Calibration of the Simulation model of the HERB building in Bologna in its Present State

In the framework of the EU-funded Project HERB (Holistic Energy-efficient Retrofitting of residential Buildings), the University of Bologna and the Municipality of Bologna are performing the energy retrofitting of a detached house with 6 apartments in Bologna. The annual saving of primary energy obtained by the retrofitting will be assessed through dynamic simulations of the building before and after retrofitting, performed through a model implemented in TRNSYS 17 and calibrated by comparison with monitoring results. In this paper, the calibration of the building model in its present state is described. The 3-D hourly simulation model of the building was obtained starting from 3-D drawings performed through Google SketchUp. Four thermal zones were considered for each small apartment (bedroom, bathroom, kitchen-living room, entrance) and five thermal zones for a larger apartment. The characteristics of the building enclosure were determined by accurate inspections and tests, which included infrared thermography, blower door tests and direct measurements of the transmittance of the external vertical wall. The building was monitored during the heating season 2013-2014. The use of natural gas for heating was determined by periodic readings of the gas meter placed in the distribution duct to the central boiler, hourly values of the internal air temperature and relative humidity in each room were recorded, while hourly weather data were taken from the Bologna Urbana Weather Station, close to the building. The internal heat gains were evaluated through measurements of the use of electricity and of natural gas in single apartments. A comparison between the measured values of the primary energy use for heating during 16 time intervals and those determined by dynamic simulation revealed that the simulation model widely fulfils the accuracy requirements stated by ASHRAE Guide 14 and yields the measured total energy use for heating during the season considered with an accuracy better than 0.4%.