[1B01-05] Advanced Simulation

[1B01] Prediction of Flow Behavior with Machine Learning Using LSTM

Application to Unsteady Backward-facing Step Flow

*Aya Kitoh¹, Kei Maeshima², Kenya Takiwaki², Hideki Horie² (1. TOSHIBA, 2. TOSHIBA ESS)

Keywords:Machine learning, Fluid analysis, Simulation, LSTM, Backward-facing step flow

Research on the application of machine learning to numerical analysis has recently made rapid progress. In this study, we have been investigating the applicability of machine learning to fluid analysis for supporting the design and development of fluid equipment. In the previous report, we showed the useful result in the case of applying Convolutional LSTM as one of time-series prediction methods. Then this time, we newly apply LSTM as one of those, and the result shows higher accuracy than that of Convolutional LSTM.