The Application of the Oxygen Saturation Level Prediction in Dam Lower Reaches Based on the Principle Component RBF Network Model

The dam operation has brought great benefits in flood control, power generation, navigation, irrigation and so on, but also has some effects on the ecological environment of reservoir area and lower reaches of the dam. When the dam discharges water, huge flood takes a lot of dissolved oxygen and rushes down to the lower riverbed, which makes the total dissolved oxygen supersaturated especially near the dam area.This has negative effect on downstream aquatic life. The gas bubble disease of the fish is the typical symptoms caused by the dissolved oxygen supersaturation. For the lower reaches of Three Gorges and Gezhouba dam region, this paper established the principle component RBF network model of dissolved oxygen saturation level, utilized the test data for network training and learning, then simulated and predicted the dissolved oxygen saturation level in lower reaches of the dam at different Sluicing flow, compared with numerical simulation finally. The conclusion shows that the principal component RBF network model gets closer to the dissolved oxygen saturation level of specific areas near the dam downstream. Curves of RBF reflects the sluicing flow´s effect on dissolved oxygen saturation level more effectively than numerical simulation.