Introduction: Digital twins are becoming a powerful tool to enhance industrial processes worldwide. This paper proposes a model for the creation of industrial processes’ digital twins using a steam distillation process for essential oil extraction as a case study. Case Description: A grey box modeling is suggested combining a machine learning based model with physical modeling to improve the process. Real time simulation and a hybrid control strategy are used, linked to reinforcement learning and proportional integral derivative control, focusing on the yield increase and optimization. Computer Vision and Artificial Intelligence enhancements were suggested. Discussion and Evaluation: Digital twins, in combination with Artificial Intelligence can be of great help to support companies with the decision-making challenges. Furthermore, some benefits that Artificial Intelligence can bring to the process were enlightened. Computer Vision approaches were also discussed. Conclusions: A creation method is elaborated to support other applications of digital twins in industrial processes in the future. In order to apply it to different processes, generalization capabilities must be proved.

Yield uncertainties characterize the industrial extraction of essential oils. The variable plant contents, undetected and uncorrected operational issues such as production in small batches, with a significant variation in the process cycles, productivity, and quality, compromise process performance, and the value of the finished product. The technological advances of industry 4.0 provide an opportunity to achieve economic, environmental, and productivity gains, even with the wide variety of products manufactured. This study aimed at improving the performance of such processes using the pillars of industry 4.0 as a method, highlighting the techniques of data acquisition and processing for intelligent analysis, based on automated and systemic learning, generating improvements for the overall performance of the operation as well as enhancing good manufacturing practices (process safety, quality, repeatability and traceability) and supply chain management. Our findings showed the adherence of 4.0 vision to this industry.

No abstract

This paper aims at the potential and unexplored improvements imbedded in the essential oil (EO) extraction processes. These processes are worldwide recognized by their low yields and high energy demands, becoming a target to be addressed to improve cost, reliability and quality indicators. A strategic thinking is presented, as a guide, with focus on enabling better business results through the TDC (Total Delivered Cost) optimization and the impact over gross margin in this promising segment. A basic process was evaluated, step by step, to then add technology improvements. The resulting impacts were assessed, improving yield from 1% to 1.2%. Such improvements are discussed in terms of digital technology contribution with the expected positive effects.

Smart sensors, self-configuration, operational flexibility, and automatic learning, among others, are technological attributes from industry 4.0 appliable to the essential oil extraction by the steam distillation process. These operations are recognized by their simplicity. Nevertheless, lack of automatic controls, process monitoring, and self-adjustment lead to uncontrolled extraction, poor yields, low quality of products. It occurs because of overexposure to high temperatures and overspending resources like energy and water. As far as capacity utilization is concerned, the optimized process is key to planning and managing the production activities.