Quantifying and Predicting Residential Building Flexibility Using Machine Learning Methods

TL;DR

This work tackles the problem of quantifying and forecasting residential building load flexibility to support grid operations amid increasing distributed energy resources. It introduces two metrics—power flexibility bounds and energy flexibility duration—and assesses multiple machine learning approaches, with LSTM delivering the best performance for day-ahead power flexibility forecasts at horizons of $4$ and $24$ hours. Energy flexibility forecasting proves more challenging, particularly for HVAC, with a strong dependence on external conditions and rapid, non-linear dynamics; forecasting water heater flexibility yields the most promising results, around $20$ minutes of maintainable duration at short horizons ($2$ hours). The findings highlight the need for physics-informed (gray-box) models and potential future inclusion of energy storage to robustly harness residential flexibility for grid services.

Abstract

Residential buildings account for a significant portion (35\%) of the total electricity consumption in the U.S. as of 2022. As more distributed energy resources are installed in buildings, their potential to provide flexibility to the grid increases. To tap into that flexibility provided by buildings, aggregators or system operators need to quantify and forecast flexibility. Previous works in this area primarily focused on commercial buildings, with little work on residential buildings. To address the gap, this paper first proposes two complementary flexibility metrics (i.e., power and energy flexibility) and then investigates several mainstream machine learning-based models for predicting the time-variant and sporadic flexibility of residential buildings at four-hour and 24-hour forecast horizons. The long-short-term-memory (LSTM) model achieves the best performance and can predict power flexibility for up to 24 hours ahead with the average error around 0.7 kW. However, for energy flexibility, the LSTM model is only successful for loads with consistent operational patterns throughout the year and faces challenges when predicting energy flexibility associated with HVAC systems.

Figures (8)

• Figure 1: Flexibility bounds for an example summer and winter day
• Figure 2: Procedure for calculating the building power flexibility bounds
• Figure 3: Energy flexibility for an example summer and winter day, 3d plot
• Figure 4: Procedure for calculating the energy flexibility
• Figure 5: Forecasting error for flexibility bounds over various forecast horizons