A Novel Damage Identification Method for Steel Catenary Risers Based on a Novel CNN-GRU Model Optimized by PSO

As a new type of riser connecting offshore platforms and submarine pipelines, steel catenary risers (SCRs) are generally subject to waves currents for long time, thus it is significant fully evaluate the SCR structure’s safety. Aiming at damage identification SCR, acceleration time series signals multiple locations taken as characteristics. The characteristics include spatial information measurement point location acquisition signal. Therefore, convolutional neural network (CNN) employed obtain information. Considering variable period gated recurrent unit (GRU) utilized study these However, neither single CNN nor GRU model can simultaneously temporal data by combining with GRU, CNN-GRU established. Moreover, hyperparameters deep learning models have influence on their performance. particle swarm optimization (PSO) applied solve hyperparameter problem CNN-GRU. Thus, PSO-CNN-GRU (PCG) Subsequently, an method based PCG presented predict degree series. By analyzing data, prediction performances PSO capacity verified. experimental results indicate that result proposed better than several existing (CNN, CNN-GRU).