اميرحسين تندري

One of the most important processes in the manufacturing industry is the turning process. Turning is an operation in which the internal or external surface of a metal or non-metal piece is grooved and chipped by a turning tool. Part of the costs of this operation is related to the turning tool, considering that the turning process or machining in general is a complex process, it is difficult to determine the time to change the tool in order to reduce the cost and waste of the tool, even for experts. The purpose of this research is to investigate the influencing parameters during the lifetime of the tool and, accordingly, its replacement time, as well as to predict the replacement time by predicting the flank wear. After collecting 1995 data from the tests performed on turning operation, we have investigated the parameters affecting the life of the tool, then using two machine learning algorithms, gradient boosting and neural network to predict the amount of flank wear. Finally, by eva‎luating the neural network and gradient boosting models, with the help of the series of features selected from the regularization and recursive elimination method, respectively, they were able to perform well in order to predict the wear rate of the side of the tool.

پيش‌بيني , ماشين كنترل عددي , پايش وضعيت ابزار , تراشكاري , سايش پهلو , يادگيري ماشين

prediction , numerical control machine , Tool condition Monitoring , turning , flank wear , Machine Learning

Amir Hossein Tondari

Dr. Mohammad Reza Mohammad Aliha