Parametric optimisation of laser beam machining processes using shuffled frog leaping algorithm

Need for generation of intricate shapes on difficult-to-machine materials, better surface finish and higher material removal rate drives towards the development of an array of non-traditional machining (NTM) processes. Laser beam machining (LBM) is one such NTM process where a laser beam is directed towards the workpiece surface for material removal. With development of different NTM processes over the years, LBM process has become the first choice for machining of metallic and non-metallic workpieces, giving rise to the requirement to optimize its input parameters to achieve the best machining performance. Variation in any of these parameters (lamp current, pulse frequency, pulse width, cutting speed, assist air pressure etc.) may result in deviation in the responses (surface finish, material removal rate, heat affected zone, conicity, kerf etc.). Since in LBM process, there is a combination of multiple parameters, change in any or all of these parameters has a remarkable effect on the responses. Thus, it becomes important to study the effects of various LBM process parameters on the responses as well as to search out the best possible combination of these parameters to attain the target results. Several mathematical tools, like Taguchi method, desirability function, grey relational analysis etc. have been proposed for parametric optimization of LBM processes. In this chapter, an almost unexplored meta-heuristic in the form of shuffled frog leaping algorithm is adopted for both single and multi-objective optimization of the responses for two LBM processes. It is a nature inspired algorithm and mimics the leaping pattern of frogs in search of food. Its major advantage is that it doesnot accumulate towards some local optima. Local exploration of frogs ensures memetic evolution within a population, whereas, global exploration makes way for a global share of information (memes) between the populations. As the leaping behavior of frogs is mimicked in this algorithm, the number of iterations required to reach the global optimum is drastically reduced. The developed scatter plots can be utilized to investigate the variationsin LBM responses with respect to changes in different process parameters involved. The derived results show a significant improvement in the response values as compared to the earlier attempts for optimization of LBM processes employing other meta-heuristics.

Download paper here