Research on Jacket Digital Twin Prediction Technology Based on LSTM

In this paper a prediction technique for deep-water jacket digital twin model based on long short term memory network (LSTM) is proposed. By studying the long and short memory network technology, the random effect of environmental parameters was studied. Combined with the environmental parameters of the area where the jacket is located and operating time database of the jacket, the digital twin model of the jacket structure is predicted. Realize the synchronization and real-time update of the digital twin physical model of the jacket. It obtain more accurate jacket structure prediction results. Compared with the simulation model database, the jacket digital twin prediction model base on LSTM technology has higher accuracy and saves training time of the model.