Hypertemporal Imaging of NYC Grid Dynamics
Abstract
Hypertemporal visible imaging of an urban lightscape can reveal the phase of the electrical grid granular to individual housing units. In contrast to in-situ monitoring or metering, this method offers broad, persistent, real-time, and non-permissive coverage through a single camera sited at an urban vantage point. Rapid changes in the phase of individual housing units signal changes in load (e.g., appliances turning on and off), while slower building- or neighborhood-level changes can indicate the health of distribution transformers. We demonstrate the concept by observing the 120 Hz flicker of lights across a NYC skyline. A liquid crystal shutter driven at 119.75 Hz down-converts the flicker to 0.25 Hz, which is imaged at a 4 Hz cadence by an inexpensive CCD camera; the grid phase of each source is determined by analysis of its sinusoidal light curve over an imaging "burst" of some 25 seconds. Analysis of bursts taken at ~15 minute cadence over several hours demonstrates both the stability and variation of phases of halogen, incandescent, and some fluorescent lights. Correlation of such results with ground-truth data will validate a method that could be applied to better monitor electricity consumption and distribution in both developed and developing cities.
Cite
@article{arxiv.1611.04633,
title = {Hypertemporal Imaging of NYC Grid Dynamics},
author = {Federica B. Bianco and Steven E. Koonin and Charlie Mydlarz and Mohit S. Sharma},
journal= {arXiv preprint arXiv:1611.04633},
year = {2016}
}
Comments
This paper uses astronomical techniques applied to the study of urban lights. This research is reproducible but the data access is restricted. A Github repository contains all code supporting this research as well as additional material: https://github.com/fedhere/detect120