Usage analysis for smart meter management

Abstract

Smart meters gather utility usage data, such as water, electricity and gas readings, by remote reporting. The flood of usage data obtained from each meter in realtime or near real-time enable data analytics and optimization tool to support smart meter management. Predictive usage analytics can provide significant benefits to both the utilities and customers. We propose statistical approaches for meter anomaly detection, usage demand forecasting and association analysis for utility companies. These analyses provide efficient ways to detect malfunctioning meters, optimize water supply in the future and understand the association factors that drive meter failures and water demand. We illustrate our methodology using the automated meter reading (AMR) database from a water utility customer. © 2011 IEEE.