Unmanned aerial vehicle-based concrete structure multi-disease collaborative diagnosis system and method

Abstract

The invention discloses a concrete structure multi-disease collaborative diagnosis system and method based on an unmanned aerial vehicle, and belongs to the technical field of structure health monitoring. The system comprises a fiber grating sensor group arranged at a key part of a structure, an unmanned aerial vehicle nest system, an edge calculation unit and a multi-modal data analysis platform. The fiber grating sensor acquires data such as strain, temperature and cracks in real time; when the data is abnormal, the system automatically dispatches the unmanned aerial vehicle for routing inspection to obtain image data; the edge calculation unit is used for realizing image recognition and crack quantification; and the multi-modal data analysis platform fuses sensor data, image data, a BIM model and a historical report to realize disease correlation analysis and risk grading and prediction. According to the invention, all-weather, wide-coverage and quick response of structure monitoring are realized, and the intelligent level and reliability of infrastructure health management are remarkably improved.

Description

[0001] This invention belongs to the interdisciplinary fields of structural health monitoring, low-altitude economy and artificial intelligence. Specifically, it relates to a collaborative diagnosis system and method for concrete structure defects that integrates fiber optic grating sensor networks, UAV mobile inspection and multimodal intelligent analysis. It is applicable to the full life cycle safety operation and maintenance management of major infrastructure such as bridges, tunnels and dams. Background Technology

[0002] In recent years, my country's infrastructure construction has continued to expand. By the end of 2023, the total number of highway bridges nationwide had exceeded one million, of which more than 40% were over 20 years old, highlighting the increasingly prominent problems of structural aging and performance degradation. Traditional structural health monitoring systems mainly rely on manual inspections and fixed sensor networks, which present the following technical bottlenecks: Limited coverage: Fixed sensor networks are usually deployed at pre-set locations, making it difficult to fully cover structural surfaces and hidden areas, and thus unable to achieve rapid response to sudden damage.

[0003] Limited data dimensions: Most systems only collect physical parameters such as strain and vibration, lacking the ability to accurately quantify apparent defects (such as cracks and spalling).

[0004] Low level of intelligence: Data analysis relies on threshold alarms and manual interpretation, lacks multi-source data fusion and autonomous diagnosis capabilities, and the false alarm rate is as high as 30% to 40%.

[0005] Slow response time: It usually takes several hours or even days from data anomaly to on-site confirmation, which cannot meet the needs of post-disaster emergency assessment and rapid decision-making.

[0006] Although existing research has attempted to introduce drones for visual inspection, the following problems still exist: The drone inspection and fixed sensor data operate independently, lacking a spatiotemporal synchronization and collaborative analysis mechanism; Visual detection algorithms suffer from poor stability under complex lighting and occlusion environments, making it difficult to achieve quantitative diagnosis. Insufficient edge computing capabilities lead to high pressure on data backhaul and make it difficult to guarantee real-time performance.

[0007] Therefore, there is an urgent need to develop a collaborative monitoring system that integrates fixed monitoring and mobile inspection, and supports multi-source data fusion and intelligent diagnosis. Summary of the Invention

[0008] Purpose of the invention This invention aims to overcome the shortcomings of existing concrete structure monitoring technologies, such as incomplete monitoring coverage, difficulty in data fusion, large response delay, and low level of intelligence. It provides a hybrid monitoring system and method based on the collaboration of UAVs and fiber Bragg grating sensors to achieve accurate identification, rapid diagnosis, and risk prediction of multiple defects in concrete structures (including internal strain anomalies, temperature deformation, surface cracks, concrete spalling, etc.), providing technical support for the safe operation and maintenance of infrastructure throughout its entire life cycle.

[0009] Technical solution The UAV-based collaborative diagnosis system for multiple defects in concrete structures described in this invention includes a fiber Bragg grating sensor network, a UAV nesting system, an edge computing unit, and a multimodal data analysis platform. The functions and technical parameters of each module are as follows: Fiber Bragg grating sensor networks: Deployed at key stress-bearing parts of concrete structures, including FBG strain gauges (accuracy ±2με), thermometers (accuracy 0.1℃), and crack gauges (resolution 0.05mm). It supports multi-point serial connection, with a maximum of 80 sensors deployed on a single optical fiber, achieving a spatial resolution of 5cm. It has temperature-strain cross-compensation function with an error of <0.5%.

[0010] Unmanned Aerial Vehicle (UAV) Nest System: Equipped with a DJI M300 RTK drone and a Zenmuse H20T payload (visible light, infrared, and lidar). Supports autonomous takeoff and landing, route planning, and multi-aircraft collaborative operations; It possesses millimeter-level positioning accuracy and centimeter-level 3D modeling capabilities.

[0011] Edge computing unit: Deploy lightweight AI algorithms based on the NVIDIA Jetson Orin NX platform; Real-time data preprocessing, anomaly detection, and image recognition (YOLOv8s, 10FPS) are achieved. Supports data compression transmission, saving ≥60% of bandwidth.

[0012] Multimodal data analysis platform: Integrating sensor data, drone imagery, BIM models, and historical data; Integrating the TransUNet segmentation model (mIoU=0.91) and the LSTM temporal prediction model (R²>0.95); It supports disease association analysis, risk classification, and autonomous decision-making.

[0013] working methods The method for collaborative diagnosis of multiple defects in concrete structures based on unmanned aerial vehicles (UAVs) as described in this invention includes the following steps: Routine monitoring: Fiber optic sensors acquire data in real time, and edge nodes perform trend analysis and anomaly detection. The drone inspection task is automatically triggered when data is abnormal.

[0014] Collaborative diagnosis: The drone collects images and 3D point cloud data along a preset flight path; Real-time crack identification and disease quantification are performed on edge units; Multimodal platforms integrate data for damage assessment and risk warning.

[0015] Adaptive optimization: Dynamically optimize inspection paths and sensor sampling frequencies based on reinforcement learning; The alarm thresholds and maintenance strategies are automatically adjusted based on historical data.

[0016] Beneficial effects Compared with the prior art, the present invention has the following significant advantages: Full-coverage monitoring capability: Combining fixed-point sensing and mobile inspection, it achieves full-area monitoring of the structural surface and interior, with no blind spots.

[0017] Multi-source data fusion: Breaking through data silos and achieving multimodal collaborative analysis of "physical parameters + visual features + three-dimensional geometry".

[0018] Real-time response and early warning: Time from abnormal trigger to drone deployment < 5 minutes, and time for generating diagnostic report < 30 minutes.

[0019] Intelligent decision support: Based on AI and digital twin technology, it enables autonomous identification of diseases, life prediction, and maintenance strategy recommendation.

[0020] Scalability and cost-effectiveness: Supports modular deployment and smooth upgrades, reducing long-term operation and maintenance costs by more than 30%. Attached Figure Description Figure 1 This is a schematic diagram of the overall architecture of the system of the present invention; Figure 2 A flowchart of the multimodal data fusion process; Figure 3 Diagram of edge-cloud collaborative computing architecture; Figure 4 This is a system workflow diagram; Detailed Implementation

[0022] Taking a simply supported beam bridge with a span of 50m and an age of 20 years (with potential cracking risks) as an example, the deployment steps are as follows: Sensor deployment: FBG sensor arrays are deployed in key parts of the bridge, such as the main beam, piers, and cables. It adopts magnetic snap-on and wire-wound flexible board packaging, which increases tensile strength by 300% and is suitable for dynamic load environments.

[0023] Drone nesting deployment: Intelligent machine nests are deployed within a 5km radius of the structure to support automatic charging and data transmission; Equip the system with a weather monitoring module to ensure flight safety.

[0024] Edge node installation: Deploy Jetson Orin NX edge servers in the on-site data center; Connect the fiber optic demodulator to the UAV data link to achieve local processing.

[0025] Platform setup: A multimodal analytics platform is built using a microservices architecture, supporting containerized deployment; Integrate BIM models with historical databases to build a digital twin foundation.

[0026] Workflow Example (Taking Bridge Crack Diagnosis as an Example) The FBG sensor detected an abnormal strain (≥50με), and the edge node triggered an early warning. The platform automatically generates drone inspection tasks, with flight paths covering abnormal areas; The drone collects high-definition images and infrared data, and the edge unit runs the YOLOv8s model in real time to identify cracks; The platform integrates strain data with visual features to assess crack length, width, and development trend; Generate diagnostic reports and recommend repair priorities and maintenance plans.

[0027] Key technical parameters Technical indicators Performance parameters Strain measurement accuracy ±2με Temperature measurement accuracy 0.1℃ Crack identification accuracy 96% (visible light), 90% (infrared) Drone response time <5 minutes Data fusion processing latency <200ms (edge), <10s (cloud) System power consumption Sensor node: 0.1W, edge unit: 15W Communication Protocol LoRaWAN / NB-IoT / 5G heterogeneous network compatibility Typical application scenarios

[0028] Bridge monitoring: Applicable to various bridge types such as beam bridges, arch bridges, and cable-stayed bridges, supporting load testing, long-term monitoring, and emergency assessment.

[0029] Tunnel monitoring: To address issues such as lining cracks, leakage, and deformation, automated inspections are implemented throughout the entire tunnel.

[0030] Dam monitoring: Combining underwater drones and fiber optic sensors to achieve multi-dimensional monitoring of the dam body and foundation.

[0031] Urban elevated roads and subways: Support real-time monitoring of structural safety in dense traffic environments.

[0032] in conclusion This invention utilizes fiber optic grating sensor networks and UAV nesting systems for deep collaboration, combined with edge computing and multimodal data analysis technologies, to construct an integrated collaborative diagnostic platform that integrates "fixed monitoring + mobile inspection," "real-time perception + intelligent analysis," and "emergency response + predictive maintenance." This platform addresses the problems of insufficient monitoring coverage, difficulties in data fusion, slow response, and low intelligence in traditional monitoring systems.

[0033] The system features high precision, all-weather operation, and full automation, which can significantly improve the level of infrastructure safety management. Its modular design makes it highly scalable and economical, and it is widely applicable to bridges, tunnels, dams and other fields. It provides key support for smart cities, smart transportation and disaster prevention and mitigation, and has broad application prospects and market value.

Claims

1. A collaborative diagnostic system for multiple defects in concrete structures based on unmanned aerial vehicles (UAVs), characterized in that, include: Fiber grating sensor arrays are deployed in key parts of concrete structures to collect strain, temperature, and crack data in real time. The drone nest system automatically takes off after receiving an abnormal trigger signal and performs image acquisition and 3D modeling of a designated area. Edge computing units, deployed in the field, are used for image recognition, crack quantification, and defect detection; A multimodal data analysis platform is used to integrate fiber Bragg grating data, UAV imagery, BIM models, and historical data to perform disease correlation analysis, risk classification, and prediction.

2. The system according to claim 1, characterized in that, The multimodal data analysis platform integrates large models for cross-modal data fusion and intelligent decision support.

3. The system according to claim 1, characterized in that, The system also includes an adaptive threshold adjustment module, which is used to dynamically adjust the anomaly trigger threshold and inspection strategy based on historical data.

4. A collaborative diagnosis method for multiple defects in concrete structures based on unmanned aerial vehicles (UAVs), characterized in that, Includes the following steps: The structural status is monitored in real time using fiber optic grating sensors; When the monitored data exceeds the preset threshold, the drone inspection task is automatically triggered; The drone collects image data and transmits it to the edge computing unit for real-time processing. The processed data is then fused and analyzed with historical data and BIM models using a multimodal approach. Output disease diagnosis results, risk level, and maintenance recommendations.

Images