Improved regressive maintenance in smart Industry 4.0 using embedded systems with machine learning techniques

Abstract

Despite extensive research in this field, fourth generation of industrial revolution or the "Industry 4.0," remains rarely used. This technology enables sectors to make intelligent on taking decision, knowledge, problemrectifying, self-diagnosis, configuration, and automation feasible. This research study proposes to transform the food sector into a smart Industry 4.0 by utilizing sensor and machine learning (ML) methods. The motive of this article is to use a variety of sensor data, such as motor temperature, oven temperature, gear pump temperature, vibration, current consumption, and motor speed, to forecast the likelihood of machine failure in the food processing business. We create models using artificial neural networks (ANN) and linear regression with the data collected from these sensors to estimate the likelihood of failure. The findings demonstrate that both of the generated models have excellent accuracy in forecasting the chance of failure. The technology could help the food sector transition to sector 4.0, where networked equipment exchanges data to maximize production efficiency. In summary, the present study showcases the potential of sensorbased analysis maintenance methods in the food sector and emphasizes the need to incorporate Industry 4.0 principles to enhance the effectiveness of food processing facilities.