The machining quality is influenced significantly by the gap discharge states in electrical discharge machining (EDM). In order to avoid abnormal discharges such as arc

the discharge states needed to be identified by the discharge state detection system according to different discharge characteristics. In order to solve this problem

the single pulse discharge method was used to explore the different characteristics of spark discharge and arc discharge. Under different experimental conditions

the two kinds of discharge voltage waveforms were respectively transformed by fast fourier transformation (FFT). It is found that the frequency amplitude of spark discharge is significantly different from that of arc discharge in the lower frequency range. Therefore

the average frequency amplitude of 0–5MHz frequency range was taken as the research object to study the influence of different variables on it. The experimental results show that the open-circuit voltage and pulse duration have little effect on the average frequency amplitude of the two discharge states

while the electrode materials and workpiece materials have great influence on them. The average frequency amplitude of the spark discharge state is higher than that of the arc discharge

–30dB can be used as the threshold to distinguish the spark discharge and the arc discharge.