State of product detection method applicable to Industry 4.0 manufacturing models with small quantities and great variety: An example with springs

摘要

A major feature of the manufacturing models in Industry 4.0 is smaller quantities and greater variety. In other words, each machine tool will produce multiple types of products, but each type in small numbers. This is just the opposite of the manufacturing models with large quantities and small variety in Industry 3.0. As a result, the conventional means of detecting the health of machines in Industry 3.0 models may not be applicable to Industry 4.0 models. This is because the conventional means involves collecting and analyzing data from numerous products and then using the results to determine the state of products in the future. However, in new manufacturing models, a production process may end before the data can even be collected, let alone the checking of machine health. For this reason, this study proposed a novel concept based on artificial recurrent neural networks to assist in the determination of machine health in Industry 4.0 manufacturing models with small quantities and increased variety. Experiments demonstrated the feasibility of the proposed approach.