Optimal machining parameters based on surface roughness experimental data and genetic search

Purpose - Surface roughness is an important parameter in manufacturing engineering with significant influence on the performance of mechanical parts. Failures, sometimes catastrophic failures, leading to high costs, have been imputed to a component's surface roughness. Owing to the need for improvement of machining parameters in order to obtain a prescribed surface roughness, new developments have been recently investigated. This work aims to report on a study of an optimisation model based on genetic algorithms (GAs). Design/methodology/approach - The developed algorithm considers a machining parameter data population obtained from experimental tests. The exchange of structured information based on natural selection principles and "survival-of-the-fittest" allows the combination of solutions in a sequence of generations leading to the best solution. Findings - Over standard experimental design methodologies the proposed GA approach shows advantages in finding the optimal conditions under the imposed constraints. Indeed the quality of the produced surface roughness cannot be evaluated using only a criterion. This GA method determines the combined effects of the input parameters to the optimal machining parameter. Originality/value - A new methodology for determining optimal machining parameters in dry turning based on the measurement of the surface roughness is proposed. The numerical and experimental developed model can be used with success on further applications with industrial interest.