Analyzing End-use Electricity Consumption Data to Investigate Residential Buildings' Consumption Patterns
The shares of total United States' retail sales of electricity in the residential sector in 2016 was 38%, and it is projected that total U.S. electricity use grows by an average of less than 1% per year from 2016 to 2040. Therefore, estimating the electricity consumption and a load of a community or neighborhood for planners, utilities, and designers become essential. Fortunately, installation of intelligent devices and smart meters lead to collecting electricity consumption data for further analysis. This dissertation implements a data analysis methodology to collect, utilize and analyze patterns of appliance electricity consumption in addition to Electric Vehicles (EV) and Photovoltaic (PV) panels for a large dataset of single-family and multi-family homes in Austin, Texas. Further, the study characterizes the significant factors that contribute to the electricity consumption in residential buildings, namely building characteristics, socio-economic factors, occupant behavior and climatic factors. Also, due to the increase in the EV market, the study also considers the EV as a plug load in the future electricity consumption portfolio of residential buildings. The micro-environment of residential buildings is created to make sense of the data by determining whether they have EV, PV, or both in more detail regarding their electricity consumption patterns and correlation with the mentioned four factors. Finally, it advances the predictive model capability by including the mentioned four factors and their correlations on appliance electricity consumption in buildings. The predictive models of the appliance electricity consumption based on socio-economic and climatic variables provide insights to the utilities and policymakers regarding the price scheme, electricity efficiency, and demand-response programs.