Steel structure bolt detection system suitable for complex environment

Abstract

The invention relates to a steel structure bolt detection system suitable for a complex environment. The magnetic type wall-climbing robot is used for acquiring pre-tightening force or loosening state data of a bolt, the unmanned aerial vehicle is used for acquiring a missing, rusting or damage state image of the bolt, the multi-modal fusion judgment unit is used for performing data fusion and intelligent analysis on detection data, and the path planning and cooperative control unit is used for performing cooperative control on the detection data. The maintenance work order generation unit is used for automatically generating a maintenance work order comprising a fault position, a fault type and suggested measures according to a comprehensive state judgment result, so that direct and quantitative detection on the mechanical state of the bolt is realized through the magnetic wall-climbing robot, and the problems of subjectivity and incapability of quantification of a manual knocking method are solved.

Description

Technical Field

[0001] This application relates to the field of building inspection technology, specifically to the technical field of a steel structure bolt inspection system suitable for complex environments. Background Technology

[0002] In the construction and operation of transportation infrastructure such as highway entrances and exits, interchanges, and service areas, steel structures such as space frames, tubular trusses, and steel frames are widely used to achieve the goals of large-span and high-efficiency construction. The safety and durability of such structures are highly dependent on the reliability of the high-strength bolt connections between components. Therefore, whether the bolt preload meets the standard, and whether there is any loosening, damage, or early corrosion, are key factors affecting the overall performance of the structure and must be strictly tested and monitored.

[0003] Currently, the inspection of steel structure bolts in the aforementioned scenarios mainly relies on manual inspection methods, such as using torque wrenches for verification, tapping with a small hammer to listen to the sound, and visual observation for judgment. However, this method poses safety hazards to the inspectors, is highly dependent on the operator's experience, is prone to misjudgment and missed detection, and cannot quantitatively assess the internal stress state of the bolts.

[0004] In response, patent CN107632071A, concerning a bolt inspection device and method, discloses a railcar-based bolt inspection system. This system uses an ultrasonic probe mounted on a moving inspection platform along a pre-set track to perform non-destructive testing on bolts along the track. This approach reduces reliance on manual labor and improves the continuity of inspection to some extent. However, there is still room for improvement in terms of flexibility (a dedicated track must be pre-laid, and the installation and disassembly process is complex) and environmental adaptability. Its use is particularly limited in complex environments.

[0005] Therefore, existing technologies lack a comprehensive steel structure bolt detection system that can achieve flexible deployment, efficient operation, multimodal perception, and intelligent decision-making in complex, high-altitude, and dangerous building environments. Summary of the Invention

[0006] This application proposes a steel structure bolt inspection system suitable for complex environments to overcome the safety and efficiency bottlenecks of manual inspection, as well as the shortcomings of traditional automated equipment in terms of flexibility, adaptability and intelligence, thereby achieving comprehensive, accurate and closed-loop management of bolt connection status.

[0007] To achieve the above objectives, the present application adopts the following technical solution: This application proposes a steel structure bolt inspection system suitable for complex environments, comprising: The magnetic wall-climbing robot is used to move by adhering to the surface of steel structures. It is equipped with mechanical sensors to perform mechanical sensing and detection on bolts and obtain data on the preload or looseness of the bolts. Drones are used to fly around steel structures and are equipped with visual sensors to visually identify and detect bolts, acquiring images of bolts that are missing, corroded, or damaged. The multimodal fusion and discrimination unit is used to receive and synchronize the detection data of the magnetic wall-climbing robot and the drone, perform data fusion and intelligent analysis, and generate a comprehensive state discrimination result of the bolt; The path planning and collaborative control unit is used to plan and detect paths for the magnetic wall-climbing robot and the drone, and coordinate their collaborative operation. The data transmission unit is used to transmit detection data and judgment results to a remote terminal in real time via a 5G network; The maintenance work order generation unit is used to automatically generate maintenance work orders containing fault location, type, and suggested measures based on the comprehensive status judgment results, thereby realizing intelligent closed-loop management of detection and maintenance.

[0008] This system utilizes a magnetically attached wall-climbing robot to directly and quantitatively inspect the mechanical state of bolts, overcoming the subjectivity and lack of quantification inherent in manual tapping methods. Combined with rapid, wide-area visual recognition of bolt surface conditions using drones, it addresses the low efficiency and high-risk nature of manual inspections, particularly at heights. This collaborative approach covers both the qualitative and quantitative dimensions of inspection. Multimodal fusion significantly improves the accuracy and reliability of the inspection, avoiding misjudgments from single-method testing. Real-time 5G transmission and automatic maintenance work order generation enable intelligent, closed-loop management from inspection to maintenance, greatly enhancing the safety, efficiency, accuracy, and intelligence of steel structure bolt inspection in complex environments.

[0009] In some possible implementations, both the magnetic wall-climbing robot and the drone are integrated with a SLAM positioning system for autonomous navigation and precise positioning in complex environments without GPS signals.

[0010] In some possible implementations, a UWB positioning base station is also included. The magnetic wall-climbing robot and the drone are equipped with UWB tags. The UWB positioning base station works in conjunction with the SLAM positioning system to provide high-precision position calibration to improve the accuracy of the detection path.

[0011] In some possible implementations, the mechanical sensor includes a torque sensor, strain gauge, or vibration sensor for directly measuring or indirectly estimating the preload of the bolt.

[0012] In some possible implementations, the magnetic wall-climbing robot uses a tracked chassis, its magnetic module is an electromagnet with adjustable magnetic force, and its front end is equipped with a retractable mechanical interface for mounting mechanical sensors.

[0013] In some possible implementations, the visual sensor includes a high-definition camera, an infrared thermal imager, or a laser scanner for acquiring visible light images, temperature distribution, or three-dimensional morphology information of the bolt.

[0014] In some possible implementations, the multimodal fusion discrimination unit employs a deep learning-based algorithm model to extract and fuse features from mechanical sensing data and visual recognition images, and outputs a probabilistic discrimination result for the bolt state.

[0015] In some possible implementations, the path planning and collaborative control unit automatically plans the optimal detection path based on the BIM model of the steel structure or a pre-scanned 3D point cloud map, and dynamically adjusts the relative positions and operation sequence of the magnetic wall-climbing robot and the drone.

[0016] In some possible implementations, the data transmission unit supports dual-mode 5G and Wi-Fi communication, and automatically switches to Wi-Fi hotspot for data backhaul when the 5G signal is weak.

[0017] In some possible implementations, the maintenance work orders generated by the maintenance work order generation unit are pushed to on-site maintenance personnel via mobile terminals, and the work order execution status is supported for feedback and closed-loop tracking. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the steel structure bolt detection system applicable to complex environments as described in this application; Figure 2 This is a schematic diagram illustrating an implementation of the magnetic wall-climbing robot in this application. Detailed Implementation

[0019] The following examples further illustrate the features of this application and other related features in detail, so as to facilitate understanding by those skilled in the art: It should be noted that the terms “front,” “back,” “left,” “right,” “up,” and “down” used in the following description refer to the directions in the attached diagrams, while the terms “bottom surface,” “top surface,” “inner,” and “outer” refer to the directions toward or away from the geometric center of a specific component, respectively.

[0020] Furthermore, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this case based on the specific circumstances.

[0021] like Figure 1 As shown, this invention provides a steel structure bolt inspection system suitable for complex environments, aiming to solve the problems of low efficiency and poor safety of manual inspection in existing technologies, as well as the lack of flexibility and low level of intelligence of traditional automated equipment. This system is particularly suitable for the construction and operation and maintenance of large steel structure buildings in scenarios such as highway entrances and exits, interchanges, and service areas, including toll booth canopies, management rooms, and pedestrian overpasses.

[0022] The core of this system lies in constructing a multimodal intelligent inspection system that utilizes a magnetic wall-climbing robot and a drone in collaborative operation. In typical steel frame or space frame structures, numerous high-strength bolts are used to connect steel beams, columns, and truss components. These connection points are widely distributed, at considerable heights, and in complex environments. Traditional manual inspections not only face significant safety risks associated with working at heights but also rely on experience for results, making quantification difficult. This system, by introducing an autonomous mobile platform, achieves automated, precise, and closed-loop management of bolt status.

[0023] Specifically, such as Figure 2 As shown, the magnetic wall-climbing robot adopts a tracked chassis design with an integrated electromagnet module at the bottom that allows it to stably adhere to steel structure surfaces of varying thicknesses and curvatures. The robot's front end is equipped with a retractable mechanical interface for mounting mechanical sensing devices such as torque sensors, strain gauges, or vibration sensors (these sensors are industry-standard technologies and are not shown in the diagram here; the illustration primarily demonstrates the feasibility of implementation via the robot). During the inspection process, the robot autonomously moves to the target bolt location according to a system-planned path, contacts the bolt head with a probe, and measures its torque, strain, or vibration response to quantitatively assess the bolt's preload or looseness. This method overcomes the subjectivity and non-quantitative shortcomings of the traditional "hammer tapping" method, achieving direct and accurate detection of the bolt's mechanical properties.

[0024] For the mechanical sensing device described above, a possible implementation is as follows: the torque sensor is based on the strain gauge measurement principle, forming a Wheatstone bridge by attaching strain gauges to an elastic body. When the bolt is subjected to torque, the elastic body undergoes slight deformation, causing a change in resistance, thereby outputting an electrical signal proportional to the torque. The strain gauges are used to directly measure the axial strain of the bolt shank or head under preload. Combined with the elastic modulus of the material and the bolt's geometric parameters, the magnitude of the preload can be calculated. The vibration sensor uses a piezoelectric accelerometer, which excites the bolt to generate free or forced vibration, collects its vibration signal, and performs spectral analysis. The degree of bolt loosening is assessed by utilizing changes in the natural frequency or damping ratio. The above-mentioned mechanical sensing devices are all known non-destructive testing techniques in the field. Their basic principles and applications have been maturely practiced in the field of structural health monitoring. This invention integrates them into a magnetic wall-climbing robot platform, realizing automated and quantitative detection of bolt connection status in complex field environments.

[0025] Meanwhile, drones autonomously inspect the steel structure, equipped with high-definition cameras, infrared thermal imagers, and laser scanners to perform wide-area, rapid visual identification of bolts. The high-definition cameras capture surface defects such as missing bolts, corrosion, and cracks; the infrared thermal imagers monitor abnormal temperature rises in connection areas, indicating potential contact defects; and the laser scanners generate high-precision 3D point clouds to aid in geometric deformation analysis. The introduction of drones significantly improves inspection coverage and efficiency, making them particularly suitable for areas difficult for humans to access, such as high altitudes and near edges.

[0026] To ensure accurate positioning of robots and drones in complex indoor or tall structures without GPS signals, the system employs a fusion strategy of SLAM (Simultaneous Localization and Mapping) and UWB (Ultra-Wideband) positioning. The SLAM system, based on LiDAR and an inertial measurement unit (IMU), builds a real-time environmental map and performs autonomous navigation, providing continuous pose estimation. Building upon this, a network of UWB positioning base stations deployed on-site provides centimeter-level absolute position correction for the mobile platform through Time-of-Flight (ToF) measurements. The two systems fuse data using an Extended Kalman Filter (EKF) algorithm. SLAM provides environmental perception and path continuity, while UWB effectively suppresses the cumulative drift of SLAM, forming a complementary and progressive high-precision positioning capability that significantly improves the accuracy of path detection and operational reliability. This SLAM / UWB fusion positioning strategy can also be used in magnetic wall-climbing robots.

[0027] The core intelligence of the system lies in its multimodal fusion discrimination unit. This unit receives and synchronously processes mechanical sensing data transmitted from the magnetic robot and visual recognition images transmitted from the drone, employing a deep learning-based dual-stream neural network model for intelligent analysis. The mechanical data stream passes through a one-dimensional convolutional neural network (1D-CNN) to extract temporal features, while the visual data stream passes through a two-dimensional convolutional neural network (2D-CNN) to extract spatial features. These two feature streams are weighted and integrated in a fusion layer, and the classifier ultimately outputs a comprehensive discrimination result for the bolt's condition (e.g., "normal," "insufficient preload," "rust," etc.) and its confidence level. For example, when the system simultaneously detects that a bolt has low preload and obvious rust on its surface, the fusion model can comprehensively determine it as a high-risk fault requiring immediate attention, avoiding misjudgments or omissions that might occur with a single detection method, significantly improving the accuracy and reliability of the diagnosis.

[0028] In another application scenario, a magnetic wall-climbing robot, while performing mechanical performance testing, discovered that although the preload of a specific bolt was within the normal range, an abnormal vibration pattern was detected in the area where the bolt was located by an integrated vibration sensor. Simultaneously, a high-definition camera on a drone detected a slight misalignment of the bolt head during visual inspection, which could be due to minor slippage caused by long-term dynamic loads.

[0029] In this situation, relying solely on a single detection method, such as considering only the preload measurement result, might misjudge the bolt as "normal" because its preload value does not exceed the set safety threshold. However, by combining the results of vibration analysis and visual recognition, the multimodal fusion discrimination unit can more comprehensively assess the bolt's condition. Abnormal frequency components in the vibration signal may indicate that the bolt and its surrounding structure are undergoing unexpected dynamic behavior, while minute shifts in the image provide direct physical evidence that the bolt may have begun to loosen.

[0030] At this point, the fusion model integrates information from different sensing channels, extracts features using deep learning algorithms, weights and integrates them, and makes a final judgment. Considering abnormal vibrations and visual deviations, even if the preload is still within the standard range, the system will still mark the bolt as "potential risk" or "requiring attention," suggesting further manual verification or proactive preventative maintenance. This approach not only avoids missed judgments that might result from single-dimensional data but also significantly improves the reliability of the overall detection system and the accuracy of fault prediction through the effective integration of multiple information sources.

[0031] The path planning and collaborative control unit automatically plans the optimal inspection path based on the BIM (Building Information Modeling) of the steel structure or a pre-scanned 3D point cloud map. The system can identify the 3D coordinates of all bolt nodes and allocate tasks and coordinate paths based on the differences in capabilities between robots and drones (e.g., robots excel at mechanical inspection but move slowly, while drones excel at visual scanning but cannot make physical contact). For example, in a large space frame structure, the system can plan for a drone to scan first, and after identifying suspected problem areas, instruct the robot to proceed for focused mechanical verification, achieving efficient and intelligent collaborative operation.

[0032] For example, in an inspection task of a steel-structured pedestrian overpass spanning a highway lane, due to the continuous flow of traffic beneath the overpass, the safety risks are high and the working window is limited. Therefore, the system initiated a fully automated inspection process. First, a drone took off from a safe area and performed rapid 3D scanning and high-definition imaging along the main beam and support nodes of the overpass, completing the first round of wide-area visual screening. During image analysis, the system identified surface corrosion signs in several bolt connection areas located at the mid-span of the main beam, and some bolt washers showed signs of loosening and falling off. The path planning and collaborative control unit then dynamically adjusted the task instructions, marking these areas as "high-concern areas," and planned an optimal path in real time, guiding a magnetic climbing robot to start from the side of the overpass and precisely climb along the web of the steel beam to the aforementioned corrosion nodes. After the robot arrived, it automatically deployed mechanical sensors to retest the preload of the relevant bolts and analyze the vibration response, confirming that the preload of two bolts had dropped to 65% of the design value, posing a structural safety hazard. This comprehensive judgment immediately triggered an alarm and generated an emergency repair work order, which was then pushed to the maintenance team. Through a collaborative model of "wide-area initial screening by drones followed by precise inspection by robots," the system can identify high-risk fault points in the shortest possible time, avoiding the inefficient task of mechanically inspecting each bolt individually. This significantly improves inspection efficiency and resource utilization while minimizing disruption to traffic.

[0033] Furthermore, all detection data is transmitted back in real time via a 5G network through a data transmission unit. This unit supports dual-mode communication of 5G and Wi-Fi, and can automatically switch to the on-site Wi-Fi hotspot when the 5G signal is weak, ensuring the continuity and stability of data transmission. Finally, based on the results of multimodal fusion discrimination, the system automatically generates structured electronic maintenance work orders by the maintenance work order generation unit. The work order content includes the precise location of the faulty bolt (based on BIM coordinates), the fault type, the risk level, the recommended maintenance measures, and is associated with the original detection images and data. The work order is pushed to on-site maintenance personnel via mobile terminals (such as WeChat Work and DingTalk), enabling on-site processing, photo feedback, and closed-loop confirmation, truly realizing intelligent, full-process closed-loop management from "detection and discovery" to "maintenance execution".

[0034] In summary, this invention, through the collaboration of a magnetic wall-climbing robot and a drone, combined with SLAM / UWB fusion positioning, multimodal data intelligent discrimination, BIM-driven path planning, and 5G real-time communication and automated work order generation, constructs a complete, efficient, and safe intelligent inspection system for steel structure bolts. This system effectively overcomes many shortcomings of existing technologies and meets the urgent needs for inspection accuracy, efficiency, and intelligence in complex environments.

[0035] As stated above, this case protects a steel structure bolt detection system suitable for complex environments, and all technical solutions that are the same as or similar to this case should be considered to fall within the scope of protection of this case.

Claims

1. A steel structure bolt inspection system suitable for complex environments, characterized in that, include: The magnetic wall-climbing robot is used to move by adhering to the surface of steel structures. It is equipped with mechanical sensors to perform mechanical sensing and detection on bolts and obtain data on the preload or looseness of the bolts. Drones are used to fly around steel structures and are equipped with visual sensors to visually identify and detect bolts, acquiring images of bolts that are missing, corroded, or damaged. The multimodal fusion and discrimination unit is used to receive and synchronize the detection data of the magnetic wall-climbing robot and the drone, perform data fusion and intelligent analysis, and generate a comprehensive state discrimination result of the bolt; The path planning and collaborative control unit is used to plan and detect paths for the magnetic wall-climbing robot and the drone, and coordinate their collaborative operation. The data transmission unit is used to transmit detection data and judgment results to a remote terminal in real time via a 5G network; The maintenance work order generation unit is used to automatically generate maintenance work orders containing fault location, type, and suggested measures based on the comprehensive status judgment results, thereby realizing intelligent closed-loop management of detection and maintenance.

2. The steel structure bolt inspection system suitable for complex environments as described in claim 1, characterized in that, Both the magnetic wall-climbing robot and the drone are equipped with SLAM positioning systems for autonomous navigation and precise positioning in complex environments without GPS signals.

3. The steel structure bolt inspection system suitable for complex environments as described in claim 2, characterized in that, It also includes a UWB positioning base station. The magnetic wall-climbing robot and the drone are equipped with UWB tags. The UWB positioning base station works in conjunction with the SLAM positioning system to provide high-precision position calibration to improve the accuracy of the detection path.

4. The steel structure bolt inspection system suitable for complex environments as described in claim 1, characterized in that, The mechanical sensors include torque sensors, strain gauges, or vibration sensors, used to directly measure or indirectly calculate the preload of the bolts.

5. The steel structure bolt inspection system suitable for complex environments as described in claim 4, characterized in that, The magnetic wall-climbing robot uses a tracked chassis, and its magnetic module is an electromagnet with adjustable magnetic force. Its front end is equipped with a retractable mechanical interface for mounting mechanical sensors.

6. The steel structure bolt inspection system suitable for complex environments as described in claim 1, characterized in that, The visual sensor includes a high-definition camera, an infrared thermal imager, or a laser scanner, used to acquire visible light images, temperature distribution, or three-dimensional morphology information of the bolt.

7. The steel structure bolt inspection system suitable for complex environments as described in claim 1, characterized in that, The multimodal fusion discrimination unit uses a deep learning-based algorithm model to extract and fuse features from mechanical sensing data and visual recognition images, and outputs a probabilistic discrimination result of the bolt state.

8. The steel structure bolt inspection system suitable for complex environments as described in claim 1, characterized in that, The path planning and collaborative control unit automatically plans the optimal detection path based on the BIM model of the steel structure or the pre-scanned 3D point cloud map, and dynamically adjusts the relative position and operation sequence of the magnetic wall-climbing robot and the drone.

9. A steel structure bolt inspection system suitable for complex environments as described in claim 1, characterized in that, The data transmission unit supports dual-mode communication of 5G and Wi-Fi, and automatically switches to Wi-Fi hotspot for data backhaul when the 5G signal is weak.

10. A steel structure bolt inspection system suitable for complex environments as described in claim 1, characterized in that, The maintenance work orders generated by the maintenance work order generation unit are pushed to on-site maintenance personnel via mobile terminals, and the work order execution status feedback and closed-loop tracking are supported.

Images