scope:

ABSTRACT

This paper investigates the accuracy of grey-box models for the quantification of the energy flexibility of a school building. The developed models are based on a data-driven grey-box model formulation and calibrated with measured data. These models, which are suitable for the context of demand-side management in smart grids, are used to quantify energy flexibility of the building.

The case study school (located in Quebec, Canada) has been in operation since 2017, and it is an all-electric building. Several sources of high-resolution data, including the building automation system (BAS) and dedicated electrical sub-meters, are available. This paper consists of three components: (1) Collection and analysis of real measured data in the school building, (2) development of thermal network Resistance- Capacitance (RC) models and calibration of these models with the collected data, (3) quantification of the energy flexibility associated with the electricity demand used for heating.

In this paper, system identification is carried out to identify suitable reduced-order models. Therefore, grey-box models from first-order to third-order model are identified on results from measured data. In this school, each classroom receives thermal energy from a geothermal water to air heat pump and has Proportional-Integral control (PI) in the local-loop control of room air temperature.

The results show that simplified third-order thermal network can capture the thermal dynamics of the building with good accuracy. This paper also illustrates how lumped parameter thermal network models of low order can be utilized to calculate the Building Energy Flexibility Index (BEFI). Results show that by lowering the setpoint temperature 2°C when required by the grid, it is possible to provide between 5-30% reduction in average power demand for the grid for duration of 1 to 8 hours.