A study on Improvement of Resource Efficiency for IoT-based Pipe Leak Detection

In managing today’s complex and aging power plant facilities safely, increasing attention has been paid to the challenges of detecting pipe leak faults quickly and accurately. This study focuses on developing a resource-efficient leak detection system using distributed acoustic sensors, in pursuit of the Internet of Things (IoT) paradigm. The proposed system extracts a small number of featured predictors from the raw acoustic signals, so the presence of leaks can be readily detected by applying machine learning classifiers while reducing the burden on data transmission, storage and computation. A system prototype is successfully evaluated through the experiments with acoustic signals measured around a laboratory scale nuclear power plant coolant pipelines, considering the ambient background and machinery noises.