|
Robert Smalls,
The height necessary to avoid ground effect that Morelli showed for a teardrop with a flattened bottom was about half its width. As I recall his data was scaled to automobile sizes and speeds, but I am away from home where that paper exists so I can not check that; and again from recollection, Morelli was a little short on detail such that some confusion remained. However, his published test curve showed that when the vehicle was half the width the ground effect had mostly tapered off. Here we were seeing a free flow Cd of .07. The flattened body is not so much flattened that the body height differs a lot from its width.
It looks like the solar cars would meet a criterion for road clearance that is based on body height, and this makes sense since airflow that would otherwise pass through the body region has to be split between top and bottom. The wing model is relevant here, and there are known wing shapes where airflow underneath is quite straight. So perhaps they could be somewhat lower and still qualify as free flow vehicles.
But I am not too interested in solar cars, being more practical by nature. So for bodies having roughly the same width and height, I reason from comparing flow patterns that would occur for a true body of revolution (BOR) that the height needed might be somewhat less than that needed for the flattened BOR of Morelli, since the impact of the ground would be reduced at points further out laterally.
It would be extremely useful to have real Computational Fluid Dynamics (CFD) modeling capability on this project. I pursued this with financial limitations, and started with a "cloud" service called CHAM or Fonix or something like that. They offered a free month trial, which I signed up for, and proceeded to validate the model for an airship body. For coarse mesh sizes, the model would run but the Cd values were way too high. As I reduced mesh sizes and flow sizes the model became more and more unstable and took huge amounts of computer time. In the end, I think I brought the servers to their knees, and the company was not even interested in selling me their service after that trial. Ultimately, the best I could do was a Cd of .12 which was a run that only worked once. That was with substantial separation where Cd should have been about .04. (The flow field looked steady at the outer edges, but even that was not satisfactorily tested.) Clearly, this would not help much with the kind of fine optimization that would be great to have. I did get some feeling for flow effects for the airship form relative to a boundary. I would not say this makes me an expert. I might search again for CFD services, since that effort was some time ago. Does anyone have suggestions? (Though Miastrada Company is a large organization -joke- it does not have a CFD analysis division.)
Given your background I imagine you perceive that this is not a fully resolved issue, and yes, I see this as a development program whereby an optimum shape would be determined before anything like production would begin. However, I feel fairly comfortable with a four foot diameter body and a two foot clearance for starters. If possible in building a full scale model, I will make provision for adjusting strut height so that some hard data can be obtained. But ultimately it is necessary to actually build and test, especially in this field, so I am trying to accomplish this.
The real hard facts are the airship test data itself, including angled measurements and measurements width fins of different sizes attached. (See Freeman 1934 from the NASA reports server or the miastrada.com site. ) This is a complicated read, but reading this, and other reports of that era that considered attached gondolas and so on, gives me some confidence that this can be made to work.
|