BASIN CLUSTERING OF TURKEY BY USE OF MONTHLY STREAM-FLOW DATA


ARSLAN Y., Birturk A. N. , EREN S.

IEEE 14th International Conference on Machine Learning and Applications ICMLA, Florida, United States Of America, 9 - 11 December 2015, pp.1169-1174 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/icmla.2015.82
  • City: Florida
  • Country: United States Of America
  • Page Numbers: pp.1169-1174
  • Middle East Technical University Affiliated: Yes

Abstract

Security of the energy supply is an important topic in energy field. It has two parts which are supply and demand. To ensure that demand is met, the supply at the specific time points has to be known or predicted. Supply is predicted by use of seasonal, yearly and regional information. The streamflow dataset resolution is monthly and it supplies the yearly and seasonal information. The only missing part for supply prediction is the regional information. The aim of this study to find the basin based regional clustering of the streams and correspondingly hydroelectric power plants. In this paper, 14 out of 26 basins of Turkey, which contain over 80% of the hydroelectric power plants of Turkey in Dispatcher Information System, are clustered by use of different clustering techniques. Results are visualized on Turkey basin map.