Performance analysis of simple grey-box RC models for simulating strategic heating reductions

The increasing reliance on air-water heat pumps for residential space heating poses significant challenges to electricity grids, particularly during winter months when both electricity production and heat pump efficiency are reduced. This research investigates the potential for peak load shifting through strategic heating reductions in dwellings, aimed at preventing grid blackouts during peak winter demand periods. Specifically, the study evaluates the impact of temporarily turning off heating systems until a lower setpoint temperature is reached.

Estimating the potential load shift for a building stock using detailed simulation tools, such as Modelica, was found impractical due to excessive computation times. Therefore, the use of simple RC models that represent a buildings thermal dynamics using a combination of resistors and capacitors, is investigated as an alternative.

The primary objective of this study is to evaluate the performance and reliability of the TiTe model in simulating the specified heating reduction scenarios for the building stock. The TiTe model, an RC model with two states representing the interior temperature (Ti) and building envelope temperature (Te), has shown strong performance in previous research.

The TiTe model was trained using a grey-box modelling approach, based on a two-week Modelica simulation incorporating a pseudorandom binary sequence signal for the heat production to optimize parameter identification.

The evaluation of this model’s performance is done based on one-week simulations, where the heat pump in a building is turned off until a lower setpoint for the interior temperature is reached. During these simulations no model-correction is applied. Ultimately, TiTe’s performance will be benchmarked against detailed Modelica simulations by calculating the Root Mean Square Error.

To be able to evaluate the performance on the level of the Flemish building stock these simulations are performed on ten distinct housing types, with six insulation levels, over three years with different temperature profiles and considering both floor heating and radiator heating systems.

Preliminary findings indicate the TiTe model’s potential of approximating the results of more detailed simulations, thus offering a viable solution for large-scale energy use studies. Although no specific results are present on this date, we are convinced that this research will underscore the importance of efficient simulation tools in developing strategies to enhance energy resilience and sustainability in residential buildings.