Industry4: Optimization of Digital Twins in Petrochemistry
The digital twin concept is widely applied in petrochemistry. A digital twin basically models the behaviour of production plants at the physical and chemical level. From our perspective, the presence of a digital twin makes it possible to enhance real world production by involving in silico optimization. We work on an optimization framework that connects to external chemical modelling software via a software interface (COM). It means that it is not necessary to work on a modeling algorithm. The framework conducts optimization on the models by instructing the modeling software to adjust the model parameters. The goal of this project is to conduct research on optimization algorithms. The bottleneck is in the modeling phase, as it is a calculation intensive task. We would like to elaborate algorithms that try to reduce the amount of model calculations while manages to identify the optimum of the model.