Author(s): Kailin Yang; Tao Wang; Yongxin Guo
Linked Author(s): Kailin Yang, Tao Wang
Keywords: Ice condition forecast; Yellow River Inner Mongolia Reach; Artificial Neural; Network; Expert system; GIS technique
Abstract: An Artificial Neural Network Expert System is developed for ice condition forecast of the Yellow River Inner Mongolia Reach. The system is based on GIS technique while its kernel is BP network arithmetic. The key parameter forecast includes the change of the water temperature with time and the beginning dates of the ice runs appearing, freeze-up and breakup, taken account of the hydraulic factors and thermodynamic factors, such as air temperature, water temperature and volumetric discharge as well as water stage and ice concentration on water surface. Using the expert system, the ice conditions in the 2004-2005 ice period and the 20052006 ice one, such as the water temperature and the beginning dates of the ice run, freeze-up and break-up etc, were forecast on-line. It has been verified that above 75 % of the forecast results are qualified or excellent while the forecast period is in 15 days.
Year: 2008