Adaptive Control for Fast EDM Drilling of High Aspect-Ratio Holes

Abstract

Fast EDM (electrical discharge machining) drilling is a powerful machining technique dedicated for small hole machining (0.2mm~3mm in diameter). It employs rotary tubular electrode and water-based dielectric fluid with high-pressure inner flushing. Owing to the enhanced inner flushing, evacuation of debris from the discharge gap becomes much easier. Therefore it can achieve higher machining efficiency than conventional sinking EDM and is capable of drilling high aspect-ratio (100:1 and higher) holes. It has now been widely used to drill small holes in manufacturing industries such as molds & dies, filters, automotive and aerospace parts. Fast EDM drilling process is a typical time-variant process, especially when drilling high aspect-ratio holes. The machining condition changes with time as the electrode goes deeper, since the evacuation of debris becomes increasingly difficult and the machining condition deteriorates. The conventional control system of the process determines the feedrate by average gap voltage with its parameters unchanged during the process. Such control systems become less efficient when drilling high aspect-ratio holes. In this paper, an adaptive control system for fast EDM drilling is proposed to deal with this complex process. The controlled plant is considered as a double-input double-output (DIDO) system. The inputs are the servo-reference voltage and the gain of the original controller, and the outputs are the ratio of normal discharge state (NDR) and its variance in a time interval. The servo-reference voltage is a key parameter for the control of the process, while the gain determines the actual feedrate along with the servo-reference voltage. The adaptive control system first estimates the parameters of the controlled plant model, and then automatically adjusts the two inputs according to current outputs and desired ones. The time-variant parameters of the controlled plant are estimated by the recursive least squares (RLS) method with a forgetting factor. And the multivariable generalized minimum variance method is used to obtain the control law of the two inputs. Experiments were conducted to verify the effectiveness of the controller, where 40-mm-deep holes were drilled by electrodes with a diameter of 0.4 mm (aspect-ratio 100:1). The results show that the machining state was improved by the adaptive controller especially in the later stage of the drilling process. NDR was higher and short circuits were suppressed. Results show that machining time was reduced by 11.3% and tool wear was also reduced by 3.0% with the adaptive controller.