Evaluating the vibration performance of a subsea pump module by numerical modelling and full-scale testing

Before installation, each subsea system has to be inspected and tested to verify whether it performs in accordance with the specifications and performance evaluation criteria. The assembled components should work in accordance with the assumptions and design criteria used in detailed engineering, including the vibration and fatigue performance. In the current study the pump module within the Åsgard subsea compression station is subject to such an (vibration) evaluation test. The vibrations are caused by flow through a complex process that is affected by different factors as for instance piping geometry, flow and operating conditions and also the fluid properties. Finally, mechanical vibrations can lead to fatigue failure. One of the major parameters affecting vibration of subsea piping is the surrounding water. It is known that surrounding water participates in some vibration modes by adding mass to the total, dynamic mass participating in the vibration. As a result structural natural frequencies of a pipe system will differ between non-submerged and submerged conditions. In addition, the surrounding water can also add damping to the vibration modes. Until now mostly for engineering purposes simplified added mass approximations are used, which in principal are only applicable for infinitely long and slender pipes. In this paper the effect of submerging the pipe system in water is quantified, by analyzing damping coefficient changes and the measured in-situ vibration of geometrically complex piping system of a subsea pump module. This is done by fullscale frequency response tests carried out on the subsea pipe system of the pump module in both non-submerged and submerged conditions. The results are used for validation of numerical models, used to quantify pipe vibrations and accompanying fatigue stresses in the submerged conditions. The paper shows that the influence of the surrounding water on the pipe vibration is different for small-bore piping than that for main piping and also deviates significantly from simple added mass approximations. Next to that different modeling approaches and general observations and trends in damping coefficients are compared with the measurements and discussed.