University of Twente Student Theses

Login

The influence of slugs on pipe supports : improving the design parameters for Caesar II

Stoutmeijer, K. (2015) The influence of slugs on pipe supports : improving the design parameters for Caesar II.

Full text not available from this repository.

Full Text Status:Access to this publication is restricted
Abstract:Slug flows cause an alternating force on the piping which depend on the velocity, composition and shape of the slug. From experience, it is known that the stress levels in the pipes stay within the acceptable limits, but large impact on pipe supports due to slug loads should be expected. The time history analysis in Caesar II is used to predict these load but the results from this analysis seem too conservative. Research is needed to accurately predict these forces. The goal of this study is to find better modeling techniques in Caesar II which will predict the reaction force in the supports more accurately. In order to calculate the transient response of a system, the natural frequencies of the piping system are calculated in the time history analysis in Caesar II. The dynamical response study investigates the impact of the dynamical input on each natural frequency up to the so called cut-off frequency. The default cut-off frequency in Caesar II is 33 Hz, which is too low for predicting the dynamic response for slug flows. The rise time of slug loads is in the order of 10ms, resulting in a minimum input frequency of 50 Hz. Increasing the cut-off frequency increases the accuracy and reduces the maximum calculated response forces in many cases. For this reason the cut-off frequency is set at 100 Hz for all Caesar II simulations. Slugs are studied using Fluent, the Computational Fluid Dynamics (CFD) solver in the finite element package ANSYS. The influence of flow velocities, pressure and pipe diameter on the liquid front are studied. These simulations show that the liquid front is steeper when the velocities are high and the pressure is low. As this yields the steepest change in linear momentum, simulations with high flow velocities and low pressure give the most conservative results. The next step is the simulation of the reaction forces in the supports. This is done by coupling the flow and structural domain with the system coupling feature in ANSYS. These results are compared with the dynamic time history analysis in Caesar II. Close to the inlet of the slug the differences between Caesar II and ANSYS results are small. Further downstream of the inlet, the differences Caesar II and ANSYS results increases because the diffusion of the slug is not taken into account in the Caesar II simulation. The input forces in Caesar II should be altered to get more realistic dynamic behavior. Mass flow rate plots in ANSYS show that the rise time entered in Caesar II should be higher to achieve more realistic slug loads. Furthermore, the ANSYS simulations show that the slug length should be significantly longer than the longest bend length, which is currently used in the Caesar II analyses. Simulations also show that forces on supports can be three to four times as big as the initial load when a pipe system has a low stiffness due to low frequency vibrations. These forces can be reduced by implementing axial stops up- and downstream of each elbow. It can be concluded that Caesar II provides a good model for calculating dynamic responses when the right cut-off frequency is used. The time dependent input forces for this model are too conservative because the diffusion of the slug is not taken into account. The expected slug velocity and density should be examined closely by process engineering. Challenging these parameters could result in significant lower expected loads resulting in lower construction costs.
Item Type:Internship Report (Master)
Clients:
Tebodin Netherlands B.V., the Netherlands
Faculty:ET: Engineering Technology
Subject:52 mechanical engineering
Programme:Mechanical Engineering MSc (60439)
Keywords:Slug load study, Fluent Multi-Phase-Flow simulation, ANSYS System Coupling, Caesar II pipe modeling
Link to this item:http://purl.utwente.nl/essays/70260
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page