PEAK: Software for Prepaid Energy Awareness Kit

Presenter: Thomas C. Murphy

Faculty Sponsor: Maitreyee Marathe

School: UMass Amherst

Research Area: Electrical and Computer Engineering

Session: Poster Session 6, 4:15 PM - 5:00 PM, Auditorium, A32

ABSTRACT

Low-to-moderate income households enrolled in prepaid electricity programs face unique challenges managing energy under strict budget constraints. Unlike traditional postpaid billing, prepaid customers must purchase credits in advance and risk immediate and automatic disconnection when balances deplete, making energy access precarious and difficult to plan around. Without automated management tools, households may struggle to track consumption and anticipate costs, increasing the likelihood of unexpected and disruptive service interruptions. Interruptions can disrupt many aspects of life, such as daily cooking and hygiene. This thesis implements and compares three algorithms that manage usage of household loads according to user-defined priorities to maximize availability of critical loads while respecting budget limits. Each algorithm represents a distinct computational approach to balancing competing energy demands under real-world constraints such as time-of-use pricing and variable household consumption patterns. The primary contribution is three prepaid energy management systems, each deploying one algorithm on resource-constrained embedded hardware suitable for installation in such households. The performance of each system will be evaluated by comparing algorithm outputs produced on high-accuracy, non-resource-constrained reference machines against those produced on the embedded systems, measuring tradeoffs between computational efficiency and solution quality attributable to hardware constraints. These results will help assess the feasibility of deploying energy management solutions for under served communities.

RELATED ABSTRACTS