Machine learning induction of a model for online parameter selection in EDM rough machining

In electrical discharge machining (EDM), appropriate average current in the gap has to be selected for the given machining surface in order to obtain the highest material removal rate at low electrode wear. Thus, rough machining parameters have to be selected according to the machining surface. In the case of sculptured features, the machining surface varies with the depth of machining. Hence, the machining parameters have to be selected on-line to obtain appropriate current density in the gap. In this paper, inductive machine learning is used to derive a model based on the voltage and current in the gap. The sufficient inputs to the model are only two discharge attributes extracted from the voltage signal in the gap. The model successfully selects between two machining parameter settings that obtain different average surface current in the gap. It requires only voltage signal acquisition during the machining process and a simple algorithm that is easy to implement on industrial machines.

[1]  Jean-Pierre Kruth,et al.  A study of pocketing EDM , 1992 .

[2]  Duc Truong Pham,et al.  Micro-EDM—recent developments and research issues , 2004 .

[3]  Bogdan Filipič,et al.  Using inductive machine learning to support decision making in machining processes , 2000 .

[4]  Sanjay S. Jagdale,et al.  Decision support for manufacturing using artificial neural-networks , 1995 .

[5]  Mukund R. Patel,et al.  Theoretical models of the electrical discharge machining process. II. The anode erosion model , 1989 .

[6]  Ivan Bratko,et al.  Identifying the grinding process by means of inductive machine learning , 1991 .

[7]  Kamlakar P Rajurkar,et al.  WEDM Identification and Adaptive Control for Variable-Height Components , 1994 .

[8]  Mihael Junkar,et al.  Machining parameters selection for varying surface in EDM , 2007 .

[9]  J. Peklenik,et al.  Statistical Model for an Adaptive Control of EDM-Process , 1985 .

[10]  Mukund R. Patel,et al.  Theoretical models of the electrical discharge machining process. I. A simple cathode erosion model , 1989 .

[11]  Manfred Weck,et al.  Analysis and Adaptive Control of EDM Sinking Process Using the Ignition Delay Time and Fall Time as Parameter , 1992 .

[12]  Sohyung Cho,et al.  Tool breakage detection using support vector machine learning in a milling process , 2005 .

[13]  Mihael Junkar,et al.  On-line selection of rough machining parameters , 2004 .

[14]  Shaw Voon Wong,et al.  Development of machine learning strategy for acquiring on-line machining skills during turning process , 2004 .

[15]  Yunn-Shiuan Liao,et al.  A neural network approach for the on-line estimation of workpiece height in WEDM , 2002 .

[16]  D. F. Dauw,et al.  Advanced Pulse Discriminating System for EDM Process Analysis and Control , 1983 .

[17]  Y. S. Tarng,et al.  A fuzzy pulse discriminating system for electrical discharge machining , 1997 .

[18]  Mihael Junkar,et al.  Detection of the eroding surface in the EDM process based on the current signal in the gap , 2006 .

[19]  Mihael Junkar,et al.  A model for detection of the eroding surface based on discharge parameters , 2004 .

[20]  Kamlakar P Rajurkar,et al.  WEDM-Adaptive Control with a Multiple Input Model for Indentification of Workpiece Height , 1997 .

[21]  Ron Kohavi,et al.  A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection , 1995, IJCAI.