Machine learning aplicado à usinagem: previsão de informações de um processo de usinagem por machine learning com dados de vibração obtidos com aplicativo Sci Journal

Abstract

Real-time process monitoring is one of Industry 4.0's key advantages and needs. This requires the acquisition of process data and industrial equipment data through sensors of various types. In the machining process, the vibration resulting from the cutting effort of the material by the tool is a source of information about the state of the process, the state of the cutting tool, the quality characteristics being generated and the state of the machine tool. To measure vibration, accelerometers are used, similar to those found in many Smartphones available on the market. This paper proposes the use of acceleration sensors present in a common mobile device to obtain data from a machining process through the Sci Journal data collection application. Then, supplying information for a Decision Tree Regression, a Machine Learning algorithm in Python language, with those data seeking to predict information about the machining process and part quality in real time based on the measured vibrations.