A Cross-Domain Data-driven Approach to Analyzing and Predicting the Impact of COVID-19 on the U.S. Electricity Sector

点击翻译按钮获取中文摘要

This project aims at developing a cross-domain, data-driven approach to tracking and measuring the impact of the ongoing COVID-19 pandemic on the U.S. electricity sector. The COVID-19 crisis has gone beyond anybody’s wildest imagination and is turning out to be a once-in-a-century societal challenge. As the lifeblood of civil society and a key enabling infrastructure system, the electricity sector is quickly adjusting to the new normal, and it is crucial to understand the severity and the resiliency of the grid in response to disruption caused by COVID-19. This project substantiates a data-driven, science-based approach to evaluating the impact of various policy options on the operation of the electric energy infrastructure. Once successfully pursued, the project will provide much needed planning decision support for the electricity sector. The research program will be tightly coupled with an educational effort to train future leaders in the electricity and public health sectors. The research team has engaged female and African American students in building the preliminary version of this data hub and to continue research on the project. The team is also working with the industry members to provide training materials to a broad set of industry affiliatesThe goal of this project is to develop a first-of-its-kind cross-domain data hub and data-driven analysis of the COVID’s impact on the U.S. electricity sector. The approach is to 1) build a comprehensive open-access data hub with quality monitoring and daily updates, 2) quantify the sensitivity of electricity consumption with respect to social distancing and public health policies by using Ensemble Backcast Models and Restricted Vector Autoregression (VAR), and 3) construct a predictive model for the electricity sector considering social distancing policies and mobility in different sectors. The contribution of this project is four-fold. First, this is a first-of-its-kind data hub that combines otherwise unrelated domains of data like electricity markets, public health, and mobility data into a coherent infrastructure. A machine learning-based cleaning and pre-processing technique is proposed. Second, a statistical approach is proposed to quantify the unique impact of a public health crisis on the electricity sector. This entails building and analyzing novel statistical models that encompass societal mobility and public health data into the regression analysis of electricity consumption in major hot spots in the U.S. Third, a novel concept of elasticity of power consumption with respect to societal mobility is proposed and substantiated as an effective indicator of the power consumption as a function of social distancing policy measures. Last but not least, this project will combine all the above three innovations to create a first-of-its-kind predictive model of electricity sector as a function of social distancing policies and public health data. Drawing upon expertise from biostatistics and electric power engineering, this project will contribute to the cross-fertilization between the public health and electric energy sectors.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.