Intelligent hoisting auxiliary method in power generation equipment manufacturing scene
Through computer-aided technology and AI, combined with real-time positioning and safety monitoring of binocular cameras, the problems of traditional lifting efficiency and safety risks are solved, and efficient and safe workpiece lifting operations are achieved.
Patent Information
- Application Number
- CN202410920957.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-10
- Publication Date
- 2024-11-08
AI Technical Summary
In the manufacturing scenario of power generation equipment, the traditional lifting workpiece model relies on manual command, is inefficient and has safety risks, and cannot effectively perceive the position and status of the workpiece.
Computer-aided technology is used to combine AI and three-dimensional engines to obtain workpiece models through binocular cameras, real-time positioning and security monitoring are achieved, rendered images and monitoring images are integrated, workpiece speed and acceleration are calculated, and safety warnings are provided.
It improves the efficiency and safety of lifting operations, realizes controllability of large workpieces, and reduces the safety risks of manual operations.
Smart Images

Figure CN118908065A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent lifting technology, and in particular to an intelligent lifting assistance method in a power generation equipment manufacturing scenario. Background Art
[0002] In the machining workshop of power generation equipment, the transportation and clamping of large workpieces mainly rely on cranes. The traditional lifting operation mode completely relies on manual command and operation. Due to the large weight and volume of the workpiece, it is not convenient to sense the position of the workpiece, resulting in low efficiency of the manual mode and certain safety risks. Summary of the invention
[0003] In view of this, the present invention provides an intelligent lifting assistance method in a power generation equipment manufacturing scenario, which combines computer-aided technology to realize intelligent lifting operation assistance, analyzes the lifting position of the workpiece, safety warnings and other information and provides them to the operating personnel, thereby improving the efficiency and safety of the lifting operation.
[0004] The present invention discloses an intelligent lifting assistance method in a power generation equipment manufacturing scenario, which comprises:
[0005] Obtain the model of the workpiece to be lifted in the power generation equipment manufacturing scenario and set its lifting position. Use a 3D engine to render the workpiece model at the lifting position of the workpiece model to be lifted, and fuse the model rendering image with the real monitoring image to output in real time for on-site workers to view. Use AI recognition and 3D reconstruction to locate the workpiece in the lifting operation in real time, and determine whether the lifting operation is in place based on the spatial position of the workpiece obtained by real-time positioning. Safe monitoring of the workpiece lifting process is achieved by calculating the speed and acceleration of the workpiece.
[0006] Furthermore, several binocular cameras are installed so that their field of view covers the lifting operation area in the workshop; the intrinsic parameter matrix of the binocular camera and the extrinsic parameter matrix relative to the ground of the lifting operation area are obtained so that the workpiece model image rendered by the 3D engine coincides with the actual monitoring image. The simplified relationship between the pixel coordinates in the camera and the real 3D space coordinates is:
[0007]
[0008] Where u and v are the pixel coordinates in the camera image, X, Y, and Z are the real three-dimensional space coordinates, f is the focal length, u0 and v0 are the pixel origin coordinates in the camera image, K is the camera intrinsic parameter matrix, R is the rotation matrix, and t is the displacement vector. The chessboard calibration method is used to fix the black and white chessboard at multiple arbitrary positions on the ground of the working area and obtain images respectively. The intrinsic parameter matrix K of the monitoring camera and the extrinsic parameter matrix relative to the ground of the working area can be calculated by the OpenCV tool. The extrinsic parameter matrix includes the rotation matrix R and the displacement vector t.
[0009] Furthermore, the workpiece model is imported into the 3D engine, the position of the workpiece model is adjusted according to the clamping position of the hoisting, and a virtual camera is set up in the 3D engine, wherein the parameters of the virtual camera in the engine include the camera's external parameter matrix, resolution, and field of view fov; the external parameter matrix and resolution directly use the parameters of the actually installed binocular camera; the field of view fov is calculated by the following formula:
[0010]
[0011] Where w is the horizontal resolution of the virtual camera.
[0012] Furthermore, the workpiece model is rendered in the virtual camera, and the appearance is rendered using the normal. The color of each pixel in the workpiece model is determined by the normal direction of its corresponding pixel position, which is calculated by the following formula:
[0013]
[0014] Where XYZ2RGB is the mapping matrix from three-dimensional space to RGB color space, color r 、color g 、color b They are the values of red, green, and blue in the RGB color space, respectively. x 、normal y 、normal z They are the x-axis coordinate value, y-axis coordinate value, and z-axis coordinate value of the model normal direction in the three-dimensional space coordinate system;
[0015] The workpiece model is rendered with a black background to facilitate the removal of the background when it is integrated with the monitoring screen.
[0016] Furthermore, the OpenCV tool is used to directly add the rendered image and the real monitoring image to achieve image fusion; or, the pixel values of all pixels in the model rendered image are halved so that the model rendered image appears semi-transparent after fusion; since the background of the rendered image is black and the pixel value is 0, the model rendered image and the real monitoring image after the pixel values of all pixels are halved are directly added and normalized to achieve image fusion without affecting the real monitoring image, that is:
[0017] Mat 融合 =Mat 监控 +Mat 渲染 / 2
[0018] Among them, Mat 融合 To render the model into an image Mat 渲染 The pixel values of all pixels are halved and compared with the real monitoring image Mat监控 The fused image is output by fusion; the fused image is output to the display terminal so that the on-site operators can view the real-time status of the lifting operation.
[0019] Furthermore, a data set is obtained: images of a period of time are retrieved from the video of the actual workpiece lifting operation area in the workshop monitored by the binocular camera, and images are sampled according to the rules of different workpieces, different times, and different weather conditions. The workpiece targets in the sampled images are marked with rectangular frames to form a workpiece target recognition data set;
[0020] Training AI model: Input the image target recognition data set into the YOLO model training to obtain the workpiece AI recognition model. The workpiece AI recognition model outputs different types of workpieces in the workshop;
[0021] Reasoning: Input the real-time image monitored by the binocular camera to the workpiece AI recognition model. At the current time t, n binocular cameras can obtain n t states, each state contains the category and rectangular frame position of each workpiece.
[0022] Furthermore, a binocular stereo matching 3D reconstruction algorithm is used to obtain a workshop point cloud; the point cloud inside the workpiece rectangular frame identified by the workpiece AI recognition model is taken to segment the workpiece point cloud from the workshop point cloud; the vertex group of the workpiece model is converted into a point cloud, and the point cloud is aligned with the segmented workpiece point cloud to obtain the spatial coordinates of the workpiece.
[0023] Furthermore, when the sum of the three-axis coordinate differences between the real-time position P of the workpiece and its clamping position D is less than the set threshold, it is considered that the lifting operation has been completed;
[0024] The continuous motion trajectory of the workpiece is regarded as a function of position to time, a polynomial fitting of the workpiece position trajectory is performed, and the polynomial is derived to obtain the velocity v and acceleration a of the workpiece;
[0025] When the speed v of the workpiece exceeds the set safety speed threshold, it means that the lifting operation may be too fast, a safety accident may occur, and the operation site needs to be controlled;
[0026] When the acceleration a of the workpiece exceeds the set safety acceleration threshold, it means that the hoisted workpiece may topple over or the hoisting rope may break, and the work site should be evacuated safely immediately.
[0027] Furthermore, the calculation process of the velocity v and acceleration a of the workpiece includes:
[0028] The positions of the i sets of workpieces obtained are: {(x0, t0), (x1, t1), …, (x i-1 ,t i-1 ),},x i-1 represents time i-1 ti-1 The position of the workpiece, the n-order polynomial of the workpiece position trajectory to be fitted is:
[0029] x=k n t n +k n-1 t n-1 +…+k1t+k0
[0030] x represents the workpiece position obtained by fitting; k n represents the nth polynomial coefficient;
[0031] Substituting the acquired i groups of workpiece positions into the polynomial, we get the linear equation system:
[0032]
[0033] Written in matrix form:
[0034] Ak=x
[0035] When i>n+1, the system of equations is overdetermined, and the least squares solution is obtained by singular value decomposition:
[0036] k=(A T A) -1 A T x
[0037] After obtaining the polynomial fitting of the workpiece position trajectory, the workpiece velocity v and acceleration a can be obtained by derivation of the polynomial:
[0038] v=nk n t n-1 +(n-1)k n-1 t n-2 +…+2k2t+k1
[0039] a=n(n-1)k n t n-2 +(n-1)(n-2)k n-1 t n-3 +…+6k3t+2k2.
[0040] Due to the adoption of the above technical solution, the present invention has the following advantages: the present invention incorporates the lifting of large workpieces into digital control through auxiliary methods such as computer vision, so that operators can intuitively and conveniently view the lifting status of workpieces and conduct operation commands, and realize safety monitoring of the lifting process, greatly improving the controllability of the lifting operation. The auxiliary system can also be expanded to unmanned warehouses, large-scale transportation and lifting, port loading and unloading and other scene applications. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.
[0042] Figure 1 A schematic flow chart of an intelligent lifting assistance method in a power generation equipment manufacturing scenario according to an embodiment of the present invention;
[0043] Figure 2 A schematic diagram of a model image rendered according to an embodiment of the present invention;
[0044] FIG3( a ) is a schematic diagram of an image before fusion according to an embodiment of the present invention;
[0045] FIG3( b ) is a schematic diagram of a fused image according to an embodiment of the present invention;
[0046] Figure 4 Schematic diagram of the recognition result of the workpiece AI recognition model according to an embodiment of the present invention. DETAILED DESCRIPTION
[0047] The present invention is further described in conjunction with the accompanying drawings and embodiments, and the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field belong to the protection scope of the embodiments of the present invention.
[0048] See also Figure 1 The present invention provides an embodiment of an intelligent lifting assistance method in a power generation equipment manufacturing scenario, which includes: obtaining a workpiece model to be lifted in the power generation equipment manufacturing scenario and setting its lifting position, using a three-dimensional engine to render the workpiece model at the lifting position of the workpiece model to be lifted, and fusing the model rendering image and the real monitoring image to output in real time for on-site workers to view; real-time positioning of the workpiece in the lifting operation through AI recognition and three-dimensional reconstruction, and judging whether the lifting operation is in place based on the spatial position of the workpiece obtained by real-time positioning; and realizing safe monitoring of the workpiece lifting process by calculating the speed and acceleration of the workpiece.
[0049] Specifically, this embodiment includes the following steps:
[0050] Install several binocular cameras so that their field of view covers the lifting operation area in the workshop; obtain the intrinsic parameter matrix of the binocular camera and the extrinsic parameter matrix relative to the ground of the lifting operation area, so that the workpiece model image rendered by the 3D engine coincides with the actual monitoring image. The simplified relationship between the pixel coordinates in the camera and the real 3D space coordinates is:
[0051]
[0052] Where u and v are the pixel coordinates in the camera image, X, Y, and Z are the real three-dimensional space coordinates, f is the focal length, u0 and v0 are the pixel origin coordinates in the camera image, K is the camera intrinsic parameter matrix, R is the rotation matrix, and t is the displacement vector. The chessboard calibration method is used to fix the black and white chessboard at multiple arbitrary positions on the ground of the working area and obtain images respectively. The intrinsic parameter matrix K of the monitoring camera and the extrinsic parameter matrix relative to the ground of the working area can be calculated by the OpenCV tool. The extrinsic parameter matrix includes the rotation matrix R and the displacement vector t.
[0053] Import the workpiece model into the 3D engine, adjust the position of the workpiece model according to the clamping position of the hoisting, and set up a virtual camera in the 3D engine. The parameters of the virtual camera in the engine include the camera's external parameter matrix, resolution, and field of view fov; the external parameter matrix and resolution directly use the parameters of the actual installed binocular camera; the field of view fov is calculated by the following formula:
[0054]
[0055] Where w is the horizontal resolution of the virtual camera.
[0056] The workpiece model is rendered in the virtual camera, and the appearance is rendered using normals. The color of each pixel in the workpiece model is determined by the normal direction of its corresponding pixel position, which is calculated by the following formula:
[0057]
[0058] Where XYZ2RGB is the mapping matrix from three-dimensional space to RGB color space, color r 、color g 、color b They are the values of red, green, and blue in the RGB color space, respectively. x 、normal y 、normal z They are the x-axis coordinate value, y-axis coordinate value, and z-axis coordinate value of the model normal direction in the three-dimensional space coordinate system;
[0059] In order to remove the background when merging with the monitoring screen, a black background is used for rendering. The rendered model image is as follows Figure 2 shown.
[0060] In order to achieve pixel-level fusion of the monitoring screen and the rendering screen, the distortion of the monitoring screen needs to be removed. The main types of distortion of monitoring cameras are barrel distortion or fisheye distortion. To remove the distortion, the intrinsic parameter matrix K and distortion parameter k of the binocular camera are required, which can be obtained together in the checkerboard calibration calculation in the first step. After obtaining the distortion parameters, the undistort() method of the OpenCV tool is used to quickly obtain the undistorted monitoring screen.
[0061] Image fusion is implemented using the OpenCV tool, which can directly add the rendered image (as shown in Figure 3(a)) and the real monitoring image (as shown in Figure 3(b)). In order to make the rendered image appear semi-transparent after fusion, the pixel values of all pixels in the rendered image are halved; since the background of the rendered image is black and the pixel value is 0, the model rendering image with the pixel values halved and the real monitoring image are directly added and normalized to achieve image fusion without affecting the real monitoring image, that is:
[0062] Mat 融合 =Mat 监控 +Mat 渲染 / 2
[0063] Among them, Mat 融合 To render the model into an image Mat 渲染 The pixel values of all pixels are halved and compared with the real monitoring image Mat 监控 The fused image is output by fusion; the fused image is output to the display terminal so that the on-site operators can view the real-time status of the lifting operation.
[0064] The image before fusion is shown in Figure 3(a), and the image after fusion is shown in Figure 3(b). The workers at the lifting site can use the fused video to intuitively and quickly view the lifting status of the workpiece and control the movement of the crane.
[0065] Use the YOLOv8 algorithm to implement AI recognition of artifacts. The steps are as follows:
[0066] Obtaining data sets: Retrieve images from the workshop monitoring video for a period of time, sample images according to the rules of different workpieces, different times, and different weather conditions, and mark the workpiece targets in the sampled images with rectangular frames to form a workpiece target recognition data set;
[0067] Training AI model: Take various large workpieces commonly seen in the workshop as classification output targets, and use the YOLO algorithm to train the workpiece image recognition AI model. Input the image target recognition data set into the YOLO model training to obtain the workpiece AI recognition model, which outputs different types of workpieces in the workshop;
[0068] Reasoning: Input the real-time monitoring video of the workshop into the workpiece AI recognition model. n surveillance cameras can obtain n t states, t is the corresponding time, and each state contains the category and rectangular frame position of each workpiece. The recognition results are as follows: Figure 4 shown.
[0069] Workpiece positioning requires the combination of 3D reconstructed point cloud data from the workshop, which is generally obtained through a binocular camera. The positioning operation process is as follows:
[0070] (1) Obtain workshop point cloud using binocular stereo matching 3D reconstruction algorithm;
[0071] (2) Take the point cloud inside the workpiece rectangular frame identified by the workpiece AI recognition model and segment the workpiece point cloud from the workshop point cloud;
[0072] (3) The vertex group of the workpiece STL model is converted into a point cloud, and the ICP precise matching is performed with the segmented workpiece point cloud to obtain the precise spatial coordinates of the workpiece.
[0073] The point cloud ICP matching algorithm process is as follows:
[0074] The ICP algorithm inputs two point clouds, the initial value of the transformation, and the iteration stop condition, and outputs the pose transformation matrix between the two point clouds. The calculation process is divided into two steps:
[0075] (1) Find the nearest corresponding point. Use the initial value of the point cloud transformation or the result of the previous iteration to transform the coordinates of the initial point cloud to obtain an intermediate transformed point cloud. Compare the intermediate point cloud with the target point cloud and regard the pair of points with the closest distance as corresponding points.
[0076] (2) Iteratively solve the transformation, solve the optimal transformation for all corresponding point pairs obtained in the previous step, and obtain the intermediate transformation matrix.
[0077] The ICP algorithm continuously iterates the above two steps until the iteration termination condition is met, and the maximum number of iterations is used to terminate the condition to ensure the stability of the algorithm. At the same time, in the continuous frames of the surveillance video, the matching result of the previous frame is used as the initial value of the transformation of the next frame, which greatly improves the efficiency of the matching algorithm.
[0078] Real-time analysis of lifting operations is mainly divided into two parts: completion prompts and safety monitoring.
[0079] Lifting completion prompt: When the sum of the three-axis coordinate differences between the real-time position P of the workpiece and the target position D is less than the set threshold, the lifting operation is considered to be in place;
[0080] Safety monitoring: It is necessary to further calculate the velocity v and acceleration a of the workpiece. The continuous motion trajectory of the workpiece is regarded as a function of position versus time, and a polynomial fitting is performed:
[0081] The positions of the i sets of workpieces obtained are: {(x0, t0), (x1, t1), …, (x i-1 ,t i-1 ),},x i-1 represents time i-1 t i-1 The position of the workpiece, the n-order polynomial of the workpiece position trajectory to be fitted is:
[0082] x=k n t n +k n-1 t n-1 +…+k1t+k0
[0083] x represents the workpiece position obtained by fitting; k n represents the nth polynomial coefficient;
[0084] Substituting the acquired i groups of workpiece positions into the polynomial, we get the linear equation system:
[0085]
[0086] Written in matrix form:
[0087] Ak=x
[0088] When i>n+1, the system of equations is overdetermined, and the least squares solution is obtained by singular value decomposition:
[0089] k=(A T A) -1 A T x
[0090] After obtaining the polynomial fitting of the workpiece position trajectory, the workpiece velocity v and acceleration a can be obtained by derivation of the polynomial:
[0091] v=nk n t n-1 +(n-1)k n-1 t n-2 +…+2k2t+k1
[0092] a=n(n-1)k n t n-2 +(n-1)(n-2)k n-1 t n-3 +…+6k3t+2k2
[0093] Finally, the workpiece speed and acceleration are safely monitored, corresponding to different states:
[0094] When the speed v exceeds the set safety threshold, it means that the lifting operation may be too fast and a safety accident may occur, so the operation site should be controlled;
[0095] When the acceleration a exceeds the set safety threshold, it means that the hoisted workpiece may topple over or the hoisting rope may break, and the work site should be evacuated immediately.
[0096] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art understand that the specific implementation modes of the present invention may still be modified or replaced by equivalents, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention shall be included in the scope of protection of the claims of the present invention.
Claims
1. An intelligent lifting assistance method in a power generation equipment manufacturing scenario, characterized in that: include: Obtain the model of the workpiece to be lifted in the power generation equipment manufacturing scenario and set its lifting position. Use a 3D engine to render the workpiece model at the lifting position of the workpiece model to be lifted, and fuse the model rendering image with the real monitoring image to output in real time for on-site workers to view. Use AI recognition and 3D reconstruction to locate the workpiece in the lifting operation in real time, and determine whether the lifting operation is in place based on the spatial position of the workpiece obtained by real-time positioning. Safe monitoring of the workpiece lifting process is achieved by calculating the speed and acceleration of the workpiece.
2. The intelligent lifting assistance method in the electrical equipment manufacturing scenario according to claim 1 is characterized in that: Install several binocular cameras so that their field of view covers the lifting operation area in the workshop; obtain the intrinsic parameter matrix of the binocular camera and the extrinsic parameter matrix relative to the ground of the lifting operation area, so that the workpiece model image rendered by the 3D engine coincides with the actual monitoring image. The simplified relationship between the pixel coordinates in the camera and the real 3D space coordinates is: Where u and v are the pixel coordinates in the camera image, X, Y, and Z are the real three-dimensional space coordinates, f is the focal length, u0 and v0 are the pixel origin coordinates in the camera image, K is the camera intrinsic parameter matrix, R is the rotation matrix, and t is the displacement vector. The chessboard calibration method is used to fix the black and white chessboard at multiple arbitrary positions on the ground of the working area and obtain images respectively. The intrinsic parameter matrix K of the monitoring camera and the extrinsic parameter matrix relative to the ground of the working area can be calculated by the OpenCV tool. The extrinsic parameter matrix includes the rotation matrix R and the displacement vector t.
3. The intelligent lifting assistance method in the electrical equipment manufacturing scenario according to claim 2 is characterized in that: Import the workpiece model into the 3D engine, adjust the position of the workpiece model according to the clamping position of the hoisting, and set up a virtual camera in the 3D engine. The parameters of the virtual camera in the engine include the camera's external parameter matrix, resolution, and field of view fov; the external parameter matrix and resolution directly use the parameters of the actual installed binocular camera; the field of view fov is calculated by the following formula: Where w is the horizontal resolution of the virtual camera.
4. The intelligent lifting assistance method in the electrical equipment manufacturing scenario according to claim 3 is characterized in that: The workpiece model is rendered in the virtual camera, and the appearance is rendered using normals. The color of each pixel in the workpiece model is determined by the normal direction of its corresponding pixel position, which is calculated by the following formula: Where XYZ2RGB is the mapping matrix from three-dimensional space to RGB color space, color r 、color g 、color b They are the values of red, green, and blue in the RGB color space, respectively. x 、normal y 、normal z They are the x-axis coordinate value, y-axis coordinate value, and z-axis coordinate value of the model normal direction in the three-dimensional space coordinate system; The workpiece model is rendered with a black background to facilitate the removal of the background when it is integrated with the monitoring screen.
5. The intelligent lifting assistance method in the electrical equipment manufacturing scenario according to claim 1 is characterized in that: Use OpenCV tools to directly add the rendered image and the real monitoring image to achieve image fusion; or, halve the pixel values of all pixels in the model rendered image so that the model rendered image appears semi-transparent after fusion; since the background of the rendered image is black and the pixel value is 0, the model rendered image with the pixel values of all pixels halved is directly added to the real monitoring image and then normalized to achieve image fusion without affecting the real monitoring image, that is: Food 融合 =Food 监控 +Food 渲染 / 2 Among them, Mat 融合 To render the model into an image Mat 渲染 The pixel values of all pixels are halved and compared with the real monitoring image Mat 监控 The fused image is output by fusion; the fused image is output to the display terminal so that the on-site operators can view the real-time status of the lifting operation.
6. The intelligent lifting assistance method in the electrical equipment manufacturing scenario according to claim 1 is characterized in that: Obtaining the data set: Retrieve images of a period of time from the video of the actual workpiece lifting operation area in the workshop monitored by the binocular camera, sample images according to the rules of different workpieces, different times, and different weather conditions, and mark the workpiece targets in the sampled images with rectangular frames to form a workpiece target recognition data set; Training AI model: Input the image target recognition data set into the YOLO model training to obtain the workpiece AI recognition model. The workpiece AI recognition model outputs different types of workpieces in the workshop; Reasoning: Input the real-time image monitored by the binocular camera to the workpiece AI recognition model. At the current time t, n binocular cameras can obtain n t states, each state contains the category and rectangular frame position of each workpiece.
7. The intelligent lifting assistance method in the electrical equipment manufacturing scenario according to claim 1 is characterized in that: The workshop point cloud is obtained by using the binocular stereo matching 3D reconstruction algorithm. The point cloud inside the workpiece rectangular frame identified by the workpiece AI recognition model is taken to segment the workpiece point cloud from the workshop point cloud. The vertex group of the workpiece model is converted into a point cloud, and the point cloud is aligned with the segmented workpiece point cloud to obtain the spatial coordinates of the workpiece.
8. The intelligent lifting assistance method in the electrical equipment manufacturing scenario according to claim 1 is characterized in that: When the sum of the three-axis coordinate differences between the real-time position P of the workpiece and its clamping position D is less than the set threshold, it is considered that the lifting operation has been completed; The continuous motion trajectory of the workpiece is regarded as a function of position to time, a polynomial fitting of the workpiece position trajectory is performed, and the polynomial is derived to obtain the velocity v and acceleration a of the workpiece; When the speed v of the workpiece exceeds the set safety speed threshold, it means that the lifting operation may be too fast, a safety accident may occur, and the operation site needs to be controlled; When the acceleration a of the workpiece exceeds the set safety acceleration threshold, it means that the hoisted workpiece may topple over or the hoisting rope may break, and the work site should be evacuated safely immediately.
9. The intelligent lifting assistance method in the electrical equipment manufacturing scenario according to claim 8 is characterized in that: The calculation process of the velocity v and acceleration a of the workpiece includes: The positions of the i groups of workpieces obtained are: {(x0, t0), (x1, t1), …, (x i-1 ,t i-1 ),},x i-1 represents time i-1 t i-1 The position of the workpiece, the n-order polynomial of the workpiece position trajectory to be fitted is: x=k n t n +k n-1 t n-1 +…+k1t+k0 x represents the workpiece position obtained by fitting; k n represents the nth polynomial coefficient; Substituting the acquired i groups of workpiece positions into the polynomial, we get the linear equation system: Written in matrix form: Ak=x When i>n+1, the system of equations is overdetermined, and the least squares solution is obtained by singular value decomposition: k=(A T A) -1 A T x After obtaining the polynomial fitting of the workpiece position trajectory, the workpiece velocity v and acceleration a can be obtained by derivation of the polynomial: v=nk n t n-1 +(n-1)k n-1 t n-2 +…+2k2t+k1 a=n(n-1)k n t n-2 +(n-1)(n-2)k n-1 t n-3 +…+6k3t+2k2。
Citation Information
Cited By
Three-dimensional visual safety monitoring method for wind power hoisting site
CN121614986A
