Application of logistic regression in industrial maintenance management

Abstract

In the area of industrial maintenance, the application of statistical methods is essential, in that sense, the purpose of this analysis is to explore logistic regression as an element of industrial maintenance management. By means of logistic regression, a predictor equation for the response variable, machine failure, is obtained by correlating it with categorical and continuous predictor variables. The continuous explanatory variables are machine age, mean time between failures, mean time to repair and the categorical ones are application of preventive and corrective maintenance. The results obtained indicate that only the explanatory variable preventive maintenance is significant to the response variable by applying the Wald test and this result was also validated with goodness-of-fit tests. Logistic regression is more used in other areas, such as health, however, in maintenance categorical variables are used such as machine with autonomous maintenance whose result is yes/no, therefore, it is important to incorporate a regression model that considers different types of independent variables, in addition to the use of emerging technologies of Industry 4.0 such as Machine Learning for the prediction of scenarios for efficient maintenance management.