Artificial Intelligence Forecasting of Digital Twin Assembly Line Performances Within Learning Factory Environment

As it is already known the main purpose of learning factory is to simulate a real industrial environment where academy personnel as well as professionals from industry can acquire skills and techniques that will help them to determine and solve problems which they have in everyday real conditions. Usually, physical resources, machines, tools, equipment etc. represent necessary prerequisites to build a tangible learning factory environment where practical investigations as similar as possible to real industrial conditions can be performed. Besides that, now-days digitalization era enables the creation of an industrial environment credible replica in the context of digital twin technology where virtual simulations and experimentations can be conducted and accordingly appropriate lessons can be derived. In this paper, Siemens Tecnomatix software was applied to create a digital twin assembly line within the Lean Learning Factory at University of Split. Assembly line consists of a few workstations where human-robot collaboration in the product assembly process will be presented. The digital twin of assembly line represents a platform where virtual experimentations were performed. Virtual experiments were conducted by varying different assembly process inputs and observing output values. The artificial intelligence (AI) fuzzy logic technique was applied to develop AI system that will be able to predict assembly process response depending on various input parameters. Digital twin AI forecasting approach presented in this paper represents a good foundation for further investigations in this area in order to establish a valuable digital twin learning factory landscape that will be supported by artificial intelligence technology.