CFD from Solid Models

[LarryMcI]

I have a very accurate 3D solid-model, of my customer's 44' sportfishing catamaran, that I designed in Pro/ENGINEER. The boat has been in service for 3+ years and, for the most part, performed well. They have encountered large head-on waves (10 foot) on several occasions which resulted in plowing. They want more bouyancy. I have modeled several add-on hull-extensions which may do the trick. Now I would like to tweak the add-on's geometries to minimize drag. We need a competent CFD analyst to take my geometry and return some results/recommendations. I can supply IGES, STEP or Pro/E file formats.

Any recommendations, referrels or comments would be appreciated.

[baeckmo]

Diving into a wave is a dynamic phenomenon more than a question of static buoyancy. You must have control over what is happening when the hull is pitching. As the trim is reduced (or going negative), the center of buoyancy moves forward, creating a static trim restoring moment. But since there is a significant vertical dynamic force as well, you must also design the bottom in such a way, that the center of the waterplane area is moving forward correspondingly!

This is often forgotten, even by experienced designers, when very sharp stems are used in combination with flat aft sections with the idea to create a soft ride. I have seen slender monohull ferries shipping green water for this reason. Slender multihulls often lack this balance, resulting in pitch-poling or serious broaching tendencies.

For instance, the first two generations of passenger cats that were introduced in Scandinavian waters long ago, were good seaboats in most wheathers. Later versions were much more slender, their pitching frequency was lower than the early ones, with the result of serious syncronous pitching problems in some sea conditions. In order to break that behaviour, the linear displacement of CofB, together with CofWPA must be linked to a non-linear change of a dynamic factor, such as an area discontinuity like a longitudinal or transverse step.

So, screw the nice CFD, go back to basic engineering and check what happens to your centers when the hull is pitching. If you really want to go hands-on: make a drop test! If the hull is dropped from a reasonable height, say half a meter above surface, it must NOT take a nose-down attitude. A good, balanced hull will stay flat or even juuuust slightly nose-up during the process.

Would be interesting to see a lines plan of the hull, together with a realistic weight and C of Mass, as built.

[yipster]

"screw the nice CFD" is a bit hearsh but i see the wisdom baeckmo
CFD seems to me often very limited, flowlines can be played with in freeship
still looking to place a nice tuftgrid over solids giving some numbers, keep me posted

[baeckmo]

Aah sorry, tried to avoid the f-word...... Maybe I'm getting too old and grumpy, but I am worried about the loss of basic engineering skills and logical reasoning that seems to come hand in hand with an increased belief in the blessings of fancy computer software. It is not that I am against the use of computerized analysis, not at all; it is rather a feeling that it is all too easy to focus "on the map instead of on the terrain".

So Larry, please don't get offended here, it's not my intention to twist your nose, but your "ouverture" (I have a very accurate CFD model...) is to me an example of "focus on the map". You are facing a physical engineering problem and want our input, but you did not give us one single engineering fact to start with. There are three cardinal rules in trouble-shooting:

• All calculations are wrong
• All measurements are wrong
• Everybody is lying or has forgotten

Unfortunately a good breakdown investigator, knowing the importance of these rules, will often be a pain in the seat of those involved in his search for a cause......., but that is a question of focus.

[Sam_BMT]

Larry,

I'd be more than happy to help you. What time frame are you looking at?

[LarryMcI]

We have in-house CNC machines (largest 5-axis has 21' x 12' x 6' envelope) and go directly from design-to-mold. Hence I never made a "lines plan", tables of offsets etc. The master 3D model has precise and accurate sub-components (engines, generator, battery banks, galley appliances, windows, anchors etc.) each with their respective mass-properties. The complete assembly's COG, CofB and dry/wet displacement are known. My question is: What format (IGES, STEP ) solid-model is preferred for CFD analysis of my existing hull and several proposed extensions?

We would like to machine the 2 mold-halves during May-June 2010, ship them to Key West by July 4th, make parts, then sub-assemblies and be ready for a 2-week R&R program in December.

[LarryMcI]

Displacement calculations matched the as-launched load balance within +/- .2 inches. b-deck.jpg is an installation drawing for the CNC-cut bulkheads and serves as a reference for the bouyancy and CG values. Other pix are just for fun. Present CofB is 281" from bow. Present CG is 252" from Bow and 5.5" below shear line.

This shallow draft charter sportsfishing boat operates in the Dry Tortugas near Key West, FL. The low bridgedeck clearance was mandated by my customer to keep his clients closer to the water. An appropriate slam-resistant clearance would hinder handling their catch. Large head-on waves are rare but, as always, happen at the worst possible time.

[baeckmo]

Ok, this is exactly the kind of hull that I was referring to, giving serious digging-in problems!!! This is the risk with cats; the great transverse stability is "luring" the designer to forget longitudinal balance! You have the tools available in your model to see what happens; a "classical" naval engineering excercise. Just move the CofG forward, plotting CofFlotation against CofB. Continue until you reach a trim angle of -10 to -15 degrees, representing a hull diving into a wave.

You will see that the CofF (or CofWaterplane Area) is hardly moving at all with these forward lines, so the dynamic lift when the hull is entering the wave is creating a nose-down trim moment! There is only the slight, linear increase in forward static lift from the diving noses, that prevents disaster. With the superstructure shown, you have a high CofG, producing additional nose-down moment, when the forward motion is retarded. In a following seaway of any significance, this boat is at risk.

In order to increase seaworthiness, you must create a hull shape that is letting the CofF follow the CofB when trimming nose-down; or better to have a slight lead on the CofB. That will generate lift where needed!

In our catamarans (both planing and semiplaning) we have used a "planing plate", ie a nearly flat surface at ~45 degrees, bridging the tunnel from stem to stem at deck level. Your design has a surface at about 20 degrees, which cannot give any lift because its angle of attack versus the wave slope is too low! In addition to that we also have generous spray strakes forward, above DWL. I'll see if I can send some photos to you in order to explain what I mean.

[daiquiri]

Nice and clear explanation, Baeckmo!

[LarryMcI]

I agree! The original 37' hull molds were for a sailboat. The customer asked me to design the transom-extensions (with propeller tunnels etc.) and build those molds. Then the boatbuilder "lengthened" the hulls by 60" at midship. Of course tangency isn't possible with such modifications, so lots of fairing was required. It is what it is, but my customer wants to squeeze another 2-3 years of service out of her before building a new vessel.

Do to the range of operating speeds, bulbous extensions do not seem optimum. I could design very long extensions that become tangent with the mid-ship extension, but the mold-cost and installation complexity don't seem justifiable.

Given that we are going to cut-away parts of the existing hulls, and that we will graft-on new, fuller bow-extensions; I want CFD analysis of the turbulence/drag/whatever-else attributable to my non-tangent interface(s) between the new bows and old hulls. If seriously non-tangent intersections (at the 161" bulkhead, for example) only cause 1-2% more drag than the longer option, we might consider the less-costly/-complex solution.

The new bows will be plumb since projecting bulbs would interfere with handling lines which, often, force the fishermen to work-around the boat during the fight.

My customer wants mostly netting between the new extensions, with only a catwalk and center gangplank for line handling. 1-2 crossbeams (probably pultrusions) are anticipated.

[daiquiri]

At first glance it looks to me like you have added more structural weight than buoyancy at the bow. But I guess I must be wrong, am I?

[LarryMcI]

The bottom-view image does show a full-span solid (actually Nida-Core) foredeck, but this will actually be mostly netting. The existing hulls were hand-layup and are very heavy ~0.0228 lb/cu-in. Molding the new bow-extensions will use a vacuum-assisted resin-infusion process less density with better physical characteristics. This project will increase LOA ~19% but weight-increase target is <13%.

[LarryMcI]

Does anyone recognize the program(s) used to create this simulation? Does it seem accurate or just a pretty, animated rendering. My customer sent this link and would like to see similar analysis of his boat but it seems like expen$ive overkill to me. I would, however, like to do this type of dynamic analysis for our next from-scratch design.
http://www.youtube.com/watch?v=Cefqx...eature=related

[HReeve]

the Youtube link has this if you expand the information box on the right.

"The analysis was done with the CFD software Tdyn (www.compassis.com) "

I'm not familiar with it.

That animation is pretty typical of what you can get from most high end CFD packages. Because the vessel is free to trim in that simulation, you are looking at some pretty serious run time to generate that animation. Probably on the order of 36 hours.

[baeckmo]

"This project will increase LOA ~19% but weight-increase target is <13%."

Afraid Daiquiri is homing in on a weakness in the design here! You had better compare what is happening forward of the junction (pos. 165?). Taking it to the extreme, there may well be a length increase without any volume/buoyancy added. You are adding length with very little "local"volume, so the above ratios are misleading. These hulls are lacking reserve buoyancy when pitching. Those nearly vertical sides have no flair, because the designer has allowed the forward panels to progress towards the stem with minimum curvature, seen from above. If you allow a bluffer stem angle up where the sides meet the deck, you will get decreasing radii (in the horizontal plane) towards the stem, and thus an increased flair.

Again, when seaworthiness is on the agenda, the still waterline shape is not the critical issue, but what is happening in extreme pitching conditions. Remember that the slender catamaran normally is faaaar softer longitudinally, than a monohull of comparable capacity; for the good and the bad of it!

Now, for the resistance, I would not bother too much about it, unless the scarfing is very rough. You have a slenderness ratio (metric, per hull) of ~6.3, which per se is ok, and the added wl length will reduce wave resistance slightly. The real problem is how to add volume along the wl forward, plus creating reserve buoyancy/flare up front. You may add considerable flare to the inner sides of the bows, without too much disturbance to the eye, though.

Years back I was in a similar predicament as you, with a customer who moved a rather heavy crane from close to transom up to the front, of course without asking me (=designer/builder) about the consequenses, so I have gone through the procedure. Hope that experience could be of use for you.