For the safety and availability of temporary facilities in the Yanqing area of the Winter Olympic Games,by sufficiently combining signal processing algorithm and deep neural network,this paper proposes a brand-new model that consists of two parts:Hilbert-Huang transform(HHT)used for signal decomposition and extraction of signal feature for time-series data,and long short-term memory(LSTM)for prediction of the operation trend of temporary facility.Based on the real vibration and tilt angle data measured while it is affected by a series of exogenous factors such as grandstand vibration induced by severe cold weather and heavy passenger flow,the model realizes effective prediction for facilities,so as to avoid safety problems and solve the problem of low prediction accuracy due to the interference of some irrelev-ant feature factors in the data.By comparing with such operational trend prediction methods as recurrent neural network(RNN),gated recurrent neural network(GRU),bi-directional RNN and bi-directional GRU,the feasibility and effective-ness of the method was demonstrated.The experimental results also show that the proposed model performs very well in such tasks.
维普
万方数据