Development of a method to predict the remaining useful life of a mill roll for grain milling

In grain processing, milling usually plays a key role. The efficiency of the milling process is key to the profitability of the process and is decisive for the quality of the final product. However, milling is often subject to wear problems. Indeed, the degradation of the rolls varies over time depending on factors that are difficult to quantify, especially in an uncertain environment. This research project will provide a data-driven method for the prediction of the remaining useful life (RUL) of rolls in grain milling. RUL prediction is used to forecast the future performance of machinery and obtain the time left before machinery loses its operation ability. The prediction of the RUL of the rolls would be beneficial to plan maintenance in advance and would allow a better control of the production process. Data-driven methods attempt to explain the wear and tear of a machine based on collected data from the shop floor using machine learning techniques. The method will specify the selection of the most relevant attributes for the prediction of the RUL of rollers in grain processing.

Faculty Supervisor:

Bruno Agard

Student:

Partner:

Centro de Inovação e Tecnologia SENAI FIEMG - Campus CETEC

Discipline:

Life Sciences

Sector:

Manufacturing and Construction; Artificial Intelligence; Agriculture and Food

University:

Polytechnique Montréal

Program: