부산대학교 도서관

To deal with the critical issue of long computational time in practical application of computational fluiddynamics (CFD), this paper presents a new approach of deep learning for voidage prediction (DeepVP) that couplesshort time CFD simulations (limited CFD iterations) with the deep learning method to accelerate the 2D voidage distributionprediction for a gas-solid fluidized bed at steady state. Short time CFD simulations are first performed toobtain a sequence of voidage distribution images containing the temporal-spatial property of a gas-solid fluidized bed ofthe early period. A deep learning model is built to predict the voidage distribution at steady state, which is achieved byimplementing multi-scale convolutional neural networks based on the sequence of voidage images. The case study resultsfor a bubbling bed show that the voidage distribution at steady state for the bubbling bed can be predicted with comparableaccuracy of conventional CFD simulations at about 1/30th computational cost. Moreover, the DeepVP methodexhibits better extrapolation capability than the deep learning approach merely based on CFD condition parameters.