Intelligent bridge reinforcement decision-making method, device and equipment based on damage identification

Abstract

The present application relates to the field of data processing technology, and discloses a method, device, and equipment for generating intelligent bridge reinforcement decisions based on damage identification. The method comprises: performing bridge damage identification and evaluation on a bridge status data set to obtain a bridge damage characteristic map; performing intelligent decision-making analysis on the bridge damage characteristic map to generate a preliminary reinforcement plan and obtain a set of candidate reinforcement strategies; performing parametric modeling and multi-scale simulation on the candidate reinforcement strategy set, and performing coupled calculation evaluation on the reinforcement effect to obtain an optimized reinforcement plan; performing virtual construction planning on the optimized reinforcement plan, generating a construction process model through an intelligent scheduling algorithm, and obtaining a target construction plan; performing real-time monitoring and feedback on the target construction plan, and dynamically optimizing the construction process through an adaptive control algorithm to obtain a target intelligent reinforcement implementation plan. The present application improves the efficiency and accuracy of intelligent bridge reinforcement decision generation based on damage identification.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a method, device and equipment for generating intelligent bridge reinforcement decisions based on damage identification. Background Art

[0002] Existing bridge reinforcement decision-making methods primarily rely on manual experience and traditional structural analysis techniques. Engineers assess the structural condition of bridges through regular inspections and non-destructive testing. They then develop reinforcement plans based on static analysis and finite element simulation. These methods typically include visual inspection, ultrasonic testing, strain measurement, and structural analysis using commercial finite element software. During the reinforcement design phase, engineers select appropriate reinforcement materials and methods based on regulatory requirements and experience, such as external prestressing, carbon fiber reinforcement, and concrete repair.

[0003] However, these traditional methods have some limitations. First, manual inspection is easily affected by subjective factors, making it difficult to detect potential structural problems in a timely manner. Second, conventional structural analysis methods find it difficult to fully consider the dynamic behavior and long-term performance evolution of bridges. Furthermore, the traditional reinforcement decision-making process lacks systematicity and intelligence, making it difficult to find the optimal balance between multiple objectives (such as safety, economy, and durability). In addition, existing construction planning and monitoring methods are often static and difficult to adapt to complex site conditions and uncertainties. These shortcomings limit the efficiency and accuracy of reinforcement decisions, which may affect the overall effect and resource utilization of the reinforcement project. Summary of the Invention

[0004] The present application provides a method, apparatus and device for generating intelligent bridge reinforcement decisions based on damage identification, which are used to improve the efficiency and accuracy of generating intelligent bridge reinforcement decisions based on damage identification.

[0005] In the first aspect, the present application provides an intelligent bridge reinforcement decision-making method based on damage identification, which includes: multi-source data collection on the bridge structure, pre-processing the collected data through a multimodal data fusion algorithm to obtain a bridge status data set; bridge damage identification and evaluation of the bridge status data set through a multi-branch neural network model to obtain a bridge damage characteristic map; intelligent decision analysis of the bridge damage characteristic map to generate a preliminary reinforcement plan to obtain a set of candidate reinforcement strategies; parametric modeling and multi-scale simulation of the candidate reinforcement strategy set, and coupled calculation evaluation of the reinforcement effect to obtain an optimized reinforcement plan; virtual construction planning of the optimized reinforcement plan, generating a construction process model through an intelligent scheduling algorithm to obtain a target construction plan; real-time monitoring and feedback of the target construction plan, and dynamic optimization of the construction process through an adaptive control algorithm to obtain a target intelligent reinforcement implementation plan.

[0006] In a second aspect, the present application provides an intelligent bridge reinforcement decision-making device based on damage identification, the intelligent bridge reinforcement decision-making device based on damage identification comprising:

[0007] The acquisition module is used to collect multi-source data of the bridge structure and pre-process the collected data through a multimodal data fusion algorithm to obtain a bridge status data set;

[0008] An evaluation module, configured to identify and evaluate bridge damage on the bridge status dataset using a multi-branch neural network model to obtain a bridge damage feature map;

[0009] An analysis module is used to perform intelligent decision-making analysis on the bridge damage characteristic map, generate a preliminary reinforcement plan, and obtain a candidate reinforcement strategy set;

[0010] A simulation module is used to perform parameterized modeling and multi-scale simulation on the candidate reinforcement strategy set, and to perform coupled calculation and evaluation on the reinforcement effect to obtain an optimized reinforcement scheme;

[0011] A planning module is used to perform virtual construction planning for the optimized reinforcement scheme, generate a construction process model through an intelligent scheduling algorithm, and obtain a target construction plan;

[0012] The feedback module is used to monitor and provide feedback on the target construction plan in real time, and dynamically optimize the construction process through an adaptive control algorithm to obtain a target intelligent reinforcement implementation plan.

[0013] The third aspect of the present application provides a computer device, wherein the memory stores machine-readable instructions executable by the processor. When the computer device is running, the processor and the memory communicate through a bus, and when the machine-readable instructions are executed by the processor, the steps of the above-mentioned intelligent bridge reinforcement decision-making method based on damage identification are performed.

[0014] The technical solution provided in this application acquires a comprehensive and high-quality bridge status dataset by collecting multi-source data from bridge structures and preprocessing it using a multimodal data fusion algorithm. This provides a reliable data foundation for subsequent analysis and effectively improves the accuracy and comprehensiveness of bridge status assessments. Secondly, a multi-branch neural network model is used to identify and assess damage in the bridge status dataset, generating a bridge damage signature map. This significantly improves the accuracy and efficiency of damage identification, enabling the timely detection of potential structural problems and providing a key basis for reinforcement decision-making. Furthermore, intelligent decision-making analysis is performed on the bridge damage signature map to generate a preliminary reinforcement plan and a set of candidate reinforcement strategies. This achieves automated and intelligent generation of reinforcement plans, significantly reducing the subjectivity and uncertainty of manual decision-making. Subsequently, an optimized reinforcement plan is obtained by performing parametric modeling and multi-scale simulation on the candidate reinforcement strategy set, followed by coupled computational evaluation. This process fully considers multiple factors, including reinforcement effectiveness, cost, and construction difficulty, ensuring the scientific and feasible nature of the reinforcement plan. Furthermore, virtual construction planning is performed on the optimized reinforcement solution. Using an intelligent scheduling algorithm, a construction process model is generated to obtain the target construction plan. This significantly improves the efficiency and rationality of construction planning, helping to reduce resource waste and time delays during construction. Finally, real-time monitoring and feedback are provided for the target construction plan, and the construction process is dynamically optimized using an adaptive control algorithm to obtain the target intelligent reinforcement implementation plan. This closed-loop control mechanism greatly enhances the adaptability and robustness of the reinforcement implementation process, effectively addressing various uncertainties and emergencies during construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 This is a schematic diagram of an embodiment of a method for generating intelligent bridge reinforcement decisions based on damage identification in an embodiment of the present application;

[0017] Figure 2 This is a schematic diagram of an embodiment of an intelligent bridge reinforcement decision-making device based on damage identification in an embodiment of the present application;

[0018] Figure 3 Schematic diagram of the structure of the computer device in the embodiment of the present application. DETAILED DESCRIPTION

[0019] The embodiments of the present application provide a method, device and equipment for generating intelligent bridge reinforcement decisions based on damage identification. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or that are inherent to these processes, methods, products or devices.

[0020] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiments of the present application, an embodiment of the intelligent bridge reinforcement decision-making method based on damage identification includes:

[0021] Step S101: collect multi-source data on the bridge structure, and pre-process the collected data using a multimodal data fusion algorithm to obtain a bridge status data set;

[0022] Step S102: performing bridge damage identification and evaluation on the bridge status data set using a multi-branch neural network model to obtain a bridge damage feature map;

[0023] Step S103: Perform intelligent decision analysis on the bridge damage characteristic map to generate a preliminary reinforcement plan and obtain a candidate reinforcement strategy set;

[0024] Step S104: performing parameterized modeling and multi-scale simulation on the candidate reinforcement strategy set, and performing coupled calculation evaluation on the reinforcement effect to obtain an optimized reinforcement scheme;

[0025] Step S105: Perform virtual construction planning on the optimized reinforcement plan, generate a construction process model through an intelligent scheduling algorithm, and obtain a target construction plan;

[0026] Step S106: Real-time monitoring and feedback are conducted on the target construction plan, and the construction process is dynamically optimized through an adaptive control algorithm to obtain a target intelligent reinforcement implementation plan.

[0027] It is understandable that the execution subject of this application can be an intelligent bridge reinforcement decision-making device based on damage identification, or a terminal or a server, which is not limited here. The embodiment of this application is described by taking the server as the execution subject as an example.

[0028] Specifically, multi-source data collection for the bridge structure is performed, including vibration signals, strain data, and displacement data. Wavelet transform is used to eliminate noise in the vibration signals, followed by fast Fourier transform (FFT) to obtain frequency domain features. Wavelet transform effectively removes high-frequency noise while preserving key signal features. Strain data is subjected to median filtering to remove outliers, followed by principal component analysis (PCA) to extract key features. Median filtering effectively removes sudden noise, while PCA reduces data dimensionality and preserves key information. Displacement data is smoothed using Kalman filtering, followed by time series analysis to obtain displacement trend features. Kalman filtering effectively estimates the state of the dynamic system and reduces the impact of measurement noise. These processed data are then multimodally fused using a tensor decomposition algorithm to form a comprehensive bridge status dataset. A multi-branch neural network model is then used to perform damage identification and assessment on the bridge status dataset. Time windowing is first performed, and the data is segmented using a sliding window algorithm. Fourier transform is then performed on each subset of time series features to obtain a frequency domain feature matrix. These frequency domain feature matrices are then processed using a convolutional neural network to extract local features, and global feature vectors are then obtained through global average pooling. Long short-term memory networks are used to capture long-term temporal features, and the self-attention mechanism further optimizes feature representation. Finally, a fully connected layer and softmax classification are used to obtain the probability distribution of damage types, and the final bridge damage feature map is obtained through threshold filtering and spatial mapping. Based on the obtained bridge damage feature map, intelligent decision analysis is performed to generate a preliminary reinforcement plan. First, a graph convolutional network is used to process the damage feature map, extract the topological relationship of the damage, and obtain damage representation vectors through graph embedding. K-means clustering analysis is performed on these vectors to group similar damages and assess their severity to determine the reinforcement priority. Applicable reinforcement technologies are then selected through knowledge graph matching to form a preliminary reinforcement plan. After these plans are checked for constraints, the decision paths for the plans are determined through multi-objective optimization and decision tree analysis. Finally, the robustness of the plans is evaluated through Monte Carlo simulation to obtain a set of candidate reinforcement strategies.

[0029] A set of candidate reinforcement strategies was subjected to parametric modeling and multi-scale simulation. First, a geometric model of the reinforcement component was generated using non-uniform rational B-spline technology, and meshing was performed to obtain a finite element model. Material properties were then assigned and boundary conditions were set to obtain a computationally ready model. Macroscale simulations were performed using nonlinear finite element methods to calculate the overall structural response. Mesoscale simulations were performed in critical areas to simulate the internal damage evolution of the material and determine the microscopic crack propagation path. Fracture mechanics analysis was used to determine the critical stress intensity factor. Finally, a reliability assessment was performed to obtain the reliability index of the reinforcement schemes, and a comprehensive ranking was performed to determine the optimized reinforcement scheme. Virtual construction planning was then performed for the optimized reinforcement schemes. The analytic hierarchy process was first used to decompose and prioritize the processes, resulting in a process dependency diagram and a key process sequence. Resource demand analysis and resource balance optimization were performed on the key process sequence to obtain a preliminary resource allocation plan. Discrete event simulation was used to assess the construction progress and generate schedule adjustment recommendations. A heuristic algorithm was used to generate multiple scheduling options, and the optimal one was determined through multi-criteria evaluation. Finally, a risk assessment was performed to determine risk mitigation strategies, which were integrated into the construction process to form a target construction plan.

[0030] Finally, the target construction plan is monitored and feedback is provided in real time. Sensor layout is planned using an optimal coverage algorithm, and the data collection frequency is set to form a real-time monitoring plan. The collected data is preprocessed using edge computing technology for anomaly detection. Bayesian network reasoning and analysis are performed on detected anomalies to generate a construction deviation report and a list of adjustment priorities. A reinforcement learning algorithm is used to generate a set of candidate adjustment plans, and simulation verification reveals the optimal adjustment plan. Finally, a specific set of construction adjustment instructions is generated and integrated into the original construction plan to achieve the target intelligent reinforcement implementation plan.

[0031] For example, consider damage identification and reinforcement decision-making for a reinforced concrete bridge. First, vibration sensors collect vibration signals from the bridge. After wavelet transform denoising, the vibration amplitude is reduced by 20%. Fast Fourier transform analysis reveals the main frequency components at 2.5 Hz, 5.8 Hz, and 9.3 Hz. Median filtering of the strain data reduces outliers by 15%, and principal component analysis retains 95% of the information. Kalman filtering of the displacement data reduces the noise variance by 30%. Multimodal data fusion yields a 50 × 100 × 3 tensor dataset. A multi-branch neural network model identifies microcracks in the main beam (confidence level 0.92) and minor corrosion in the bearings (confidence level 0.87). Based on these damage signatures, the intelligent decision-making system generates three candidate reinforcement strategies: carbon fiber reinforcement, external prestressing, and concrete repair. Multiscale simulations confirm that the carbon fiber reinforcement solution has the highest reliability index, reaching 0.95. Virtual construction planning broke down the entire reinforcement process into 15 steps, with seven steps on the critical path, and an estimated construction period of 45 days. Real-time monitoring detected a 2% delay in progress on the third day. The system automatically adjusted resource allocation for subsequent steps, ultimately ensuring the actual construction period matched the plan and achieving the desired reinforcement results.

[0032] In this embodiment of the present invention, by collecting multi-source data from bridge structures and preprocessing it using a multimodal data fusion algorithm, a comprehensive and high-quality bridge status dataset is obtained. This provides a reliable data foundation for subsequent analysis and effectively improves the accuracy and comprehensiveness of bridge status assessments. Secondly, a multi-branch neural network model is used to identify and assess damage in the bridge status dataset, generating a bridge damage signature map. This significantly improves the accuracy and efficiency of damage identification, enabling the timely detection of potential structural problems and providing a key basis for reinforcement decision-making. Furthermore, intelligent decision-making analysis is performed on the bridge damage signature map to generate preliminary reinforcement plans and a set of candidate reinforcement strategies. This automates and intelligentizes reinforcement plan generation, significantly reducing the subjectivity and uncertainty of manual decision-making. Subsequently, an optimized reinforcement plan is obtained by performing parametric modeling and multi-scale simulation on the candidate reinforcement strategy set, followed by coupled computational evaluation. This process fully considers multiple factors, including reinforcement effectiveness, cost, and construction difficulty, ensuring the scientific and feasible nature of the reinforcement plan. Furthermore, virtual construction planning is performed on the optimized reinforcement solution. Using an intelligent scheduling algorithm, a construction process model is generated to obtain the target construction plan. This significantly improves the efficiency and rationality of construction planning, helping to reduce resource waste and time delays during construction. Finally, real-time monitoring and feedback are provided for the target construction plan, and the construction process is dynamically optimized using an adaptive control algorithm to obtain the target intelligent reinforcement implementation plan. This closed-loop control mechanism greatly enhances the adaptability and robustness of the reinforcement implementation process, effectively addressing various uncertainties and emergencies during construction.

[0033] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0034] (1) Collect vibration signals from the bridge structure, eliminate noise from the vibration signals through wavelet transform to obtain noise-reduced vibration data, and perform fast Fourier transform on the noise-reduced vibration data to obtain frequency domain features;

[0035] (2) Strain data of the bridge structure is collected, outliers are removed from the strain data through median filtering to obtain filtered strain data, and principal component analysis is performed on the filtered strain data to obtain the principal components of strain;

[0036] (3) Collect displacement data of the bridge structure, smooth the displacement data through Kalman filtering to obtain smoothed displacement data, and perform time series analysis on the smoothed displacement data to obtain displacement trend characteristics;

[0037] (4) Data standardization is performed on the frequency domain features, strain principal components, and displacement trend features to obtain standardized feature data, and multimodal fusion of the standardized feature data is performed through a tensor decomposition algorithm to obtain a fused feature tensor;

[0038] (5) Perform dimensionality reduction processing on the fused feature tensor to obtain the reduced-dimensional feature vector, and organize the reduced-dimensional feature vector into a time series to obtain the bridge status dataset.

[0039] Specifically, vibration signals from the bridge structure are collected using high-precision accelerometers at key locations on the bridge. The raw vibration signals collected often contain various noise factors, necessitating noise removal. Wavelet transform is an effective signal denoising method. It decomposes the signal into wavelet coefficients of varying scales, removes the noise coefficients through thresholding, and then reconstructs the signal to produce de-noised vibration data. Fast Fourier transform is then performed on the de-noised vibration data to convert the time-domain signal into frequency-domain features, thereby obtaining important information such as the bridge structure's natural frequency and modes. Strain data from the bridge structure is collected using fiber Bragg grating (FBG) strain sensors deployed at key stress-bearing locations. Since strain data may contain outliers, median filtering is required to remove them. Median filtering is a nonlinear filtering method that replaces the center point value with the median value of the data within a sliding window, effectively removing sudden noise. After filtering, smoother and more reliable strain data is obtained. The filtered strain data is subjected to principal component analysis. By calculating the covariance matrix of the data, solving the eigenvalues ​​and eigenvectors, and selecting the main eigenvectors as principal components, the data dimension is reduced and the most representative strain features are extracted.

[0040] At the same time, displacement data of the bridge structure is collected, using laser displacement sensors or GNSS systems to measure the displacement of key bridge components. Displacement data often contains measurement noise and random errors, necessitating smoothing using a Kalman filter. The Kalman filter is a recursive estimation algorithm that continuously optimizes the state estimate through two steps: prediction and update, resulting in smoothed displacement data. Time series analysis, including trend analysis, periodicity analysis, and autocorrelation analysis, is performed on the smoothed displacement data to identify displacement trend characteristics that reflect the long-term deformation trend of the bridge. To comprehensively utilize different types of data, data standardization is required for frequency domain features, strain principal components, and displacement trend features. Standardization eliminates dimensional differences between features. Z-score normalization is typically used to transform each feature into a standard normal distribution with a mean of 0 and a variance of 1. The standardized feature data is then subjected to multimodal fusion using a tensor decomposition algorithm. Tensor decomposition is a multidimensional data processing technique that decomposes high-dimensional data into a product of low-dimensional factors. Common methods include Tucker decomposition and CP decomposition. Tensor decomposition captures correlations between data from different modalities, generating a fused feature tensor.

[0041] Finally, the fused feature tensor undergoes dimensionality reduction, using commonly used methods such as principal component analysis (PCA) or t-distributed stochastic neighbor embedding (t-SNE). Dimensionality reduction further reduces the data dimension and extracts the most important feature information. The reduced feature vectors are organized chronologically to form a time series, ultimately yielding a bridge status dataset reflecting the overall condition of the bridge. For example, consider the health monitoring of a 200-meter steel box girder bridge. Three accelerometers are placed at each of the 1 / 4, 1 / 2, and 3 / 4 spans of the bridge, sampling at a 100 Hz frequency. Vibration data is collected continuously for 24 hours. The raw data contains vibrations caused by both ambient noise and traffic loads. Denoising is achieved through wavelet transform, using the db4 wavelet. A five-layer decomposition of the signal is performed, and the signal is reconstructed after soft thresholding. The noise amplitude is reduced by approximately 40%. Fast Fourier transform (FFT) of the denoised data identifies the bridge's first three natural frequencies as 0.52 Hz, 1.18 Hz, and 2.35 Hz, respectively.

[0042] Twenty fiber Bragg grating strain sensors were placed on the bottom plate of the bridge's main beam, sampling at a frequency of 10 Hz. The raw strain data contained some sudden outliers. Using a median filter with a window size of 5 reduced the number of outliers by 95%. Principal component analysis was performed on the filtered strain data, and the first three principal components were selected, resulting in a cumulative explained variance of 92%. A GNSS receiver was installed on the top of the bridge tower, sampling at a frequency of 1 Hz, to measure the three-dimensional displacement of the tower. The raw displacement data contained centimeter-level random errors. Using a Kalman filter, the standard deviation of the displacement data was reduced from 3.2 cm to 0.8 cm. Time series analysis of the smoothed displacement data revealed a small periodic oscillation of the tower, with a period of approximately one hour and an amplitude within ±5 mm.

[0043] The processed frequency domain features (9 features, 3 measurement points × 3rd-order frequency), strain principal components (3 features), and displacement trend features (6 features, displacement and velocity in the X, Y, and Z directions) were combined into an 18-dimensional feature vector. These 18 features were Z-score standardized, and then a 18×24×3 third-order tensor (18 features, 24 hours, 3 data types) was constructed. The tensor was decomposed using Tucker decomposition to obtain a 5×6×2 core tensor and three factor matrices. Finally, the core tensor was flattened and reorganized in chronological order to obtain a 60×24 bridge status dataset, with one state vector per hour and 60 fused features. This dataset comprehensively reflects the changes in the health status of the bridge within 24 hours, providing a reliable data foundation for subsequent damage identification and reinforcement decisions.

[0044] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0045] (1) The bridge status dataset is divided into time windows, and the data is segmented using a sliding window algorithm to obtain a time series feature subset. The time series feature subset is then Fourier transformed to obtain a frequency domain feature matrix.

[0046] (2) Perform convolutional neural network processing on the frequency domain feature matrix, extract local features through multi-layer convolution and pooling operations to obtain a convolution feature map, and perform global average pooling on the convolution feature map to obtain a global feature vector;

[0047] (3) The global feature vector is processed by the long short-term memory network, and the long-term features are captured by temporal dependency modeling to obtain the temporal feature vector. The temporal feature vector is then processed by the self-attention mechanism to obtain the attention-weighted feature;

[0048] (4) The attention-weighted features are processed by a fully connected layer, and feature mapping is performed through a multi-layer perceptron to obtain a high-dimensional feature representation. The high-dimensional feature representation is then classified by softmax to obtain the probability distribution of the damage type.

[0049] (5) Threshold filtering is performed on the probability distribution of damage types. High-confidence predictions are screened by setting a confidence threshold to obtain the damage type determination results. The damage type determination results are spatially mapped to obtain a bridge damage characteristic map.

[0050] Specifically, the bridge status dataset is divided into time windows and segmented using a sliding window algorithm. The sliding window algorithm sets a fixed-size time window, such as one hour, and slides it along the time axis in steps of a certain size, such as 10 minutes, to segment the continuous time series data into multiple overlapping time series feature subsets. This processing method captures the local temporal characteristics of the data while maintaining data continuity. A Fourier transform is performed on each time series feature subset, converting the time domain signal into a frequency domain representation, resulting in a frequency domain feature matrix. The Fourier transform reveals the frequency composition of the signal, helping to identify the vibration characteristics and potential abnormal patterns of the bridge structure. The frequency domain feature matrix is ​​then processed using a convolutional neural network (CNN). CNNs extract local features through multiple layers of convolution and pooling operations. The convolution operation filters the input data using different convolution kernels to extract features of different scales and types. The pooling operation reduces the data dimensionality through downsampling, improving computational efficiency. This process generates a convolution feature map, which is then globally averaged and pooled. The average value of each feature map is used as the global representation of the feature, resulting in a global feature vector. Global average pooling can effectively reduce the number of parameters while retaining important feature information.

[0051] The global feature vector is then input into a long short-term memory (LSTM) network for processing. LSTM is a special type of recurrent neural network that effectively processes long-sequence data and captures long-term dependencies. LSTM controls the flow of information through a gating mechanism, including input, forget, and output gates, enabling it to learn long-term temporal features in the data. The resulting temporal feature vector is then processed by a self-attention mechanism. This mechanism allows the model to focus on different parts of the input sequence and weight them based on their relevance to the current moment, thereby generating attention-weighted features. This mechanism captures long-range dependencies in the sequence and improves the model's sensitivity to important information. The attention-weighted features are processed by a fully connected layer, which is essentially a multi-layer perceptron (MLP). It maps and transforms features using nonlinear activation functions to produce a high-dimensional feature representation. This step further extracts abstract features and enhances the model's expressive power. The high-dimensional feature representation is finally processed by a softmax classifier to obtain a probability distribution of damage types. The softmax function converts the high-dimensional features into probability values ​​for each damage type. The sum of these probabilities is 1, reflecting the model's confidence in the prediction of each damage type.

[0052] Finally, the damage type probability distribution is threshold-filtered, with a confidence threshold (e.g., 0.8) set to select highly reliable predictions. Only predictions exceeding the threshold are considered reliable damage type determinations. These determinations are then spatially mapped to the actual structural location of the bridge, yielding an intuitive bridge damage signature map.

[0053] For example, consider damage identification for a 500-meter-long steel box girder bridge. The bridge status dataset consists of 7 days of monitoring data from 200 sensor points, sampled at 100 Hz. First, the data is segmented using a sliding window with a length of 1 hour and a step size of 30 minutes, resulting in 336 time series feature subsets. A Fourier transform is performed on each subset, and frequency components in the 0-50 Hz range are selected to obtain a 200×500 frequency domain feature matrix (200 sensor points, 500 frequency components). These frequency domain feature matrices are input to a CNN consisting of four convolutional layers, each using 64 3×3 convolution kernels, followed by ReLU activation and 2×2 max pooling. After CNN processing, a 12×12×64 feature map is obtained, which is then subjected to global average pooling to obtain a 64-dimensional global feature vector.

[0054] The global feature vector is input to a two-layer LSTM network with 128 hidden units per layer. The 128-dimensional temporal feature vector output by the LSTM is passed through a multi-head self-attention layer with eight attention heads, resulting in 512-dimensional attention-weighted features. The attention-weighted features are processed through a four-layer fully connected network with hidden layer sizes of 256, 128, 64, and 32, respectively, using the Reluctant Unit (ReLU) activation function. The output of the final layer undergoes softmax classification, assuming six predefined damage types (e.g., cracks, corrosion, fatigue, deformation, support damage, and prestress loss).

[0055] The model outputs a probability distribution of six damage types, for example, [0.03, 0.02, 0.89, 0.01, 0.04, 0.01]. The confidence threshold is set to 0.85, so only the third damage type (fatigue) with a probability of 0.89 is determined to be a reliable result. Finally, the fatigue damage judgment result is mapped to the bridge structure diagram, marking the specific locations where fatigue damage may occur, such as a welded joint area in the middle span of the main beam. In this way, the model successfully identifies the fatigue damage of the bridge and accurately locates the damage location. This result provides timely early warning information for bridge managers, enabling them to take targeted reinforcement measures, such as local reinforcement of fatigue damaged areas or replacement of high fatigue strength materials, thereby effectively extending the service life of the bridge and ensuring the safe operation of the bridge.

[0056] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0057] (1) The bridge damage feature map is processed by graph convolutional network, and the topological relationship of the damage is obtained by spatial feature extraction to obtain a damage association map. The damage association map is then embedded into a graph to obtain a damage representation vector.

[0058] (2) Cluster analysis is performed on the damage representation vectors. Similar damages are grouped using the K-means algorithm to obtain a set of damage categories. The severity of the damage category set is then evaluated to obtain a reinforcement priority ranking.

[0059] (3) Perform knowledge graph matching on the reinforcement priority ranking, select the applicable reinforcement technology through semantic similarity calculation, obtain the preliminary reinforcement plan, and check the constraints of the preliminary reinforcement plan to obtain a set of feasible reinforcement plans;

[0060] (4) Perform multi-objective optimization on the feasible reinforcement scheme set, balance the reinforcement effect and cost through genetic algorithm, obtain the Pareto optimal solution set, and perform decision tree analysis on the Pareto optimal solution set to obtain the scheme decision path;

[0061] (5) Monte Carlo simulation is performed on the decision path of the scheme, and the robustness of the scheme is evaluated through random sampling to obtain the scheme reliability index. The scheme reliability index is then sorted and screened to obtain a set of candidate reinforcement strategies.

[0062] Specifically, the bridge damage feature map is processed using a graph convolutional network (GCN), extracting spatial features to capture the topological relationships of damage. GCN views the bridge structure as a graph, where nodes represent structural elements and edges represent connections between elements. GCN updates the features of the central node by aggregating information from neighboring nodes, thereby capturing the spatial distribution and mutual influence of damage within the structure. This process generates a damage association graph, which reflects the relationships between different damage points. Subsequently, the damage association graph is subjected to graph embedding, compressing the high-dimensional graph structure information into a low-dimensional space to generate a damage representation vector. Graph embedding techniques such as Node2Vec or GraphSAGE can be used for this step, as they preserve the graph's structural information and node features. Cluster analysis is performed on the damage representation vector, using the K-means algorithm to group similar damages. The K-means algorithm, through iterative optimization, partitions data points into K clusters, ensuring that each data point belongs to the nearest cluster center. This process generates a set of damage categories, grouping similar damage patterns. The damage categories are then evaluated for severity, taking into account factors such as damage severity, location importance, and development trends. The comprehensive score is used to rank reinforcement priorities.

[0063] The reinforcement priority ranking is matched against a knowledge graph, and appropriate reinforcement technologies are selected through semantic similarity calculation. The knowledge graph contains information on various reinforcement technologies, applicable conditions, and effectiveness evaluation. Semantic similarity calculations can use methods such as cosine similarity or Word2Vec to match damage characteristics with reinforcement technology descriptions, resulting in a preliminary reinforcement plan. Constraints on these preliminary reinforcement plans are then checked, considering factors such as structural safety, construction feasibility, and economic efficiency. A set of feasible reinforcement solutions that meet all constraints is selected. Multi-objective optimization is then performed on these feasible reinforcement solutions, using a genetic algorithm to balance reinforcement effectiveness and cost. Genetic algorithms mimic biological evolution, continuously optimizing the solution space through selection, crossover, and mutation operations. Optimization objectives include improving structural safety, extending service life, and construction cost. This process yields a Pareto optimal solution set—a set of non-dominated solutions that balances multiple objectives. Decision tree analysis is then performed on this Pareto optimal solution set, constructing a hierarchical decision-making structure that considers the optimal option under different conditions and determines the decision path for the solution.

[0064] Finally, a Monte Carlo simulation is performed on the solution decision path, assessing the robustness of the solution through random sampling. Monte Carlo simulations simulate uncertainty through a large number of random samples, evaluating the performance of the solution under different scenarios. This process yields a solution reliability index, reflecting its stability under various possible scenarios. The solution reliability indexes are then sorted and screened to ultimately generate a set of candidate reinforcement strategies, providing decision makers with multiple high-quality alternatives.

[0065] For example, consider damage identification and reinforcement decision-making for a sea-crossing bridge. The bridge damage feature map contains 50 damage points, each with 10 feature dimensions. A 50×64 damage association map is generated using a three-layer GCN with 64 hidden units per layer. Node2Vec is used for graph embedding, mapping each damage point to a 32-dimensional vector space. K-means clustering (K=5) classifies the damage into five categories: cracks, corrosion, fatigue, deformation, and support damage. Severity assessment considers damage severity (40% weight), location importance (30% weight), and development trend (30% weight), with a score ranging from 0 to 100. The knowledge graph contains 20 reinforcement technologies, and semantic similarity is calculated using Word2Vec to select the top three applicable technologies for each damage category. Constraint checks, considering a budget cap of 50 million yuan and a construction period of no more than 6 months, screen out 10 feasible reinforcement solutions. Multi-objective optimization is performed with a population size of 100 and 500 iterations, resulting in 30 Pareto optimal solutions. Decision tree analysis considers three key nodes: budget, construction period, and traffic impact, and constructs a decision path.

[0066] Monte Carlo simulations were performed using 10,000 random sampling cycles, accounting for uncertainties such as material price fluctuations (±10%), construction delays (0-30 days), and environmental factors (±5%). Reliability indicators were calculated based on performance within a 90% confidence interval. Five candidate reinforcement strategies were ultimately selected, including carbon fiber beam reinforcement (reliability 0.92), external prestressing reinforcement (reliability 0.89), and cathodic protection (reliability 0.95).

[0067] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0068] (1) Parametric modeling is performed on the candidate reinforcement strategy set. The geometric model of the reinforcement component is generated by the non-uniform rational B-spline technology to obtain the reinforcement component model set. The reinforcement component model set is then meshed to obtain the finite element model;

[0069] (2) Assign material properties to the finite element model, select appropriate material parameters through the constitutive relationship database to obtain a complete structural model, and set boundary conditions for the complete structural model to obtain a calculation-ready model;

[0070] (3) Perform macro-scale simulation on the computationally ready model, calculate the overall structural response using the nonlinear finite element method, obtain macro-mechanical performance indicators, and perform local stress analysis on the macro-mechanical performance indicators to obtain stress distribution in key areas;

[0071] (4) Perform mesoscopic simulation of the stress distribution in the key area, simulate the internal damage evolution of the material using the discrete element method, obtain the micro crack propagation path, and perform fracture mechanics analysis on the micro crack propagation path to obtain the critical stress intensity factor;

[0072] (5) The reliability of the critical stress intensity factor is evaluated, and the structural failure probability is calculated by the first-order and second-order moment method to obtain the reliability index of the reinforcement scheme. The reliability index of the reinforcement scheme is comprehensively ranked to obtain the optimized reinforcement scheme.

[0073] Specifically, a set of candidate reinforcement strategies is parametrically modeled, and the geometric models of the reinforcement components are generated using non-uniform rational B-spline (NURBS) technology. NURBS is a mathematical method for accurately representing complex surfaces, defining geometric shapes through control points, node vectors, and weights. This technology can accurately describe the shapes of various reinforcement components, such as carbon fiber plates and steel plates. The generated reinforcement component model set contains the geometric information of all possible reinforcement solutions. Subsequently, the reinforcement component model set is meshed to obtain a finite element model. During the meshing process, adaptive meshing technology is used to adopt a finer mesh in stress concentration areas to improve calculation accuracy. Material properties are assigned to the finite element model. Parameters suitable for various materials, such as elastic modulus, Poisson's ratio, and yield strength, are selected using a pre-established constitutive relationship database. This constitutive relationship database contains mechanical property data for common reinforcement materials, such as high-strength concrete, carbon fiber, and steel. After the material properties are assigned, a complete structural model is obtained. Boundary conditions are then set for the complete structural model, including support constraints and loading conditions, to obtain a calculation-ready model.

[0074] A macroscale simulation is performed on the computationally ready model, and the overall structural response is calculated using the nonlinear finite element method. Nonlinear analysis takes into account factors such as the nonlinear behavior of the material and large geometric deformations, and can more accurately simulate actual structural behavior. By solving the nonlinear equilibrium equations, information such as the structural displacement field and stress field is obtained, and macroscopic mechanical performance indicators such as load-bearing capacity and deformation are calculated. Local stress analysis is performed on the macroscopic mechanical performance indicators to identify stress concentration areas and obtain the stress distribution in critical areas. The stress distribution in critical areas is simulated at the mesoscale, and the discrete element method (DEM) is used to simulate the internal damage evolution of the material. DEM treats the material as composed of discrete particles and describes the material behavior by simulating the interactions between particles. This method can simulate microscopic processes such as crack initiation and propagation, and obtain the microscopic crack propagation path. Fracture mechanics analysis is performed on the microscopic crack propagation path to calculate the stress intensity factor. The crack propagation rate is predicted using methods such as the Paris law, and the critical stress intensity factor is ultimately obtained.

[0075] Finally, the critical stress intensity factor was subjected to reliability assessment, and the structural failure probability was calculated using the first-order and second-order moment method. The first-order and second-order moment method is a commonly used structural reliability analysis method that estimates the failure probability by calculating the mean and variance of the limit state function. This process yields the reliability index of the reinforcement scheme, reflecting its safety and reliability. The reliability indexes of the reinforcement schemes were comprehensively ranked, taking into account multiple factors such as reliability, cost-effectiveness, and construction difficulty, ultimately resulting in the optimal reinforcement scheme.

[0076] For example, consider the reinforcement design for a prestressed concrete continuous beam bridge. Candidate reinforcement strategies include external carbon fiber plate attachment, external prestressing, and steel plate reinforcement. First, NURBS technology is used for modeling. The carbon fiber plate model uses a 20×20 control point grid to accurately describe its curved surface shape at the bottom of the beam. The external prestressing steel strands are described using a cubic NURBS curve with 15 control points. Meshing uses a 0.1m grid size at the edges of the carbon fiber plate and in the prestressing anchorage area, and 0.5m in other areas, resulting in approximately 500,000 elements in total. Material properties are assigned from the constitutive relationship database, with an elastic modulus of 230 GPa and an ultimate strength of 3500 MPa for the carbon fiber plate; and an elastic modulus of 195 GPa and a yield strength of 1860 MPa for the prestressing steel strands. Boundary conditions include fixed constraints on the piers, temperature loads (±25°C), and vehicle loads (standard vehicle model).

[0077] The macro-scale nonlinear analysis used the Newton-Raphson iteration method, with convergence criteria set to displacement error less than 0.1mm and force balance error less than 1kN. The analysis results showed that the carbon fiber plate reinforcement scheme reduced the deflection at mid-span by 12mm and the tensile stress at the bottom of the main beam by 15MPa. Local stress analysis revealed stress concentration at the ends of the carbon fiber plate, with the maximum shear stress reaching 3.2MPa. DEM analysis was performed on this stress concentration area, simulating a 1cm 3 The volume contains 100,000 discrete element particles. Simulation results show the formation of microcracks at the interface, with a maximum crack length of 0.5 mm. Fracture mechanics analysis calculates the critical stress intensity factor (KIC) to be 0.8 MPa·m^(1 / 2).

[0078] Using the first-order and second-order moment methods for reliability assessment, accounting for uncertainties such as material strength and load, the reliability index β of the reinforced structure was calculated to be 4.2, corresponding to a failure probability of 1.33 × 10^-5. Taking into account reliability, cost, and construction difficulty, the carbon fiber plate reinforcement solution was ultimately determined to be the optimal option, with an estimated lifespan extension of 20 years.

[0079] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0080] (1) Decompose the optimized reinforcement scheme into processes, determine the process priority through the hierarchical analysis method, obtain the process dependency graph, and perform critical path analysis on the process dependency graph to obtain the key process sequence;

[0081] (2) Analyze the resource requirements of the key process sequence, optimize resource allocation through resource balancing algorithm, obtain a preliminary resource allocation plan, and perform conflict detection on the preliminary resource allocation plan to obtain a feasible resource plan;

[0082] (3) Conduct construction simulation on feasible resource plans, evaluate the construction progress through discrete event simulation technology, obtain construction progress forecast data, and conduct deviation analysis on the construction progress forecast data to obtain progress adjustment suggestions;

[0083] (4) Perform intelligent scheduling optimization on the schedule adjustment suggestions, generate multiple scheduling schemes through heuristic algorithms, obtain a candidate scheduling set, and perform multi-criteria evaluation on the candidate scheduling set to obtain the optimal scheduling scheme;

[0084] (5) Conduct risk assessment on the optimal scheduling plan, analyze potential risk factors through Monte Carlo simulation, obtain risk response strategies, and integrate risk response strategies into the construction process to obtain the target construction plan.

[0085] Specifically, the optimized reinforcement plan was decomposed into process steps, and the process priorities were determined using the Analytic Hierarchy Process (AHP). The AHP method decomposes complex problems into a hierarchical structure. A judgment matrix is ​​established through expert scoring, and eigenvectors are calculated to determine the weights of each process, thereby determining the priority. This process generates a process dependency graph, which reflects the logical relationships between processes. Subsequently, a critical path analysis is performed on the process dependency graph. Using forward and backward traversal algorithms, the earliest start time, latest start time, and total time difference of each process are calculated to identify the critical process sequence with the greatest impact on the total construction period. Resource requirements are analyzed for this critical process sequence, and resource allocation is optimized using a resource balancing algorithm. The resource balancing algorithm considers resource constraints such as manpower, equipment, and materials. By adjusting the start times of non-critical processes, the resource utilization curve is stabilized, avoiding excessive resource concentration or idleness. This process generates a preliminary resource allocation plan. This preliminary resource allocation plan is then subjected to conflict detection to check for spatial and temporal resource conflicts, such as multiple processes using the same equipment or overlapping work surfaces. Conflicts are resolved by adjusting the process sequence or resource allocation, resulting in a feasible resource plan.

[0086] Construction simulations are conducted on feasible resource plans, and discrete event simulation technology is used to assess construction progress. Discrete event simulation abstracts the construction process into a series of discrete events. By simulating the occurrence and interaction of these events, dynamic changes throughout the construction process are predicted. The simulation process takes into account random factors such as weather and equipment failures, resulting in construction progress forecasts. Deviation analysis is performed on the construction progress forecasts, comparing the predicted progress with the planned progress, identifying potential delay risks, and formulating schedule adjustment recommendations. Intelligent scheduling optimization is then performed on the schedule adjustment recommendations, generating multiple scheduling solutions using heuristic algorithms. Heuristic algorithms, such as genetic algorithms or ant colony algorithms, simulate the principles of biological evolution or swarm intelligence to search for near-optimal solutions in the solution space. This process generates multiple candidate scheduling solutions, forming a candidate scheduling set. This candidate scheduling set undergoes a multi-criteria evaluation, considering multiple objectives such as construction period, cost, quality, and safety. Multi-criteria decision-making methods, such as the TOPSIS method (ranking by near-ideal solutions), are used to comprehensively evaluate the pros and cons of each solution and determine the optimal scheduling solution.

[0087] Finally, the optimal scheduling plan undergoes a risk assessment, analyzing potential risk factors through Monte Carlo simulation. Monte Carlo simulation uses a large number of random samples to simulate the impact of various uncertainties on the project, such as schedule delays, cost overruns, and quality issues. Based on the simulation results, the probability and impact of the risk are calculated, and risk response strategies are developed. These strategies include risk avoidance, risk transfer, risk mitigation, and risk acceptance. Risk response strategies are integrated into the construction process, adjusting construction plans and resource allocation, and ultimately achieving the target construction plan.

[0088] For example, suppose a reinforced concrete arch bridge is being reinforced. The optimized reinforcement plan includes carbon fiber reinforcement of the main arch ribs and bearing replacement. A process breakdown yields 20 key steps, including traffic control, scaffolding erection, surface treatment, carbon fiber paving, maintenance, and bearing replacement. AHP analysis determined that carbon fiber paving has a weight of 0.25, bearing replacement has a weight of 0.2, and the weights of other steps range from 0.02 to 0.1. Critical path analysis identified seven key steps, with a total estimated construction period of 60 days. Resource requirement analysis revealed that the carbon fiber paving step required 10 specialized workers and two pieces of specialized equipment, while bearing replacement required five workers and a large crane. After resource balancing, non-critical steps, such as traffic control, were moved forward by five days to avoid a peak in manpower exceeding 20 people. Conflict detection revealed a one-day overlap in crane usage, which was resolved by adjusting the order of bearing replacement.

[0089] Discrete event simulation, factoring in a 30% probability of rain, resulted in an average construction duration of 63.5 days and a standard deviation of 2.1 days after 1,000 simulations. Deviation analysis indicated that carbon fiber curing time may be insufficient, recommending an additional one-day buffer. Intelligent scheduling optimization employed a genetic algorithm with a population size of 100 and 500 iterations, resulting in 20 candidate solutions. A multi-criteria evaluation, considering construction duration (weight 0.4), cost (weight 0.3), quality (weight 0.2), and safety (weight 0.1), yielded the optimal solution with an estimated construction duration of 62 days, a 2% increase in cost, and a 5% improvement in quality and safety scores. A risk assessment, conducted through 10,000 Monte Carlo simulations, identified weather risk (30% probability of impact, average delay of 2 days) and material supply risk (10% probability of impact, average delay of 3 days) as the primary risks. Response strategies included adding a 15% construction buffer and establishing alternate material supply channels. The final target construction plan adjusted the construction duration to 65 days, incorporating a pre-delivery inspection procedure for materials before key processes.

[0090] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0091] (1) Plan the sensor layout for the target construction plan, determine the monitoring point locations through the optimal coverage algorithm, obtain the monitoring network topology, and set the data collection frequency for the monitoring network topology to obtain a real-time monitoring plan;

[0092] (2) Collect data for the real-time monitoring scheme, pre-process the collected data through edge computing technology to obtain an effective monitoring data stream, and perform anomaly detection on the effective monitoring data stream to obtain abnormal events in the construction process;

[0093] (3) Diagnose the causes of abnormal events in the construction process, analyze the root causes of abnormalities through Bayesian network reasoning, obtain a construction deviation report, and conduct an impact assessment on the construction deviation report to obtain an adjustment priority list;

[0094] (4) Generate a plan for the adjustment priority list, formulate an adjustment strategy through a reinforcement learning algorithm, obtain a set of candidate adjustment plans, and simulate and verify the candidate adjustment plan set to obtain the optimal adjustment plan;

[0095] (5) Generate execution instructions for the optimal adjustment plan, convert it into specific operation instructions through semantic parsing technology, obtain a construction adjustment instruction set, and generate a target intelligent reinforcement implementation plan based on the construction adjustment instruction set.

[0096] Specifically, sensor layout planning is performed for the target construction plan, and monitoring point locations are determined using an optimal coverage algorithm. Based on a greedy strategy, the optimal coverage algorithm iteratively selects sensor locations that maximize coverage until the preset coverage requirement is met. This process results in a monitoring network topology, which reflects the spatial distribution and connectivity of the sensors. Subsequently, the data acquisition frequency is configured within the monitoring network topology, taking into account the characteristics of different sensor types and the monitoring objectives. For example, strain sensors may require a higher sampling frequency (e.g., 100 Hz), while displacement sensors may require a lower sampling frequency (e.g., 1 Hz). This results in a comprehensive real-time monitoring solution. Data for this real-time monitoring solution is collected and pre-processed using edge computing technology. Edge computing offloads some data processing tasks to edge nodes close to the data source, reducing data transmission volume and improving response speed. Pre-processing includes data cleaning, outlier detection, and feature extraction, resulting in a valid monitoring data stream. Anomaly detection is performed on this valid monitoring data stream, using algorithms such as isolation forests or local outlier factors (LOFs) to identify data points that significantly deviate from the normal pattern, thereby identifying abnormal events in the construction process.

[0097] The cause of abnormal events during the construction process is diagnosed, and the root cause of the anomaly is analyzed using Bayesian network reasoning. A Bayesian network is a probabilistic graphical model that can represent conditional dependencies between variables. By constructing a Bayesian network with nodes including construction processes, environmental factors, and equipment status, and setting a conditional probability table based on historical data and expert knowledge, the most likely cause of the abnormal event can be inferred. This process generates a construction deviation report that details the nature, location, and possible causes of the anomaly. An impact assessment is conducted on the construction deviation report, considering the severity, development trend, and potential risks of the deviation to generate a priority list of adjustments. Plans are generated based on the priority list, and an adjustment strategy is formulated using a reinforcement learning algorithm. Reinforcement learning learns optimal strategies through interaction between an intelligent agent and its environment, making it suitable for handling complex decision-making problems. Specifically, a deep Q-network (DQN) or policy gradient method can be used. It takes the current state (including construction progress, resource allocation, and abnormal conditions) as input and outputs the optimal adjustment action. This process generates a set of candidate adjustment plans, consisting of multiple possible adjustment strategies. Simulations are conducted on these candidate adjustment plans, using discrete event simulation or system dynamics models to evaluate the effectiveness and feasibility of each plan, ultimately determining the optimal adjustment plan.

[0098] Finally, execution instructions are generated for the optimal adjustment plan. Semantic parsing technology is used to translate high-level adjustment strategies into specific operational instructions. Based on natural language processing and domain knowledge graphs, semantic parsing technology understands the semantic content of the adjustment plan and generates detailed instructions that conform to construction specifications and operational procedures. This process generates a set of construction adjustment instructions, encompassing specific process adjustments, resource reallocation, and quality control measures. Based on this set of construction adjustment instructions and combined with the original construction plan, an updated target intelligent reinforcement implementation plan is generated, enabling dynamic optimization and intelligent control of the construction process.

[0099] For example, consider the carbon fiber reinforcement of the main beam of a sea-crossing bridge. First, using an optimal coverage algorithm, 50 strain sensors and 20 accelerometers were deployed along the 1,000-meter-long main beam, achieving 95% coverage. The sampling frequency for the strain sensors was set at 100 Hz, and for the accelerometers at 200 Hz. During construction, edge computing nodes performed noise reduction and compression on the raw data, reducing the data volume by 60%. An anomaly detection algorithm detected an abnormal event in the middle section of the bridge, where the strain value suddenly increased by 20%. Bayesian network analysis indicated an 80% probability that this anomaly was caused by bubbles during the carbon fiber laying process. An impact assessment prioritized this issue as the highest priority, requiring immediate action. Based on historical data and current status, a reinforcement learning algorithm generated three candidate adjustment options: 1) partial rework and re-laying; 2) increasing vacuum pressure; and 3) extending the curing time. Simulation verification showed that option 1 had the highest success rate, reaching 95%. The semantic parsing system translated the optimal solution into specific instructions: 1) Stop the current process; 2) Define the rework area (pile numbers K15+200 to K15+220); 3) Remove the defective carbon fiber; 4) Re-apply surface treatment; 5) Re-lay the carbon fiber according to standard procedures; 6) Increase the curing time by two hours. These instructions were integrated into the original construction plan, forming a new intelligent reinforcement implementation plan. This real-time monitoring and dynamic adjustment enabled the timely identification and resolution of potential quality issues, ensuring the overall effectiveness of the reinforcement project.

[0100] The above describes the intelligent bridge reinforcement decision-making method based on damage identification in the embodiment of the present application. The following describes the intelligent bridge reinforcement decision-making device based on damage identification in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of an intelligent bridge reinforcement decision-making device based on damage identification includes:

[0101] The acquisition module 201 is used to collect multi-source data of the bridge structure and pre-process the collected data using a multimodal data fusion algorithm to obtain a bridge status data set;

[0102] An evaluation module 202 is configured to identify and evaluate bridge damage on the bridge status dataset using a multi-branch neural network model to obtain a bridge damage feature map;

[0103] An analysis module 203 is configured to perform intelligent decision-making analysis on the bridge damage characteristic map, generate a preliminary reinforcement plan, and obtain a candidate reinforcement strategy set;

[0104] The simulation module 204 is used to perform parameterized modeling and multi-scale simulation on the candidate reinforcement strategy set, and perform coupled calculation evaluation on the reinforcement effect to obtain an optimized reinforcement solution;

[0105] Planning module 205, for performing virtual construction planning for the optimized reinforcement scheme, generating a construction process model through an intelligent scheduling algorithm, and obtaining a target construction scheme;

[0106] The feedback module 206 is used to monitor and provide feedback on the target construction plan in real time, and dynamically optimize the construction process through an adaptive control algorithm to obtain a target intelligent reinforcement implementation plan.

[0107] Through the collaborative efforts of these components, multi-source data collection and pre-processing of bridge structure data using a multimodal data fusion algorithm yielded a comprehensive, high-quality bridge condition dataset. This dataset provides a reliable foundation for subsequent analysis and effectively improves the accuracy and comprehensiveness of bridge condition assessments. Secondly, a multi-branch neural network model was used to identify and assess damage within the bridge condition dataset, generating a bridge damage signature map. This significantly improved the accuracy and efficiency of damage identification, enabling timely identification of potential structural issues and providing a critical basis for reinforcement decision-making. Furthermore, intelligent decision-making analysis of the bridge damage signature map was performed to generate preliminary reinforcement plans and a set of candidate reinforcement strategies. This automated and intelligent generation of reinforcement plans significantly reduced the subjectivity and uncertainty inherent in manual decision-making. Subsequently, the candidate reinforcement strategies were subjected to parametric modeling, multi-scale simulation, and coupled computational evaluation to determine the optimal reinforcement plan. This process fully considered multiple factors, including reinforcement effectiveness, cost, and construction difficulty, ensuring the scientific and feasible nature of the reinforcement plan. Furthermore, virtual construction planning is performed on the optimized reinforcement solution. Using an intelligent scheduling algorithm, a construction process model is generated to obtain the target construction plan. This significantly improves the efficiency and rationality of construction planning, helping to reduce resource waste and time delays during construction. Finally, real-time monitoring and feedback are provided for the target construction plan, and the construction process is dynamically optimized using an adaptive control algorithm to obtain the target intelligent reinforcement implementation plan. This closed-loop control mechanism greatly enhances the adaptability and robustness of the reinforcement implementation process, effectively addressing various uncertainties and emergencies during construction.

[0108] Based on the same technical concept, the embodiment of the present application also provides an electronic device. Figure 3 3 is a schematic diagram of the structure of an electronic device 300 provided in an embodiment of the present application, including a processor 301, a memory 302, and a bus 303. The memory 302 is used to store execution instructions and includes a memory 3021 and an external memory 3022. The memory 3021 is also referred to as internal memory and is used to temporarily store operation data in the processor 301 and data exchanged with an external memory 3022 such as a hard disk. The processor 301 exchanges data with the external memory 3022 through the memory 3021. When the electronic device 300 is running, the processor 301 and the memory 302 communicate via the bus 303.

[0109] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0110] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0111] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent bridge reinforcement decision-making method based on damage identification, characterized in that: This involves collecting multi-source data on bridge structures, pre-processing the collected data using a multimodal data fusion algorithm to obtain a bridge status dataset; and identifying and evaluating bridge damage using a multi-branch neural network model to obtain a bridge damage feature map. Conduct intelligent decision-making analysis on the bridge damage characteristic map to generate preliminary reinforcement plans and obtain a set of candidate reinforcement strategies; Perform parametric modeling and multi-scale simulation on the candidate reinforcement strategy set, and conduct coupled computational evaluation on the reinforcement effect to obtain an optimized reinforcement scheme. This includes parametric modeling of the candidate reinforcement strategy set, generating a geometric model of the reinforcement component using non-uniform rational B-spline technology to obtain a reinforcement component model set, and meshing the reinforcement component model set using adaptive meshing technology to obtain a finite element model. Assign material properties to the finite element model, select appropriate material parameters through the constitutive relationship database to obtain a complete structural model, and set boundary conditions for the complete structural model to obtain a computationally ready model. Perform macroscale simulation on the computationally ready model, calculate the overall structural response using the nonlinear finite element method, and obtain macroscopic mechanical performance indicators. Perform local stress analysis on the macroscopic mechanical performance indicators to obtain stress distribution in key areas. Perform mesoscopic simulation of stress distribution in key areas, simulate internal damage evolution of the material using the discrete element method, obtain microscopic crack propagation paths, and perform fracture mechanics analysis on the microscopic crack propagation paths to obtain the critical stress intensity factor; Conduct reliability assessment on the critical stress intensity factor, calculate the structural failure probability using the first-order and second-order moment methods, obtain the reliability index of the reinforcement scheme, and comprehensively rank the reliability index of the reinforcement scheme to obtain the optimized reinforcement scheme; Conduct virtual construction planning for the optimized reinforcement scheme, generate a construction process model through an intelligent scheduling algorithm, and obtain a target construction scheme, including process decomposition of the optimized reinforcement scheme, determining process priorities through the hierarchical analysis method, obtaining a process dependency graph, and performing critical path analysis on the process dependency graph to obtain a key process sequence; conduct resource demand analysis on the key process sequence, optimize resource allocation through a resource balancing algorithm, obtain a preliminary resource allocation scheme, and perform conflict detection on the preliminary resource allocation scheme to obtain a feasible resource scheme; conduct construction simulation on the feasible resource scheme, evaluate the construction progress through discrete event simulation technology, obtain construction progress forecast data, and perform deviation analysis on the construction progress forecast data to obtain progress adjustment suggestions; perform intelligent scheduling optimization on the progress adjustment suggestions, generate multiple scheduling schemes through a heuristic algorithm, obtain a candidate scheduling set, and perform multi-criteria evaluation on the candidate scheduling set to obtain the optimal scheduling scheme; Conduct risk assessment on the optimal scheduling plan, analyze potential risk factors through Monte Carlo simulation, and obtain risk response strategies. These strategies are then integrated into the construction process to obtain the target construction plan. The target construction plan is monitored and fed back in real time, and the construction process is dynamically optimized through an adaptive control algorithm to obtain the target intelligent reinforcement implementation plan.

2. The intelligent bridge reinforcement decision-making method based on damage identification according to claim 1 is characterized in that: Multi-source data collection is performed on the bridge structure. The collected data is pre-processed using a multimodal data fusion algorithm to obtain a bridge status dataset, including: Vibration signals of bridge structures are collected and noise is eliminated through wavelet transform to obtain de-noised vibration data. Fast Fourier transform is then performed on the de-noised vibration data to obtain frequency domain features. Strain data of the bridge structure is collected, outliers are removed from the strain data through median filtering to obtain filtered strain data, and principal component analysis is performed on the filtered strain data to obtain the principal components of strain; The displacement data of the bridge structure is collected and smoothed by Kalman filtering to obtain the smoothed displacement data. The smoothed displacement data is then subjected to time series analysis to obtain the displacement trend characteristics. The frequency domain features, strain principal components and displacement trend features are normalized to obtain standardized feature data, and the standardized feature data are multimodally fused using a tensor decomposition algorithm to obtain a fused feature tensor. The fused feature tensor is subjected to dimensionality reduction to obtain the reduced feature vector, which is then organized into a time series to obtain the bridge status dataset.

3. The intelligent bridge reinforcement decision-making method based on damage identification according to claim 1 is characterized in that: Bridge damage identification and assessment are performed on the bridge status dataset using a multi-branch neural network model to obtain a bridge damage feature map, including: The bridge status dataset is divided into time windows, and the data is segmented using a sliding window algorithm to obtain a time series feature subset. The time series feature subset is then Fourier transformed to obtain a frequency domain feature matrix. The frequency domain feature matrix is ​​processed by convolutional neural network, and local features are extracted through multi-layer convolution and pooling operations to obtain a convolution feature map. The convolution feature map is then globally averaged and pooled to obtain a global feature vector. The global feature vector is processed by the long short-term memory network, and the long-term features are captured by temporal dependency modeling to obtain the temporal feature vector. The temporal feature vector is then processed by the self-attention mechanism to obtain the attention-weighted feature. The attention-weighted features are processed with a fully connected layer and feature mapped using a multi-layer perceptron to obtain a high-dimensional feature representation. Softmax classification is then performed on the high-dimensional feature representation to obtain the probability distribution of the damage type. The probability distribution of damage types is filtered by thresholds, and high-confidence predictions are screened by setting confidence thresholds to obtain damage type determination results. The damage type determination results are spatially mapped to obtain a bridge damage characteristic map.

4. The intelligent bridge reinforcement decision-making method based on damage identification according to claim 1 is characterized in that: Intelligent decision-making analysis is performed on the bridge damage characteristic map to generate a preliminary reinforcement plan and obtain a set of candidate reinforcement strategies, including: The bridge damage feature map is processed using a graph convolutional network. The topological relationship of the damage is obtained through spatial feature extraction, and a damage association graph is obtained. The damage association graph is then embedded into a graph to obtain a damage representation vector. Cluster analysis is performed on the damage representation vectors. Similar damages are grouped using the K-means algorithm to obtain a set of damage categories. The severity of the damage category set is then evaluated to determine the reinforcement priority ranking. Perform knowledge graph matching on the reinforcement priority ranking, select the applicable reinforcement technology through semantic similarity calculation, obtain the preliminary reinforcement plan, and check the constraints of the preliminary reinforcement plan to obtain a set of feasible reinforcement plans; Perform multi-objective optimization on the feasible reinforcement scheme set, balance the reinforcement effect and cost through genetic algorithm, obtain the Pareto optimal solution set, and perform decision tree analysis on the Pareto optimal solution set to obtain the scheme decision path; Monte Carlo simulation is performed on the decision path of the scheme, and the robustness of the scheme is evaluated through random sampling to obtain the scheme reliability index. The scheme reliability index is then sorted and screened to obtain a set of candidate reinforcement strategies.

5. The intelligent bridge reinforcement decision-making method based on damage identification according to claim 1 is characterized in that: The target construction plan is monitored and fed back in real time, and the construction process is dynamically optimized through adaptive control algorithms to obtain the target intelligent reinforcement implementation plan, including: Plan the sensor layout for the target construction plan, determine the monitoring point locations through the optimal coverage algorithm, obtain the monitoring network topology, and set the data collection frequency for the monitoring network topology to obtain a real-time monitoring plan; Collect data for real-time monitoring solutions, pre-process the collected data using edge computing technology to obtain effective monitoring data streams, and perform anomaly detection on the effective monitoring data streams to obtain abnormal events in the construction process; Diagnose the causes of abnormal events during the construction process, analyze the root causes of the abnormalities through Bayesian network reasoning, obtain construction deviation reports, and conduct impact assessments on the construction deviation reports to obtain a list of adjustment priorities; Generate solutions for the adjustment priority list, formulate adjustment strategies through reinforcement learning algorithms, obtain a set of candidate adjustment solutions, and perform simulation verification on the candidate adjustment solution set to obtain the optimal adjustment solution; Generate execution instructions for the optimal adjustment plan, convert them into specific operation instructions through semantic parsing technology, obtain a construction adjustment instruction set, and generate a target intelligent reinforcement implementation plan based on the construction adjustment instruction set.

6. An intelligent bridge reinforcement decision-making device based on damage identification, used to implement the intelligent bridge reinforcement decision-making method based on damage identification as described in any one of claims 1 to 5, characterized in that: The intelligent bridge reinforcement decision-making device based on damage identification includes: The acquisition module is used to collect multi-source data of the bridge structure and pre-process the collected data through a multimodal data fusion algorithm to obtain a bridge status data set; An evaluation module is used to identify and evaluate bridge damage on a bridge status dataset using a multi-branch neural network model to obtain a bridge damage feature map; The analysis module is used to perform intelligent decision-making analysis on the bridge damage characteristic map, generate preliminary reinforcement plans, and obtain a set of candidate reinforcement strategies; The simulation module is used to perform parametric modeling and multi-scale simulation on the candidate reinforcement strategy set, and to perform coupled calculation and evaluation on the reinforcement effect to obtain the optimized reinforcement plan; The planning module is used to conduct virtual construction planning for the optimized reinforcement scheme, generate a construction process model through an intelligent scheduling algorithm, and obtain the target construction plan; The feedback module is used to monitor and provide feedback on the target construction plan in real time, and dynamically optimize the construction process through an adaptive control algorithm to obtain the target intelligent reinforcement implementation plan.

7. A computer device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the computer device is running, the processor and the memory communicate via the bus. When the machine-readable instructions are executed by the processor, the steps of the intelligent bridge reinforcement decision-making method based on damage identification as described in any one of claims 1 to 5 are performed.

Images