Ultrasonic-based leak detection in factories with spatial mapping

Energy consumption of compressed air in the factory raises a high cost, while a significant part of energy loss is caused by leakage. Traditional methods involve labor-intensive processes using handhold detectors to detect ultrasound generated by compressed air leakages. However, the best practice requires ongoing leak detection and repair. Crucial to this is the visualization of ultrasonic distribution in the factory based on ultrasound data at various positions. Although various automatic ultrasonic air leak detection methods are applied in small pneumatic systems, there are few methods available for large factory environments. This paper proposes a leak detection method for the factory to identify the leaks at an early stage via spatial mapping. The optimal mapping method is ordinary kriging with a spherical variogram, exhibiting a relative error of 4.7%. In pursuit of practical industrial application, we conducted case studies on the impact of background noise, measuring directivity as well as susceptibility to machine interference. The results proved the applicability of this method for industry to proactively seek for air leaks and reduce energy loss.