Machine Failure Analysis Using Multinomial Logistic Regression

Abstract

Purpose: The main purpose of this study is to carry out a failure analysis of a filling machine by applying multinomial logistic regression.Design/methodology/approach:For this purpose,data related to failure mode, product, scrap rate, and shift parameters were collected from the machine and analysis was conducted by establishing two multinomial logistic regression models.Findings: Statistical results suggest that a hydraulic failure must be expected while filling the mix product. Besides, it is highly probable that a final folder failure will occur while filling the cherry product. Paper failure stands out while filling the apple product compared to other products. In addition, it is likely that a final folder failure will occur while filling this product. Photocell failure is common while filling the peach product. Results of the study show that the odd for low-level scrap is high when there is any failure in the machine.Discussion: A more effective analysis can be performed by collecting parameters that may affect the position of machinery such as vibration, humidity, temperature and pressure through the sensors to be installedon various units of the filling machine and adding them into the models developed under the study.