Leakage detection method of acoustic emission pipeline based on deep convolutional neural network transfer learning

Key Points

• The central aim is to develop a robust method for detecting pipeline leaks using deep convolutional neural networks with transfer learning.
• Implemented a deep convolutional neural network architecture for leak detection.
• Utilized transfer learning to enhance model performance with limited data.
• Gathered acoustic emission data from pipelines as input for the model.
• The proposed method shows improved accuracy in identifying leaks compared to traditional techniques.
• Achieved significant reduction in false positive rates during leak detection.
• Demonstrated the ability to detect leaks at early stages, which can prevent larger failures.