Ultra-Precision Diamond Turning Error Compensation via Iterative Learning from On-machine Measured Data

In ultra-precision diamond turning, the reduction of machining form errors can generally be achieved through on-machine measurement and compensation. However, the efficiency of conventional compensation methods is often insufficient, particularly when high form accuracy is required or when intricate surface topography and microstructures need to be machined. Consequently, this research proposes a novel machining error compensation method based on iterative learning from on-machine measured data to enhance the machining accuracy and compensation efficiency. The on-machine measurement system and cutting path generation algorithm are introduced first. Then, the compensation method via iterative learning is presented theoretically, demonstrating a higher convergence order compared to the conventional method. Finally, machining experiments involving the cutting of cosine surfaces are conducted, followed by measurements of the processed workpieces. The experimental results indicate that after four rounds of compensation using the conventional method, the peak-to-valley (PV) value of the form error is reduced to 0.1134 mu m. In contrast, employing the proposed method, a similar value of 0.1156 mu m is achieved after only two rounds of compensation. This highlights the significant reduction in compensation time facilitated by the proposed method. Furthermore, the measurement results verify that the proposed compensation method maintains excellent surface quality.