“20th PSCC 2018 papers submission and review platform

Full Program »

Selecting and Evaluating Representative Days for Generation Expansion Planning

This work applies feature engineering and machine learning to design a rigorous algorithm to select representative daily profiles of net load for Generation Expansion Planning (GEP). This algorithm incorporates Principal Component Analysis and clustering techniques to produce multiple sets of representative days. Guidelines to determine the proper number of representative days and the number of times the clustering process should be repeated are also proposed. These sets are then assessed using a rolling horizon unit commitment (RHUC) on a test system. Using RHUC as a metric to select the best set of representative days ensures that the selection is based on a criterion that closely approximates the operation of a power system with a large penetration of renewable generation.

Author(s):

Abeer Almaimouni    
University of Washington
United States

Atinuke Ademola-Idowu    
University of Washington
United States

J. Nathan Kutz    
University of washington
United States

Ahlmahz Negash    
Tacoma Power
United States

Daniel Kirschen    
University of Washington
United States

 

Powered by OpenConf®
Copyright ©2002-2014 Zakon Group LLC