| CPC B25J 9/1664 (2013.01) [B23K 26/03 (2013.01); B23K 26/0884 (2013.01); B23K 26/34 (2013.01); B25J 9/1684 (2013.01); B25J 9/1694 (2013.01); C23C 24/10 (2013.01)] | 6 Claims |
|
1. A posture optimization and following control method for a robot used in freeform surface repair, wherein the method comprises the following steps:
Step 1, collecting a contour of a worn freeform surface, performing reverse modeling on the worn freeform surface to obtain a reverse-reconstructed worn surface, and obtaining a size, a shape and microscopic surface topography characteristics of the worn freeform surface;
Step 2, performing denoising and sparsification on the reverse-reconstructed worn surface, and performing statistical calculation to obtain a distribution condition of original normal vectors of the reverse-reconstructed worn surface;
Step 3, designing, based on the reverse-reconstructed worn surface, machining trajectories for repairing the worn freeform surface, and combining, according to the distribution condition of the original normal vectors, positions and initial normal vectors of discrete trajectory points of the machining trajectories, to obtain 6-dimensional machining position and normal vector trajectory data;
Step 4, establishing a trajectory plane of each of the machining trajectories, and obtaining, by introducing a normal vector rotation angle and using a polynomial fitting noise reduction method with a sliding window, optimized machining position and normal vector trajectory data, comprising:
establishing, based on the 6-dimensional machining position and normal vector trajectory data, the trajectory plane of each of the machining trajectories;
performing planar projection on the initial normal vectors of the discrete trajectory points of the machining trajectories;
calculating, by introducing the normal vector rotation angle, initial normal vector rotation angles at the discrete trajectory points;
performing fitting denoising on the initial normal vector rotation angles to remove jitter of the initial normal vectors caused by microscopic pits and protrusions on the worn freeform surface; and
reverse-solving optimized normal vector rotation angles after regression denoising, and obtaining the optimized machining position and normal vector trajectory data; Step 5, fitting, by introducing a 6-dimensional non-uniform rational b-splines (NURBS) curve, discrete position and posture data of the robot, to generate high-precision, smooth and synchronous motion position and posture trajectories of the robot;
Step 6, adaptively segmenting, based on curvatures of the motion position and posture trajectories in a position space and a posture space, the motion position and posture trajectories into hazardous segments and safe segments, and assigning processing velocities to the hazardous segments and the safe segments;
Step 7, performing, based on a position and posture synchronous look-ahead algorithm based on S-curve acceleration and deceleration, a backward deceleration iteration and a forward acceleration iteration to calculate node velocities of the hazardous segments and node velocities of the safe segments, to realize rapid and smooth transition of the processing velocities of the safe segments and the dangerous hazardous segments in the position space and the posture space; and
Step 8, performing, based on the node velocities of the hazardous segments and the node velocities of the safe segments, a laser repair process on the worn freeform surface, dynamically monitoring an actual movement velocity of the robot during the laser repair process, and adjusting laser cladding process parameters online to ensure thermodynamic stability of a laser cladding molten pool in the laser repair process and improve quality of laser cladding repair.
|