“20th PSCC 2018 papers submission and review platform

Full Program »

A Machine Learning Method Creating Network Models Based on Measurements

Network models are essential to perform power flow analyses. In this paper a supervised regression method creating simplified network models using measurements is presented. It is an iterative method creating a network model by minimizing the difference between measurements and obtained power flow using measured net-exchanges for each node. The method is tested in a case study for the Nordic Synchronous Area considering each bidding zone as a node. The simplified network model is created using a training set and is validated using various validation methods. The obtained reactances are not correct in absolute terms; however results indicate that the obtained power flows using the created network model are accurate enough for several different applications.

Author(s):

Martin Nilsson    
Svenska kraftnät
Sweden

Lennart Söder    
KTH Royal Institute of Technology
Sweden

Jon Olauson    
Svenska kraftnät
Sweden

Lars Nordström    
KTH Royal Institute of Technology
Sweden

Robert Eriksson    
Svenska kraftnät
Sweden

Göran Ericsson    
Svenska kraftnät
Sweden

 

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