Predictive maintenance for paper and pulp production processes

This project intends to develop a data-driven technique for predicting the remaining useful life of paper and pulp manufacturing equipment. 

Over 20 sensors and maintenance records were collected from two vacuum blower processes over a two-year period. It is meant to anticipate the remaining useful life for about 40 to 60 days before failure. 

We tested the method on both known and unknown failure time data sets, and the prediction results showed that our performance met the requirements.