Monitoring hole quality in a drilling process using a fuzzy subtractive clustering-based system identification method

in this study, a subtractive clustering fuzzy identification method and a Sugeno-type fuzzy inference system are used to monitor the hole quality in a drilling. The model for the hole quality is identified by using the hardness of the workpiece, the cutting speed, and the cutting feed as input data and the hole quality features of hole roughness, roundness error, and oversize error as the output data. The process of model building is carried out by using subtractive clustering in both the input and output spaces. A minimum error model is obtained through exhaustive search of the clustering parameters. The fuzzy model obtained is capable of predicting the hole quality for a given set of inputs (hardness of the workpiece, the cutting speed, and the cutting feed). Therefore, one can predict the quality of the drilled hole for a given set of working parameters. The fuzzy model is verified experimentally using different sets of inputs. This study deals with the experimental results obtained during drilling on medium carbon steel (AISI 1060), aluminum, and brass.