Abstract — Conventionally the selection of parameters depends intensely on the operator’s experience or conservative technological data provided by the EDM equipment manufacturers that assign inconsistent machining performance. The parameter settings given by the manufacturers are only relevant with common steel grades. A single parameter change influences the process in a complex way. Hence, the present research proposes artificial neural network (ANN) models for the prediction of surface roughness on first commenced Ti-15-3 alloy in electrical discharge machining (EDM) process. The proposed models use peak current, pulse on time, pulse off time and servo voltage as input parameters. Multilayer perceptron (MLP) with three hidden layer feedforward networks are applied. An assessment is carried out with the models of distinct hidden layer. Training of the models is performed with data from an extensive series of experiments utilizing copper electrode as positive polarity. The predictions based on the above developed models have been verified with another set of experiments and are found to be in good agreement with the experimental results. Beside this they can be exercised as precious tools for the process planning for EDM. Keywords— Ti-15l-3, surface roughness, copper, positive polarity, multi-layered perceptron. I. INTRODUCTION The recent developments in the field of EDM have progressed due to the growing application of EDM process and the challenges being faced by the modern manufacturing industries, from the development of new materials that are hard and difficult-to-machine [1]. These materials like tool steels, composites, ceramics, super alloys, hastalloy, nitralloy, waspalloy, nemonics, carbides, stainless steels, heat resistant steel, etc. being widely used in die and mould making industries, aerospace, aeronautics, and nuclear industries. Many of these materials also find applications in other Md. Ashikur Rahman Khan is PhD Student with Faculty of Mechanical Engineering, Universiti Malaysia Pahang, 26300 Kuantan, Pahang, Malaysia (e-mail: ashik_nust@yahoo.com). M. M. Rahman is Associate Professor with Faculty of Mechanical Engineering and Deputy Director, Automotive Engineering Centre, Universiti Malaysia Pahang, 26300 Kuantan, Pahang, Malaysia (corresponding author: 609-4242246; fax: 609-4242202; e-mail: mustafizur@ump.edu.my). K. Kadirgama is Lecturer with Faculty of Mechanical Engineering, Universiti Malaysia Pahang, 26300 Kuantan, Pahang, Malaysia (e-mail: kumaran@ump.edu.my). Rosli A. Bakar is Professor with Faculty of Mechanical Engineering, Universiti Malaysia Pahang, 26300 Kuantan, Pahang, Malaysia (e-mail: rosli@ump.edu.my) M. A. Maleque is Associate Professor with Department of Materials and Manufacturing Engineering, International Islamic University Malaysia, Gombak, Selangor, Malaysia (email: Maleque@iiu.edu.my) industries owing to their high strength to weight ratio, hardness and heat resisting qualities. Ti–15–3 (Ti–15V–3Cr–3Al–3Sn) alloy is a kind of metastable -titanium alloy. Ti-15-3 alloy is used for springs such as clock-type springs due to its strip producibility [2]. There are several areas besides springs where Ti-15-3 is being used in current generation aircraft. One of the large users is the Boeing 777, and the big item is for ECS ducting. To produce fire extinguisher bottles, Ti-15-3 is consumed in place of 21-6-9 steel, providing a weight savings of about 23 kg per airplane. As well as it is used for numerous clips and brackets in the floor support structure and other areas of the 777. In spite of its more advantages and increased utility of titanium alloys, the capability to produce parts products with high productivity and good quality becomes challenging. Owing to their poor machinability, it is very difficult to machine titanium alloys economically with traditional mechanical techniques [3]. Thus, titanium and titanium alloy, which is difficult-to-cut material, can be machined effectively by EDM [4]. The various machining characteristics used to evaluate the performance of EDM such as material removal rate (MRR), tool wear rate (TWR), relative wear ratio (WR) and surface roughness (SR ) [5,6]. The important variables that affect the performance of EDM are peak current, pulse-on time, pulse- off time, the polarity of the electrode, nozzle flushing etc [7]. The thermodynamic and physical properties of the tool and the work-piece also influence the electrical discharge machining performance [8]. The selection of appropriate machining conditions for EDM characteristics, such as material removal rate, is based on the analysis relating the various process parameters to material removal [9]. Artificial neural network models were proposed for the prediction of surface roughness from roughing to near- finishing conditions considering workpiece material, pulse current and pulse duration as the input parameters of the models [10]. Particularly mild steel (St 37), alloyed steels (C 45 and 100Cr6), high strength low alloyed (HSLA) steels such as a microalloyed (Mic/al 1) steel and dual-phase (DP1) steel were tested employing electrolytic copper as tool electrode of positive polarity. Assarzadeh and Ghoreishi [11] were presented back-propagation neural network approach for prediction and optimal selection of process parameters in support of MRR and surface roughness in die sinking EDM. BD3 steel work piece and commercial cylindrical copper as tool with positive polarity were used throughout the experiments. Chattopadhyay et al. [12] investigated machining characteristics of EN-8 steel and also developed empirical models for prediction of output parameters using linear Artificial Intelligence Model to Predict Surface Roughness of Ti-15-3 Alloy in EDM Process Md. Ashikur Rahman Khan, M. M. Rahman, K. Kadirgama, M.A. Maleque and Rosli A. Bakar World Academy of Science, Engineering and Technology International Journal of Mechanical and Mechatronics Engineering Vol:5, No:2, 2011 503 International Scholarly and Scientific Research & Innovation 5(2) 2011 scholar.waset.org/1307-6892/5730 International Science Index, Mechanical and Mechatronics Engineering Vol:5, No:2, 2011 waset.org/Publication/5730