Research on Ship Fault Identification Based on Optimized Convolutional Neural Network Algorithm

The identification of abnormal vibration noise of underwater ships is of great significance for troubleshooting ship faults to prevent damage deterioration and ensure its vitality. Aiming at the problems of many parameters and low recognition efficiency of existing intelligent diagnostic models, an underwater ship abnormal vibration noise fault identification method based on convolutional neural network is proposed, which optimizes the traditional neural network (BP), enhances the nonlinear expression ability of the convolutional neural network model by introducing convolutional kernel; and replaces the fully connected layer in the traditional convolutional neural (CNN) network with the average pooling layer to reduce the model parameters and computation amount while preventing the overfitting phenomenon. parameters and computation amount, while preventing overfitting phenomenon. The abnormal vibration noise of underwater ships is tested using the rolling bearing vibration noise dataset for alternative experiments to verify the feasibility of the method. The experimental results show that the method can accurately identify different fault states and is faster and more accurate than the traditional intelligent diagnostic model, which has certain potential for practical engineering applications.