Robotic grinding of complex surfaces with an internal structured compliant tool: Multi-performance optimization in confined spaces

Robotic Compliant grinding of complex surfaces with limited space is a crucial and challenging process for the production of the high-performance components such as blisk. The trajectory-varying tool attitude, resulting from the complex structure, introduces an inherent complexity to the contact state for material removal. This, in turn, influences grinding performance and tool life, as well as robot kinematic performance. Thus, a quantitative model between the processing parameters and convoluted material removal process was established. A multi-indicator optimization (MIO) model considering robot kinematics, grit trajectory continuity and tool longevity was developed and solved by the adaptively controlled differential evolutionary (ACDE) algorithm. The results demonstrate that the root mean square error (RMSE) of the predicted contour is 6.45–9.48 μm with different dwell times featuring maximum depths from 47.62 to 158.31 μm. Meanwhile, the RMSE for the three-dimensional morphology with varying tool attitude was less than 12.75 μm. Furthermore, the tool attitudes employing MIO enabled the smoothing grinding of an blisk part featuring a channel as narrow as 23.41 cm, which avoids robot joint mutations and reduces the jerk variation from 0.81 to 0.04 rad/s³. The uniform material removal achieved over the entire annular root surface exhibited a mean error of 0.013–0.016 mm for a pre-set removal depth of 0.3 mm.