Simulation model calibration for a multi-purpose building on an hourly basis

Abstract

When developing digital twins for buildings, the calibration of simulation models on an hourly basis is essential to maintain the fidelity of the virtual representation and to enable real-time monitoring and analysis in the operational phase. Achieving such a high accuracy in building performance simulations (BPS) calls for novel calibration strategies with enhanced effectiveness. In this regard, this paper outlines a calibration strategy that makes use of hourly measurements to improve the fidelity of energy simulation models. The proposed approach includes a hierarchical structure involving data acquisition and management, setting unknown weather parameters, sensitivity analysis, calibration of fixed parameters, and hourly calibration of dynamic variables. Here, acquired data from the building’s sensor are refined to enable hourly demand calibration, and an accurate weather data file is gathered. Next, sensitivity analysis is conducted to identify the key fixed parameters for the calibration process. Following the calibration of these fixed parameters, the final level involves the calibration of dynamic variables to achieve a robust hourly agreement between simulated and measured data. The developed strategy is implemented in a multi-purpose building located in the Aalto University campus in Finland. The building is simulated as a simplified five-zone model developed in the whole-building simulation software IDA-ICE, including various educational sections, workshops, a shopping center, and a metro station. Sensors and meters are used to measure the hourly indoor air temperature by zone, whereas the calibration aims at minimizing the difference between measured and simulated heating and cooling energy demands. In conclusion, the proposed calibration strategy appears to be successful in facilitating hourly synchronization between simplified simulation models and multi-purpose buildings.