An artificial neural network & machine learning approach for predicting the performance of cutting machine
Un réseau de neurones artificiels et une approche d'apprentissage automatique pour prédire la performance d'une machine de coupe
Résumé
In production systems, repeated failures must be taken into account with great importance. The pursuit of longdysfunctional states as well as temporary interventions involve excessive costs in terms of time and money.
Industry 4.0 technologies make extensive use of the big real-time data collected from machines, allowing address and resolve potential problems before they become an avalanche for the company. Permanent solutions can be ensured, and thus production efficiency can be established.
In this paper, we will use the values of average time before failure (MTBF), the average time to repair (MTTR) and the history of past failures of cutting machines to ensure the model learning. Indeed, a network model of Artificial Neurons (ANN) and the Random Forest algorithm are established for the prediction of system failures.
Contents I. Introduction II. Materials and Methods III. Results IV. Conclusion V. References