微细电火花加工间隙放电状态智能检测方法的研究

The Study on the Discrimination of Electrical Discharging States in Micro-EDM

针对微细电火花加工中的高频、微能、单一信号严重畸变等特点,提出了利用电压、电流信号的互补性特点,运用模糊逻辑控制规则,得出采样点的状态值,然后采用LVQ神经网络的分类功能,将采样点的状态值转化成加工间隙状态中所属的状态矢量,最后通过统计模型,得出电火花加工极间放电状态。

High frequency and small electric power of micro-EDM cause the waveforms of voltage and current highly distorted, thus indistinguishable by the commonly used EDM discriminating methods. Upon such knowledge of this, a new method is presented in this paper, where the fuzzy logic rules are used to combine the complementary signals from voltage and current taken as the two inputs to the fuzzy system and deduce a value in a range representing the discharging state of the sampled point. Learning vector quantification （LVQ） neural network architecture is adopted to convert this value of the sampled point deduced from the fuzzy system to the corresponding discharging state vector. After the statistical generalization of the points in a set of pulses, the ratio between the elements in the vector clarifies the discharging state of the gap between electrode and workpiece.