Device and method for disorderly grabbing multiple layers of steel balls in ball milling workshop

Abstract

The invention provides a disordered grabbing device and method for multiple layers of steel balls in a ball milling workshop. The device comprises a steel ball storage pool (1), a mechanical arm (2), a visual module (3) and a control module. The mechanical arm, the visual module and the control module are electrically connected, and the visual module and the mechanical arm are arranged on the two sides of the steel ball storage pool respectively. The steel ball storage pool is used for storing multiple layers of disorderly arranged steel balls; the visual module is used for performing point cloud identification on the positions of the steel balls in the stock bin; the mechanical arm grabs the calibrated steel balls through an end effector electromagnet (12) and places the calibrated steel balls at designated points of the conveying belt. The control module is used for controlling the vision module to obtain the 3D coordinates of the multiple layers of stacked steel balls and controlling the pose, the tail end path and the grabbing rhythm of the mechanical arm.

Description

Technical Field

[0001] The invention belongs to the technical field of ore grinding, and in particular relates to a device and method for randomly grabbing multi-layer steel balls in a ball mill. Background Art

[0002] In the grinding process, the ball mill is a key equipment. Its important role is to ensure the particle size of the grinding product and to promote the ideal monomer dissociation state of the selected minerals. Usually, steel balls are selected as the main grinding medium for grinding operations inside the ball mill. During the operation cycle of the equipment, the steel balls will be constantly worn, causing dynamic changes in their total number and mass ratio. In order to maintain the filling rate and ideal ratio of the steel balls inside the ball mill, the steel balls must be replenished in real time to ensure the stability of the grinding efficiency and production capacity.

[0003] The traditional method of adding steel balls mainly relies on manual operation. The operator first loads the steel balls into a temporary ball adding bucket, and then uses a crane or other lifting equipment to transfer the steel balls to the ball adding port of the ball mill for adding. However, this method has the limitations of high labor intensity and poor ball adding accuracy. Summary of the invention

[0004] Aiming at the technical difficulty of fully automatic, effective and accurate grasping of steel balls in multi-layer disorderly stacked steel balls, the present invention provides a multi-layer disorderly grasping device and method platform for ball mill workshop, which includes a grasping module: steel ball storage tank, mechanical arm; visual module: industrial 3D intelligent visual module (industrial-grade 3D camera, visual guidance platform); control module: PLC control cabinet, industrial computer; steel ball transportation module: speed-adjusting belt. Among them, the steel ball storage tank is used to store multi-layer disorderly arranged steel balls; the industrial 3D intelligent visual module performs point cloud recognition and calibration on the position of the steel balls in the silo; the six-axis mechanical arm handling device grasps the calibrated steel balls through the end magnetic gripper and places them at the designated point of the transport belt; the PLC control cabinet and industrial computer realize the overall electrical control of the operation system.

[0005] The present invention also provides a master control program and terminal path planning algorithm for the automation platform. The master control program can adjust the steel ball grabbing beat, steel ball transport belt speed, and belt tension according to the working conditions fed back in real time by the ball mill load detection system to ensure the stability of the ball mill load. The self-developed intelligent terminal path planning algorithm reduces the vibration and impact of the robot arm through the trajectory interpolation method.

[0006] The present invention can adjust the frequency of adding balls in real time according to the operation of the ball mill, accurately identify and stably grasp the target points of multiple layers of disorderly stacked steel balls, and fundamentally solve the problem of inaccurate and discontinuous steel ball grasping process. It is conducive to realizing intelligent control of the steel ball grasping process, thereby ensuring the rationality of the steel ball ratio, and can effectively reduce the energy consumption of the ball mill operation and unnecessary steel ball loss.

[0007] The technical solution specifically adopted by the present invention to solve the technical problem is:

[0008] A device for grabbing multiple layers of disordered steel balls in a ball mill workshop comprises: a steel ball storage tank (1), a mechanical arm (2), a visual module (3) and a control module; the mechanical arm, the visual module and the control module are electrically connected, and the visual module and the mechanical arm are respectively arranged on both sides of the steel ball storage tank; the steel ball storage tank is used to store multiple layers of disordered steel balls; the visual module is used to perform point cloud recognition on the positions of the steel balls in the material bin; the mechanical arm grabs the calibration steel ball through the end effector electromagnet (12) and places it at a designated point on a conveyor belt; the control module is used to control the visual module to obtain the 3D coordinates of the multiple layers of stacked steel balls, and control the posture, end path and grabbing rhythm of the mechanical arm.

[0009] Furthermore, the steel ball storage pool is shaped like a concave "bowl" with a circular cross-section, and is used to store multiple layers of disorderly stacked Ф90-Ф120 steel balls, and each storage pool only stores steel balls of one size.

[0010] Furthermore, the pulse current generator (10) of the robotic arm is connected to the end effector electromagnet to output the magnetizing current intensity according to the distance between the end effector electromagnet and the steel ball to be taken, so as to ensure that only one steel ball is taken in one operation, and the grabbing cycle is ≤30s; the weighing sensor (11) of the robotic arm is connected to the control module for real-time signal transmission, and when it is detected that the end effector electromagnet holds the steel ball, the robotic arm performs the steel ball transporting action.

[0011] Furthermore, the control module is connected to the load detection module (7) of the ball mill and the online detection module (8) of the slurry fineness, and is used to obtain the working status information of the ball mill; the load detection module of the ball mill detects the noise signal, vibration signal, inlet and outlet pressure difference, and spindle power change during the operation of the ball mill, and forms a function mapping with the load of the ball mill to form a characteristic parameter fusion model; the online detection module of the slurry fineness is a stirring detection device, which converts the shear force of the stirring blade on the steady flow of the slurry per unit time into real-time slurry concentration and fineness.

[0012] Furthermore, the 3D coordinates of the steel balls within the identifiable range in the steel ball storage pool are first obtained through the visual module, and then the grabbing order is determined according to the mutual occlusion relationship of the steel ball point cloud and the position of the robotic arm to avoid the overlapping steel balls from being selected for grabbing; the robotic arm controls the end to move to the designated point for grabbing according to the acquired grabbing steel ball coordinates, and determines whether the grabbing is successful according to the data sent back by the weighing sensor, otherwise repeats the above steps; after successful grabbing, the steel ball is placed at the designated position of the conveyor belt.

[0013] Furthermore, the method for deciding the grasping order according to the mutual occlusion relationship of the steel ball point cloud and the position of the robotic arm is to scan the point cloud data of the steel ball directly above the ball pool and score the exposed surface area of ​​the steel ball. The larger the exposed area, the higher the score and the higher the priority to be grasped.

[0014] Furthermore, the grabbing frequency and grabbing sequence of the steel balls are dynamically adjusted according to the actual working conditions of the ball mill: when the load of the ball mill exceeds the threshold, the grabbing of the balls is reduced or stopped; when the load of the ball mill does not reach the threshold, the grabbing frequency is increased.

[0015] Furthermore, during the transport process from when the steel ball is grabbed to when the steel ball is released, the terminal path planning algorithm is executed by the robot arm. Based on the quintic polynomial interpolation function, a circular arc trajectory interpolation algorithm is adopted to make the velocity function and acceleration function of each joint of the robot arm change continuously, so as to reduce the vibration and impact of the movement process while ensuring the operation accuracy.

[0016] Furthermore, the quintic polynomial interpolation function is a function of the joint angle θ with respect to the movement time t;

[0017] The arc trajectory interpolation algorithm obtains a two-dimensional plane through three points in space, establishes a two-dimensional coordinate system XY, transforms the three-dimensional space points into two-dimensional plane points, calculates the center coordinates and arc angle of the arc surrounded by the three points, and finally transforms the two-dimensional coordinates (x, y) into three-dimensional coordinates (u, v, w) through a transformation matrix.

[0018] Furthermore, the quintic polynomial interpolation function is a transformation function of the joint angle θ over time t:

[0019] θ(t)=a0+a1t+a2t 2 +a3t 3 +a4t 4 +a5t 5 +a6t 6

[0020] The angular velocity is obtained by taking the first derivative:

[0021] θ'(t)=a1+2a2t+3a3t 2 +4a4t 3 +5a5t 4 +6a6t 5

[0022] The angular acceleration is obtained by taking the second derivative:

[0023] θ”(t)=2a2+6a3t+12a4t 2 +20a5t 3 +30a6t 4

[0024] The coordinate transformation in the arc trajectory interpolation algorithm satisfies the following relationship:

[0025] AT=B

[0026] Among them, A is the three 3D space coordinates of circular interpolation; B is the transformed 2D coordinates; T is the end effector space transformation matrix, which satisfies the following relationship:

[0027] T=T1·T2·T3·T4·T5·T6

[0028] Among them, T1, T2, T3, T4, T5, and T6 are the posture transformation matrices of the six joints respectively;

[0029] The objective function of the iterative optimization algorithm is:

[0030]

[0031] I i is the moment of inertia of the i-th joint; ω i is the angular velocity of the i-th joint; Error(x,y,z) is a function of the terminal coordinates, which represents the Euclidean distance positioning error of the terminal; α and β are the weight coefficients of the energy consumption term and the positioning error term, respectively, which are solved by iteration.

[0032] Compared with the prior art, the innovative features of the present invention and its preferred solution include at least:

[0033] 1. It can adjust the frequency and sequence of adding steel balls in real time according to the operating status of the ball mill, which fundamentally solves the problems of low accuracy and poor continuity in traditional manual or mechanical ball adding methods, improves the intelligence level of steel ball adding operations, and effectively ensures the rationality of the steel ball filling rate and ratio in the ball mill, thereby reducing energy consumption and steel ball loss, which is beneficial to improving grinding efficiency and overall grinding quality.

[0034] 2. The problem of accurate identification and grasping of multi-layer stacked steel ball targets is solved by reconstructing the steel ball storage pool and the end-effector electromagnet.

[0035] 3. Through the coordinated control of the PLC control cabinet and the industrial computer, accurate and automated management of the entire process is achieved, greatly improving the intelligence and stability of the ball mill operation.

[0036] 4. The end robot arm path planning algorithm provided can reduce the vibration and impact during the movement process while ensuring the operation accuracy, thereby extending the service life of the robot. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0038] Figure 1 This is a schematic diagram of the overall framework of the device system of an embodiment of the present invention;

[0039] Figure 2 It is a schematic diagram of the arrangement of the device according to the embodiment of the present invention;

[0040] Figure 3 This is a diagram of a handling device using a six-axis industrial robot arm as an example in an embodiment of the present invention;

[0041] Figure 4 This is a schematic diagram of the interaction between the lifting electromagnet and the steel ball of the packaged end effector according to the embodiment of the present invention;

[0042] Figure 5 This is a structural diagram of the end effector and the retractable part of an embodiment of the present invention;

[0043] Figure 6 This is a flowchart of the crawling operation according to an embodiment of the present invention;

[0044] Figure 7 A comparison diagram of the displacement S-time t function of the fifth-order polynomial path planning before and after optimization in an embodiment of the present invention;

[0045] Figure 8 The three-point position and motion trajectory diagram of the terminal space after arc trajectory interpolation during the handling process of the embodiment of the present invention;

[0046] Fig. 9 It is a schematic diagram of the overall scheme of an embodiment of the present invention.

[0047] Numbers in the figure: 1- steel ball storage tank, 2- robotic arm, 3- visual module, 4- PLC control cabinet, 5- industrial computer, 6- speed belt, 7- load detection module, 8- slurry fineness online detection module, 9- robotic arm body, 10- pulse current generator, 11- weighing sensor, 12- end effector electromagnet, 13- telescopic rod, 14- spring. DETAILED DESCRIPTION

[0048] In order to make the features and advantages of the present invention more clearly understood, the following embodiments are specifically described in detail as follows:

[0049] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in this specification have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0050] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0051] like Figure 1-Figure 9 As shown, an embodiment of the present invention provides a multi-layer steel ball disorderly grasping device for a ball mill workshop, which is applied to the field of grinding technology and comprises: a grasping module, a visual module (3), a control module, and a transportation module (6); and is linked with a load detection module (7) of a ball mill and an online detection module (8) for slurry fineness.

[0052] The grabbing module includes a steel ball storage tank (1) and a mechanical arm (2); the visual module preferably adopts an industrial 3D intelligent visual module (3), including an industrial-grade 3D camera and a visual guidance platform; the control module includes a PLC control cabinet (4) and an industrial computer (5); the steel ball transport module adopts a speed-adjusting belt (6), which is independent of the ore transport belt; and the ball mill load detection module (7) includes a sound sensor, a vibration acceleration sensor, a pressure sensor, and an active power sensor. As a preferred design, the ball mill load detection module forms a function mapping with the ball mill load by detecting the noise signal, vibration signal, inlet and outlet pressure difference, and spindle power change during the operation of the ball mill, and forms a characteristic parameter fusion model. Multi-source signal fusion detection is realized to minimize the mill load detection error. The slurry fineness online detection module is a stirring detection device. The real-time slurry concentration and fineness are obtained by measuring the shear force of the stirring blade on the steady flow of the slurry per unit time. The grabbing module, the visual module, and the main control module are electrically connected, and the visual module and the mechanical arm are arranged on both sides of the steel ball storage tank respectively. The control module is used to control the visual module to obtain the 3D coordinates and visual images of the multi-layer stacked steel balls, control the robot arm posture, end path, grasping rhythm, and speed and tension of the speed-adjusting belt. The robot arm is a six-axis or higher robot arm, supporting hand-eye calibration with the eye outside the hand.

[0053] As a preferred solution of this embodiment, the steel ball storage tank (1) is semi-underground, shaped like a concave "bowl", with a circular cross section, a diameter d of 2 to 4 m, and an angle θ of 15° to 45°. It can be used to store multiple layers of disorderly stacked Φ90 to Φ120 steel balls, and each storage tank can only store steel balls of one size. When the steel balls in the middle are taken out, the ball pile collapses to the bottom due to fluidity, so that the steel balls at the bottom are more evenly distributed.

[0054] As a preferred solution of this embodiment, the robot arm (2) adopts an industrial robot arm with more than six joints. Taking a six-axis robot arm as an example, it includes a robot arm body (9); a pulse current generator (10), a weighing sensor (11), and an end effector (12), which are connected to and controlled by a control system and are used to grab and place steel balls; the arm span can cover the diameter d range of the steel ball storage tank.

[0055] Preferably, the shell material of the electromagnet end effector is austenitic stainless steel, the iron core material inside the suction cup is ferrite, and the coil material is copper wire. Epoxy resin is poured in the gap between the coil, the iron core and the shell. For the end effector electromagnet (12) of the robotic arm, it weighs about 5kg. The shape is a cylinder with a spherical groove at the bottom. The ratio of the radius of the bottom spherical groove to the radius of the cylindrical bottom surface r_c:r=1:1~1:1.2; the ratio of the height of the cylinder to the radius r:h=7~9; the z-direction adjustment distance of the groove z_c=0.8~1.1r_c. The top of the end electromagnet can be welded with a telescopic rod (13) of a certain length according to actual needs. The telescopic rod is connected to the joint of the robotic arm through a spring (14), and the length can be adjusted by the spring to ensure that it can accurately grab the steel ball at the bottom of the material pool according to actual conditions.

[0056] For the pulse current generator (10) of the robot arm, the robot arm (2) is precisely controlled according to the position information provided by the visual module, grabs the steel ball from the storage bin and puts it to the designated point. The pulse current module can control the intensity of the magnetizing current according to the distance between the end effector electromagnet (12) and the steel ball to be taken, ensuring that only one steel ball is taken in one operation, and the grabbing cycle is ≤30s.

[0057] The weighing sensor (11) of the mechanical arm transmits real-time signals to the industrial computer. When it is detected that the steel ball is held, the mechanical arm moves the steel ball;

[0058] When the sensor detects that the steel ball is not held:

[0059] If there are no steel balls in the steel ball storage tank (1), an alarm is activated.

[0060] If there is a steel ball in the steel ball storage tank (1), the 3D vision module will give the position coordinates of the steel ball to be grasped again.

[0061] As a preferred solution of this embodiment, the installation height of the industrial 3D intelligent vision module (3) is located 1200 to 3000 mm above the silo. Including an industrial-grade 3D camera and a visual guidance platform, the industrial 3D intelligent vision module (3) captures and analyzes the position information of the steel balls in the steel ball storage tank in real time, generates point cloud data, scores according to the size of the scanned area, and intelligently sorts and accurately calibrates the order of grabbing the steel balls. The visual area can cover the surface of the storage silo, with a working distance of 1200 to 3000 mm; a resolution of 30 to 50 million; and a single-point repeatability accuracy of the Z axis: 0.32@2.4mThe steel ball positioning error is ±2mm. The device of this embodiment can dynamically adjust the grabbing frequency and grabbing sequence of the steel balls at different stages of the ball mill operation according to the actual working conditions of the ball mill. When the load of the ball mill exceeds the threshold, the grabbing of the balls is reduced or stopped; when the load of the ball mill does not reach the threshold, the grabbing frequency is increased.

[0062] For the visual guidance platform, the camera viewer, graphical machine vision software, drag-and-drop algorithm execution module, hand-eye calibration software, point cloud editor, and flowchart visual development module are integrated. The visual guidance platform is compatible with multiple brands of cameras and mainstream brand robots; it supports fast hand-eye calibration. The camera viewer supports camera parameter adjustment, texture images, depth images, and point cloud images in real time; the graphical machine vision software and drag-and-drop algorithm execution module algorithms support secondary development (no advanced development language is required).

[0063] As a preferred solution of this embodiment, the control system works in coordination with the PLC control cabinet (4) through the industrial computer (5), and plans the grabbing sequence, that is, the machine vision scans the point cloud data of the steel balls directly above the ball pool. The exposed surface area of ​​the steel balls is scored, and the larger the exposed area, the higher the score, and the priority is to be grabbed. Avoid the overlapped steel balls from being grabbed first. Real-time monitoring of the ball mill operation is achieved, and the steel ball placement strategy is adjusted accordingly to ensure the rationality of the steel ball ratio and the ball mill operation efficiency, thereby effectively reducing the ball mill operation energy consumption and abnormal steel ball loss.

[0064] As a preferred solution of this embodiment, the ball mill load detection module (7) is used to detect data such as the active power of the ball mill, the medium filling rate, the steel ball ratio, etc., and can provide real-time feedback of the ball mill spindle power and the steel ball filling rate to the industrial computer to prevent hunger or bloating.

[0065] The slurry fineness online detection module (8) is connected to the main control module and can provide the slurry fineness in real time.

[0066] Based on the above device design of the embodiment of the present invention, the following provides a working control method under normal working conditions:

[0067] Step 1: Get the 3D coordinates of the steel ball;

[0068] Step 2: Decide the crawling order;

[0069] Step 3: According to the given grabbing steel ball coordinates, the end point is controlled to move to the designated point for grabbing;

[0070] Step 4: Determine whether the grab is successful based on the data sent back by the weighing sensor. If unsuccessful, repeat the above process;

[0071] Step 5: Place the steel ball at the designated position of the belt;

[0072] Step 6: Adjust the belt tension and transport speed.

[0073] like Figure 6 As shown, its working logic is as follows:

[0074] Step 1: Start the system, and the 3D camera will take pictures of the steel ball above the storage tank to generate point cloud data.

[0075] Step 2: Check whether there is a steel ball above the material pool. If so, the system will give the three-dimensional geometric coordinates of the steel ball to be taken; if not, it will pause and give a prompt to add the ball.

[0076] Step 3: The robot arm adjusts its position so that the end moves to the top of the steel ball to be picked up, and then powers on to suck the ball.

[0077] Step 4: The weighing sensor detects whether the steel ball is sucked up. If not, the system gives the three-dimensional geometric coordinates of the next steel ball to be picked up and repeats step 3; if yes, the steel ball is moved to the designated placement point, unloaded, and the control system automatically counts.

[0078] The cycle steps are as above.

[0079] As a preferred solution of this embodiment, Figure 7 , Figure 8 As shown in the figure, based on the quintic polynomial interpolation function, the circular arc trajectory interpolation algorithm is adopted to make the velocity function and acceleration function of each joint of the robot arm change continuously, thereby reducing the vibration and impact of the movement process while ensuring the operation accuracy, and extending the service life of the robot.

[0080] The terminal path planning algorithm is aimed at the transportation process from "steel ball being grabbed" to "steel ball being released".

[0081] The quintic polynomial interpolation function is the transformation function of the joint angle θ over time t,

[0082] θ(t)=a0+a1t+a2t 2 +a3t 3 +a4t 4 +a5t 5 +a6t 6

[0083] Find the first derivative of angular velocity:

[0084] θ'(t)=a1+2a2t+3a3t 2 +4a4t 3 +5a5t 4 +6a6t 5

[0085] Find the second derivative and get the angular acceleration:

[0086] θ”(t)=2a2+6a3t+12a4t 2 +20a5t 3 +30a6t 4

[0087] The arc trajectory interpolation algorithm obtains a two-dimensional plane through three points in space, establishes a two-dimensional coordinate system XY, transforms the three-dimensional space points into two-dimensional plane points, calculates the center coordinates and arc angle of the arc surrounded by the three points, and finally transforms the two-dimensional coordinates (x, y) into three-dimensional coordinates (u, v, w) through a transformation matrix.

[0088] The coordinate transformation in the arc trajectory interpolation algorithm satisfies the following relationship:

[0089] AT=B

[0090] Among them, A is the three 3D space coordinates of circular interpolation; B is the transformed 2D coordinates; T is the end effector space transformation matrix, which satisfies the following relationship:

[0091] T=T1·T2·T3·T4·T5·T6

[0092] Among them, T1, T2, T3, T4, T5, and T6 are the pose transformation matrices of the six joints respectively.

[0093] The objective function of the iterative optimization algorithm is:

[0094]

[0095] I i is the moment of inertia of the i-th joint; ω i is the angular velocity of the i-th joint; Error(x,y,z) is a function of the terminal coordinates, which represents the Euclidean distance positioning error of the terminal; α and β are the weight coefficients of the energy consumption term and the positioning error term, respectively, which are solved by iteration.

[0096] The motion process is optimized through algorithm iteration to achieve the lowest energy consumption of each joint and the highest end-operation accuracy.

[0097] Based on the device size provided in this embodiment, the total length of the interpolation path through the above path planning is 150mm-350mm, the starting speed and the ending speed are 0, and the acceleration is 15-20mm / s 2 , deceleration is 7~12mm / s 2 .

[0098] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0099] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.

[0100] The present invention is not limited to the above-mentioned optimal implementation mode. Anyone can derive other various forms of a device and method for disorderly grabbing multi-layer steel balls in a ball mill workshop under the inspiration of the present invention. All equal changes and modifications made within the scope of the invention application of the present invention should be covered by the scope of the present invention.

Claims

1. A device for grabbing disordered multi-layer steel balls in a ball mill, characterized in that: include: A steel ball storage tank (1), a mechanical arm (2), a visual module (3) and a control module; the mechanical arm, the visual module and the control module are electrically connected, and the visual module and the mechanical arm are respectively arranged on both sides of the steel ball storage tank; the steel ball storage tank is used to store multiple layers of disorderly arranged steel balls; the visual module is used to perform point cloud recognition on the positions of the steel balls in the silo; the mechanical arm grabs the calibrated steel ball through the end effector electromagnet (12) and places it at a designated point on the conveyor belt; the control module is used to control the visual module to obtain the 3D coordinates of the multiple layers of stacked steel balls, and control the posture, end path and grabbing rhythm of the mechanical arm.

2. The device for grabbing multi-layer steel balls in a ball mill according to claim 1, characterized in that: The steel ball storage pool is in the shape of a concave "bowl" with a circular cross section, and is used to store multiple layers of disorderly stacked Φ90-Φ120 steel balls, and each storage pool only stores steel balls of one size.

3. The device for grabbing multi-layer steel balls in a ball mill according to claim 1, characterized in that: The pulse current generator (10) of the mechanical arm is connected to the end effector electromagnet to output the magnetizing current intensity according to the distance between the end effector electromagnet and the steel ball to be taken, so as to ensure that only one steel ball is taken in one operation, and the grabbing cycle is ≤30s; the weighing sensor (11) of the mechanical arm is connected to the control module for real-time signal transmission, and when it is detected that the end effector electromagnet holds the steel ball, the mechanical arm performs the steel ball transporting action.

4. The device for grabbing multi-layer steel balls in a ball mill according to claim 1, characterized in that: The control module is connected to the load detection module (7) of the ball mill and the online detection module (8) of the slurry fineness, and is used to obtain the working state information of the ball mill; the load detection module of the ball mill detects the noise signal, vibration signal, inlet and outlet pressure difference, and spindle power change during the operation of the ball mill, and forms a function mapping with the load of the ball mill to form a characteristic parameter fusion model; the online detection module of the slurry fineness is a stirring detection device, which converts the shear force of the stirring blade on the steady flow of the slurry per unit time to obtain the real-time slurry concentration and fineness.

5. The working method of the device for grabbing disorderly multi-layer steel balls in a ball mill workshop according to claim 4 is characterized in that: First, the 3D coordinates of the steel balls within the identifiable range in the steel ball storage pool are obtained through the visual module, and then the grasping order is determined according to the mutual occlusion relationship of the steel ball point cloud and the position of the robotic arm to avoid the overlapping steel balls from being selected for grasping; the robotic arm controls the end to move to the designated point for grasping according to the acquired grasped steel ball coordinates, and determines whether the grasping is successful based on the data sent back by the weighing sensor, otherwise repeat the above steps; after successful grasping, the steel ball is placed at the designated position of the conveyor belt.

6. The working method of the device for grabbing disorderly multi-layer steel balls in a ball mill workshop according to claim 5 is characterized in that: The method for deciding the grasping order according to the mutual occlusion relationship of the steel ball point cloud and the position of the robotic arm is to scan the steel ball point cloud data directly above the ball pool and score the exposed surface area of ​​the steel ball. The larger the exposed area, the higher the score and the higher the priority to be grasped.

7. The working method of the device for grabbing disorderly multi-layer steel balls in a ball mill workshop according to claim 5 is characterized in that: According to the actual working conditions of the ball mill, the grabbing frequency and grabbing order of the steel balls are dynamically adjusted: when the ball mill load exceeds the threshold, the grabbing of the balls is reduced or stopped; when the ball mill load does not reach the threshold, the grabbing frequency is increased.

8. The working method of the device for grabbing disorderly multi-layer steel balls in a ball mill workshop according to claim 5 is characterized in that: The terminal path planning algorithm is executed in the handling process from when the steel ball is grabbed to when the steel ball is released by the robot arm. Based on the quintic polynomial interpolation function, a circular arc trajectory interpolation algorithm is adopted to make the speed function and acceleration function of each joint of the robot arm change continuously, so as to reduce the vibration and impact of the movement process while ensuring the operation accuracy.

9. The working method of the device for grabbing disorderly multi-layer steel balls in a ball mill workshop according to claim 8 is characterized in that: The quintic polynomial interpolation function is a function of the joint angle θ with respect to the motion time t; The arc trajectory interpolation algorithm obtains a two-dimensional plane through three points in space, establishes a two-dimensional coordinate system XY, transforms the three-dimensional space points into two-dimensional plane points, calculates the center coordinates and arc angle of the arc surrounded by the three points, and finally transforms the two-dimensional coordinates (x, y) into three-dimensional coordinates (u, v, w) through a transformation matrix.

10. The working method of the device for grabbing disorderly multi-layer steel balls in a ball mill workshop according to claim 9 is characterized in that: The fifth-order polynomial interpolation function is a transformation function of the joint angle θ over time t: θ(t)=a0+a1t+a2t 2 +a3t 3 +a4t 4 +a5t 5 +a6t 6 The angular velocity is obtained by taking the first derivative: <h2 style=";text-align:left;direction:ltr">θ'(t) = a1+2a2t+3a3t<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +4a4t<h2 style=";text-align:left;direction:ltr"> 3 <h2 style=";text-align:left;direction:ltr"> +5a5t<h2 style=";text-align:left;direction:ltr"> 4 <h2 style=";text-align:left;direction:ltr"> +6a6t<h2 style=";text-align:left;direction:ltr"> 5 The angular acceleration is obtained by taking the second derivative: θ″ (t) =2a2+6a3t+12a4t 2 +20a5t 3 +30a6t 4 The coordinate transformation in the arc trajectory interpolation algorithm satisfies the following relationship: AT=B Among them, A is the three 3D space coordinates of circular interpolation; B is the transformed 2D coordinates; T is the end effector space transformation matrix, which satisfies the following relationship: T=T1·T2·T3·T4·T5·T6 Among them, T1, T2, T3, T4, T5, and T6 are the posture transformation matrices of the six joints respectively; The objective function of the iterative optimization algorithm is: I i is the moment of inertia of the i-th joint; ω i is the angular velocity of the i-th joint; Error(x,y,z) is a function of the terminal coordinates, which represents the Euclidean distance positioning error of the terminal; α and β are the weight coefficients of the energy consumption term and the positioning error term, respectively, which are solved by iteration.

Images