Urban Energy Models Validation in Data Scarcity Context: Case of the Electricity Consumption in the French Residential Sector

Urban energy models (UEM) are useful to evaluate energy efficiency policies at district or city scale and to make the best decisions in terms of financial and environmental impacts. Probabilistic and simple physical models coexist in the UEM often with several dozens of parameters per building. Parameters coming from different databases with no consistency in the levels of uncertainty make the necessary validation difficult. This article proposes a method as a first attempt to validate UEM in a data scarcity context which requires only the annual electricity consumption of several carefully chosen cities. The first step aims to identify one parameter per energy end-use which has the largest impact on the energy consumption and the highest uncertainty. The second step aims to calibrate the model at country scale or at large territory scale to set parameters' values. The third step consists in selecting several cities (inside the territory previously used for calibration) to evaluate particular aspects of the UEM. The last step aims to simulate the calibrated model on the selected cities and to compare the simulation with the energy consumption information. The method is illustrated through the case of the electricity consumption in the French residential building sector with Smart-E, a bottom-up UEM tool.