Abstract
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. In this paper, an ultra-fast one-minute resolution Hybrid Machine Learning Model (HMLM) is proposed as part of a novel contribution in the field of residential physics-based smart home surrogate modeling. Emulation of white box models, or digital twins, with editable parameters through machine learning (ML) meta-modeling serves as an alternative to wide-spread experimental big data collection. The HMLM employs combined k-means clustering with multiple linear regression (MLR) to emulate minutely HVAC power timestep-to-timestep with satisfactory nRMSE error of less than 10% across an entire year test set. An approach is also described to characterize HVAC systems as generalized storage (GES) devices to unify household appliance and virtual power plant (VPP) controls in accordance with industry Communication Technology Association (CTA) 2045 protocol and Energy Star metrics. Synthetic output data from experimentally calibrated EnergyPlus models for three existing smart homes managed by the Tennessee Valley Authority (TVA) is employed in residential case studies and a discussion provided for the large-scale application to hundreds of homes.
Original language | English |
---|---|
Journal | IEEE Transactions on Industry Applications |
DOIs | |
State | Accepted/In press - 2024 |
Bibliographical note
Publisher Copyright:© 1972-2012 IEEE.
Keywords
- ANSI/CTA-2045-B
- Demand Response (DR)
- Energy Star
- Energy Take
- General Energy Storage (GES)
- Heating Ventilation and Air Conditioning (HVAC)
- Home Energy Management (HEM)
- Machine Learning (ML)
- Smart Grid
- Smart Homes
- Surrogate Model
ASJC Scopus subject areas
- Control and Systems Engineering
- Industrial and Manufacturing Engineering
- Electrical and Electronic Engineering