Session: 07-17-03: Machine Learning and Artificial Intelligence in Dynamics, Vibrations and Control

Paper Number: 146002

146002 - A Comparative Classification Study on the Use of Audible Acoustic Emission Signals for Surface Roughness Condition Monitoring in Shoulder Milling of Steel 

Despite the current advances in additive manufacturing, subtractive manufacturing methods such as Machining is still taking a major share in modern industry. Surface roughness of the machined product is a crucial parameter that impacts the functionally, assembly, and service life, of the final product. The Surface finish texture of machined components could be too complex for accurate prediction when using analytical techniques or computer simulation techniques. This is because there are numerous parameters relating to the material of the work part, cutting tool, and machining process conditions.  Therefore, machine learning (ML) techniques are becoming more popular in designing smart models that are capable of providing more reliable real time surface quality predication. The aim of this study is to develop a machine learning model to predict the surface roughness in shoulder milling of steel parts. The process milling is controlled by the feed per tooth, cutter radial immersion, axial depth of cut, and spindle speed.  The surface roughness measure is presented by the parameter Ra in this study. The milling tests are carried out on a three-axis vertical milling Haas CNC Minimill2 machine using a square face milling cutter.  Microphones when used properly, can pick up valuable acoustic data that is highly correlated to the acoustic waves produced by the machining process. These sound measuring devices are non-invasive and can be easily integrated within the machining envelope without disrupting or stopping the machining process. These features or indicators include averaged wavelet decomposition quantities, statistical quantities, and filtered time signatures of the sound waves. The features extracted from the audible sound signals are (1) used in training model to identify important features or a combination of features that have highly corelated to surface finish, and (2) develop a learning model to use these features to predict the surface roughness in shoulder milling. In this study, we use and compare a variety of dimensionality reduction algorithms in the training phase. For the learning model, we use both supervised and unsupervised classification models and compare their performance using nonparametric statistical modeling methods.  The overall goal is to develop a reliable and robust predictive tool with potential for practical implementation in a real-time industrial machine tool installation for process monitoring.

Presenting Author: Amit Banerjee Penn State Harrisburg

Presenting Author Biography: Amit Banerjee is an Associate Professor of Mechanical Engineering at Pennsylvania State University Harrisburg. His research interests include artificial intelligence, computational intelligence in manufacturing, evolutionary computation, machine learning, robotics and automation, and application of computational technologies in engineering education. He teaches courses in robotics, manufacturing, design and mechanics. He has a Ph.D. in Mechanical Engineering from the New Jersey Institute of Technology and a Master's degree in Product Design and Engineering from the Indian Institute of Science, Bangalore. He has also been a postdoctoral research fellow at the Evolutionary Computational Systems Lab in the department of Computer Science and Engineering at the University of Nevada Reno.

Authors:

Amit Banerjee Penn State Harrisburg
Issam Abu-Mahfouz Penn State Harrisburg
Esfakur Ahm Rahman Penn State Harrisburg