IAHR Document Library

« Back to Library Homepage « Proceedings of the 39th IAHR World Congress (Granada, 2022)

Research on Practical Predictions of Dam Inflow Based on the Sparse Modeling Method

Author(s): Makoto Nakatsugawa; Tomohiro Sando; Yosuke Kobayashi

Linked Author(s): Makoto Nakatsugawa

Keywords: Dam inflow prediction; Sparse modeling method; Elastic Net; Multipurpose dam; Preliminary discharge

Abstract: This study proposes a method to obtain the dam inflow predictions applicable to multipurpose dams. In recent years, in response to the frequent floods that have occurred in Japan as well as worldwide, enhancing the flood control function of dams by leveraging preliminary discharge prior to a severe flood has attracted considerable attention. The flood control function of dams has occasionally been lost due to inflows beyond their accommodation capacity, and thus, functional enhancement is required. In addition, flood control through effective use of the dam’s storage function is desired for even water-use dams. However, because specific flood control operations have not been established, there is concern regarding the adverse effects of such actions on water use. For instance, water storage capacity would be lost if much rainfall was predicted but less rainfall occurred. Therefore, difficult judgments on the implementation of preliminary discharge values prior to a severe flood are required, and this increases the burden on managers. Additionally, for effective disaster prevention, it is also important to predict situations that have never been experienced. Thus, improving the accuracy of inflow prediction for multipurpose dams with functions of flood control as well as water use is an important endeavor. This study developed a regression model capable of predicting the inflow of dams based on an analysis using the Elastic Net, a sparse modeling method utilized to identify relationships between data from small amounts of information. In summary, Elastic Net is a practical method that is useful for making proper predictions, even in circumstances with limited information provided by observations.


Year: 2022

Copyright © 2024 International Association for Hydro-Environment Engineering and Research. All rights reserved. | Terms and Conditions