Predictive maintenance is a preventive maintenance strategy that uses AI methods to optimize the management of every type of resource. Generally, supervised methods are used to monitor the resources, unsupervised methods to control their status and forecasting methods, like time series approach, to estimate the Remaining Useful Life of them. This course consists in the explanation of an R project developed to analyze a milling case study.

Audience (and prerequisites)

Anyone with a basic knowledge in statistics and programming who are interested in programming tools with R for the industrial sector

Approaches (Objective)

Students will follow the various steps of a data-driven framework to estimate the tool wear status and predict its remaining useful life by using sensors measurements.

The first part is dedicated to sensor data preprocessing and feature engineering, the second one to compare Neural Network performances with standard regression methods and the last one to use time series forecasting to predict the Remaining Useful Life of the tool.