High-resolution residential feeder load characterization and variability modelling

Abstract

Data from of a highly instrumented residential feeder in Ota City, Japan was used to determine 1 second load variability for the aggregation of 50, 100, 250, and 500 homes. The load variability is categorized by binning the data into seasons, weekdays vs. weekends, and time of day to create artificial sub-15-minute variability estimates for modeling dynamic load profiles. An autoregressive, AR(1) function along with a high pass filter was used to simulate the high resolution variability. The simulated data were validated against the original 1-second measured data.