(86c) Plant-Wide Digital Twinning of Surface Finishing for Sustainable Manufacturing

The surface finishing industry is critical to many manufacturing industries, such as automotive, aerospace, electronics, defense, as well as a variety of OEMs. This is because the finish on a surface can make a huge effect on the performance, durability, and/or aesthetic appearance of the parts that are used in those industries. The surface finishing industry has been highly regulated by EPA due to its significant use of numerous toxic/hazardous chemicals and the generation of huge amounts of waste in various forms; this could be very harmful to the environment, human health and communities, as well as facilities’ financial performance.

While significant progress has been made to improve the sustainability performance in the industry over the past two decades, the consumptions of chemical, water, and energy are still too high, there still exist various environmental and health risks, and profit margin is still very low in a large number of plants. One of the solution approaches is firstly to characterize the dynamic behavior of electroplating lines under different operating conditions with uncertainty. Digital twinning is a powerful technique in Industry 4.0. A digital twin is a virtual representation that serves as the real-time digital counterpart of a physical process system. In this paper, we introduce a framework for generating a set of neural-networks-based digital twins for characterizing pre-surface preparation, surface finishing, and post-surface processing of a zinc electroplating line. Using large sets of plant data, the digital twins are created by resorting to a supervised learning technique; they will be updated as new plant data become available. We will show how the developed digital twins can be employed to analyze plant’s sustainability performance, and identify performance improvement opportunities.