The Metroscope methodology uses Modelica based digital twins of power units as a basis. Such component-based models need to be calibrated to find the value of the component parameters (heat exchange coefficients, efficiencies…), to fit the process best. Such a calibration is extremely dependant on the total power produced by the installation : at 100% or at 30% results are extremely different. Our method today focuses on full load.
The goal of the internship is to extend Metroscope’s solution to a new perimeter : improving Metroscope’s solution for partial load. The goal being to study the effect of load variation on parameter calibration, then design a multi-load calibration, suitable for a broad range of operating conditions.
Once introduced to Metroscope’s solution for modeling industrial installations, you will :
You will work on a real usecase from one of our clients. If time allows it, you will also explore the effect of other boundary conditions (for instance external air temperature) on the calibration.
The tools that you will use for this work include Modelica (and the use of our internal modeling library) and Python (for calibration and result analysis).