Application of Deterministic Chaos and Neural Networks in Water Reservoir Management

摘要

The two-layer hierarchical technique with different prediction methods was applied to a part of the Wupper Reservoir System. The reservoir system consists of two reservoirs in series with additional inflow to the lower reservoir. The tasks of these reservoirs are flood control, recreation, hydropower and low flow augmentation with the aim of water quality improvement. Attention is focused on the implementation of anticipatory methods based on Deterministic Chaos and Artificial Neural Networks prediction of the inflows to the system that result in different operation rules. It is shown that the introduced optimisation concept improves considerably the system performance in comparison with the Standard Operation Rule.