A torque limiter intelligent debugging system and method

Abstract

The present invention discloses an intelligent debugging system for a torque limiter, including a weight, a 3D sensor, an operation controller and a torque limiter; the 3D sensor detects the weight number on the weight, and the 3D sensor can simultaneously output the image and the distance information of the corresponding pixels of the image area; the operation controller processes the visual signal and / or distance information of the 3D sensor and is electrically connected to the 3D sensor through a high-speed communication bus; the operation controller is electrically connected to the torque limiter through the communication bus, and the operation controller transmits the weight and amplitude information of the hoisting weight to the torque limiter. And discloses an intelligent debugging method for a torque limiter. The present invention relies on the linkage of the 3D sensor and the torque limiter to automatically identify the weight and amplitude of the weight, automatically debug and calibrate the torque limiter, avoid errors caused by human measurement and operating errors, improve the working environment and intensity of on-site debugging personnel, and improve work efficiency.

Description

Technical Field

[0001] The invention relates to the field of crane control, and in particular to an intelligent debugging system and method for a torque limiter. Background Art

[0002] The crane torque limiter is an essential safety device for truck cranes and crawler cranes. Before such cranes leave the factory, the torque limiter must be strictly debugged and calibrated. During the debugging process of the crane torque limiter, the weight of the weight and the lifting range need to be constantly changed, and then the lifting range and weight need to be manually measured and entered into the torque limiter debugging system for calibration.

[0003] The existing technology manually measures the hoisting amplitude of each debugging point, and then the personnel calculate the actual hoisting weight based on the actual weight data on the hoisting weight, and manually input the two data into the corresponding parameter entry of the torque display debugging interface. The torque limiter corrects the model according to the input hoisting amplitude and actual hoisting weight to ensure that the hoisting amplitude and weight display of the debugging point are accurate. Since the hoisting amplitude of the existing technology needs to be manually measured on site, the hoisting weight needs to be manually calculated and manually input. The whole process may cause errors due to human factors, and someone needs to be specifically responsible for measuring the amplitude. Crane debugging is carried out outdoors, and severe cold and heat make the working environment of the measurement personnel poor and the intensity is high.

[0004] Therefore, it is necessary to develop intelligent debugging technology for torque limiters in a targeted manner to avoid errors caused by human measurement and operational errors, improve the working environment and intensity of on-site debugging personnel, and improve work efficiency. Summary of the invention

[0005] Purpose of the invention: In view of the shortcomings and defects of the prior art, the present invention provides an intelligent debugging system and method for a torque limiter, which relies on the linkage of a 3D sensor and a torque limiter to automatically identify the weight and amplitude of the weight, automatically debug and calibrate the torque limiter, avoid errors caused by human measurement and operational errors, improve the working environment and intensity of on-site debugging personnel, and improve work efficiency.

[0006] Technical solution: The invention provides an intelligent debugging system for a torque limiter, characterized in that it includes weights, a 3D sensor, an operation controller and a torque limiter; the 3D sensor detects the weight numbers on the weights, and the 3D sensor can simultaneously output the image and the distance information of the pixels corresponding to the image area; the operation controller processes the visual signal and / or distance information of the 3D sensor and is electrically connected to the 3D sensor through a high-speed communication bus; the operation controller is electrically connected to the torque limiter through the communication bus, and the operation controller transmits the weight and lifting amplitude information of the hoisting weight to the torque limiter.

[0007] Wherein, the 3D sensor, operation controller and torque limiter are all arranged on the crane.

[0008] The 3D sensor is installed on the boom or turntable of the crane, and the 3D sensor covers the hoisting area during the debugging of the torque limiter.

[0009] Wherein, the weight is hung on the hook of the crane through a wire rope during debugging.

[0010] The intelligent debugging method for the torque limiter of the present invention is characterized in that: during debugging, it is divided into no-load debugging and loaded debugging; no-load debugging means that no weight is hung on the hook of the crane, and loaded debugging means that weight is hung on the hook of the crane.

[0011] Wherein, the no-load debugging comprises the following steps:

[0012] 1) Select no-load debugging on the torque limiter interface;

[0013] 2) The computing controller processes the image signal of the 3D sensor and identifies the hook position and type through a deep learning algorithm;

[0014] 3) Calculate the three-dimensional distance information of the hook through the two-dimensional hook position information, and then calculate the hook amplitude information;

[0015] 4) The calculation controller sends the hook amplitude information and hook type to the torque limiter;

[0016] 5) Storing weight information of different hook types in the torque limiter;

[0017] 6) The torque limiter looks up the table according to the hook type sent by the operation controller and gives the current hook weight;

[0018] 7) The torque limiter is automatically debugged and calibrated according to the collected hook weight and amplitude information.

[0019] Furthermore, the algorithm in step 2) is a YOLO series algorithm, and the YOLO v5 algorithm can be preferably used.

[0020] The on-load debugging includes the following steps:

[0021] 1) The crane has lifted the weight to be debugged, and select load debugging on the torque limiter interface;

[0022] 2) The computing controller processes the image signal of the 3D sensor and identifies the hook position and type through a deep learning algorithm;

[0023] 3) Calculate the three-dimensional distance information of the hook through the two-dimensional hook position information, and then calculate the hook amplitude information;

[0024] 4) The operation controller processes the image signal of the weight collected by the 3D sensor and identifies the weight of the weight through the digital OCR algorithm;

[0025] 5) The operation controller sends the hook amplitude information and hook type to the torque limiter;

[0026] 6) Storing weight information of different hook types in the torque limiter;

[0027] 7) The torque limiter looks up the table according to the hook type sent by the operation controller and gives the current hook weight;

[0028] 8) The torque limiter adds the weight of the weight sent by the operation controller and the weight of the hook obtained by looking up the table to obtain the load weight;

[0029] 9) The torque limiter is automatically debugged and calibrated according to the collected load weight and amplitude information.

[0030] Furthermore, the YOLO series algorithm identifies the hook position and hook type: put multi-angle and multi-position pictures of all hooks of the crane to be debugged into the YOLO data set, and replace multiple classified pictures in its pre-trained data set; then use the trained model data in the operation controller to identify the pictures containing the hook to effectively mark the type and position of the hook; read the corresponding three-dimensional information within the calibrated hook position range and find the closest point cloud clustering with the Euclidean clustering algorithm, and calculate the crane amplitude direction average value of the point cloud clustering as the sensor to the horizontal hook distance L1 detected by the 3D sensor, and then calculate the hook amplitude L=L1+L2 based on the measured distance L2 from the installation position of the 3D sensor to the rotation center.

[0031] Furthermore, the size specified by the weight information of the weight is printed in a specific area of ​​the weight in a computer standard font. The operation controller collects image information from the 3D sensor to mark the position of each weight, and then performs OCR text detection near the center of each weight. Each detected weight corresponds to a weight value, and then all weight values ​​are added together to obtain the total weight of the weight.

[0032] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: The present invention relies on the linkage of 3D sensor and torque limiter to automatically identify the weight and amplitude of the weight, automatically debug and calibrate the torque limiter, avoid errors caused by human measurement and operation errors, improve the working environment and intensity of on-site debugging personnel, and improve work efficiency. The present invention realizes automatic calibration and debugging of the torque limiter on site, improves work efficiency and reduces labor intensity. The weight system only needs to be sprayed in a standardized manner, which is easy to implement. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a structural schematic diagram of the present invention;

[0034] Figure 2 It is a schematic diagram of the installation structure of the 3D sensor of the present invention;

[0035] Figure 3 It is a schematic diagram of the hook amplitude calculation of the present invention;

[0036] Figure 4 It is a schematic diagram of the weight identification of the present invention;

[0037] In the figure, 1 is the hook; 2 is the boom; 3 is the 3D sensor; 4 is the turntable; and 5 is the rotation center. DETAILED DESCRIPTION

[0038] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and specific implementation methods.

[0039] The invention discloses an intelligent debugging system and method for a torque limiter.

[0040] A torque limiter intelligent debugging system includes an operation controller, a 3D sensor 3, a torque limiter and a weight, wherein the operation controller is responsible for processing the visual signal and / or distance information of the 3D sensor 3 and is electrically connected to the 3D sensor 3 through a high-speed communication bus. The operation controller is electrically connected to the torque limiter through a communication bus, and can transmit the weight of the hoisting weight and the hoisting amplitude information to the torque limiter. The 3D sensor 3 can simultaneously output the image and the distance information of the pixels corresponding to the image area. The prior art has two forms: 1. A binocular structure composed of two cameras; 2. A device after a camera and a laser radar are jointly calibrated. The torque limiter is a system installed on the vehicle and communicates with the operation controller through the bus. The weight is an essential component for crane debugging. This system can use conventional weights. It only needs to ensure that the weight of the weight is clearly and accurately marked on the weight, and the weight number of the weight can be detected by the 3D sensor 3. The system is arranged on the crane except for the weights, wherein the 3D sensor is installed on the crane turntable 4 or the boom 2, and the installation position satisfies that the 3D sensor 3 can cover the hoisting area during the debugging of the torque limiter, and the weights are hung on the crane hook 1 by wire ropes during debugging.

[0041] An intelligent debugging method for a torque limiter: The debugging follows the common debugging method for the torque limiter, which is divided into no-load debugging and loaded debugging. No-load debugging means that no weight is mounted on the crane hook 1, and loaded debugging means that a weight is mounted on the crane hook 1.

[0042] No-load debugging method: The crane keeps the hook empty and does not lift any weight. 1. Select no-load debugging on the torque limiter interface; 2. The operation controller processes the image signal of the 3D sensor 3, and identifies the hook position and hook type through a deep learning algorithm, such as the YOLO series algorithm; 3. The three-dimensional distance information of the hook is calculated through the two-dimensional hook position information, and then the hook amplitude information is calculated; 4. The operation controller sends the hook amplitude information and hook type to the torque limiter; 5. The weight information of different hook types is stored in the torque limiter; 6. The torque limiter looks up the table according to the hook type sent by the operation controller and gives the current hook weight; 7. The torque limiter automatically debugs and calibrates according to the collected hook weight and amplitude information.

[0043] On-load debugging method: The crane has lifted the weight to be debugged. 1. Select on-load debugging on the torque limiter interface; 2. The operation controller processes the image signal of the 3D sensor 3, and identifies the hook position and hook type through a deep learning algorithm, such as the YOLO v5 algorithm; 3. The three-dimensional distance information of the hook is calculated through the two-dimensional hook position information, and then the hook amplitude information is calculated; 4. The operation controller processes the image signal of the weight collected by the 3D sensor 3, and identifies the weight of the weight through a digital OCR algorithm; 5. The operation controller sends the hook amplitude information and hook type to the torque limiter; 6. The weight information of different hook types is stored in the torque limiter; 7. The torque limiter looks up the table according to the hook type sent by the operation controller to give the current hook weight; 8. The torque limiter adds the weight of the weight sent by the operation controller and the hook weight obtained by looking up the table to obtain the load weight; 9. The torque limiter automatically debugs and calibrates according to the collected load weight and amplitude information.

[0044] Further:

[0045] The YOLO v5 algorithm identifies the hook position and type of hook, and puts all the multi-angle and multi-position pictures of the crane to be debugged into the YOLO v5 data set, and replaces the pictures of multiple categories in its pre-trained data set. For example, if the vehicle to be debugged has 3 types of hooks, then the corresponding 3 labels are used to store the pictures of the 3 types of hooks respectively, and multiple rounds of training are performed to achieve the highest recognition accuracy; then the trained model data is used in the operation controller to identify the pictures containing the hooks to effectively mark the type and position of the hook. The corresponding three-dimensional information within the calibrated hook position range is read and the closest point cloud cluster is found by the Euclidean clustering algorithm, and the crane amplitude direction average value of the point cloud cluster is obtained as the sensor to the horizontal hook distance L1 detected by the 3D sensor 3, and then the hook amplitude L=L1+L2 can be calculated based on the measured distance L2 from the installation position of the 3D sensor 3 to the rotation center 5.

[0046] The weight information of the weight can be printed in a specified size in a standard computer font in a specific area of ​​the weight, such as the center position of the side facing the 3D sensor; the weight image is trained using the YOLO v5 algorithm program in a method similar to the above description, and the operation controller collects the image information of the 3D sensor 3 to mark the position of each weight, and then performs OCR text detection near the center of gravity of each weight, and corresponds to a weight value for each detected weight, and then all weight values ​​are added together to obtain the total weight of the weight.

[0047] The method of debugging and calibrating the torque limiter according to the collected load weight and amplitude information is consistent with the existing method of manually inputting the load weight and amplitude information to calibrate the torque limiter.

[0048] The present invention relies on the linkage of 3D sensor and torque limiter to automatically identify the weight and amplitude of the weight, automatically debug and calibrate the torque limiter, avoid errors caused by human measurement and operation errors, improve the working environment and intensity of on-site debuggers, and improve work efficiency. The present invention realizes automatic calibration and debugging of the torque limiter on site, improves work efficiency and reduces labor intensity. The weight system only needs to be sprayed in a standardized manner, which is easy to implement.

Claims

1. A torque limiter intelligent debugging system, characterized in that: The invention comprises a weight, a 3D sensor (3), an operation controller and a torque limiter; the 3D sensor (3) detects the weight number on the weight, and the 3D sensor (3) can simultaneously output an image and distance information of pixels corresponding to the image area; the operation controller processes the visual signal and / or distance information of the 3D sensor (3) and is electrically connected to the 3D sensor (3) via a high-speed communication bus; the operation controller is electrically connected to the torque limiter via the communication bus, and the operation controller transmits the weight and hoisting amplitude information of the hoisting weight to the torque limiter.

2. The torque limiter intelligent debugging system according to claim 1 is characterized in that: The 3D sensor (3), operation controller and torque limiter are all arranged on the crane.

3. The torque limiter intelligent debugging system according to claim 2 is characterized in that: The 3D sensor (3) is installed on the boom (2) or the turntable (4) of the crane, and the 3D sensor (3) covers the hoisting area during the debugging of the torque limiter.

4. The torque limiter intelligent debugging system according to claim 3 is characterized in that: The weight is hung on the hook (1) of the crane via a steel wire rope during debugging.

5. A debugging method using the torque limiter intelligent debugging system according to any one of claims 1 to 4, characterized in that: During debugging, there are no-load debugging and loaded debugging; no-load debugging means that no weight is hung on the hook (1) of the crane, and loaded debugging means that weight is hung on the hook (1) of the crane.

6. The intelligent debugging method of torque limiter according to claim 5 is characterized in that: The no-load debugging includes the following steps: 1) Select no-load debugging on the torque limiter interface; 2) The computing controller processes the image signal of the 3D sensor and identifies the hook position and type through a deep learning algorithm; 3) Calculate the three-dimensional distance information of the hook through the two-dimensional hook position information, and then calculate the hook amplitude information; 4) The calculation controller sends the hook amplitude information and hook type to the torque limiter; 5) Storing weight information of different hook types in the torque limiter; 6) The torque limiter looks up the table according to the hook type sent by the operation controller and gives the current hook weight; 7) The torque limiter is automatically debugged and calibrated according to the collected hook weight and amplitude information.

7. The intelligent debugging method of torque limiter according to claim 6 is characterized in that: The algorithm in step 2) is a YOLO series algorithm.

8. The intelligent debugging method of torque limiter according to claim 5 is characterized in that: The on-load debugging includes the following steps: 1) The crane has lifted the weight to be debugged, and select load debugging on the torque limiter interface; 2) The computing controller processes the image signal of the 3D sensor and identifies the hook position and type through a deep learning algorithm; 3) Calculate the three-dimensional distance information of the hook through the two-dimensional hook position information, and then calculate the hook amplitude information; 4) The operation controller processes the image signal of the weight collected by the 3D sensor and identifies the weight of the weight through the digital OCR algorithm; 5) The operation controller sends the hook amplitude information and hook type to the torque limiter; 6) Storing weight information of different hook types in the torque limiter; 7) The torque limiter looks up the table according to the hook type sent by the operation controller and gives the current hook weight; 8) The torque limiter adds the weight of the weight sent by the operation controller and the weight of the hook obtained by looking up the table to obtain the load weight; 9) The torque limiter is automatically debugged and calibrated according to the collected load weight and amplitude information.

9. The intelligent debugging method of torque limiter according to claim 7, characterized in that: The YOLO series algorithm identifies the hook position and hook type: put all the multi-angle and multi-position pictures of the crane to be debugged into the YOLO data set, and replace the pictures of multiple categories in its pre-trained data set; then use the trained model data in the operation controller to identify the pictures containing the hook to effectively mark the type and position of the hook; read the corresponding three-dimensional information within the calibrated hook position range and find the closest point cloud cluster using the Euclidean clustering algorithm, and calculate the crane amplitude direction average value of the point cloud cluster as the sensor to the horizontal hook distance L1 detected by the 3D sensor, and then calculate the hook amplitude L=L1+L2 based on the measured distance L2 from the 3D sensor installation position to the rotation center (5).

10. The intelligent debugging method of torque limiter according to claim 9, characterized in that: The weight information of the weight is printed in a specific area of ​​the weight in a computer standard font. The operation controller collects the image information of the 3D sensor to mark the position of each weight, and then performs OCR text detection near the center of gravity of each weight. Each detected weight is assigned a weight value, and then all weight values ​​are added together to obtain the total weight of the weight.

Images