A Prediction Model for Fault Detection in Molding Process Based on Logistic Regression Technique

Conference paper > Full Text · 2019

Authors: İnanç KABASAKAL, Fatma DEMİRCAN KESKİN, Aydın KOÇAK, Haluk SOYUER

Lecture Notes in Mechanical Engineering, Austria, Vienna, International, Verbal, Full Text, 2019-2019, Scopus, pp 351-360

DOI: 10.1007/978-3-030-31343-2_31

Summary

<p>Industry&nbsp;4.0&nbsp;is&nbsp;often&nbsp;described&nbsp;as&nbsp;a&nbsp;wave&nbsp;of&nbsp;transformation&nbsp;that&nbsp;enforces&nbsp;the&nbsp;digitalization&nbsp;of&nbsp;processes&nbsp;to&nbsp;create&nbsp;integrated&nbsp;and&nbsp;autonomous&nbsp;systems.&nbsp;In&nbsp;this&nbsp;regard,&nbsp;the&nbsp;collection&nbsp;of&nbsp;process&nbsp;data&nbsp;is&nbsp;a&nbsp;necessity&nbsp;to&nbsp;analyze&nbsp;data&nbsp;with&nbsp;advanced&nbsp;techniques&nbsp;for&nbsp;various&nbsp;purposes.&nbsp;Statistical&nbsp;techniques&nbsp;in&nbsp;machine&nbsp;learning&nbsp;might&nbsp;provide&nbsp;solutions&nbsp;for&nbsp;fault&nbsp;detection&nbsp;and&nbsp;other&nbsp;tasks&nbsp;in&nbsp;manufacturing&nbsp;processes.&nbsp;In&nbsp;our&nbsp;study,&nbsp;a&nbsp;learning&nbsp;model&nbsp;is&nbsp;proposed&nbsp;for&nbsp;a&nbsp;fault&nbsp;prediction&nbsp;task&nbsp;with&nbsp;the&nbsp;use&nbsp;of&nbsp;Logistic&nbsp;Regression.&nbsp;The&nbsp;data&nbsp;used&nbsp;for&nbsp;the&nbsp;analysis&nbsp;involve&nbsp;measurements&nbsp;from&nbsp;sequential&nbsp;processes&nbsp;carried&nbsp;out&nbsp;in&nbsp;a&nbsp;large-scale&nbsp;wheel&nbsp;rim&nbsp;manufacturer.&nbsp;The&nbsp;pre-processing&nbsp;and&nbsp;analysis&nbsp;of&nbsp;process&nbsp;data&nbsp;was&nbsp;introduced&nbsp;along&nbsp;with&nbsp;a&nbsp;case&nbsp;study.&nbsp;Moreover,&nbsp;findings&nbsp;of&nbsp;the&nbsp;model&nbsp;were&nbsp;presented&nbsp;and&nbsp;the&nbsp;potential&nbsp;use&nbsp;of&nbsp;the&nbsp;model&nbsp;will&nbsp;be&nbsp;discussed.</p>

View details