Lawrence Berkeley National Laboratory

The expansion of commercial building demand response (DR) as a demand-side management resource for the electric grid necessitates new decision support resources for customers to rapidly assess the benefit-risk tradeoffs of candidate load flexibility strategies. This work develops surrogate models of load flexibility impacts on office building electricity demand and indoor temperature. The surrogate models are fit to a large synthetic database generated via whole building simulations of multiple flexibility strategies under a variety of conditions; the models are translated to a Bayesian framework to allow straightforward communication of uncertainty and parameter updating given new evidence. The strong predictive performance of the models underscores their potential utility in guiding DR decision-making in office settings.