Intelligent hoisting method and monitoring system for large-span cable crane system

Abstract

The invention relates to the technical field of hoisting construction, particularly discloses an intelligent hoisting method for a large-span cable hoisting system and a monitoring system, and aims to solve the problems of low construction efficiency, difficulty in prevention and control of collision risks and low hoisting in-place efficiency in traditional cable hoisting construction. The system comprises a cable hoisting module, a visual measurement module, a reference positioning module, a main cable state recognition module, a data processing module, an execution control module and an obstacle avoidance early warning module, wherein the visual measurement module, the reference positioning module, the main cable state recognition module, the execution control module and the obstacle avoidance early warning module are electrically connected with the data processing module; the cable hoisting module comprises a winch group and a cable tower, the cable comprises a main cable, a traction cable and a hoisting cable, a sports car is arranged on the main cable, the sports car is connected with a hoisting point, the hoisting point is connected with a hoisting object, the sports car, the hoisting point, the hoisting object and the cable tower are all provided with mark points, and the execution control module is electrically connected with the winch group. The problems that the construction efficiency is low, the collision risk is difficult to prevent and control and the hoisting in-place efficiency is low in traditional cable hoisting construction are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of hoisting construction, and in particular to an intelligent hoisting method and monitoring system for a large-span cable hoisting system. Background Art

[0002] With the continuous development of infrastructure construction, especially in bridge and building projects, cable hoisting systems have been widely used as an important construction technology. This is especially true in complex mountainous areas with many V-shaped valleys and undulating terrain. Traditional support construction methods are difficult to implement, while cable hoisting systems can flexibly adapt to these complex environments.

[0003] However, large-scale cable hoisting systems still rely heavily on traditional hoisting controls, combined with manual communication and command operations. This not only requires a large number of personnel and high-quality operators, but also requires repeated adjustments before the hoist is in place, resulting in low efficiency. Furthermore, the numerous ropes in cable hoisting systems for long-span bridges make collisions with objects being hoisted in mid-air difficult to prevent, posing a significant safety hazard. Collisions are particularly common in cable hoisting systems within the cable-stayed, buckled systems of long-span arch bridges, severely impacting construction safety.

[0004] Although there have been certain innovations in the technology of intelligent lifting construction in recent years, such as the use of technologies such as the Internet of Things, big data, and cloud computing to monitor and analyze the lifting process, there has been insufficient consideration of the recognition of the main cable force and cable shape, autonomous control of the position of the hoisted object, automatic obstacle avoidance and warning of the hoisted object, and direct improvement of the lifting efficiency. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent lifting method and monitoring system for a large-span cable hoisting system, aiming to improve the problems of low construction efficiency, difficult to prevent and control collision risks, and low lifting efficiency in traditional cable hoisting construction.

[0006] The technical solution adopted by the present invention to solve the above technical problems is:

[0007] According to a first aspect of the present invention, the present invention provides an intelligent hoisting and detection system for a large-span cable hoisting system, comprising a cable hoisting module, wherein the cable hoisting module comprises a hoisting unit, a cable tower, a main cable, a traction cable, a lifting cable, and a running car;

[0008] The sports car is connected to a lifting point, the lifting point is connected to a hoisting object, and marking points are provided on the sports car, the lifting point, the hoisting object and the cable tower;

[0009] Also includes:

[0010] The visual measurement module includes at least three industrial cameras, which are used to capture images of each marker point and calculate the three-dimensional coordinates;

[0011] The benchmark positioning module, including a total station, is used to establish a global three-dimensional coordinate benchmark. It cooperates with the visual measurement module to measure the world coordinates of preset position markers during the camera calibration phase, providing a coordinate benchmark for the visual measurement module's external parameter matrix solution. When the industrial camera's angle changes, the markers are remeasured for external parameter correction and calibration of the visual measurement system.

[0012] The main cable state identification module is used to iteratively calculate the stress-free length of the main cable based on the finite element model and solve the main cable sag and inclination in real time;

[0013] The data processing module is used to integrate multi-source data, generate and dynamically update the digital twin model, realize the conversion between world coordinates and local coordinates, solve the posture of the hoisted object, and retrieve the posture after the marker point fails. Based on the real-time posture data of the hoisted object, the main cable status parameters and the preset control logic, the control instructions of the winch group are generated and sent to the execution control module;

[0014] The execution control module is used to receive the control instructions of the hoisting unit generated by the data processing module, and dynamically adjust the retraction and release rates of the traction rope and lifting rope of the hoisting unit based on the instructions to achieve accurate transportation and positioning of the hoisted objects;

[0015] The obstacle avoidance warning module includes a ranging sensor installed on the hoisted object, which is used to realize early warning of hoisted object collision through rigid body distance judgment based on the digital twin model and ranging data provided by the data processing module; the ranging sensor can be any one of millimeter wave radar, lidar, and visual sensor.

[0016] According to a second aspect of the present invention, there is provided a method for intelligent lifting of a large-span cable hoisting system, wherein the method uses the intelligent lifting and detection system for the large-span cable hoisting system to lift a hoisted object, comprising the following steps:

[0017] Step 1: Arrange markers and calibrate industrial cameras;

[0018] Step 2: Iteratively calculate the stress-free length of the main cable to determine the reference stress-free length and generate the sag and inclination reference parameters;

[0019] Step 3: Perform 3D scanning and modeling of the hoisted object, record the spatial association between key feature points and marker points, and provide a geometric basis for pose calculation;

[0020] Step 4: Store the local coordinates of the hoisted object, establish the local coordinate system of the hoisted object, and realize the flexible conversion between world coordinates and local coordinates;

[0021] Step 5: Target height setting and adjustment: clarify the height benchmark for hoisted objects, dynamically adjust the target height parameters, and provide accurate data basis for lifting rope retraction and extension control;

[0022] Step 6: The sports car moves and tracks the target, tracking the position of the hoisted object in real time and providing the position information of the hoisted object including position, inclination, and speed for dynamic control;

[0023] Step 7: Calculate the main cable state and control the winch to ensure the hoisted object maintains a constant posture and moves at a constant speed.

[0024] Step 8: Obstacle avoidance warning and travel decision-making: Real-time detection of the distance between the hoisted object and the obstacle. When the distance is ≤ the warning threshold, adaptive deceleration and hovering operations are performed to identify collision risks in advance and ensure hoisting safety.

[0025] Step 9: Automatically adjust the inclination angle of the hoisted object to accurately adjust the posture of the hoisted object to meet the adaptation requirements of the installation posture;

[0026] Step 10: Lower the hoisted object into place.

[0027] Furthermore, in the step 1, the industrial camera calibration includes the calibration of the intrinsic parameter matrix and the extrinsic parameter matrix. The intrinsic parameter matrix is ​​calibrated before the camera is installed to correct the radial distortion and tangential distortion of the image. When the industrial cameras are arranged, at least two industrial cameras should be located on opposite sides of the lifting operation area and in a position where the line of sight is not blocked. The remaining cameras are arranged in a clear imaging area and are not in a straight line with the above two industrial cameras. The extrinsic parameter matrix is ​​calibrated in combination with the reference positioning module under the preset environmental stability conditions. The preset environmental stability conditions include wind speed ≤ 2m / s, ambient temperature Fluctuation ≤ ±2℃ / h; when calibrating the camera extrinsic parameter matrix, a heavy object is hung on the hook of the hanging point, and the total station measurement and industrial camera shooting are carried out simultaneously; no less than 3 marking points are fixed on the surface of the hanging object and the sides of the two sports cars. During calibration, the sports car and the hanging point stop at five different positions with a certain distance between them. During the stop, the world coordinates of the center of the marking point are measured using the total station, and 5 sets of known marking point world coordinate data are formed at the above 5 positions; then the pixel coordinates of all marking points are obtained using the images taken at the above stop positions, and the camera extrinsic parameter matrix is ​​obtained using the P3P or EPnP algorithm.

[0028] Furthermore, in Step 2, the main cable state identification module screens the data of the marking points on the sports car based on the world coordinates of the five sets of marking points obtained during the external parameter matrix calibration in Step 1, and establishes a finite element model for five working conditions; with the constraints of "equal cable forces at the main cable nodes on both sides of the saddle and unchanged sum of the stress-free lengths S of the main cables at the reference temperature", combined with the temperature data, the stress-free linear shape of the main cables is iteratively adjusted to determine the reference stress-free length.

[0029] Furthermore, in Step 3, a high-precision photographic 3D reconstruction algorithm is used to capture multiple continuous images at different angles. Through a process including feature matching, sparse reconstruction, BA optimization, and triangulated dense reconstruction, the spatial relationship between the key feature points of the dense point cloud on the surface of the hoisted object and any three marking points is obtained. The remaining marking points are used as backup points to prevent the position information from being retrieved after the marking points are destroyed during construction. The key feature points of the point cloud can be marked as the focus of the rectangles on the surfaces of the two ends of the hoisted object, or the spatial center of the circle. The spatial relationship is the coordinates of the hoisted object point cloud in the local coordinate system composed of the marking points, and the conversion method is as follows:

[0030] Take the center of any one mark point as the reference point of the local coordinate system and establish the local coordinate system Coor( , , ), as shown in formulas (1~3):

[0031] (1)

[0032] (2)

[0033] (3)

[0034] Among them, Norm means normalization 、 、 They represent the world coordinates x, y, and z of the center point of the jth target on the tube body, A vector is obtained by cross-producting two vectors.

[0035] Furthermore, in Step 4, the world coordinates of the key feature points of the hoisted object are converted to the local coordinate system of the hoisted object, as shown in formulas (4-5); the same method is used to establish a local coordinate system between the backup marking point and any two pre-selected marking points, and the corresponding coordinate conversion is performed and kept for backup. If the pre-selected marking point is damaged, the local coordinates of the hoisted object are retrieved using the remaining three backup marking points.

[0036] (4)

[0037] in (5)

[0038] Among them, lowercase x, y, and z represent world coordinates, and uppercase X, Y, and Z represent local coordinates. Represents the target coordinate system transformation matrix of the i-th steel pipe segment.

[0039] Furthermore, in Step 6, the execution control module drives the winch assembly to pull the sports car along the main cable, and the visual measurement module calls the target detection algorithm to process the collected images, identify the marker targets therein, and then fit the pixel coordinates of the centers of the identified markers. On this basis, the spatial coordinates of the markers in multiple views are triangulated using known internal and external parameters of the camera; the distance between the known markers and the distance between the three-dimensional reconstructed point clouds of the corresponding marker centers are divided to obtain a scale factor, and the markers are combined in pairs to obtain multiple groups of scale factors. Finally, the scale factors are averaged and the scale factors that differ from the mean by more than 5% are removed. The final mean of the remaining scale factors is calculated; the absolute scale of the point cloud and the spatial coordinates of the marker centers are restored based on the average scale factor;

[0040] Using the spatial coordinates of the marker points on the hoisted object, the world coordinates of the dense point cloud recorded on the hoisted object are calculated using the inverse transformation. In the scanning stage of Step 3 above, only the local coordinates of the feature points of the hoisted object are retained:

[0041] (6)

[0042] The lowercase x, y, and z represent world coordinates, and the uppercase X, Y, and Z represent local coordinates. is the inverse transformation matrix, which can be obtained by calculating the inverse matrix using formula (5).

[0043] Furthermore, in the step 7, the target elevation of the hoisted object is recorded as h, and the sag of the main cable is recorded as f. The main cable state identification module calculates the sag f and the inclination angle θ of the main cable in real time based on the finite element model; the data processing module integrates the hoisted object posture data and the main cable state parameters, and generates the traction cable and lifting cable retraction and release rate instructions according to the preset function; the execution control module drives the winch unit and adjusts the retraction and release rate to ensure that the posture of the hoisted object is constant and the speed is uniform.

[0044] Furthermore, the control method of the traction rope retraction rate is as follows: define the traction rope change length as Δl, the horizontal position change of the suspension point as Δx, and the ratio of the two as the traction rope retraction rate v l The horizontal speed of the sports car v x The ratio is equal to the cosine value of the main cable inclination angle θ; the main cable inclination angle θ is obtained by finite element model calculation, and then the v of the coordinate x position of any hanging point is extracted. l / v x , the cosine values ​​at different temperatures, different lifting weights, and different x-coordinates are fitted into a continuous function, and finally v l / v x=g1(x, T, G), where T is the main cable temperature. The g1(x, T, G) function is input into the PCL control terminal of the traction cable winch. The x coordinate of the sports car measured above, combined with the main cable temperature sensor data and the load G, is imported into the execution control module, so that the sports car can move according to the preset horizontal speed v x Automatically control the traction rope retraction rate v l .

[0045] Furthermore, the method for controlling the rate of retraction of the hoisting rope is as follows:

[0046] When the hoisted object is at the target height during lifting, define two adjacent positions 1 and 2. If the length of the lifting rope remains unchanged, the height difference of the hoisted object at the two x positions is the difference between the sag f1 and the sag f2. According to different temperatures T and different lifting weights G, the variation law of the sag f along the x coordinate of different bridge span positions is analyzed, and the function f=g2(x, T, G) is fitted. The derivative of x is obtained to obtain the change rate of the sag f df / dx=dg2(x, T, G) / dx, and the crane rope retraction speed v is converted to the value of the sag f. d =n*dg2(x, T, G) / dx is imported into the PCL control terminal, where n is the number of rope segments at the lifting point, and controls the rope retraction speed at any position;

[0047] When the hoisted object is not at the target height, first adjust the hoisted object to the target height, and then adjust and control the rope retraction rate according to the above method.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] 1. The present invention relies on an empty cable car to identify the stress-free linear shape of the main cable of the cable crane system with high accuracy, provides the stress-free length as an initial parameter for the calculation of the cable force in the lifting condition, and facilitates the real-time feedback of the subsequent cable crane main cable force and tower deviation.

[0050] 2. The present invention provides data support for real-time adjustment of the winch speed and rope retraction length by calculating the maximum sag of the cable hoist main rope and the change of the target length of the hoist rope in real time, ensuring that the hoisted object moves forward at a constant speed at a fixed elevation.

[0051] 3. The present invention uses a small number of marker ball measurements and local coordinate translation and rotation transformations of the rigid body to achieve rapid feedback on the position and posture of the hoisted object; using multi-point displacement control and hoisting adjustment in two stages, it achieves one-click positioning and autonomous posture adjustment of the hoisted object, reducing the original 3-4 hours of single steel truss hoisting time to 0.5 hours.

[0052] 4. The present invention uses digital twins of the hoisted object and the original structure on the travel path, combined with rigid body distance judgment, to achieve early warning of hoisted object collisions and provide data support for adjusting the hoisting height. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a schematic diagram of cable lifting;

[0054] Figure 2 Scanning diagram for hoisting objects and marking points;

[0055] Figure 3 This is a schematic diagram of hoisting into place;

[0056] Figure 4 This is a flow chart of the intelligent lifting method for a large-span cable crane system;

[0057] Figure 5 This is a flowchart for hoisting obstacle avoidance;

[0058] Figure 6 Calculation structure and mechanical model diagram for the stress-free length of the main cable;

[0059] Figure 7 Flow chart for calculating the unstressed length of the main cable;

[0060] Figure 8 Main cable status analysis diagram.

[0061] In the figure: 1. Main rope; 2. Traction rope; 3. Hoisting rope; 4. Sports car; 5. Lifting point; 6. Hoisted object; 6-1. Preset lifting end position; 7. Marking point; 8. Cable tower; 9. Industrial camera; 10. Winch unit. DETAILED DESCRIPTION

[0062] In the present invention, unless otherwise specified or limited, the terms "installed," "connected," "connect," "fixed," etc. should be understood broadly. For example, they may refer to fixed or detachable connections, mechanical or electrical connections, and direct or indirect connections through an intermediary. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0063] The following is a further description with reference to the accompanying drawings and specific embodiments:

[0064] Example 1

[0065] like Figure 1-Figure 3As shown, this embodiment provides an intelligent lifting and detection system for a large-span cable hoist system, comprising a cable hoisting module, which includes a hoisting unit 10 and a cable tower 8 for supporting the cable. The cable includes a main cable 1, a traction cable 2, and a lifting cable 3. A sports car 4 is provided on the main cable 1, which is connected to a lifting point 5, which is connected to a hoisted object 6. Marking points 7 are provided on the sports car 4, the lifting point 5, the hoisted object 6, and the cable tower 8. Marking points 7 can be targets such as reflective sheets, QR codes, or checkerboards, and are rectangular in size of 1m*1m. The target panel can be made of materials such as plastic reflective film, frosted acrylic, or diffuse matte paint. The back of the panel is fixed to a steel frame welded with angle steel to maintain a flat surface. The traction cable 2 and the lifting cable 3 are both connected to the hoisting unit 10, with the traction cable 2 connected to the sports car 4 and the lifting cable 3 connected to the hoisted object 6.

[0066] The intelligent lifting and detection system for large-span cable crane systems also includes a visual measurement module, a reference positioning module, a main cable 1 status recognition module, a data processing module, an execution control module and an obstacle avoidance warning module. The visual measurement module, the reference positioning module, the main cable 1 status recognition module, the execution control module and the obstacle avoidance warning module are all electrically connected to the data processing module. The execution control module includes electrical components such as PLC, transformer, intermediate relay, contactor, push button switch, etc. The execution control module is electrically connected to the winch unit 10.

[0067] Among the modules mentioned above, the visual measurement module includes three industrial cameras 9, which are used to capture images of each marker point 7 and calculate their three-dimensional coordinates. The reference positioning module includes a total station, which is used to establish a global three-dimensional coordinate reference. It cooperates with the visual measurement module to measure the world coordinates of the preset marker points 7 during the camera calibration phase, providing a coordinate reference for the visual measurement module's external parameter matrix calculation. When the industrial camera angle changes, the marker points are remeasured for external parameter correction and calibration of the visual measurement system. The main cable 1 state recognition module iteratively calculates the stress-free length of the main cable 1 based on the finite element model and calculates the sag and inclination of the main cable 1 in real time. The data processing module integrates multi-source data to generate and dynamically update a digital twin model, enabling conversion between world coordinates and local coordinates, calculating the pose of the hoisted object 6, and recovering the pose of the marker point 7 after failure. Based on the real-time pose data of the hoisted object 6, the main cable 1 state parameters, and the preset control logic, control instructions for the hoisting unit 10 are generated and sent to the execution control module. The control instructions generated by the data processing module for the hoisting unit 10 include posture alignment instructions, deviation correction instructions, and obstacle avoidance-related instructions. The execution control module receives control commands for the hoist assembly 10 generated by the data processing module and dynamically adjusts the retraction and release rates of the traction rope 2 and lifting rope 3 of the hoist assembly 10 based on these commands, ensuring precise transportation and positioning of the hoisted object 6. The obstacle avoidance and warning module, which includes a ranging sensor mounted on the hoisted object 6, uses the digital twin model and radar ranging data provided by the data processing module to determine the rigid body distance and provide early warning of collisions with the hoisted object 6. The ranging sensor can utilize millimeter-wave radar, lidar, vision sensor, or a combination of these sensors.

[0068] Example 2

[0069] This embodiment provides a large-span cable hoisting system intelligent hoisting method, using the large-span cable hoisting system intelligent hoisting and detection system provided in Example 1 to hoist the hoisting object 6, such as Figures 1-4 As shown, the following steps are included:

[0070] Step 1: Arrange the marking points 7, and then calibrate the industrial camera 9 through the collaboration of the visual measurement module and the reference positioning module.

[0071] Step 2: The data processing module collaborates with the main cable 1 state identification module to iteratively calculate the stress-free length of the main cable 1, determine the reference stress-free length, and generate the sag and inclination reference parameters.

[0072] Step 3: The visual measurement module collaborates with the data processing module to perform three-dimensional scanning and modeling of the hoisted object 6, record the spatial association between key feature points and marker points 7, and provide a geometric basis for pose calculation.

[0073] Step 4: The local coordinates of the hoisted object 6 are stored through the data processing module, and the local coordinate system of the hoisted object 6 is established to realize the flexible conversion between the world coordinates and the local coordinates;

[0074] Step 5: Set and adjust the target height through the data processing module, clarify the height reference of the hoisted object 6, and dynamically adjust the target height parameters to provide accurate data basis for the control of the retraction and extension of the lifting rope 3.

[0075] Step 6: The visual measurement module collaborates with the data processing module to track and identify the movement of the sports car 4 and the target of the marker point 7, and to track the posture of the hoisted object 6 in real time, providing posture information of the hoisted object 6 including position, inclination, and travel speed for dynamic control.

[0076] Step 7: Through the coordinated action of the main cable 1 state recognition module, data processing module, and execution control module, the main cable 1 state is calculated and the winch is controlled, so that the hoisted object 6 maintains a constant posture and moves at a uniform speed.

[0077] Step 8: Through the coordinated action of the obstacle avoidance warning module, data processing module, and execution control module, obstacle avoidance warning and travel decision-making are carried out, and the distance between the hoisted object 6 and the obstacle is detected in real time. When the distance is ≤ the warning threshold, adaptive deceleration and hovering operations are performed to identify collision risks in advance and ensure hoisting safety.

[0078] Step 9: Through the data processing module and the execution control module, the inclination angle of the hoisted object is automatically adjusted, and the posture of the hoisted object 6 is accurately adjusted to meet the adaptation requirements of the posture required for installation.

[0079] Step 10: Under the joint action of the data processing module, the execution control module, the visual measurement module and the reference positioning module, the hoisting object 6 is lowered into place.

[0080] Step 11: Hand over to the manual worker. If the posture requirements are met, the lifting is completed.

[0081] like Figure 4 As shown in the figure, the intelligent lifting method of the large-span cable crane system mainly includes three stages and a total of 11 steps.

[0082] In Step 1, if Figure 1As shown, the three industrial cameras 9 should be arranged as far apart as possible to form a triangle with the largest possible area to improve measurement accuracy. At least two industrial cameras 9 should be located on both sides of the bridge, as close to the piers as possible, without obstructing the view toward the center of the bridge span. The other camera should be located in a position where the image is clear. Calibration of industrial camera 9 includes calibration of the camera's intrinsic parameter matrix and extrinsic parameter matrix. Calibration of the intrinsic parameter matrix is ​​performed before the camera is installed. The intrinsic parameter matrix of industrial camera 9 (including parameters such as focal length, principal point coordinates, and distortion coefficient) is determined by the inherent characteristics of the camera's optical lens and imaging chip. Due to factors such as lens processing accuracy and assembly errors, the original image captured by the camera may have distortion problems such as radial distortion and tangential distortion (for example, a straight line will appear curved after shooting).

[0083] Intrinsic calibration accurately determines these intrinsic parameters through image analysis of preset targets (such as a checkerboard). Distortion is then corrected during image preprocessing, ensuring that the pixel coordinates of the markers (7) captured by the nine industrial cameras accurately reflect physical imaging principles, providing undistorted raw image data for subsequent pose calculations. Without intrinsic calibration, image distortion can lead to a direct mismatch between pixel coordinates and actual spatial locations, fundamentally reducing measurement accuracy. Calibration of the extrinsic parameter matrix must be performed under favorable weather conditions, defined as no wind and no significant temperature fluctuations. No wind refers to a wind speed of ≤2 m / s, and no significant temperature fluctuations refer to ambient temperature fluctuations of ≤±2°C / h. Ensure that the cable crane carriage (4) is stable. To ensure the sling point (5) or hanging basket remains as stable as possible, a heavy object (the total weight of the hook, carriage (4), sling, and hoisted object (6) should be weighed in grams) should be attached to the hook during camera extrinsic parameter matrix calibration. Total station measurements and camera capture should be performed simultaneously, as much as possible. At least three marking points 7 are fixed on the surface where the heavy objects are mounted and on the sides of the two sports cars 4 (the marking points 7 can be targets such as reflective sheets, QR codes, checkerboards, etc., with a size of 1m*1m rectangle. The target panel material can be plastic reflective film, frosted acrylic or diffuse matte paint, etc. The back of the plate is fixed to a steel frame welded with angle steel to keep the target in a plane). During calibration, the sports car 4 and the hanging point 5 stop near the left bank, 1 / 4 span, 1 / 2 span, 3 / 4 span, and near the right bank respectively. When stopping, the world coordinates of the center of the marking point 7 are measured using a total station, and 5 sets of known world coordinate data of the marking point 7 are formed at the above 5 positions. Then, the pixel coordinates of all the marking points 7 are obtained using the images taken at the above-mentioned stop positions, and the camera extrinsic parameter matrix is ​​obtained using the P3P or EPnP algorithm.

[0084] In Step 2, accurate identification of the linear shape of the main cable 1 is a difficult point in the lifting operation, especially in long-distance and large-span lifting projects. The deformation and stress state of the cableway have an important impact on the lifting accuracy and the stability of the hoisted object 6. During the lifting process, since the sag of the main cable 1 caused by the lifting point 5 at different bridge span positions is difficult to predict, it is often necessary to repeatedly adjust the height of the lifting point 5, which consumes a lot of time. Therefore, this technology pre-calculates the stress-free length of the main cable 1 as the basis for predicting the linear shape of the main cable 1 under any lifting condition. The stress-free length calculation is performed based on the world coordinates of 5 sets of marker points 7 during calibration. First, the world coordinates of the marker points 7 on the sports car 4 are selected, and the finite element models under 5 working conditions are established respectively. By adjusting the stress-free linear shape of the main cable 1 and combining the import of temperature data, the stress-free length of the main cable 1 is iteratively determined, providing initial parameters for the prediction of the linear shape and tower deviation of the main cable 1 under subsequent lifting conditions. The specific process for calculating the stress-free length S of the main cable 1 at the reference temperature (referred to as the reference stress-free length) is as follows: Figure 6 and Figure 7 shown.

[0085] The iterative adjustment of the main cable 1 of the cable hoist must meet two conditions at the same time: (1) the main cable 1 node forces on both sides of each saddle are equal; (2) the total stress-free length S of the main cable 1 at the reference temperature remains unchanged.

[0086] In Step 3, the geometric shape of the hoisting object 6 is scanned and recorded, and a high-precision photographic 3D reconstruction algorithm (such as Colmap, Glomap, etc., or the MVG multi-view reconstruction algorithm) is used. Multiple continuous images are taken at different angles (the overlap rate of adjacent images is about 50%). Through feature matching, sparse reconstruction, BA optimization, triangulation dense reconstruction and other processes, the spatial relationship between the key feature points of the dense point cloud on the surface of the hoisting object 6 and any three marking points 7 is obtained. The remaining marking points 7 are used as spare points to prevent the position information from being retrieved after the marking points 7 are destroyed during construction. The key feature points of the point cloud can be marked as the focus of the rectangles on the surfaces of the two ends of the hoisting object 6, or the spatial center of the circle. The spatial relationship is the coordinate of the point cloud of the hoisting object 6 in the local coordinate system composed of the marking points 7, and the conversion method is as follows:

[0087] First, take the center of any one mark point 7 as the reference point of the local coordinate system and establish the local coordinate system Coor( , , ), such as formulas (1~3) and Figure 1 、 2 shown.

[0088] (1)

[0089] (2)

[0090] (3)

[0091] Among them, Norm means normalization 、 、 They represent the world coordinates x, y, and z of the center point of the jth target on the tube body, A vector is obtained by cross-producting two vectors.

[0092] In Step 4, the world coordinates of the centers of the two ends of the steel pipe are converted to the local coordinate system of the pipe body, as shown in Formulas (4-5). Using the same method, a local coordinate system is established between the backup marker point 7 and any two preselected marker points 7, and the corresponding coordinate conversion is performed. This is kept for backup. If the preselected marker point 7 is damaged, the remaining three backup marker points 7 are used to retrieve the local coordinates of the hoisted object 6.

[0093] (4)

[0094] in (5)

[0095] Among them, lowercase x, y, and z represent world coordinates, and uppercase X, Y, and Z represent local coordinates. Represents the target coordinate system transformation matrix of the i-th steel pipe segment.

[0096] Code implementation: Convert world coordinates to local coordinates of target coordinate system %P0 is the world coordinate of the target selected as the origin, PT is the transformed coordinate T=[V0,-V0*P0';0,0,0,1];%Transformation matrix, the transformed coordinate = T*the coordinate before transformation fori=1:n%1~n points P=C(i,1:3);%Store xyz values ​​by column PT=T*[P';1];%Local coordinates of the nozzle end

[0097] Entering the second stage, in Step 6, since the camera pose remains unchanged, there is no need to repeat the camera setup and calibration of the intrinsic and extrinsic parameter matrices in Step 1. If the camera pose changes, the intrinsic and extrinsic parameter matrices of each industrial camera 9 must be recalibrated. Next, the center points of the marker points 7 on the hoisted object 6 are photographed and three-dimensionally measured. To this end, an open-source YOLO or CNN target detection algorithm is pre-used to identify the marker points 7 in the captured image. The pixel coordinates of the centers of the identified marker points 7 are then fitted. On this basis, the known camera intrinsic and extrinsic parameters are used to triangulate the spatial coordinates of the marker points 7 in multiple views. This method is a well-established method in multi-view reconstruction and will not be described in detail here. The distances between the known marker points 7 and the distances between the 3D reconstructed point clouds of the corresponding marker points 7 centers are divided to obtain scale factors. The marker points 7 are then combined in pairs to obtain multiple sets of scale factors. Finally, the scale factors are averaged, and scale factors that differ by more than 5% from the mean are removed. The remaining scale factors are used to calculate the final mean. Based on the average scale factor, the absolute scale of the point cloud and the spatial coordinates of the center of marker point 7 are restored.

[0098] Using the spatial coordinates of the marker point 7 on the hoisting object 6, the world coordinates of the dense point cloud recorded on the hoisting object 6 are solved by inverse transformation. In order to improve the solving efficiency, in the scanning stage of the aforementioned Step 3, only the local coordinates of the feature points of the end of the hoisting object 6 (such as the coordinates of the port centroid) are retained:

[0099] (6)

[0100] The lowercase x, y, and z represent world coordinates, and the uppercase X, Y, and Z represent local coordinates. is the inverse transformation matrix, which is obtained by calculating the inverse matrix using formula (5). The code implementation principle is the same as the above method and will not be repeated here.

[0101] By using formula (6) to solve the spatial coordinates of the centroids at both ends of the hoisted object 6 in real time, its position, inclination, travel speed and other information can be easily calculated.

[0102] In Step 7-8, Figure 4 、 Figure 5 and Figure 8 As shown, the target elevation of the hoisted object 6 is denoted as h, and the sag of the main cable 1 is denoted as f. During the movement of the sports car 4 and the hoisted object 6, the centroids of the two ends must be kept at the same height to maintain a stable posture and prevent tilt. However, the sag f of the main cable 1 continuously changes during the movement of the sports car 4, necessitating the constant automatic adjustment of the lengths of the slings at the front and rear suspension points 5. Furthermore, the speed of the sports car 4 is controlled by the traction cable 2, which always aligns with the main cable 1, not horizontally. Therefore, even if the traction cable 2 maintains a uniform retraction speed, the horizontal projection of the retraction speed (i.e., the speed of the sports car 4) is not constant due to the constantly changing angle of the main cable 1. Therefore, the sag height and retraction speed of the traction cable 2 must be adjusted in real time based on the current sag and inclination of the main cable 1 to ensure a stable posture and uniform travel of the hoisted object 6. Otherwise, the hoisted object 6 is prone to flipping or swaying, posing a collision safety hazard. In the past, this operation was usually completed through manual visual observation and intercom command, which made it difficult to ensure the efficiency and safety of the lifting. The key to automatic control lies in the determination of the state of the main cable 1, including f and inclination, which can be obtained in advance by fast finite element analysis of the cable lifting system, such as Figure 8 shown.

[0103] Control method of the traction rope 2 retraction rate: For a certain main rope 1, when the traction rope 2 changes in length Δl and the horizontal position of the hanging point 5 changes by Δx, the ratio of the two is the ratio of the traction rope 2 retraction rate to the horizontal speed of the sports car 4 (v l / v x ). In two extremely short time intervals, the inclination angles of the main cable 1 in adjacent states are almost equal, so v l / vx This is the cosine value of the inclination angle. This angle can be obtained from the finite element model calculation, and finally the v of the x position of any hanging point 5 is extracted. l / v x , the cosine values ​​at different temperatures (affecting the stress-free state), different lifting weights, and different x-coordinates are fitted into a continuous function, and finally v l / v x =g1(x, T, G), where T is the temperature of the main cable 1. The g1(x, T, G) function is input into the execution control module, specifically the PCL control end of the traction cable 2 winch. The x coordinates of the sports car 4 measured above, combined with the temperature sensor data of the main cable 1 and the load G, are imported into the PCL control end, so that the sports car 4 can move according to the preset horizontal speed v x Automatically control the retraction rate v of the traction rope 2 l .

[0104] Rope retraction rate control method: Case 1: The hoisted object 6 is already at the target height when it is hoisted. At the two adjacent positions 1 and 2, if the length of the hoisting rope (i.e., the lifting rope 3) remains unchanged, the height difference of the hoisted object 6 at the two x positions is the difference between the sag f1 and the sag f2. In order to ensure horizontal movement, it is necessary to analyze the change law of the sag f along the x coordinate of different bridge span positions according to different temperatures T and different hoisting weights G, and fit the function f=g2(x, T, G). By taking the derivative with respect to x, the change rate of the sag f is df / dx=dg2(x, T, G) / dx. Since the crane rope retraction speed v d =n*dg2(x, T, G) / dx is imported into the PCL control terminal (n is the number of rope segments at lifting point 5, that is, twice the number of movable pulleys at lifting point 5), and the rope retraction speed at any position can be controlled. Case 2: The hoisted object 6 is not at the target height. In this case, the difference between the measured centroid height of the hoisted object 6 end and the target height is directly returned to the PCL control terminal of the rope winch. At the given rope retraction rate, the centroid height of the hoisted object 6 end reaches the target height. Thereafter, the hoist retraction speed v continues according to Case 1. d Take control.

[0105] In Steps 9 to 11, the posture of the previous hoisting object 6 that has been installed and the posture of the current hoisting object 6 are obtained using formula (6). In order to ensure that the hoisting is in place at one time, the posture of the splicing end of the current hoisting object 6 needs to be controlled so that the splicing end is aligned with the splicing end of the installed hoisting object 6. At this time, the hoisting object 6 has been hoisted above the target position and is waiting to be lowered by the lifting rope. At this time, the inclination angle of the current hoisting object 6 needs to be placed according to the inclination angle of the previous hoisting object 6 that has been identified. The implementation method is to first keep the rope at the higher end stationary and lower the rope at the lower end of the set height. When it is identified that the inclination angle of the current hoisting object 6 meets the inclination error requirement of 1%, the rope is controlled to be lowered according to the same set falling speed until the height of the splicing end of the hoisting object 6 reaches the set height, and then the lowering is stopped. During this falling process, the x, y, and z values ​​of the mark point 7 of the hoisted object 6 are tracked in real time, and the centroid positions of the two ends of the hoisted object 6 are calculated in real time. Once the x deviation occurs at both ends, the travel direction of the sports car 4 and the retraction / release of the sports car 4 are adjusted according to the deviation, so that the x value error of the hoisted object 6 is always kept within ±10cm. After it is lowered to the target height, it is handed over to manual control to fine-tune the posture of the hoisted object 6.

[0106] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0107] 1. The present invention relies on the empty cable car 4 to identify the stress-free linear shape of the main cable 1 of the cable crane system with high accuracy, provides the stress-free length as an initial parameter for the cable force calculation of the lifting condition, and facilitates the subsequent real-time feedback of the cable force of the cable crane main cable 1 and the tower deviation of the cable tower 8.

[0108] 2. The present invention provides data support for real-time adjustment of the winch speed and the length of the reel by calculating the maximum sag of the cable hoist main rope 1 and the change of the target length of the hoist rope in real time, thereby ensuring that the hoisted object 6 moves forward at a constant speed at a fixed elevation.

[0109] 3. The present invention uses a small number of marker ball measurements and local coordinate translation and rotation transformations of the rigid body to achieve rapid feedback on the position and posture of the hoisted object 6; and uses multi-point displacement control and hoisting adjustment in two stages to achieve one-click positioning and autonomous posture adjustment of the hoisted object 6, which shortens the original 3-4 hours of single steel truss hoisting time to 0.5 hours.

[0110] 4. The present invention realizes early warning of collision of hoisted objects through digital twins of the hoisted object 6 and the original structure on the travel path, combined with rigid body distance judgment, and provides data support for adjusting the hoisting height.

[0111] The above is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. An intelligent lifting and detection system for a large-span cable crane system, comprising a cable lifting module, the cable lifting module including a hoisting unit, a cable tower, a main cable, a traction cable, a lifting cable, and a running car; It is characterized by: The sports car is connected to a lifting point, the lifting point is connected to a hoisting object, and marking points are provided on the sports car, the lifting point, the hoisting object and the cable tower; Also includes: The visual measurement module includes at least three industrial cameras, which are used to capture images of each marker point and calculate the three-dimensional coordinates; The benchmark positioning module, including a total station, is used to establish a global three-dimensional coordinate benchmark. It cooperates with the visual measurement module to measure the world coordinates of preset position markers during the camera calibration phase, providing a coordinate benchmark for the visual measurement module's external parameter matrix solution. When the industrial camera's angle changes, the markers are remeasured for external parameter correction and calibration of the visual measurement system. The main cable state identification module is used to iteratively calculate the stress-free length of the main cable based on the finite element model and solve the main cable sag and inclination in real time; The data processing module is used to integrate multi-source data, generate and dynamically update the digital twin model, realize the conversion between world coordinates and local coordinates, solve the posture of the hoisted object, and retrieve the posture after the marker point fails. Based on the real-time posture data of the hoisted object, the main cable status parameters and the preset control logic, the control instructions of the winch group are generated and sent to the execution control module; The execution control module is used to receive the control instructions of the hoisting unit generated by the data processing module, and dynamically adjust the retraction and release rates of the traction rope and lifting rope of the hoisting unit based on the instructions to achieve accurate transportation and positioning of the hoisted objects; The obstacle avoidance warning module includes a ranging sensor installed on the hoisted object, which is used to achieve early warning of hoisted object collision through rigid body distance judgment based on the digital twin model and ranging data provided by the data processing module.

2. A method for intelligent lifting of a large-span cable hoist system, using the intelligent lifting and detection system for a large-span cable hoist system according to claim 1 to lift a hoisted object, characterized in that: The steps include: Step 1: Arrange markers and calibrate industrial cameras; Step 2: Iteratively calculate the stress-free length of the main cable to determine the reference stress-free length and generate the sag and inclination reference parameters; Step 3: Perform 3D scanning and modeling of the hoisted object, record the spatial association between key feature points and marker points, and provide a geometric basis for pose calculation; Step 4: Store the local coordinates of the hoisted object, establish the local coordinate system of the hoisted object, and realize the flexible conversion between world coordinates and local coordinates; Step 5: Target height setting and adjustment: clarify the height benchmark for hoisted objects, dynamically adjust the target height parameters, and provide accurate data basis for lifting rope retraction and extension control; Step 6: The sports car moves and tracks the target, tracking the position of the hoisted object in real time and providing the position information of the hoisted object including position, inclination, and speed for dynamic control; Step 7: Calculate the main cable state and control the winch to ensure the hoisted object maintains a constant posture and moves at a constant speed. Step 8: Obstacle avoidance warning and travel decision-making: Real-time detection of the distance between the hoisted object and the obstacle. When the distance is less than or equal to the warning threshold, adaptive deceleration and hovering operations are performed to identify collision risks in advance and ensure hoisting safety. Step 9: Automatically adjust the inclination angle of the hoisted object to accurately adjust the posture of the hoisted object to meet the adaptation requirements of the installation posture; Step 10: Lower the hoisted object into place.

3. The intelligent hoisting method for a large-span cable hoist system according to claim 2, characterized in that: In the step 1, the industrial camera calibration includes the calibration of the intrinsic parameter matrix and the extrinsic parameter matrix. The intrinsic parameter matrix is ​​calibrated before the camera is installed to correct the radial distortion and tangential distortion of the image. When the industrial cameras are arranged, at least two industrial cameras should be located on opposite sides of the lifting operation area and in a position where the line of sight is not blocked. The remaining cameras are arranged in a clear imaging area and are not in a straight line with the above two industrial cameras. The extrinsic parameter matrix is ​​calibrated in combination with the reference positioning module under the preset environmental stability conditions. The preset environmental stability conditions include wind speed ≤ 2m / s, ambient temperature fluctuations, etc. ≤±2℃ / h; when calibrating the camera extrinsic parameter matrix, a heavy object is mounted on the hook of the hanging point, and the total station measurement and industrial camera shooting are carried out simultaneously; no less than 3 marking points are fixed on the surface of the hanging object and the sides of the two sports cars. During calibration, the sports cars and the hanging point stop at five different positions with a certain distance between them. During the stop, the world coordinates of the center of the marking point are measured using the total station, and 5 sets of world coordinate data of known marking points are formed at the above 5 positions; then, the pixel coordinates of all marking points are obtained using the images taken at the above stop positions, and the camera extrinsic parameter matrix is ​​obtained using the P3P or EPnP algorithm.

4. The intelligent hoisting method for a large-span cable hoist system according to claim 3, characterized in that: In Step 2, the main cable state identification module selects the data of the marker points on the sports car based on the world coordinates of the five sets of marker points obtained during the external parameter matrix calibration in Step 1, and establishes a finite element model for five working conditions. The main cable state identification module uses the temperature data to iteratively adjust the stress-free linear shape of the main cable based on the constraints of "equal cable forces at the main cable nodes on both sides of the saddle and a constant sum of the stress-free lengths S of the main cables at the reference temperature" to determine the reference stress-free length.

5. The intelligent hoisting method for a large-span cable hoist system according to claim 4, characterized in that: In Step 3, a high-precision photographic 3D reconstruction algorithm is used to capture multiple consecutive images at different angles. The spatial relationship between the key feature points of the dense point cloud on the surface of the hoisted object and any three marker points is obtained through a process including feature matching, sparse reconstruction, BA optimization, and triangulation dense reconstruction. The remaining marker points are used as backup points to prevent the position information from being retrieved after the marker points are destroyed during construction. The key feature points of the point cloud can be marked as the focus of the rectangles on the surfaces of the two ends of the hoisted object, or the spatial center of the circle. The spatial relationship is the coordinates of the hoisted object point cloud in the local coordinate system composed of the marker points, and the conversion method is as follows: Take the center of any one mark point as the reference point of the local coordinate system and establish the local coordinate system Coor( , , ), as shown in formulas (1~3): （1）； （2）； （3）； Among them, Norm means normalization 、 、 They represent the world coordinates x, y, and z of the center point of the jth target on the tube body, A vector is obtained by cross-producting two vectors.

6. The intelligent hoisting method for a large-span cable hoist system according to claim 5, characterized in that: In Step 4, the world coordinates of the key feature points of the hoisted object are converted to the local coordinate system of the hoisted object, as shown in formulas (4-5); the same method is used to establish a local coordinate system between the backup marking point and any two pre-selected marking points, and the corresponding coordinate conversion is performed and kept as a backup. If the pre-selected marking point is damaged, the remaining three marking points are used as backup records to retrieve the local coordinates of the hoisted object; （4）； in (5); Among them, lowercase x, y, and z represent world coordinates, and uppercase X, Y, and Z represent local coordinates. Represents the target coordinate system transformation matrix of the i-th steel pipe segment.

7. The intelligent hoisting method for a large-span cable hoist system according to claim 6, characterized in that: In Step 6, the execution control module drives the winch assembly to pull the sports car along the main cable. The visual measurement module calls the target detection algorithm to process the collected images, identify the marker targets therein, and then fit the pixel coordinates of the centers of the identified markers. On this basis, the spatial coordinates of the markers in multiple views are triangulated using known internal and external parameters of the camera. The distance between the known markers and the distance between the three-dimensional reconstructed point clouds of the corresponding marker centers are divided to obtain scale factors. The markers are combined in pairs to obtain multiple groups of scale factors. Finally, the scale factors are averaged and the scale factors that differ from the mean by more than 5% are removed. The remaining scale factors are used to obtain the final mean. Based on the average scale factor, the absolute scale of the point cloud and the spatial coordinates of the center of the marker are restored; Using the spatial coordinates of the marker points on the hoisted object, the world coordinates of the dense point cloud recorded on the hoisted object are calculated using the inverse transformation. In the scanning stage of Step 3 above, only the local coordinates of the feature points of the hoisted object are retained: （6）； The lowercase x, y, and z represent world coordinates, and the uppercase X, Y, and Z represent local coordinates. is the inverse transformation matrix, which can be obtained by calculating the inverse matrix using formula (5).

8. The intelligent hoisting method for a large-span cable hoist system according to claim 7, characterized in that: In the step 7, the target elevation of the hoisted object is recorded as h, the sag of the main cable is recorded as f, and the main cable state recognition module calculates the sag f and the inclination angle θ of the main cable in real time based on the finite element model; The data processing module integrates the posture data of the hoisted object and the state parameters of the main rope, and generates the retraction and release rate instructions of the traction rope and the lifting rope according to the preset function; the execution control module drives the winch unit and adjusts the retraction and release rate to ensure that the posture of the hoisted object remains constant and moves at a uniform speed.

9. The intelligent hoisting method for a large-span cable hoist system according to claim 8, characterized in that: The control method of the traction rope retraction rate is as follows: define the traction rope change length as Δl and the horizontal position change of the suspension point as Δx, and the ratio of the two is the traction rope retraction rate v l The horizontal speed of the sports car v x The ratio is equal to the cosine value of the main cable inclination angle θ; the main cable inclination angle θ is obtained by finite element model calculation, and then the v of the coordinate x position of any hanging point is extracted. l / v x , the cosine values ​​at different temperatures, different lifting weights, and different x-coordinates are fitted into a continuous function, and finally v l / v x =g1(x, T, G), where T is the main cable temperature. The g1(x, T, G) function is input into the PCL control terminal of the traction cable winch. The x coordinate of the sports car measured above, combined with the main cable temperature sensor data and the load G, is imported into the execution control module, so that the sports car can move according to the preset horizontal speed v x Automatically control the traction rope retraction rate v l .

10. The intelligent hoisting method for a large-span cable hoist system according to claim 9, characterized in that: The method for controlling the rate of retraction of the hoisting rope is as follows: When the hoisted object is at the target height during lifting, define two adjacent positions 1 and 2. If the length of the lifting rope remains unchanged, the height difference of the hoisted object at the two x positions is the difference between the sag f1 and the sag f2. According to different temperatures T and different lifting weights G, the variation law of the sag f along the x coordinate of different bridge span positions is analyzed, and the function f=g2(x, T, G) is fitted. The derivative of x is obtained to obtain the change rate of the sag f df / dx=dg2(x, T, G) / dx, and the crane rope retraction speed v is converted to the value of the sag f. d =n*dg2(x, T, G) / dx is imported into the PCL control terminal, where n is the number of rope segments at the lifting point, and controls the rope retraction speed at any position; When the hoisted object is not at the target height, first adjust the hoisted object to the target height, and then adjust and control the rope retraction rate according to the above method.

Images