Eventos Anais de eventos
COBEM 2023
27th International Congress of Mechanical Engineering
EXPLORATORY DATA ANALYSIS APPLIED TO BEARING MANUFACTURING PROCESS IN THE AUTOMOTIVE FIELD
Submission Author:
Isabelle Therezinha Simão , PR , Brazil
Co-Authors:
Alan Lopes, Isabelle Therezinha Simão, LUIZ EDUARDO THOMAZ, Viviana Mariani, Leandro dos Santos Coelho
Presenter: Isabelle Therezinha Simão
doi://10.26678/ABCM.COBEM2023.COB2023-0823
Abstract
. Bearings are mechanical transmission elements used to reduce friction between two surfaces in relative motion. It is used in a variety of engineering applications, and are widely used in transportation equipment such as cars, trucks, trains and airplanes to help ensure that the movement of parts is smooth and accurate and reduce vibration and noise. Bearings are composed of series of balls, rollers, or needles arranged in a mechanical separator and an inner and outer ring. However, over time, these elements can exhibit defects that impair their performance and can even lead to equipment failure. Therefore, research involving the control of the constructive characteristics of bearings is indispensable. In bearing manufacturing characteristics studies, correlation analysis can be used to evaluate the relationship between different variables that affect bearing quality and performance. Therefore, this research aims to evaluate two case studies derived from experimental data of geometric characteristics generated through the manufacturing process of automotive bearings, containing data of bearings without defects and with defects, respectively. Through correlation analysis, the goal is to verify how the variables behave and if there are any relationship between the variables. In both case studies, values above 0.5 were obtained, therefore positive correlations. From Phik's correlation the positive correlations were obtained with values close to one. A difference was observed in the values obtained for the correlations using bearings with defects when compared to the data for bearings without defects. It is worth noting that this is a preliminary study applied to the automotive industry. Thus, it was possible to identify the existence of the relationship between the constructive variables, such as vibration, roughness, and shape errors, derived from the manufacturing process of the bearings.
Keywords
Correlation analysis, Bearing manufacturing, feature engeneering, Regression, Random Forest
