Modelling of the hole quality characteristics by Extreme Learning Machine in fiber laser drilling of Ti-6Al-4V

Abstract The hole circularity, taper angle, and spatter formation area at drilling Ti-6AI-4V alloys by fiber laser are the basic properties that determine the quality of the product. These properties are directly related to cutting parameters such as laser power, cutting speed and gas pressure. In this study, a total of 27 experiments were performed with different cut parameter combinations to provide sufficient data at the level at which the parameters could determine the effect on product quality. Spatter field formed in uncontrolled diffused and complex shapes during cutting were calculated using image processing technique. The data obtained as a result of the measurements were modeled by using a new modeling method called Extreme Learning Machine (ELM) based on Artificial Intelligence (AI) and Artificial Neural Networks (ANN) to predict hole diameter, taper angle and spatter forming area on the fiber laser drilled surface. The estimation models of ELM were compared with ANN models. When both methods are compared in terms of modeling performances of training and test phases, it is found that ELM method performs faster than ANN method and shows higher performance with smaller error margin. As a result, ELM method has been demonstrated to be able to be used safely to develop a prediction model for studies in a variety of manufacturing areas where a large number of experiments are required.