A Digital Twin Case Study Factoring in Human Dynamics in Predicting Throughput of an Industrial Production Line

This paper presents a case study to demonstrate the use of a digital twin to reliably predict the real-time productivity throughput, in the presence of random uncertainties, for a recently commissioned industry 4.0-based manufacturing system. In a complex industrial environment comprised of autonomous and symbiotic (human-machine) work cells, variations in human performance is one of the principal reasons for productivity fluctuations. Traditionally historical data, experience, and established metrics are used to estimate the expected throughput. As this is a recently developed manufacturing facility, human metrics data for this system is not available. This paper is an attempt to explore the possibility of quantifying the cumulative throughput in a work shift, prior to production, in the absence of historical reference data of human performance for this setup. A digital twin has been used to capture real-time human performance, thus enabling proactive, timely intervention, when necessary, to achieve the set productivity target. Additionally, this visualisation could be used for monitoring the health and safety conditions of the operators. The case study begins with the design of a basic digital twin, followed by improving it to generate a hi-fidelity model to reproduce a realistic physical system and finally integrating the digital twin to physical assets to observe its response to real-time inputs. A combination of event-driven synthetic data for the automated machine-based operations and real-time data for the human operations in the symbiotic (human-collaborative robot) cell are used as inputs to the digital twin. It is envisaged that this case study would serve as a prototype and demonstrate a holistic and reliable visualisation of end-to-end factory floor processes. It is most suitable for both fast-paced and precision manufacturing industries which rely heavily on automation systems and humans for their operation such as aerospace, and ongoing quality improvements in a highly variable and unpredictable environment such as blood product manufacturing.