An innovative humidity Enhancement-RC-mapping model with tailored identification framework for building HVAC demand response applications

Abstract

The increasing adoption of renewable energy sources highlights the pressing demand for greater grid flexibility. Under such circumstances, grey-box modeling offers a practical approach for estimating building flexibility and facilitating demand response control. However, existing models still present limitations: First, conventional RC models are primarily designed to capture sensible heat dynamics, while failing to represent latent heat dynamics. In fact, latent heat changes can indirectly affect indoor temperature variation during demand response. Second, existing RC models primarily focus on buildings' thermal storage while overlooking air conditioning systems' thermal storage. To address these limitations, this study proposes a Humidity Enhancement (HE)-RC-Mapping model. This model introduces a formulation for the ratio of latent to sensible heat during demand response and considers the air conditioning systems' thermal storage, besides the buildings’ thermal storage. Dedicated to the proposed model, a tailored parameter identification framework incorporating a multi-condition stepwise identification strategy and a dual-objective function is introduced. Additionally, to ensure a more equitable and rigorous assessment of model performance under different temperature fluctuation ranges during demand response, this study proposes a new performance evaluation metric, the Relative RMSE Index (RRI). Experiments in a large public building demonstrate that the proposed model significantly outperforms the conventional RC model in predicting indoor temperature under cooling load reduction during demand response. Specifically, the RMSE of indoor temperature prediction is reduced from 0.737 °C to 0.118 °C, while the RRI is reduced from 173.24 % to 21.31 %.