I'm having a few arguments with the main algorithm of my spreadsheet to make sure it's all up to a very good spec. There are a some small polishes to it I have yet to make which might change the end results very slightly.jimmy101 wrote:Ragnarok: How accurate are your Cds and velocity and range calculations? How acurately do you thing they predict actual performance?
I've tried to consider as many variables as possible - for instance, the variable air density with altitude, and how the speed of sound varies with this. That affects drag due to both air density, and mach number.
It also means if fired sufficently high and with the right drag characteristics, things are capable of hitting the ground at greater than their terminal velocity at ground level air density would suggest, as the projectile can build up a higher velocity in lower density air.
A semi complete list of advanced features beyond simple drag and gravity:
Earth Curvature; Air Density variations (currently good to about 11 km up, I'm considering working on sorting it to 71+ km); Variable Cd vs. Mach; Gravity Variation (if the user is really planning on shooting the moon); and for the user that's really feeling daft, the gravity of the Moon and Sun.
All of these advanced features can be turned off if the user desires, and that last one is probably going to be removed (because users will probably find having to put in a date, time and global co-ordinates to be a bit too bizarre), but I do things like that when I'm very bored. I blame _Fnord for suggesting a variable like that needed to exist in the first place.
I've not currently got the controlled spin features (like back spin) modelled, and I'm also trying to work out if I want to include the small natural lift of anything moving through the atmosphere, and buoyancy - which of course varies with air density and altitude. Probably both moot, but I'll likely do it anyway.
At the moment, I know I should working on cleaning the project up for release, but I'm wiling away time adding features I know really don't need to be there, because I want to see how much they don't need to be there. In many cases, the differences are less than a percent of a percent, but by then, I've already coded it.
Air density variation and Drag vs. Mach are two features that are reasonably worth having activated though, the latter being the more important.
Air density variation tends to increase the end result by a couple of percent (It only ever really increases the results, as it reduces drag at higher altitudes. I suppose it is possible a change in Mach number to increase Cd and overall lose out, but I doubt it)
Assuming the data the spreadsheet is given is all correct, and obviously allowing for environmental variations like the fact that wind speed varies considerably with altitude (a situation, that although it can be modelled, is very hard for the user to actually acquire the pre-requisite data to model it. I'm trying to find a way to approximate this.)
Essentially, there is less room for error in the calculation than the natural error of the cannons in question.
All the drag models I'm talking about are based on many hundreds of actual firings and experimental data, rather than scientific calculations. Indeed, the original version of the GI model was created before the scientific principles behind air flows were fully understood.My understanding was that precise ab initio calculations for the aerodynamic performance of shape is basically impossible.
I am wasting further time on a system that's capable of predicting drag based on the projectile's shape, although as you say, an ab initio calculation of this that is totally accurate is impossible. However, again - it should be more accurate than the cannons.
