Data-Driven Evaluation of Building Demand Response Capacity

TL;DR

This paper introduces a data-driven method to estimate building demand response capacity using sensor data, enabling more accurate and confidence-based assessments for participation in demand response programs.

Contribution

It presents a novel probabilistic formula and a framework that links sensor data to demand response capabilities, improving over manual audit methods.

Findings

Identifies key control parameters for demand response

Provides insights into site-specific demand response strategies

Demonstrates effectiveness in two real-world buildings

Abstract

Before a building can participate in a demand response program, its facility managers must characterize the site's ability to reduce load. Today, this is often done through manual audit processes and prototypical control strategies. In this paper, we propose a new approach to estimate a building's demand response capacity using detailed data from various sensors installed in a building. We derive a formula for a probabilistic measure that characterizes various tradeoffs between the available demand response capacity and the confidence level associated with that curtailment under the constraints of building occupant comfort level (or utility). Then, we develop a data-driven framework to associate observed or projected building energy consumption with a particular set of rules learned from a large sensor dataset. We apply this methodology using testbeds in two buildings in Singapore: a unique net-zero energy building and a modern commercial office building. Our experimental results identify key control parameters and provide insight into the available demand response strategies at each site.