Séminaire équipe MAGE

Title: Calibration and uncertainties of a semi-virtual environment for  performance evaluation of technologies for building

Keywords: Building Simulation, Hardware in the Loop, Sequence Reduction, Uncertainty Analysis

Supervisors : Stéphane PLOIX (G-Scop), Benoit DELICNHANT (G2Elab), Cristian MURESAN (ENGIE)

Abstract: Hardware in Loop testing (HIL) under dynamic conditions is becoming a very interesting area of study especially in the field of building related energy studies. The main problem with such HIL simulations are time constraints if at all one needs to test the products annual efficiency. In such cases a set of representative days can be chosen from a given year for which one can obtain a value of efficiency from the short testing sequence, which could be extrapolated to a whole year. The formation of a short testing sequence  in real time forms the primary objective of the thesis which is the time. In order to achieve this, different clustering algorithms are studied and used accordingly finally a new Meta-Heuristic method is proposed depending upon industrial need. After obtaining the annual efficiency of the product it is very important to quantify the uncertainty related to the test bench that was used for the HIL. The main part of the analysis includes both the parametric and experimental uncertainty of the test bench. In order to achieve this one can develop the model of the HIL simulation setup so that parametric uncertainties can be quantified by Monte Carlo Simulations(MC Simulations) as an initial study to find if at all the variations in the test bench parameters could cause considerable uncertainty on the efficiency being measured. Finally depending upon the  experiments one can quantify the experimental uncertainties. In short the idea is to make a short testing sequence to measure the seasonal / yearly efficiency of a product and then quantify the uncertainty related to the Hardware in Loop testing.