Skip to main navigation Skip to search Skip to main content

Solar forecasting by K-Nearest Neighbors method with weather classification and physical model

  • State University of New York (SUNY)

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

39 Scopus citations

Abstract

With the increasing penetration of solar photovoltaic (PV) generation in the power system, the reliability of the distribution system and efficiency of PV systems have garnered increasing attention in recent years. Forecasting the PV output is one way to decrease the uncertainty of such power systems. In this study, we present a K-Nearest Neighbors algorithm based forecasting model, which can provide the estimated PV output by utilizing numerical weather and solar irradiance prediction data. This forecasting model also includes a weather condition classification process and a physical model of PV units. Numerical results are evaluated by using data from an existing 128kW rooftop PV system.

Original languageEnglish
Title of host publicationNAPS 2016 - 48th North American Power Symposium, Proceedings
EditorsDavid Wenzhong Gao, Jun Zhang, Amin Khodaei, Eduard Muljadi
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509032709
DOIs
StatePublished - Nov 17 2016
Event48th North American Power Symposium, NAPS 2016 - Denver, United States
Duration: Sep 18 2016Sep 20 2016

Publication series

NameNAPS 2016 - 48th North American Power Symposium, Proceedings

Conference

Conference48th North American Power Symposium, NAPS 2016
Country/TerritoryUnited States
CityDenver
Period09/18/1609/20/16

Keywords

  • k-nearest neighbors
  • photovoltaic power system
  • physical model
  • solar forecasting
  • weather condition classification

Fingerprint

Dive into the research topics of 'Solar forecasting by K-Nearest Neighbors method with weather classification and physical model'. Together they form a unique fingerprint.

Cite this