Brain picking...
 

Hi.

I'm a 16 year old student from the UK and I'm taking part in a project to solve a problem that a manafacturer of large boats has. The problem is that as the boat travels along as some speed the gases from the exhaust are sucked back onto the rear deck and also sooty black marks are made around the sides of the boat.

we were thinking about using CFD to try and understand the problem better but we have no experience of it.

We have access to some software though..

So you think it's a good idea to try and use CFD or is it really as complicated and hard to learn as some say it is....??

Anyway thanx... !!

Possible but not easy
 

Hi!,

I think it possible, but needs lots of time. It will be difficult for you since you probably do not take any course on CFD.

The problem is very difficult: 3d, turbulent, two-phases, Chemical reactions, soot formation ...... seems impossible to be solved.

Sorry for telling you the truth.

Ray

Re: Brain picking...
 

I agree with Ray, using CFD is probably too difficult if you have no previous experience from it - this is a complex 3D problem which is turbulent and probably transient. If you can get access to a wind-tunnel, or even better, a water tank/tunnel, you can probably learn a lot more about how the fluid dynamics works by doing some experimental testing - build a scale model and, while you have flow passing over the vessel, eject something colorful where the exhaust pipe comes out and see where it goes. The best way to to this is probably in a water tunnel - your model can be small and still give a correct Reynolds numbers it is also easy to find a suitable dye to eject. Then you can try different solutions, like adding "spoilers", moving the exhaust, etc...

Re: Brain picking...
 

if you'd like to do it yourself it might take awhile because you'll have to learn CFD theory. obtain and learn to use a CFD code (which would probably entail installing new computer hardware ), then you'd need CAD models of your boat, generating a good mesh before you could start solving the problem. if you'd like to use CFD a good alternative is to hire a consultant who is familiar with marine problems. the reason i say this is that it isn't cost effective to acquire all the capabilities listed above if you aren't going to use them regularly. the experimental options are probably best from the point of view of having fast turn around times

Re: Brain picking...
 

(1). It is a common phenomenon for flow over a body with non-streamlined rear end shape. (2). You can see the black marks on the rear end of a bus, or a van. So, in many cases, it is equipped with a rear window wiper. (3). If you sit in the last rows of an old bus, you can also smell the bad disel fuel exhaust. (4). I think, it is a good project to look into the flow field and come up with some solutions. You can either move the source of exhaust away from the rear end area or make the rear end an streamline shape. Or you can guide the air into the rear end area and try to push the recirculation region away. This has the similar effect as the streamlined rear end. (5). To see the flow field using CFD is not a real tough job, in your case, It would be practical to work with someone with the access to a CFD code. Although 3-D solution will provide more accurate solution, a 2-D solution should provide enough information about the rear end flow field. That is taking a horizontal cut through the boat and compute the 2-D flow to give you the rear end flow field. (6). Such 2-D solution will be fairly easy to obtain, as long as you can find some one willing to work with you. (all you need to do is to setup the boat shape in 2-D, specify the inflow condition, the farfield condition, and the downstream condition. That's about all one needs to know.

Thanx guys....
 

Thanx very much for all your help and comments. It looks increasingly likely that we will jsut go with the diea of using a wind tunnel as our main means of experimentation.

We have a scale model already and if, as you say, the results you get from a model can be applied to a full size boat (can they?) this sounds a reasonable approach.

The university is laying on all the equipment inclduing a CFD prog I think but we have no computer model - perhaps we may try a 2d model as was suggested.

Once again thanx very much guys... ANything else you want to add would be appreciated...

Thanx...!

Re: Thanx guys....
 

You will get the same flow over the scale model if you make sure that it has the same Reynolds number as the real case. The Reynolds number is:

Re=u*L/nu

where u=velocity relative to boat, L=length of boat, nu=kinematic viscosity of fluid passing over boat

If you use the same medium as the real case (air) and have a smaller model you will have to increase the velocity in order to get the same Reynolds number. Another possibility is to use another medium. Water, for example, has a kinematic viscosity which is about 15 times lower than air - that means that you can have a 15 times smaller model and still use the same free-stream velocity. Water is nice since it is easy to visualize the flow with dye. Smoke in air is more difficult to see. I know that people, for example, have "driven" Formula 1 cars on the bottom of water tunnels in order to test their aerodynamics at correct Re numbers but at lower speeds.

If you can't get exactly the right Re number you can probably still learn a lot from an experiment.

A 2D simulation would be a nice complement and it should be possible to do if you can get some guidance from someone to get started.

Re: Thanx guys....
 

(1). You have just mentioned the key issue of "using wind tunnel VS CFD", that is "the ability to simulate the real world problem directly". (2). In the CFD simulation, there is no limit on the physical parameters one can use in the calculation. One can easily specify the temperature at several thousand degrees by just adding just a few zeros to the inlet condition. This is not possible for wind tunnel testing. You will have to build a special facility to do that. (3). In other words, there is no limit on the Reynolds number, Mach number, etc. in the CFD simulation. And the model shape can also be easily changed on computer. (it shouldn't be difficult at all if you had three years of vocational school training in engineering and CAD. ) (4). So, the CFD approach is the ideal way to go. (5). Whether you should take the wind tunnel approach or CFD approach depends on the resources you have in your workplace. And naturally, it also depends on your background. At your age, it will change very quickly in the next two years. (6). You will have to learn from somewhere, and from someone as well.

Re: Thanx guys....
 

I have seen a number of problems that are similar to the one you are describing. In simple terms, the exhaust from the engines is getting trapped in region of disturbed air behind the moving boat. That air will find it's way into the cabin of the boat over time. If the intake for the ventilation system is near the aft end of the boat, the air-exhaust mixture will be actually draw into the cabin. (I did an analysis of building that was designed with the exhaust of a diesel generator and the intake for the ventilation system in exactly the wrong locations. Every time the diesel was tested, the exhaust was draw right into the air conditioning system.)

You could, as some have suggested, try to solve the problem including every detail, including tracking the fraction of exhaust in the air, however this is overkill. You'll learn the most about the fundamental nature of the problem by looking at the air flow pattern around the boat. The goal is to get air to flow smoothly around the engine exhaust and away from the boat. The exhaust should be prevented from getting into a region where it could be drawn into the ventilation system.

A wind tunnel test using an existing scale model will give you good information about the current geometry. The problem is that you know that the current geometry leads to exhaust odors in the cabin. You are going to have to change the geometry to fix the problem.

I would recommend using CFD on this problem (either yourself, or by contracting with an outside expert) because CFD will be very helpful for experimenting with various fixes. It takes far less time to modify the geomerty and mesh of a CFD model, and reanalyze it (especially a 2D model) than it does to change a physical scale model and re-test. When the CFD model demonstrates that a particular concept seems to solve the problem, then you can modify the scale model and run a test (for some reason members of the species we call management like to see CFD results that match scale model results).

Best of luck with your analysis!