234 TEM Journal – Volume 2 / Number 3 / 2013. www.temjournal.com An ANFIS-Based Approach for Predicting the Surface Roughness of Cold Work Tool Steel in WEDM Kemal Aldaş 1 , İskender Özkul 1, Adnan Akkurt 2 1 Aksaray University, Faculty of Engineering, 68100, Aksaray, Turkey 2 Gazi University, Technical Education Faculty Department of Industrial Arts, 06500, Teknikokullar-Ankara, Turkey Abstract – Wire electrical discharge machining known as non-traditional machining processes, has a significant role in the manufacturing industry. Conductive materials, which can have intricate and complex forms, can be obtained regardless of hardness. In this study, the surface roughness of Sleipner cold work steel is evaluated under various machining process parameters in the WEDM process. In the experiments the feed rate, current, and pulse on time are used as independent variables. In order to predict the surface roughness, an Adaptive Neuro-Fuzzy Inference system was applied based on experimental data. Keywords- ANFIS, WEDM, cold work tool steel, surface roughness. 1. Introduction In the manufacturing industry, wire electrical discharge machining (WEDM) is accepted as a non- traditional manufacturing method. The WEDM operation is based on thermoelectric energy transferring between a wire electrode and the work piece loaded in the anode and cathode [1]. The continuously travelling wire electrode, upon passing the current, shapes the work piece using the electron discharge as spark(s) between the wire and work piece. The work piece and wire electrode are covered by die-electric fluid during operation [2]. Fig.1 shows the basic schematic of the WEDM process. In this study, surface roughness, considering as output parameter, was investigated in WEDM. During machining, the feed rate (fr), current (I), and pulse on time (P On ) are used as independent variables. Four different parameters were chosen for all the input variables. The Adaptive Neuro-Fuzzy Inference system (ANFIS) prediction method was implemented under these machining conditions and achieved the predicted results for surface roughness. Two different membership functions were used during modelling the ANFIS for the experimental data and a Sleipner cold work tool steel alloy was chosen as work piece material. Sleipner tool steel is especially used in die making industry. However, it can be adopted for various applications such as base materials. Fig. 1 Schematic of WEDM process. WEDM machining characteristics have been investigated with different parameters on various materials in order to evaluate the optimal machining