Experimental study on subsea pipeline bedding condition assessment using wavelet packet transform

Free span along subsea pipelines is a general problem both in pipeline design phase and during operation of pipelines. Traditionally, subsea pipeline system is inspected over the entire pipeline length with inspection tools, such as side-scan sonar and video/camera carried by remote controlled vehicle. The obvious disadvantage is that this kind of inspection can only be carried out at scheduled intervals. In a previous study [1], a damage indicator, WPECR¯, was proposed based on wavelet packet transform analysis to identify free-spanning damage along subsea pipelines. The reliability and efficiency of the proposed damage indicator is proven with numerical simulated vibration data of pipeline under ambient sea wave forces. To further verify the proposed method, in this paper, two scaled pipe models with different boundary conditions were designed and tested subject to random wave forces in a large wave tank. Waterproof accelerometers were used to collect pipeline responses which were subsequently analyzed for pipeline condition monitoring. The results show that using output-only dynamic responses of pipeline subject to random ambient wave excitations can accurately identify free span locations and length. With a vibration based method, real time monitoring of scouring damage to subsea pipelines is achievable so that any free span longer than allowable span length can be detected at an early stage.

► Two scaled pipeline models were designed and tested in a large wave tank.
► Pipe models were excited by the generated random wave force.
► Vibration based method was used to detect pipeline free spans location and severity.
► Wavelet packet transform was used to process the collected dynamic data.