Typhoon events occur frequently in Taiwan resulting in flood-related disasters. A well-operated reservoir can reduce the severity of a disaster. This study incorporates a genetic algorithm, a river hydraulic model, an artificial neural network and a simulation model of Tseng-Wen Reservoir to propose a real-time flooding operation model. The model includes two parts: an optimal flooding operation model (OFOM) and a reservoir inflow forecasting. Given an inflow condition, the OFOM is run based on the safety of the dam structure, reservoir flooding operation rule, and minimization of the downstream loss due to flood. A simple and robust model for reservoir inflow forecasting, which automatically chooses the most similar event from a typhoon event database as the future inflow, is developed. This study compares the model results with the real operations during Typhoons Sepat, Krosa, Kalmaegi, Fung-wong, Sinlaku, and Jangmi. This study compares the performances of the proposed model with the practical operation operated by the management center of Tseng-Wen Reservoir. The proposed model indicates shorter flooding duration in the downstream area. For example, the flood durations of the model output are 4 and 3 hours shorter during Typhoon Krosa and Sinlaku, respectively, than the practical operations.
The development of a real-time flooding operation model in the Tseng-Wen Reservoir
Yu-Wen Chen, Jui-Pin Tsai, Liang-Cheng Chang, Chih-Chao Ho, You-Cheng Chen; The development of a real-time flooding operation model in the Tseng-Wen Reservoir. Hydrology Research 1 June 2014; 45 (3): 490–503. doi: https://doi.org/10.2166/nh.2013.301
Download citation file: