Structure damage identification system and method fusing damage mechanism forward research and monitoring data reverse inversion

Abstract

The invention discloses a structure damage identification system and method fusing damage mechanism forward research and monitoring data reverse inversion. A railway station building structure is prone to damage accumulation under the action of multiple factors such as extreme loads, fatigue and corrosion, operation safety is threatened, and an efficient and accurate health monitoring technology is needed urgently. The method comprises the following steps: analyzing a forward damage mechanism; constructing a sensor network; performing noise reduction, filtering and normalization processing on the original data acquired by the sensor, and storing the preprocessed data in a database; carrying out reverse damage identification analysis; and performing fusion evaluation and verification. According to the method, the accuracy, the reliability and the efficiency of damage identification are improved by fusing a forward mechanism and reverse analysis.

Description

Technical Field

[0002] This invention belongs to the field of civil engineering structural health monitoring and safety assessment technology, specifically involving a structural damage identification system and method that integrates forward research on damage mechanisms and inverse inversion of monitoring data. Background Technology

[0003] With the large-scale construction and operation of railway station buildings in my country, their structures are prone to damage accumulation under extreme loads, fatigue, corrosion and other factors, threatening operational safety. Therefore, efficient and accurate health monitoring technology is urgently needed.

[0004] Currently, methods in the field of structural health monitoring can be mainly divided into two categories: one is forward analysis based on physical mechanisms; the other is reverse identification based on monitoring data. Forward analysis relies on establishing accurate theoretical models (such as finite element models) to predict damage evolution, but its accuracy heavily depends on the completeness of the model itself and the rationality of the boundary conditions. For statically indeterminate, multi-degree-of-freedom complex structures such as railway station buildings, the model has significant uncertainty and is difficult to fully reflect the actual state of the structure. Reverse identification, on the other hand, starts directly from sensor data and identifies damage through signal processing and algorithms. However, traditional data-driven methods often lack clear physical meaning, and are prone to false alarms and missed alarms when monitoring data is limited or there is significant noise interference. They also have poor generalization ability and are difficult to explain the mechanical mechanisms of damage. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this invention provides a structural damage identification system and method that integrates forward research on damage mechanisms with inverse inversion of monitoring data. By integrating forward mechanisms and inverse analysis, the accuracy, reliability, and efficiency of damage identification are improved.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A structural damage identification method that integrates forward research on damage mechanisms with inverse inversion of monitoring data includes the following steps: Step 1: Analysis of the forward damage mechanism; Step 2: Sensor Network Construction; Step 3: Perform noise reduction, filtering, and normalization on the raw data collected by the sensor, and store the preprocessed data in the database; Step 4: Reverse damage identification and analysis; Step 5: Integration Assessment and Validation.

[0007] Furthermore, step one specifically includes: Step 1.1: Establish an integrated damage mechanics model based on structural design information; Based on the station building's structural design information, an integrated damage mechanics model was established to simulate static loads, dynamic loads, and the structural response under these loads. Step 1.2: Identify and define the types of durability loads faced by the structure, and integrate the durability loads in the established integrated damage mechanics model based on the material service environment and performance degradation data. Step 1.3: Based on the structural response data under different load conditions obtained in the above steps, analyze the explicit relationship between load and damage; simulate the state of key components under different stiffness reduction degrees using finite element software, analyze the variation law of structural natural frequency, mode shape and strain energy parameters, and determine the damage index most sensitive to this structure. Step 1.4: Apply phonon crystal theory for analysis.

[0008] Furthermore, step 1.4 includes the following steps: Step 1.4.1: Structure simplification and unit cell extraction; Step 1.4.2: Bandgap characteristic calculation and analysis; Step 1.4.3: Analyze the bandgap frequency distribution obtained in Step 1.4.2, and adjust the geometric parameters or material parameters of the unit cell to make the structural bandgap cover the dominant frequency band of environmental fluctuation loads and seismic motion. Step 1.4.4: Monitoring hotspot prediction and sensor deployment.

[0009] Furthermore, step two specifically involves: The sensors include an accelerometer, a strain sensor, and an ambient temperature and humidity sensor. The accelerometer is arranged based on the damage-sensitive areas determined by forward analysis and the monitoring hotspots predicted by phonon crystal theory. The strain sensor is arranged on the surface of key components in the damage-sensitive area to collect the static and dynamic strain responses of the structure. The ambient temperature and humidity sensor is distributed throughout the structural space to monitor the service environment parameters of the structure.

[0010] Furthermore, step four specifically includes: Step 4.1: Construction of multi-scale model and surrogate model; Step 4.2: Damage identification based on monitoring data.

[0011] Furthermore, step 4.1 is detailed as follows: Step 4.1.1: Establish a multi-scale finite element model; Step 4.1.2: Model correction; Step 4.1.3: Construct the proxy model and parameter-response database.

[0012] Furthermore, step 4.2 specifically includes: Based on the parameter-response database constructed in step 4.1 and the real-time monitoring data, a reverse analysis algorithm is used to identify the damage state. For different load conditions, the following methods are employed: 4.2.1 Damage identification based on monitoring data during normal use: 4.2.1.1 Constructing a probabilistic framework: The surrogate model obtained in step 4.1 is used as the forward prediction model; the likelihood function is constructed based on the difference between real-time monitoring data and model prediction values, and the stiffness reduction coefficient of the member is used as the damage parameter to be identified; 4.2.1.2 Bayesian Inversion and Sampling: By introducing a Bayesian inference model, the damage identification problem is transformed into solving the posterior probability distribution of damage parameters; the Metropolis-Hastings algorithm in the Markov chain Monte Carlo sampling method is used to search in the parameter space to solve the posterior probability distribution of the above damage parameters. 4.2.1.3 Damage state determination: The combination of stiffness reduction coefficients with the highest probability in the posterior probability distribution corresponds to the most likely damage state of the structure; 4.2.2 Damage identification from monitoring data of earthquakes, accidental or extreme load conditions: 4.2.2.1 Triggering and Signal Analysis; 4.2.2.2 Defect state detection and damage localization.

[0013] Furthermore, step five specifically includes: The damage locations identified through reverse engineering are cross-validated with the damage-sensitive areas determined through forward analysis. Simultaneously, the damage degree quantitatively identified through reverse analysis is fed back into the integrated damage mechanics model to update the parameters of the corresponding parts of the model.

[0014] A structural damage identification system that integrates forward damage mechanism research and inverse inversion of monitoring data includes a sensor network, a data acquisition and preprocessing module, a forward damage mechanism analysis module, an inverse damage identification and analysis module, and a fusion evaluation module.

[0015] The beneficial effects of this invention are: 1) This invention clarifies physical laws through forward damage mechanism research, guides the reverse analysis process, overcomes the shortcomings of the purely data-driven method with unclear physical meaning, and significantly improves the reliability and interpretability of damage identification results; 2) This invention effectively utilizes monitoring data by applying multi-scale model correction and Bayesian theory in reverse analysis, reducing the dependence on the absolute accuracy of the mechanism model, and can quantitatively handle uncertainties and noise in actual engineering, thereby improving the robustness and accuracy of the method. 3) This invention introduces phonon crystal theory to handle extreme load conditions, solving the problem that traditional methods are prone to failure under such events, and realizing damage identification under all working conditions and multiple disasters. 4) This invention forms a complete technical route from mechanism research, data identification to comprehensive evaluation, and is especially suitable for health monitoring and status assessment of complex structures such as railway stations, and has important engineering application value. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the operation of the present invention. Figure 2 This is a technical roadmap for the present invention. Detailed Implementation

[0017] The present invention will now be described in detail with reference to specific embodiments.

[0018] This invention employs an integrated damage mechanics model (LDM) to clarify the explicit relationship between load and damage, guiding the placement of monitoring points and extraction of sensitive parameters. It combines a multi-scale finite element model and an approximate surrogate model to construct an implicit relationship database between structural response and component performance. Furthermore, it innovatively introduces Bayesian theory to process data under normal service conditions and phonon crystal theory to process data under extreme loads. Finally, through the deep fusion of forward mechanism and reverse identification, it achieves accurate and efficient identification and assessment of damage in complex structures.

[0019] like Figure 1 , 2 As shown, the present invention provides a structural damage identification method that integrates forward research on damage mechanisms with inverse inversion of monitoring data, comprising the following steps: Step 1: Analysis of the positive damage mechanism, specifically: Step 1.1: Based on structural design information, establish an integrated damage mechanics model (LDM). Based on the station building's structural design information (including structural design drawings and BIM models), an integrated damage mechanics model (LDM) was established using finite element software. This model is characterized by its compatibility with static and dynamic load conditions and its ability to characterize the explicit relationship between load and damage. It simulates static loads (roof dead load, equipment weight), dynamic loads (train vibration, wind load), and the structural response under these loads, including the structure's natural frequency, mode shape, displacement, strain, and acceleration response. Step 1.2: Identify and define the types of durability loads faced by the structure, such as corrosion, freeze-thaw cycles, and material aging; based on material service environment and performance degradation data, in the established integrated damage mechanics model, by introducing time-varying material performance models (such as steel corrosion rate models and concrete section loss models) or stiffness / strength degradation models, quantify the long-term impact of durability loads on mechanical parameters such as stiffness and strength of structural members, and embed them into the LDM in the form of time-varying damage factors to achieve the integration of durability loads; Step 1.3: Based on the structural response data under different load conditions obtained in the above steps, analyze the explicit relationship between load and damage; simulate the state of key components (such as beams and columns) under different stiffness reduction degrees (e.g., reduction of 20%, 40%, and 60%) using finite element software, analyze the variation law of the structure's natural frequency, mode shape, and strain energy parameters, determine the damage indicators most sensitive to this structure, such as the rate of change of the third-order frequency and the strain energy difference in the mid-span region, and accurately locate the damage-sensitive areas of the structure (such as certain key nodes and connection parts); these sensitive indicators and areas will serve as the key basis for subsequent deployment of monitoring systems and damage identification.

[0020] Step 1.4: Apply phonon crystal theory for analysis, including the following steps: Step 1.4.1: Structural simplification and unit cell extraction: The regular column grid and roof grid structure in the station building are simplified into a two-dimensional or three-dimensional periodic lattice model; based on this, the representative unit cell structure (i.e., "unit cell") is extracted as the basic analysis unit. Step 1.4.2: Bandgap characteristic calculation and analysis: Based on the above unit cell model, the finite element method is used, combined with Bloch's theorem and periodic boundary conditions, to calculate the band structure of the structure; by analyzing the band structure, the complete bandgap or directional bandgap (i.e. "forbidden band") that exists in a specific frequency range (e.g., 5-15Hz) is determined. Step 1.4.3: Analyze the bandgap frequency distribution obtained in Step 1.4.2. By adjusting the geometric parameters of the unit cell (such as the cross-section of the rod and the mass of the nodes) or the material parameters, make the structural bandgap cover the dominant frequency bands of the main environmental fluctuation loads (such as train vibration) and earthquake motion. Based on this, optimize the scheme in the structural design stage to improve the structure's dual control performance of vibration and earthquake (vibration-earthquake dual control). Step 1.4.4: Monitoring Hotspot Prediction and Sensor Deployment: Based on bandgap analysis, point defects or local resonant units are introduced into the periodic structure to calculate and analyze the defect state frequency of the structure and its corresponding vibration modes; the regions in the vibration modes where the displacement response or strain energy is significantly increased are identified as monitoring hotspots; when actually deploying sensors, monitoring equipment such as accelerometers are mainly placed in these predicted "monitoring hotspot" regions to preferentially capture abnormal signals related to damage.

[0021] Step 2: Sensor network construction, specifically: The sensors include accelerometers, strain sensors, and ambient temperature and humidity sensors. The accelerometers are arranged based on damage-sensitive areas determined by forward analysis and monitoring hotspots predicted by phonon crystal theory. They are primarily used to collect vibration acceleration signals of the structure under load to analyze its dynamic characteristics (such as frequency and mode shape). The strain sensors are arranged on the surfaces of key components in the damage-sensitive areas to collect the static and dynamic strain responses of the structure. The ambient temperature and humidity sensors are distributed throughout the structural space to monitor the structure's service environment parameters. All sensors are connected to the data acquisition and preprocessing module via wired or wireless transmission.

[0022] Step 3: Perform noise reduction, filtering, and normalization on the raw data collected by the sensor, and store the preprocessed data in the database; Step 4: Reverse damage identification and analysis, specifically: Step 4.1: Construction of multi-scale models and surrogate models, aiming to establish a surrogate model database that can efficiently and accurately reflect structural performance. The specific operations are as follows: Step 4.1.1: Establish a multi-scale finite element model: Based on the structural design drawings, establish an overall macroscopic model of the station structure, and establish local refined models for key stress nodes (such as beam-column connections, supports, etc.) to form a multi-scale finite element model; Step 4.1.2: Model Correction: Using construction acceptance data (such as material strength and geometric dimensions) and initial monitoring data (such as initial frequency and strain), the boundary conditions (such as support constraints) and material parameters (such as elastic modulus and density) of the multi-scale finite element model are corrected to reduce the difference between the model and the actual structure. Step 4.1.3: Constructing the surrogate model and parameter-response database: Using the Stochastic Response Surface Method, the stiffness reduction factor of the sensitive members is used as the input parameter, and the sensitive damage index (such as the natural frequency of the structure and the strain at a specific measuring point) is used as the output response. Parametric analysis is performed based on the modified multi-scale finite element model to construct a high-precision approximate surrogate model (i.e., meta-model). This process generates a parameter-response database covering various potential damage states, which establishes the mapping relationship between the structural physical parameters and response characteristics.

[0023] Step 4.2: Damage identification based on monitoring data, specifically: Based on the parameter-response database constructed in step 4.1 and the real-time monitoring data, a reverse analysis algorithm is used to identify the damage state. For different load conditions, the following methods are employed: 1) 4.2.1 Damage identification from normal operating condition monitoring data (based on Bayesian theory): 4.2.1.1 Constructing a probabilistic framework: The surrogate model obtained in step 4.1 is used as the forward prediction model; the likelihood function is constructed based on the difference between real-time monitoring data (such as frequency and strain) and model prediction values, and the stiffness reduction factor of the member is used as the damage parameter to be identified; 4.2.1.2 Bayesian Inversion and Sampling: By introducing a Bayesian inference model, the damage identification problem is transformed into solving the posterior probability distribution of damage parameters. The Metropolis-Hastings (MH) algorithm in the Markov Chain Monte Carlo (MCMC) sampling method is used to efficiently search in the parameter space and solve for the posterior probability distribution of the above damage parameters. 4.2.1.3 Damage State Determination: The combination of stiffness reduction coefficients with the highest probability in the posterior probability distribution corresponds to the most likely damage state of the structure (including damage location and degree). The feasibility and accuracy of this damage identification method are verified through a finite element model. 2) 4.2.2 Damage identification from monitoring data of earthquake, accidental, or extreme load conditions (based on phonon crystal theory): 4.2.2.1 Triggering and Signal Analysis: When the monitoring system detects extreme events such as earthquakes, this analysis process is automatically triggered; the structure is subjected to full-band excitation covering key band gaps (such as 5-15Hz) (environmental excitation or active excitation devices can be used), and the vibration response signal of the structure is collected through a sensor network; 4.2.2.2 Defect State Detection and Damage Location: The above response signal is compared with the reference bandgap spectrum pre-established based on phononic crystal theory under the structural health state; if a new conduction peak (i.e. "defect state") appears in the original forbidden band (bandgap) frequency range, it is determined that the vibration mode corresponding to this frequency has changed, and then the approximate area where the damage is located is located.

[0024] Step 5: Integration, evaluation, and verification, specifically: The damage locations identified through reverse engineering are cross-validated with the damage-sensitive areas determined by forward analysis. If they match, the reliability of the results is significantly improved. At the same time, the damage degree (such as stiffness reduction factor) quantitatively identified by reverse analysis is fed back into the integrated damage mechanics model (LDM) to update the parameters of the corresponding parts of the model, making the model closer to the actual state of the structure. This completes a closed-loop process of "forward guidance - reverse identification - feedback update", ultimately achieving accurate identification and comprehensive assessment of damage location, degree and type.

[0025] This invention also provides a structural damage identification system that integrates forward research on damage mechanisms and inverse inversion of monitoring data, including a sensor network, a data acquisition and preprocessing module, a forward damage mechanism analysis module, an inverse damage identification and analysis module based on a multi-scale finite element model and an approximate surrogate model, and a fusion evaluation module.

[0026] 1) Data Acquisition and Preprocessing Module The data acquisition and preprocessing module receives raw data from the sensors, performs noise reduction, filtering and normalization, and stores the preprocessed data in the database.

[0027] 2) Forward Damage Mechanism Analysis Module The forward damage mechanism analysis module is used to establish damage evolution models of structures under static, dynamic, and durability loads, and to identify damage-sensitive locations and sensitive characteristic parameters. Applying phonon crystal theory, it simplifies the crystal lattice based on the structure's periodic characteristics and guides the design of structural vibration-seismic performance and the placement of monitoring hotspots through bandgap analysis. In this module, the integrated damage mechanics model (LDM) must be compatible with both static and dynamic load conditions and be verified and optimized using experimental data. The module applies phonon crystal theory, including calculating the forbidden zone of the structure through harmonic superposition solutions and the Bloch boundary assumption. It artificially designs cell parameters to position the structural bandgap within the environmental fluctuation load frequency band and the low-frequency band of seismic waves, thereby improving the structure's vibration-seismic dual control performance. Based on the characteristics of the forbidden zone and defect states, it predicts the vibration input amplification region and identifies it as a monitoring hotspot.

[0028] 3) Reverse Damage Recognition and Analysis Module This module is used to establish and correct multi-scale finite element models of structures using design drawings and measured data, and to construct an implicit relationship database between structural response and component performance. Based on this database, Bayesian theory algorithms are used to process normal service condition monitoring data, and phonon crystal defect state analysis methods are used to process seismic, accidental, or extreme load state monitoring data to achieve damage identification. In the reverse damage identification analysis module, the construction of the multi-scale finite element model requires correction using design drawings and measured data. Specifically, for normal service condition monitoring data, the damage identification based on Bayesian theory in the reverse damage identification analysis module involves: using matrix perturbation, least squares, or stochastic response surface methods to determine the relationship between structural parameters and model response; using the MH sampling method in Markov chain Monte Carlo (MCMC) sampling to simplify calculations; introducing a Bayesian model based on the likelihood function and sampling method to achieve damage identification; and verifying the feasibility and accuracy of the method through finite element models. In the reverse damage identification and analysis module, for monitoring data of earthquake, accidental or extreme load states, the damage identification based on phonon crystal theory is specifically as follows: using the periodicity of structural column grids, the propagation of traffic vibration waves and seismic waves in the structure is periodically reduced and simplified to obtain the unit cell; through harmonic superposition solutions and the Bloch boundary assumption, forbidden zones and defect states are obtained; the full frequency band is input to the structure, and the difference with the response of a healthy structure is compared to realize vibration-seismic damage identification.

[0029] 4) Integration Assessment Module It is used to compare, verify and fuse the sensitive parameters determined by forward analysis with the results identified by reverse analysis, and output accurate identification results and comprehensive evaluation conclusions of damage location, degree and type.

[0030] The system in this embodiment includes a sensor network, a data acquisition and preprocessing module, a data processing center, and a status assessment and early warning module. The data acquisition and preprocessing module receives raw data from the sensors, performs noise reduction, filtering, and normalization, and stores the preprocessed data in a database. The data processing center has built-in algorithms for executing the method of this invention, including an integrated damage mechanics model analysis program, a multi-scale finite element model correction program, a Bayesian inference analysis program, and a phonon crystal theory analysis program. Based on the analysis results from the data processing center, the status assessment and early warning module outputs damage identification results, performs a safety level assessment, and issues early warning information when thresholds are exceeded.

[0031] This invention is specifically applied to railway station building structures, and its technical route includes, in sequence: monitoring equipment layout, monitoring equipment preprocessing, monitoring data storage and analysis, and structural performance evaluation.

[0032] The content of this invention is not limited to the embodiments listed. Any equivalent modifications made by those skilled in the art to the technical solutions of this invention by reading this specification are covered by the claims of this invention.

Claims

1. A structural damage identification method that integrates forward research on damage mechanisms with inverse inversion of monitoring data, characterized in that: Includes the following steps: Step 1: Analysis of the forward damage mechanism; Step 2: Sensor Network Construction; Step 3: Perform noise reduction, filtering, and normalization on the raw data collected by the sensor, and store the preprocessed data in the database; Step 4: Reverse damage identification and analysis; Step 5: Integration Assessment and Validation.

2. The structural damage identification method according to claim 1, which integrates forward research on damage mechanisms and inverse inversion of monitoring data, is characterized in that: Step one specifically involves: Step 1.1: Establish an integrated damage mechanics model based on structural design information; Based on the station building's structural design information, an integrated damage mechanics model was established to simulate static loads, dynamic loads, and the structural response under these loads. Step 1.2: Identify and define the types of durability loads faced by the structure, and integrate the durability loads in the established integrated damage mechanics model based on the material service environment and performance degradation data. Step 1.3: Based on the structural response data under different load conditions obtained in the above steps, analyze the explicit relationship between load and damage; simulate the state of key components under different stiffness reduction degrees using finite element software, analyze the variation law of structural natural frequency, mode shape and strain energy parameters, and determine the damage index most sensitive to this structure. Step 1.4: Apply phonon crystal theory for analysis.

3. The structural damage identification method according to claim 2, which integrates forward research on damage mechanisms and inverse inversion of monitoring data, is characterized in that: Step 1.4 includes the following steps: Step 1.4.1: Structure simplification and unit cell extraction; Step 1.4.2: Bandgap characteristic calculation and analysis; Step 1.4.3: Analyze the bandgap frequency distribution obtained in Step 1.4.2, and adjust the geometric parameters or material parameters of the unit cell to make the structural bandgap cover the dominant frequency band of environmental fluctuation loads and seismic motion. Step 1.4.4: Monitoring hotspot prediction and sensor deployment.

4. The structural damage identification method according to claim 3, which integrates forward research on damage mechanisms and inverse inversion of monitoring data, is characterized in that: Step two specifically involves: The sensors include an accelerometer, a strain sensor, and an ambient temperature and humidity sensor. The accelerometer is arranged based on the damage-sensitive areas determined by forward analysis and the monitoring hotspots predicted by phonon crystal theory. The strain sensor is arranged on the surface of key components in the damage-sensitive area to collect the static and dynamic strain responses of the structure. The ambient temperature and humidity sensor is distributed throughout the structural space to monitor the service environment parameters of the structure.

5. The structural damage identification method according to claim 4, which integrates forward research on damage mechanisms and inverse inversion of monitoring data, is characterized in that: Step four specifically involves: Step 4.1: Construction of multi-scale model and surrogate model; Step 4.2: Damage identification based on monitoring data.

6. The structural damage identification method according to claim 5, which integrates forward research on damage mechanisms and inverse inversion of monitoring data, is characterized in that: Step 4.1 is as follows: Step 4.1.1: Establish a multi-scale finite element model; Step 4.1.2: Model correction; Step 4.1.3: Construct the proxy model and parameter-response database.

7. The structural damage identification method according to claim 6, which integrates forward research on damage mechanisms and inverse inversion of monitoring data, is characterized in that: Step 4.2 specifically involves: Based on the parameter-response database constructed in step 4.1 and the real-time monitoring data, a reverse analysis algorithm is used to identify the damage state. For different load conditions, the following methods are employed: 4.2.1 Damage identification based on monitoring data during normal use: 4.2.1.1 Constructing a probabilistic framework: The surrogate model obtained in step 4.1 is used as the forward prediction model; the likelihood function is constructed based on the difference between real-time monitoring data and model prediction values, and the stiffness reduction coefficient of the member is used as the damage parameter to be identified; 4.2.1.2 Bayesian Inversion and Sampling: By introducing a Bayesian inference model, the damage identification problem is transformed into solving the posterior probability distribution of damage parameters; the Metropolis-Hastings algorithm in the Markov chain Monte Carlo sampling method is used to search in the parameter space to solve the posterior probability distribution of the above damage parameters. 4.2.1.3 Damage state determination: The combination of stiffness reduction coefficients with the highest probability in the posterior probability distribution corresponds to the most likely damage state of the structure; 4.2.2 Damage identification from monitoring data of earthquakes, accidental or extreme load conditions: 4.2.2.1 Triggering and Signal Analysis; 4.2.2.2 Defect state detection and damage localization.

8. The structural damage identification method according to claim 7, which integrates forward research on damage mechanisms and inverse inversion of monitoring data, is characterized in that: Step five specifically involves: The damage locations identified through reverse engineering are cross-validated with the damage-sensitive areas determined through forward analysis. Simultaneously, the damage degree quantitatively identified through reverse analysis is fed back into the integrated damage mechanics model to update the parameters of the corresponding parts of the model.

9. A structural damage identification system that integrates forward research on damage mechanisms with inverse inversion of monitoring data, characterized in that: It includes a sensor network, a data acquisition and preprocessing module, a forward damage mechanism analysis module, a reverse damage identification and analysis module, and a fusion evaluation module.

Images