristics of multi-parameter responses to form an anomaly pattern feature library; Decompose abnormal responses in monitoring data into three parts: environmental factors, load factors, and structural performance changes. Quantitatively analyze the contribution of each factor to the abnormality, and identify and evaluate the true abnormality caused by structural performance changes. Based on the differences in the time scale of structural responses caused by different factors, the relationship between different time patterns and environmental disturbances, load changes and structural damage is analyzed; By analyzing the spatial distribution characteristics of the structural response, the spatial propagation pattern of the abnormal response is identified, and the measured spatial distribution is compared with the expected pattern to evaluate the localization or global nature of the anomaly; Calculate the confidence and uncertainty of the cause of the anomaly, and determine the root cause based on the set rules to make anomaly attribution decisions.

8. The method for detecting anomalies of road and bridge parameters according to claim 1, characterized in that: The construction of the abnormal pattern feature library includes: Collect and organize historical abnormal event data, including abnormalities caused by environmental factors, abnormalities caused by excessive loads, abnormalities caused by structural damage, and abnormalities caused by sensor failures; For each type of abnormal event, its time domain feature vector FT = {duration, rise time, recovery time, peak intensity, root mean square value, kurtosis, skewness, and form factor} is extracted to describe the time evolution characteristics of the abnormality; Extract the frequency domain feature vector FF = {spectral center of gravity, frequency band energy distribution, main frequency offset, harmonic ratio, spectrum entropy} of the abnormal event to describe the frequency characteristics of the abnormality; Extract the spatial distribution feature vector FS = {influence range, attenuation gradient, propagation speed, spatial correlation, principal component direction} of the abnormal event to describe the spatial diffusion characteristics of the abnormality; Summarize the recovery characteristics of different types of anomalies, including fully recoverable anomalies, partially recoverable anomalies and unrecoverable anomalies, and define the recovery index IR∈[0,1] to quantify the degree of recovery; Classify historical abnormal events into patterns and form an abnormal pattern set M = {M1, M2, ..., M k }, each pattern Mg contains a feature vector [FT, FF, FS] and a category label; Establish a mapping table R(M→C) between abnormal patterns and possible causes to provide attribution reference for newly discovered abnormalities; Design a similarity-based anomaly pattern matching mechanism to compare newly detected anomalies with the patterns in the feature library to identify the most likely anomaly type.

9. The method for detecting anomalies of road and bridge parameters according to claim 1, characterized in that: Step 4 includes the following steps: Based on the time series data of monitoring points, statistical feature analysis, trend analysis and change point detection are carried out to extract time domain features to identify long-term change trends, sudden damage events and dynamic response transmission laws; Convert the time domain signal to the frequency domain, extract the dynamic characteristic parameters of the structure, analyze the frequency drift, modal changes and high-order harmonic components, identify stiffness changes and local damage, and construct a spectrum feature vector for structural parameter change identification; Establish a quantitative evaluation index system for sudden damage identification, combined with the spatial distribution characteristics of the damage location, to achieve rapid response and accurate positioning of sudden damage; Extract the slowly varying characteristics of structural parameters, identify the overall stiffness attenuation and material degradation process, calculate the degradation rate and cumulative effect, and predict the future evolution trend of structural performance; Construct a unified feature space that integrates multi-dimensional features to achieve collaborative characterization and effective differentiation of sudden damage and slow degradation; Based on the above analysis results, the impact of damage and degradation on structural safety is quantitatively evaluated, the structural risk index is calculated, and a comprehensive assessment report containing qualitative judgments and quantitative indicators is generated.

10. The method for detecting anomalies of road and bridge parameters according to claim 1, characterized in that: Step 5 includes the following steps: Based on multi-physics coupling theory and structural characteristic parameters, a high-precision finite element reference model is constructed that comprehensively considers geometry, material properties, boundary conditions, and environmental influences to truly reflect the service behavior of the bridge; Comparing and analyzing measured data with the reference model predictions, a multi-dimensional deviation index system was established that included frequency domain characteristics, time domain characteristics, and energy distribution characteristics to fully capture abnormal bridge conditions. Based on the spatiotemporal distribution of deviation indicators at each monitoring point, a comprehensive scoring mechanism considering structural importance, parameter sensitivity, and anomaly persistence was constructed, and the multidimensional deviation indicators were mapped into a standardized health index in the range of [0-100]. Combine seasonal changes with long-term deterioration factors to make probabilistic predictions of the future state of the bridge; The overall health index is decomposed into key component-level and functional unit-level indicators to achieve health status mapping from macro to micro, accurately locate potential problem areas, and quantitatively evaluate their impact on the overall structural safety.

Images