Abstract
Given the goal of more efficient parameter design in the drilling of small holes by an electrical discharge machine with an electrode diameter on the order of 0.1 mm, the strategy adopted in this study was to use Taguchi’s T-method(1) to predict the quality characteristic of the output from the input energy waveform. When members with matching rises in their discharge waveforms were taken, the voltage and current peaks both failed to align on the time axis and the predicted and measured amounts of material removed had a low correlation coefficient. When a correction was made to allow for the different discharge durations of the members and align their peak positions(maximum values)on the time axis, with voltage used as the energy characteristic, the correlation coefficient improved from 0.518 7 to 0.749 0. This was considered to indicate that the signal-to-noise ratio had improved because singular value effects had been eliminated by the correction. Nevertheless, adequate predictive accuracy was not obtained in the factor effect diagram of the removed volume, indicating that further work will be needed before predicted values can be used in practical parameter design.