Possible changes to modelling of hollow spar buoysWednesday, October 24th, 2007 by Peter Quiggin
This article proposes some possible changes that we are considering, to the way OrcaFlex treats the water inside a hollow spar buoy, i.e. where a spar buoy object in an OrcaFlex model has a non-zero ID specified for one or more of its cylinders. We would welcome feedback from OrcaFlex users (and potential users) on whether the change that we are considering would be a useful.
The existing OrcaFlex treatment of hollow spar buoys is aimed at modelling hollow axi-symmetric structures that flood and drain very freely in the axial direction. But it is less suitable for hollow objects that do not flood and drain so easily, and these proposals aim to address this.
The OrcaFlex help file gives details of the existing model. OrcaFlex assumes that each hollow cylinder in the buoy is flooded with water (up to the water line if it is only part submerged). The aspect of the model that we are considering changing is the way OrcaFlex models any enclosed water when the buoy moves axially.
At the moment OrcaFlex assumes that the enclosed water is constrained to move with the cylinder if it moves normal to its axis, but not constrained to move with the cylinder if it moves axially. So for directions normal to the cylinder axis, OrcaFlex adds the inertia of the trapped water to that of the buoy, but for the axial direction OrcaFlex does not add the trapped water inertia to the buoy.
This is not very suitable for hollow objects that do not flood and drain easily, as illustrated by the following example.
Suction Anchor Example
Recently a customer was modelling a suction anchor, which is essentially a large hollow cylinder that is open at the lower end but who’s upper end is capped. A small vent in the end cap allows water (or air) to flow in and out, but this is closed off once the anchor has been embedded in the seabed.
When the anchor is lowered through the water surface, the vent is open, and the water floods into the anchor from below quite easily, since the air trapped inside compresses and also escapes through the vent relatively easily. So at this stage of deployment the existing OrcaFlex model described above is reasonable.
But once the top of the anchor is below the water line and all the air has escaped, then the water inside the anchor is much less free to move axially relative to the anchor. This is because there is no air left to compress, and although the water can flow through the vent, it will probably do so rather less freely than air. So now the water inside is much more constrained to move axially with the buoy, as well as being constrained to move with any buoy motion normal to its axis.
To try to cater for this situation, as well as other applications, we are therefore considering the following change to how OrcaFlex handles the axial inertia of any enclosed water.
Consider the water inside any given cylinder, say one whose inner diameter has been specified by the user as D, say. Then the idea is to consider the enclosed water as being made up of two parts, an inner core cylinder of water of diameter d (less than or equal to D) and an outer annulus between inner part diameter d and cylinder inner diameter D.
The inner core part represents the part of the enclosed water that is free to move axially relative to the buoy. OrcaFlex would take this inner core diameter d to be the smallest ID of any of the cylinders in the buoy that are below the water line. And it would add this part’s inertia to the buoy inertia normal to the buoy axis, but would not add this part’s inertia to the axial inertia of the buoy. The underlying assumption here is that this inner core of enclosed water is largely free to move axially relative to the buoy, because all of the cylinders below the water line are hollow over this cross-section.
The remaining outer annulus part, between diameters d and D, would be treated as being constrained to move with the buoy both axially and in the normal direction. That is, its inertia would be added to the inertia of the buoy for all directions. The assumption here is that although this part is not constrained by this cylinder, it cannot flow freely in and out of the buoy because of a smaller ID constriction that is further along the buoy and still below the water line.
The inner part diameter d would not be affected by cylinders that are above the water line. The assumption here is that cylinders above the water line won’t act as a constriction, since the air in them can compress or expand, and also the air inside them will move fairly freely though them.
This proposed model behaves the same as the existing OrcaFlex model If all the cylinders in the spar buoy have the same ID. It only changes the model if some cylinders have a smaller ID than others.
For the suction anchor, it would allow the user to model the anchor as a spar buoy with two cylinders, one representing the main anchor body, with a large ID, and a short one representing the top end cap, with a small ID to represent the vent. When the top end cap is above the water line the axial inertia of the enclosed water will not be added to the axial inertia of the buoy. But when the end cap is below the water line most of the axial inertia of the enclosed water will be added to the buoy. This is a more reasonable model of the suction anchor that the existing OrcaFlex model gives.
At the moment all this is just a proposal, aimed at making OrcaFlex better for modelling a wider range of systems. But what do you think?
Would it help you model cases where the existing OrcaFlex model is less suitable? Or would it cause any problems for models where you have previously used hollow spar buoys?