심층 신경망을 이용한 철도 연결기용 점탄성 완충기의 동적 특성 예측 모델

초록

An energy absorbing device is installed in a railroad car coupler to absorb kinetic energy in a collision such as during shunting, and safety is verified using dynamic simulation in accordance with railway safety standards. In hydrostatic buffers, which have recently seen use as shock absorbers for couplers, the buffering characteristics change depending on the operating speed, but there is no suitable material model that can take into account the speed-dependent characteristics in the commercial software used for crash analysis, and it is difficult to find analysis cases that consider the speeddependent characteristics of hydrostatic buffers. In this study, we construct a model that predicts the dynamic characteristics of the hydrostatic buffer based on collision test data; this is done by using a deep neural network to consider the speeddependent characteristics of hydrostatic buffers. The accuracy of the model was verified through the collision test simulation.

키워드