Addressing Big Data Issues in Load Shape Profiling for Utility Customers

This paper presents two Advanced Metering Infrastructure (AMI) clustering methods, one in the time domain and another in the frequency domain, to analyze the large-scale smart meter data collected by an Arizona utility provider. These methods aim to efficiently generate load shapes across a large number of customers to inform action on Time of Use, Energy Efficiency, Demand Response, Electric Vehicle, and Distributed Energy Resources. The proposed methods are evaluated based on their robustness, computational complexity, and practical utility application. Results show that the algorithms are effective in analyzing full-scale energy consumption data and can be directly applied by utility providers for industry application. Furthermore, the results indicate possible routes for improvement over existing AMI clustering methods by clustering daily load shapes and using dynamic time warping distance.