Modelling of Virtual Sensors for Manufacturing Process using Gradient Boosting Technique

This paper proposes a learning architecture approach for creating a virtual sensor model that detects product failures by using machine operating data from a discrete manufacturing process based on Gradient Boosting and Random Forest algorithms. The main contribution of this work is to propose a methodology for creating the virtual sensor to predict the manufactured product quality with precision equivalent to that obtained by actual sensors. Simulation results showed that the proposed virtual sensor detects precisely manufacturing failures caused by supplement position to produce the target products.