Building modeling, specifically heating, ventilation, and air conditioning (HVAC) load and equivalent energy storage calculations, represent a key focus for decarbonization of buildings and smart grid controls. Widely used white box models, due to their complexity, are too computationally intensive to be employed in high resolution distributed energy resources (DER) platforms without simulation time delays. In this paper, an ultra-fast one-minute resolution Hybrid Machine Learning Model (HMLM) is proposed as part of a novel procedure to replicate white box models as an alternative to widespread experimental big data collection. Synthetic output data from experimentally calibrated EnergyPlus models for three existing smart homes managed by the Tennessee Valley Authority is used. The HMLM employs combined k-means clustering and multiple linear regression (MLR) models to predict minutely HVAC power with satisfactory nRMSE error of less than 10% across an entire year test set. An approach is provided to characterize HVAC systems through the newly proposed hybrid model as a generalized storage (GES) device suitable for DER control and event types in accordance with the Communication Technology Association (CTA) 2045 standard and Energy Star metrics such as “energy take”, currently developed by industry, to unify household appliance controls.
Digital Object Identifier (DOI)
Alden, Rosemary E.; Jones, Evan S.; Gong, Huangjie; Hadi, Abdullah Al; and Ionel, Dan, "Digital Twin for HVAC Load and Energy Storage based on a Hybrid ML Model with CTA-2045 Controls Capability" (2022). Power and Energy Institute of Kentucky Faculty Publications. 85.