Indian Journal of Engineering & Materials Sciences Vol. 14, December 2007, pp. 403-407 Optimization of tool wear in turning using genetic algorithm C Felix Prasad, S Jayabal & U Natarajan Department of Mechanical Engineering, A C College of Engineering and Technology, Karaikudi 630 004, India Received 25 January 2007; accepted 11 December 2007 Optimization has significant practical importance particularly for operating in machines. In order to increase the accuracy of finishing product the tool must be in good condition always as much as possible. To achieve good condition of tool the machining parameters like speed, feed and depth of cut should be optimized 2-4 . This paper aims to increase the condition of tool, i.e., minimization of tool wear by applying the optimized input parameters using Genetic Algorithm technique. For solving the machining problems various traditional techniques such as Integer programming, Quadratic programming and Simplex method have been used so far. But results of these various methods vary to some extent. Genetic Algorithm helps in obtaining better results than of traditional methods available. Keywords: Optimization, Tool wear, Genetic Algorithm, Input parameters The growing demand for product quality and economy necessity have forced incorporation of monitoring the process parameters in automated manufacturing system. The greatest limitation of automation in the machining operation is tool wear. If the tool wear increases then the tool life will be minimum. So the tool wear has been optimized by selecting the optimal cutting parameters such as speed and depth of cut. The main machining parameters which are to be considered as variables of the optimization are speed, feed and depth of cut the required output is minimum tool wear 3 . The optimum set of these three parameters are determined for a particular job-tool combination of High Carbon Steel-Tungsten Carbide during turning which optimizes the tool life and tool wear. Many graphical and analytical techniques were initially used for optimization 4 . Some of them are overcome by numerical optimization methods which may be either Simplex or Gradual based methods. These and other similar numerical optimization techniques have been found to give reasonably accurate results only for simple problems 1 . But in case of highly non-linear optimization problems, the numbers of iterations are required to get optimum point, thus consuming significant computation time 2-4 . So an alternative method for optimization is probabilistic methods. Genetic Algorithm is one of the example of probabilistic methods. Genetic Algorithm searches from a population and gives a set of reasonably good solution to the user. Prediction Techniques The mathematical modeling is done using simple probabilistic considerations and design of experiments 5,6 . Design of experiments The statistical design of experiments is the process of planning the experiments so that appropriate data could be collected which may be analyzed by statistical methods resulting in valid objective conditions 3-6 . Here central composite method 3 3 factorial design is used to develop a model for the tool wear. In this work three factors are used, each factor is at three levels, arranged in a factorial experiment, i.e., 3 3 factorial designs. The three levels of factors are referred to as low (-1), medium (0) and high (+1). In the present work cutting speed, feed and depth of cut have been used as input variables. Genetic algorithm Genetic algorithm is a global population search technique based on the operations of natural genetics and mimics the natural biological process 8,9 . In this work for a given optimization problem Genetic Algorithm first encodes all variables into a finite bit binary string called as chromosomes. Chromosomes represent a possible solution to the optimization problem. A population of chromosomes is formed. Each chromosome is decoded and is evaluated according to the fitness function. Fitness function The fitness function plays very important roll in Genetic algorithm. Genetic algorithms are naturally