In the attached article, to be submitted to the IEEE ETFA 2020 conference, we introduce an open-source Python-based framework, called ModelConductor, for real-time input-output data management in digital twin applications. For data interexchange, ModelConductor supports periodic polling of a database (such as MySQL) or a TCP socket connection. For simulation model interfacing, ModelConductor supports Functional Mock-up Units (FMUs), as well as machine learning models implemented using the scikit-learn library. We demonstrate the use of ModelConductor in machine learning based on-line prediction of diesel engine nitrogen oxide NOx emissions, using simulated live data from a real engine.
The work is based on Panu Aho’s M.Sc. thesis. The open source library is located on Github.