Estimating air conditioning energy consumption of residential buildings using hourly smart meter data

Accurate information on air conditioning (AC) energy consumption in residential buildings is critical for enhancing the efficiency of demand response (DR) programs and promoting energy conservation. However, financial constraints and privacy issues hinder the direct measurement of residential AC energy consumption, necessitating indirect estimation from smart meter data. To address this, this paper proposes a quantile change point (QCP) model for disaggregating hourly total energy consumption, recorded by smart meters, into AC energy consumption and base energy consumption. The QCP model synergistically integrates quantile theory with a change point model, enabling a comprehensive capture of the variability in base energy consumption and the temperature sensitivity of AC energy consumption. Rather than assuming a constant base energy consumption, the QCP model dynamically estimates AC energy consumption by subtracting the varying base energy consumption from total energy consumption. Numerical simulations based on ground truth data show that the proposed QCP model reduces the root mean square error of AC energy consumption estimation by over 20% compared to traditional methods. This improvement facilitates optimized participant selection and fair load adjustment, allowing for effective implementation of DR programs.