Forecasting plug-in electric vehicles load profile using artificial neural networks

Delshad Panahi, Sara Deilami, Mohammad A. S. Masoum, Syed M. Islam

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

42 Citations (Scopus)

Abstract

Plug-in electric vehicles (PEVs) are becoming very popular these days and consequently, their load management will be a challenging issue for the network operators in the future. This paper proposes an artificial intelligence approach based on neural networks to forecast daily load profile of individual and fleets of randomly plugged-in PEVs, as well as the upstream distribution transformer loading. An artificial neural network (ANN) model will be developed to forecast daily arrival time (Ta) and daily travel distance (Dtr) of individual PEV using historical data collected for each vehicle in the past two years. The predicted parameters are then will be used to forecast transformer loading with PEV charging activities. The results of this paper will be very beneficial to coordination and charge/discharge management of PEVs as well as demand load management, network planning and operation proposes. Detailed simulations are presented to investigate the feasibility and accuracy of the proposed forecasting strategy.

Original languageEnglish
Title of host publication2015 Australasian Universities Power Engineering Conference (AUPEC)
Place of PublicationPiscataway, NJ
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages1-6
Number of pages6
ISBN (Electronic)9781479987252, 9781479987245
DOIs
Publication statusPublished - 2015
Externally publishedYes
Event25th Australasian Universities Power Engineering Conference, AUPEC 2015 - Wollongong, Australia
Duration: 27 Sept 201530 Sept 2015

Other

Other25th Australasian Universities Power Engineering Conference, AUPEC 2015
Country/TerritoryAustralia
CityWollongong
Period27/09/1530/09/15

Fingerprint

Dive into the research topics of 'Forecasting plug-in electric vehicles load profile using artificial neural networks'. Together they form a unique fingerprint.

Cite this