As modern power systems increasingly depend on demand-side flexibility, accurately modeling building multi-energy demand under uncertainty has become essential for achieving reliable and flexible grid operation. This study proposes an agent-based framework to conduct uncertainty-aware modeling of building multi-energy demand and to assess demand-side flexibility under different demand response mechanisms. Firstly, an agent-based modeling framework is established to connect occupant activities, electrical appliance usage, and building thermal dynamics, characterizing the explicit relationship between Markovian behavioral uncertainties and multi-energy demands. Secondly, an integrated thermal load model is constructed based on a resistance–capacitance network, coupled with behavior-driven internal heat gains and building morphology-driven shading and radiative microclimate conditions. Then, the flexibility potential of electrical and thermal loads is quantified at both individual and aggregated scales. Finally, the demand response flexibilities of the multi-energy loads were assessed under price-based self-scheduling and incentive-based centralized optimization scenarios. The results demonstrate that the proposed approach effectively captures behavior-driven uncertainties and their impacts on the temporal pattern and magnitude of building energy demand, as well as on the resulting demand-side flexibility. In addition, the proposed demand response strategies effectively reduce electricity costs and achieve peak shaving and valley filling, while maintaining schedulable flexibility within acceptable operational limits.
Wang et al. (Sat,) studied this question.