Bridge health maintenance intelligent decision platform system based on deep learning

Abstract

The invention relates to the technical field of bridge maintenance management, in particular to a bridge health maintenance intelligent decision platform system based on deep learning, which comprises a data acquisition module used for acquiring real-time monitoring data of a target bridge and a current detection report and a maintenance strategy; the deep learning analysis module is used for analyzing the collected data based on a pre-trained deep learning model to obtain a damage mode and a future damage trend of the target bridge; the intelligent decision support module is used for analyzing the current health state of the target bridge based on the deep learning analysis result and generating targeted maintenance suggestions and priority ranking; and the quality control and feedback optimization module is used for auditing the deep learning analysis result, checking whether detection omission or detection abnormal data exists or not, and feeding back the detection omission or detection abnormal data to the data acquisition module and the intelligent decision support module for optimization. According to the invention, comprehensive monitoring, intelligent evaluation and maintenance strategy recommendation of the steel cable bridge can be realized, and the efficiency and accuracy of bridge maintenance are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of bridge maintenance management, and in particular to a bridge health maintenance intelligent decision-making platform system based on deep learning. Background Art

[0002] Currently, the safety of cable bridges relies on regular manual inspections and emerging automated monitoring technologies, such as the application of cable-climbing robots and sensor systems, to identify structural damage and performance degradation. However, these data are often analyzed independently and lack the ability to comprehensively evaluate and predict the health status of bridges, resulting in maintenance decisions that are not timely and accurate. Therefore, it is particularly important to develop an intelligent platform that can integrate multi-source data, conduct in-depth analysis, and autonomously make maintenance recommendations. Summary of the invention

[0003] The purpose of the present invention is to provide a bridge health maintenance intelligent decision-making platform system based on deep learning, to solve the problems of decentralized and inefficient bridge health monitoring and maintenance decision-making in the prior art, to achieve comprehensive monitoring, intelligent evaluation and maintenance strategy recommendation of cable bridges, and to improve the efficiency and accuracy of bridge maintenance.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] The intelligent decision-making platform system for bridge health maintenance based on deep learning includes: data acquisition module, deep learning analysis module, intelligent decision support module, quality control and feedback optimization module;

[0006] The data acquisition module is used to collect real-time monitoring data of the target bridge as well as current inspection reports and maintenance strategies;

[0007] The deep learning analysis module is used to analyze the collected data based on the pre-trained deep learning model to obtain the damage pattern and future damage trend of the target bridge;

[0008] The intelligent decision support module is used to analyze the current health status of the target bridge based on the deep learning analysis results and generate targeted maintenance suggestions and priority rankings;

[0009] The quality control and feedback optimization module is used to review the deep learning analysis results, check and determine whether there are any detection omissions or abnormal detection data, and feed back to the data acquisition module and intelligent decision support module for optimization.

[0010] Optionally, the data acquisition module includes: a current data acquisition unit, a data processing and fusion unit;

[0011] The current data acquisition unit is used to collect real-time monitoring data of sensors of the target bridge, video image data, and file data of current inspection reports and maintenance strategies;

[0012] The data processing and fusion unit is used to perform standardization processing on the collected data and to perform fusion processing on the standardized data.

[0013] Optionally, the real-time monitoring data of the sensor includes temperature, humidity, cable tension, various structural stresses, vibration, deflection, and degree of corrosion; the video image data includes images and video data collected by traffic monitoring cameras, rope-climbing robots, and flexible detection tools; the file data of the current inspection report and maintenance strategy includes manually processed, confirmed inspection reports, and maintenance response strategy documents.

[0014] Optionally, the standardization processing includes normalization and feature encoding processing, and the fusion processing includes time series fusion, multimodal data fusion, and heterogeneous data isomorphism processing.

[0015] Optionally, the deep learning analysis module includes: a model training unit and a data analysis unit;

[0016] The model training unit is used to train a deep learning model based on a pre-built database, wherein the database contains historical sensor data, historical detection reports, and historical maintenance strategy data processed by the data processing and fusion unit;

[0017] The data analysis unit is used to input the collected data into the trained deep learning model, output the overall health status, damage pattern, future damage trend and historical maintenance strategy tracking of the target bridge, and generate a new inspection report.

[0018] Optionally, the deep learning model includes time series analysis, image recognition algorithm and multimodal fusion network.

[0019] Optionally, the intelligent decision support module includes: a health assessment unit and an auxiliary decision unit;

[0020] The health assessment unit is used to assess the health level and potential risks of the target bridge based on the new inspection report;

[0021] The auxiliary decision-making unit generates targeted maintenance suggestions and priority rankings based on the health level and potential risks of the target bridge and the effectiveness of historical maintenance strategies.

[0022] Optionally, the quality control and feedback optimization module includes: a result review unit and a feedback optimization unit;

[0023] The result review unit is used to review new test reports, obtain data detection omissions or inconsistencies, and generate adjustment suggestions;

[0024] The feedback optimization unit is used to feed back the adjustment suggestions to the data acquisition module and the intelligent decision support module to optimize the system.

[0025] The beneficial effects of the present invention are:

[0026] The present invention realizes comprehensive, real-time monitoring and intelligent evaluation of the health status of bridges, significantly improving the scientific nature and timeliness of maintenance decisions; deeply mines data value through deep learning algorithms, improving the pertinence and effectiveness of maintenance measures; automated decision support reduces the workload of manual analysis and improves the efficiency and quality of bridge maintenance management; the intelligent audit mechanism ensures the accuracy and completeness of the inspection report and enhances the systematicness and continuity of maintenance work. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0028] Figure 1 This is a schematic diagram of the structure of a bridge health maintenance intelligent decision-making platform system based on deep learning according to an embodiment of the present invention;

[0029] Figure 2 The following is a working diagram of a data acquisition module according to an embodiment of the present invention;

[0030] Figure 3 A processing flow chart of a detection report and a maintenance strategy according to an embodiment of the present invention;

[0031] Figure 4 This is a workflow diagram of a deep learning analysis module according to an embodiment of the present invention;

[0032] Figure 5 A flowchart of the intelligent decision support module according to an embodiment of the present invention;

[0033] Figure 6 is a workflow diagram of the quality control and feedback optimization module of an embodiment of the present invention;

[0034] Figure 7 The platform structure and workflow diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0037] This embodiment provides a bridge health maintenance intelligent decision-making platform system based on deep learning, such as Figure 1 As shown, it includes: data acquisition module, deep learning analysis module, intelligent decision support module, quality control and feedback optimization module;

[0038] Data collection module, used to collect real-time monitoring data of the target bridge as well as current inspection reports and maintenance strategies;

[0039] Deep learning analysis module, which is used to analyze the collected data based on the pre-trained deep learning model to obtain the damage pattern and future damage trend of the target bridge;

[0040] Intelligent decision support module, which is used to analyze the current health status of the target bridge based on the results of deep learning analysis and generate targeted maintenance recommendations and priority ranking;

[0041] The quality control and feedback optimization module is used to review the deep learning analysis results, check to determine if there are any missed detections or abnormal detection data, and provide feedback to the data acquisition module and intelligent decision support module for optimization.

[0042] Furthermore, the data acquisition module includes: a current data acquisition unit, a data processing and fusion unit;

[0043] The current data acquisition unit is used to collect the real-time monitoring data of the target bridge's sensors, video image data, and file data of the current inspection report and maintenance strategy;

[0044] The data processing and fusion unit is used to standardize the collected data and fuse the standardized data.

[0045] Specifically, the real-time monitoring data of the sensors include but are not limited to temperature, humidity, cable tension, various structural stresses, vibration, deflection, corrosion degree, etc.

[0046] Video image data includes but is not limited to data from fixed surveillance cameras, such as traffic monitoring cameras, rope-climbing robots, and images and video data collected by flexible detection tools;

[0047] The file data of the current test report and maintenance strategy are the test reports, maintenance response strategy documents, etc. that are automatically generated by the system and are manually processed and confirmed.

[0048] Standardization of collected data includes data cleaning and standardization. Standardization includes:

[0049] The first is normalization: data of the same type but in different ranges are converted to the same scale, such as the interval [0,1] or [-1,1]. Common methods include Min-Max Scaling and Z-score standardization.

[0050] Feature encoding: For categorical data (such as material type, damage type), One-Hot Encoding, Label Encoding and other technologies are used to convert them into numerical data.

[0051] The standardized data is fused, and the following processing is used for multi-modal and multi-temporal heterogeneous data:

[0052] Time series fusion: For multi-temporal data, time series analysis techniques (such as ARIMA, state-space models) or deep learning models (such as LSTM, GRU) can be used to capture the trends and periodicity of data over time.

[0053] Multimodal data fusion: Use multimodal fusion networks in deep learning (such as Multi-Modal Fusion Networks, Attention-based Fusion Models) to weight the contribution of different data sources through attention mechanisms or gating mechanisms to achieve more effective information integration.

[0054] Homogeneity of heterogeneous data: Map different types of data to the same feature space through mapping learning (such as Autoencoder, Generative Adversarial Networks) to facilitate subsequent analysis and fusion.

[0055] like Figure 2 As shown in the figure, the data acquisition module regularly collects the real-time monitoring data of the target bridge through the real-time information acquisition input interface, video image data, and file data of the current inspection report and maintenance strategy, and saves them to the database as historical data after standardization and fusion processing. The specific workflow is as follows:

[0056] For real-time monitoring data from sensors, the signal acquisition terminal calls the interface and format provided by the system through the network and sends the data to the system platform in real time. After receiving the data, the system platform processes and fuses it through the data processing and fusion unit and saves it to the historical sensor database according to the predetermined data structure;

[0057] like Figure 3 As shown, for the current inspection report, the inspection report usually exists in the form of a document, and the system provides a file upload interface. And for the inspection report of the cable-stayed bridge, there are corresponding industry professional vocabulary. After the file is uploaded, natural language processing is integrated into the data processing and fusion unit to recognize the text of the report, and extract the key report parameters (such as whether there is damage, damaged parts, location, severity, etc.), store them according to the predetermined data structure, and store them in the historical inspection report database for subsequent deep learning module use;

[0058] The current maintenance strategy also exists in the form of a document. Its processing form is similar to that of the inspection report, except that the corresponding key parameters are different. After processing, it is stored according to the predetermined data structure and stored in the historical maintenance strategy database for use by subsequent deep learning modules.

[0059] Furthermore, the deep learning analysis module includes: a model training unit, a data analysis unit;

[0060] A model training unit, used to train a deep learning model based on a pre-built database, wherein the database contains historical sensor data, historical detection reports, and historical maintenance strategy data processed by the data processing and fusion unit;

[0061] The data analysis unit is used to input the collected data into the trained deep learning model, output the overall health of the target bridge, damage patterns, future damage trends, and tracking of historical maintenance strategies, and generate a new inspection report.

[0062] Specifically, Figure 4 As shown, the current data of the target bridge is input into the deep learning model trained by the database, and the overall health status of the target bridge, the ranking of damage items, future development trends, and the tracking of the effectiveness of maintenance of damage items in previous years are output through the deep learning analysis output interface.

[0063] The deep learning model combines time series analysis, image recognition algorithms and multimodal fusion networks. Time series analysis (such as LSTM) analyzes the temporal changes of temperature data, cable tension, various structural stresses, bridge deflection detection data, and traffic flow. Image recognition algorithms (such as CNN) analyze the images and video data collected by the cable-climbing robot and flexible detection tools to obtain the damage of the target structure, such as the identification of cracks and notches in a cable. Multimodal fusion networks (such as Multimodal Fusion Networks) can analyze the mutual influence between temperature, cable tension, structural stress, deflection changes, traffic flow, and target structural damage. For example, the stress change patterns of locations with structural damage and locations without structural damage under the same temperature and traffic flow, or the change patterns of the same location under different temperatures and traffic flows, as well as the stress changes of various components and the changes in bridge deflection. Combined with the basic information data of bridge structure materials, we can further analyze and judge under what kind of damage, stress, traffic flow, temperature and other comprehensive conditions a specific structure may break; or based on the current rules, predict the damage law of a certain structure under the current comprehensive conditions, and further predict the service life of a certain structure; we can also analyze based on the stress changes of each component and the law of bridge deflection changes, under what kind of damage, stress, traffic flow, temperature and other comprehensive conditions the bridge will have excessive deflection or even the risk of sudden collapse.

[0064] Furthermore, the intelligent decision support module includes: a health assessment unit, an auxiliary decision unit;

[0065] A health assessment unit, which is used to assess the health level and potential risks of the target bridge based on the new inspection report;

[0066] The auxiliary decision-making unit generates targeted maintenance recommendations and priority rankings based on the health level and potential risks of the target bridge and the effectiveness of historical maintenance strategies.

[0067] Specifically, Figure 5 As shown in the figure, the intelligent decision support module defines a fixed analysis report format template based on the analysis results of the deep learning model and the laws of mutual influence and change, combined with the national standards for bridge health analysis ("Technical Condition Assessment Standard for Highway Bridges" (JTG / TH21-2011), "Technical Standard for Urban Bridge Maintenance" (CJJ 99-2017), and "Highway Bridge Seismic Design Code" (JTGB02-01-2008)), and automatically generates a bridge health inspection report. At the same time, based on the relationship between historical damage and response strategies, it automatically identifies and generates the strategies that should be adopted based on the analysis results, as well as the priority ranking of the strategies.

[0068] Furthermore, the quality control and feedback optimization module includes: a result review unit and a feedback optimization unit;

[0069] The result review unit is used to review new test reports, obtain data test omissions or inconsistencies, and generate adjustment suggestions;

[0070] The feedback optimization unit is used to feed back adjustment suggestions to the data acquisition module and the intelligent decision support module to optimize the system.

[0071] Specifically, Figure 6 As shown, the quality control and feedback optimization module conducts intelligent review of the deep learning analysis results, that is, the output new test report, compares historical data, checks whether there are any test omissions or inconsistencies, generates adjustment suggestions and feeds back to the data acquisition module and the intelligent decision support module, and continuously optimizes the system accuracy.

[0072] Furthermore, if Figure 7 As shown, through the system development platform proposed in this embodiment, the platform includes a data acquisition layer, a data layer, a core processing layer, and an application layer; the data acquisition layer collects video surveillance cameras, regular detection, various sensor data, maintenance strategies, etc., inputs the data into the core processing layer for processing, and then saves it into the database of the data layer, and then trains the deep learning model through the database, performs deep learning analysis on the real-time data based on the trained deep learning model, and outputs the overall health status of the bridge, the ranking of damage items and future development trends, the tracking of the effective maintenance of damage items in previous years, and intelligent decision support, etc. at the application layer, and performs other functions such as quality control and feedback optimization.

[0073] This embodiment realizes comprehensive, real-time monitoring and intelligent assessment of the health status of the bridge, significantly improving the scientific nature and timeliness of maintenance decisions; deeply mining the value of data through deep learning algorithms improves the pertinence and effectiveness of maintenance measures; automated decision support reduces the workload of manual analysis and improves the efficiency and quality of bridge maintenance management; the intelligent audit mechanism ensures the accuracy and completeness of the inspection report and enhances the systematicness and continuity of maintenance work.

[0074] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. The bridge health maintenance intelligent decision-making platform system based on deep learning is characterized by: include: Data acquisition module, deep learning analysis module, intelligent decision support module, quality control and feedback optimization module; The data acquisition module is used to collect real-time monitoring data of the target bridge as well as current inspection reports and maintenance strategies; The deep learning analysis module is used to analyze the collected data based on the pre-trained deep learning model to obtain the damage pattern and future damage trend of the target bridge; The intelligent decision support module is used to analyze the current health status of the target bridge based on the deep learning analysis results and generate targeted maintenance suggestions and priority rankings; The quality control and feedback optimization module is used to review the deep learning analysis results, check and determine whether there are any detection omissions or abnormal detection data, and feed back to the data acquisition module and intelligent decision support module for optimization.

2. The bridge health maintenance intelligent decision-making platform system based on deep learning according to claim 1 is characterized in that: The data acquisition module includes: a current data acquisition unit, a data processing and fusion unit; The current data acquisition unit is used to collect real-time monitoring data of sensors of the target bridge, video image data, and file data of current inspection reports and maintenance strategies; The data processing and fusion unit is used to perform standardization processing on the collected data and to perform fusion processing on the standardized data.

3. The deep learning-based bridge health maintenance intelligent decision-making platform system according to claim 2 is characterized in that: The real-time monitoring data of the sensor include temperature, humidity, cable tension, various structural stresses, vibration, deflection, and corrosion degree; the video image data include images and video data collected by traffic monitoring cameras, rope-climbing robots, and flexible detection tools; the file data of the current inspection report and maintenance strategy include manually processed, confirmed inspection reports, and maintenance response strategy documents.

4. The bridge health maintenance intelligent decision-making platform system based on deep learning according to claim 2 is characterized in that: The standardization processing includes normalization and feature encoding processing, and the fusion processing includes time series fusion, multimodal data fusion, and heterogeneous data isomorphism processing.

5. The deep learning-based bridge health maintenance intelligent decision-making platform system according to claim 1 is characterized in that: The deep learning analysis module includes: a model training unit and a data analysis unit; The model training unit is used to train a deep learning model based on a pre-built database, wherein the database contains historical sensor data, historical detection reports, and historical maintenance strategy data processed by the data processing and fusion unit; The data analysis unit is used to input the collected data into the trained deep learning model, output the overall health status, damage pattern, future damage trend and historical maintenance strategy tracking of the target bridge, and generate a new inspection report.

6. The deep learning-based bridge health maintenance intelligent decision-making platform system according to claim 5 is characterized in that: The deep learning model includes time series analysis, image recognition algorithm and multimodal fusion network.

7. The bridge health maintenance intelligent decision-making platform system based on deep learning according to claim 1 is characterized in that: The intelligent decision support module includes: a health assessment unit and an auxiliary decision unit; The health assessment unit is used to assess the health level and potential risks of the target bridge based on the new inspection report; The auxiliary decision-making unit generates targeted maintenance suggestions and priority rankings based on the health level and potential risks of the target bridge and the effectiveness of historical maintenance strategies.

8. The deep learning-based bridge health maintenance intelligent decision-making platform system according to claim 1 is characterized in that: The quality control and feedback optimization module includes: a result review unit and a feedback optimization unit; The result review unit is used to review new test reports, obtain data detection omissions or inconsistencies, and generate adjustment suggestions; The feedback optimization unit is used to feed back the adjustment suggestions to the data acquisition module and the intelligent decision support module to optimize the system.

Images