An artificial neural network & machine learning approach for predicting the performance of cutting machines

In production systems, repeated failures must be taken into account with great importance. The pursuit of long dysfunctional 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.