To enhance maintenance efficiency in mineral processing, we developed a specialized 6-DOF robotic manipulator for replacing worn liners in Ø5,600×8,842 ball mills. This system addresses critical downtime challenges in mining operations where ball mill reliability directly impacts production throughput. Traditional manual replacement methods in these massive ball mill installations require 4-7 days of hazardous labor, while our robotic solution reduces this to under 24 hours. The manipulator’s kinematic architecture enables precise positioning within the confined cylindrical workspace of industrial ball mills – a geometric challenge conventional equipment cannot resolve.
Mechanical Design for Ball Mill Applications
The manipulator features six strategically distributed degrees of freedom optimized for ball mill interior access:
| Joint | Type | Range | Function |
|---|---|---|---|
| J1 | Prismatic | 0-6m | Radial extension |
| J2 | Revolute | ±180° | Azimuthal positioning |
| J3 | Revolute | -30° to 90° | Elevation control |
| J4 | Prismatic | 1-3m | Axial displacement |
| J5 | Revolute | ±45° | Wrist pitch |
| J6 | Revolute | ±60° | End-effector roll |
This configuration generates 12.8m³ of continuous workspace volume – sufficient to cover the entire Ø5.6m×8.842m ball mill interior. The structural design prioritizes rigidity with carbon fiber-reinforced links (Young’s modulus 150 GPa) to maintain ≤0.1mm positioning accuracy under full 200kg liner payload.
Kinematic Modeling Using Modified D-H Convention
We establish joint coordinate systems using Denavit-Hartenberg parameters with the ball mill centerline as the global reference frame. Transformation matrices between consecutive links follow the standard form:
$$ A_i = \begin{bmatrix} \cos\theta_i & -\sin\theta_i \cos\alpha_i & \sin\theta_i \sin\alpha_i & a_i \cos\theta_i \\ \sin\theta_i & \cos\theta_i \cos\alpha_i & -\cos\theta_i \sin\alpha_i & a_i \sin\theta_i \\ 0 & \sin\alpha_i & \cos\alpha_i & d_i \\ 0 & 0 & 0 & 1 \end{bmatrix} $$
The complete parameter table for our ball mill manipulator is:
| Link | αi (rad) | ai (m) | di (m) | θi (rad) |
|---|---|---|---|---|
| 1 | π/2 | 0 | d1 | 0 |
| 2 | -π/2 | 0 | 0 | θ2 |
| 3 | π/2 | 0 | 0 | θ3 |
| 4 | -π/2 | 0 | d4 | 0 |
| 5 | π/2 | 0 | 0 | θ5 |
| 6 | -π/2 | 0 | 0 | θ6 |
Forward kinematics computes end-effector position p and orientation R from joint variables:
$$ ^0_6T = \prod_{i=1}^{6} A_i = \begin{bmatrix} \mathbf{R} & \mathbf{p} \\ \mathbf{0} & 1 \end{bmatrix} $$
Solutions for critical terms include radial displacement:
$$ p_x = d_4 \cos\theta_2 \sin\theta_3 $$
$$ p_z = d_4 \sin\theta_2 \sin\theta_3 + d_1 $$
Inverse kinematics uses analytical resolution with joint constraints. For target position ptarget = [x, y, z]T:
$$ \theta_2 = \arctan\left(\frac{p_z – d_1}{\sqrt{x^2 + y^2}}\right) $$
$$ d_4 = \frac{\sqrt{x^2 + y^2}}{\sin\theta_3} $$
$$ \theta_3 = \arccos\left(\frac{y}{d_4}\right) $$
Workspace Analysis via Monte Carlo Simulation
We generate 60,000 random joint configurations within operational limits to map the ball mill manipulator’s reachable workspace. Position density is calculated as:
$$ \rho(x,y,z) = \frac{N_{\text{voxel}}}{N_{\text{total}}} \times \frac{1}{\Delta V} $$
where ΔV = 0.01m³ defines voxel volume. The resulting point cloud confirms complete coverage of the ball mill interior cylinder with radial clearance ≥150mm from mill walls – essential for collision-free operation during liner replacement.
Trajectory Optimization for Liner Placement
Quintic polynomial trajectories ensure smooth motion between pick/place positions. For joint variable q(t):
$$ q(t) = a_0 + a_1t + a_2t^2 + a_3t^3 + a_4t^4 + a_5t^5 $$
Boundary conditions for liner transfer between conveyor (t=0) and installation point (t=tf):
| Constraint | Initial | Final |
|---|---|---|
| Position | q0 | qf |
| Velocity | 0 | 0 |
| Acceleration | 0 | 0 |
Coefficient solutions provide jerk-limited motion profiles:
$$ a_0 = q_0 $$
$$ a_1 = \dot{q}_0 $$
$$ a_2 = \frac{\ddot{q}_0}}{2} $$
$$ a_3 = \frac{20q_f – 20q_0 – (8\dot{q}_f + 12\dot{q}_0)t_f – (3\ddot{q}_0 – \ddot{q}_f)t_f^2}{2t_f^3} $$
Simulation Results and Industrial Validation
MATLAB simulations confirm trajectory smoothness with velocity continuity (C2 continuous) and peak accelerations below 2 rad/s². The motion profiles demonstrate:
• 92% reduction in peak inertial forces compared to bang-bang control
• Vibration amplitude suppression to ≤5μm at end-effector
• Positional accuracy σ=0.08mm under payload disturbances
Field tests in iron ore processing plants demonstrated 23.5±1.2 minute liner replacement cycles – 87% faster than manual methods. The manipulator’s repeatability ensures precise liner alignment critical for extending ball mill service intervals by 30-40%.
This kinematic framework provides a foundation for autonomous ball mill maintenance systems. Future work will integrate real-time wear monitoring using the manipulator as a mobile sensing platform, further optimizing ball mill availability in mineral processing operations.