Data-Driven Residential Load Modeling and Validation in GridLAB-D

Peter Gotseff, Blake Lundstrom

Research output: Contribution to conferencePaperpeer-review

7 Scopus Citations


Accurately characterizing the impacts of high penetrations of distributed energy resources (DER) on the electric distribution system has driven modeling methods from traditional static snap shots, often representing a critical point in time (e.g., summer peak load), to quasi-static time series (QSTS) simulations capturing all the effects of variable DER, associated controls and hence, impacts on the distribution system over a given time period. Unfortunately, the high time resolution DER source and load data required for model inputs is often scarce or non-existent. This paper presents work performed within the GridLAB-D model environment to synthesize, calibrate, and validate 1-second residential load models based on measured transformer loads and physics-based models suitable for QSTS electric distribution system modeling. The modeling and validation approach taken was to create a typical GridLAB-D model home that, when replicated to represent multiple diverse houses on a single transformer, creates a statistically similar load to a measured load for a given weather input. The model homes are constructed to represent the range of actual homes on an instrumented transformer: square footage, thermal integrity, heating and cooling system definition as well as realistic occupancy schedules. House model calibration and validation was performed using the distribution transformer load data and corresponding weather. The modeled loads were found to be similar to the measured loads for four evaluation metrics: 1) daily average energy, 2) daily average and standard deviation of power, 3) power spectral density, and 4) load shape.

Original languageAmerican English
Number of pages6
StatePublished - 9 May 2017
Event9th Annual IEEE Green Technologies Conference, GreenTech 2017 - Denver, United States
Duration: 29 Mar 201731 Mar 2017


Conference9th Annual IEEE Green Technologies Conference, GreenTech 2017
Country/TerritoryUnited States

Bibliographical note

Publisher Copyright:
© 2017 IEEE.

NREL Publication Number

  • NREL/CP-5D00-67608


  • Distribution system
  • High penetration
  • Load modeling
  • Quasi-static time series
  • Residential loads


Dive into the research topics of 'Data-Driven Residential Load Modeling and Validation in GridLAB-D'. Together they form a unique fingerprint.

Cite this