[5F1-GS-10h] AI application

9:30 AM - 9:45 AM JST(12:30 AM - 12:45 AM UTC)

[5F1-GS-10h-03] A Preliminary Study on the Utilization of Neural ODE for Improving Dam Inflow Forecasting

〇Satoshi Kiryu¹, Yuuto Osada¹, Masahiro Oshima¹, Mayuko Kawaguchi¹, Satoshi Suzuki¹, Tatsuya Iizaka¹ (1. Fuji Electric Co., Ltd.)

Keywords:

AI,Machine Learning

Research is being conducted on inflow forecasting methods to support dam operations. In particular, machine learning-based forecasting techniques that do not require extensive domain knowledge of the target system have been proposed in recent years. Meanwhile, in the field of machine learning, Neural ODE have been introduced as a method to approximate the underlying differential equation of a target system using neural networks, demonstrating high prediction accuracy. Therefore, this paper proposes an approach utilizing Neural ODE to improve the accuracy of dam inflow forecasting. Using real-world dam inflow data, we compared the proposed method with a decision tree-based machine learning method and validated the effectiveness of the Neural ODE-based approach.