主催: 一般社団法人 日本機械学会
会議名: 第27回 動力・エネルギー技術シンポジウム
開催日: 2023/09/20 - 2023/09/21
The feasibility of the broadband noise prediction generated from a horizontal axis wind turbines based on the machine learning was examined. Based on the comparison between the broadband noise predicted using the analytical results of the blade element momentum theory and the actual measured value, we discussed the issues related to its prediction in the machine learning. In the prediction of the aerodynamic noise generated from a flat plate by the machine learning, it could predict not only the broadband noise but also the discrete frequency noise with Karman vortex shedding. The spectrum distribution by the machine learning could not predict the narrowband noise centered at 6000Hz. We indicated that the prediction of the aerodynamic noise by the machine learning overestimated the noise in the low-velocity region on the hub side.