Guidelines or formulas for how much water to dump in narrow or weakthermal racing?

Ben:
The best advice I ever got on the subject was, "If you are thinking about dumping the water, DUMP IT!" And that has remained pretty good advice.

The question you have to ask yourself is "why am I thinking about dumping water?" Is it because the last few thermals were weaker? (Dump it). Or, "I am perilously low and I need to float to a save?" (Dump it) Or, "The guys without water are seriously out climbing me and running away" (you guessed it - start dumping).

I think the only time to hold on to the water (when doubts emerge about its helpfulness) is when you are reasonably certain that you will get back to better conditions. This may be because you have started early in the morning before it cooks up or when you have drifted down to the weak area of the height band and can climb back up to the strong lift. But it's not a science - it's an educated guess about the future.

ROY

I understand...BUT...the leading pilots in Condor simulator racing are all, as best I can tell, holding onto their water as long as possible, i.e. they aren't leaning toward the conservative side. This is probably because it is virtual racing, and no one is going to break a real glider (or their real bodies) if they make a water mistake. If I dump water and they don't, they invariably beat me in the cruise segments and especially in final glide..

That's why I was looking for rules of thumb and formulas when racing, as opposed to cross-country.

Ben

Regarding Dave Springford's comment that both need to be minimized:

Yes, that's pretty much what my original question is about. How do you arrive at that "sweet spot" between the two, given all the many variables?

Thinking about Tim Taylor's comment: Yes, it feels like this is something that software would be good at solving. (I'm a 35 year software engineer, BTW...just not an aeronautical engineer :-) )

On Saturday, May 16, 2020 at 1:59:37 PM UTC-4, Charles Ethridge wrote:
Thinking about Tim Taylor's comment: Yes, it feels like this is something that software would be good at solving. (I'm a 35 year software engineer, BTW...just not an aeronautical engineer :-) )

All that is good but the software doesn't know your climb rate, only your guess.
Most people when answering honestly will tell you they held their water too long waiting for it to get good.
Waibel used to say below 2 knots achieved climb, go dry. Above 3 fill it up..
Other variables like streeting, height band, gustiness, and shears that need to be climbed through also affect how much weight you can carry in addition to the usual climb rate and thermal diameter.
Complex equation in may variables
UH

On Saturday, May 16, 2020 at 10:59:37 AM UTC-7, Charles Ethridge wrote:
Thinking about Tim Taylor's comment: Yes, it feels like this is something that software would be good at solving. (I'm a 35 year software engineer, BTW...just not an aeronautical engineer :-) )

A few years ago I wrote a script that computes the climb rate of a glider based on the factors you mention plus air density (thermals widen significantly at altitude) and the glider's performance polar. The calculations for climb rate given thermal strength & width are actually much simpler than you might expect, the vast majority is just trigonometry.
Modeling thermals is a little more tricky and the lift distribution varies drastically depending on which model you believe.
The model I have is by no means perfect for several reasons, a big one being that glider polars are measured in straight flight and as such, they are not necessarily representative of performance during circling flight (this effect is amplified for gliders with significant di/polyhedral and/or winglets). At some point it might be worthwhile to actually measure circling performance to avoid misappropriating polar data for this. Plus, this model does not take into account individual gliders' handling characteristics at different wing loadings, which is a non-negligible issue in real life.
To be clear, I don't suggest blindly following the numbers provided below.

If you aren't a flight dynamics expert but have some fundamental understanding of aerodynamics, I highly suggest reading the following source if you wish to learn more.
https://booksite.elsevier.com/978012...SAILPLANES.pdf

The results for a few gliders we
Discus 2a:
For lift stronger than 5kn, max gross.
For lift [3kn, 5kn], half ballast.
For lift weaker than 3kn, no ballast.

Standard Cirrus:
For lift stronger than 3kn, max gross.
For lift [2kn, 3kn], half ballast.
For lift weaker than 2kn, no ballast.

ASW-22BL (750kg):
For lift stronger than 4kn, max gross.
For lift [2kn, 4kn], half ballast.
For lift weaker than 2kn, no ballast.

This is only useful for average climbrate throughout your entire flight. If you're hauling for most of your flight with strong conditions and fail to switch gears with changing conditions, get low and have to take a weak climb to get back up, you might as well throw this information out the window.

The non-binary decision-making involved in resolving situations like this one or (avoiding it altogether) is what separates the great pilots from the decent ones, and the results of such decisions are reflected in big differences in average speed.

I am open to discussion regarding sharing my code (written in MATLAB) but in its current state it may not be easily comprehensible to an average user.

Thomas

On Saturday, May 16, 2020 at 2:49:34 PM UTC-4, wrote:
All that is good but the software doesn't know your climb rate, only your guess.
UH

True, but neither does our McReady (computer or whiz wheel or whatever), and we dial in our observed/expected climb rate to compute our McReady cruise speed and final glide speed. No essential difference - just would be a more complex algorithm, yes? And like McReady speed-to-fly, by quickly doing the math grunt work, it would not be perfect, but would help us make a better decision to up the CHANCE of winning (not the guarantee of winning).

Hmmm...maybe could be as simple as just adding the water weight effect on the polar to the McReady formula? Sort of a "super-McReady"?

I'm no mathematician...just a dreamer :-)

Ben

On Saturday, May 16, 2020 at 2:53:29 PM UTC-4, Thomas Greenhill wrote:
If you aren't a flight dynamics expert but have some fundamental understanding of aerodynamics, I highly suggest reading the following source if you wish to learn more.
https://booksite.elsevier.com/978012...SAILPLANES.pdf

The results for a few gliders we
Discus 2a:
For lift stronger than 5kn, max gross.
For lift [3kn, 5kn], half ballast.
For lift weaker than 3kn, no ballast.

Thomas

Thanks, Thomas. I read the article, and being a former CFI-I/MEI, I get the concepts. Being a current software engineer (vs a math engineer), with a bit of help breaking down the formula(s) into non-MATLAB software syntax, I can put formulas into LARGE bodies of software at proper places, without fully understanding the formulas. I'm working in the Kotlin language for Android phones presently.

What you show for the Diana2 fits for me, since that's about how much water I've had to lose to keep up with the leader in a flatland 3-5kt narrow thermal day race in Condor (running against his Condor "ghost" over and over).

He told me he thinks there's a "sweet spot" for water ballast given all the parameters, but I don't think he knows the formula(s) either. Just does it intuitively after years and years of experience (which I don't have).

Bottom line: Yes, I'd love to see your code. Maybe we could make an Android and iPhone app and give it out to the Condor racing community for evaluation? Nobody's going to die using it in Condor, if we "put a decimal point in the wrong place or something" (software engineer in the Office Space movie :-) )

Ben

My view: The main issue is not so much climbing as centering. As others have noticed, around 2-3 knots if you're flying with a gaggle that's doing all the work of finding thermals, water vs. no water is about a wash. It feels bad to watch them glide past you, but you do gain it back again in the climbs.

The real issue comes when you're out there alone, low, and you have to center a punchy thermal. It can literally be impossible to climb. You flail around in the nasty bits and pieces watching the fleet go by overhead. Eventually you give in and dump the water and now magically you climb at 5 knots.

Flying with water you're going faster, so thermal circles are larger, and you're less maneuverable.

So really, yes, the main consideration is how punchy the thermals will be and how often you're going to have to find and center one on your own.

Many of my if Idas are memories of great days blown by getting low and not dropping the water fast enough

John Cochrane BB

On Friday, May 15, 2020 at 8:55:57 PM UTC-7, wrote:
The biggest challenge (for me at least) is deciding on the amount of ballast to carry when launching early, when the lift hasn't yet developed to the strength I hope to see later in the day when it reaches its maximum power.. Many times I have been forced to "curb my enthusiasm" by dumping part or all of my water to remain aloft, only to curse that decision when the lift peaks midafternoon and I am flying lighter than I want.

Another challenge is trying to figure out what CG range works best for the type of conditions I will encounter. Many manufacturers and pilots suggest about 80% aft in the recommended CG range, but that seems to work best with a climb/cruise ratio of 50%. If I am in strong, streeting lift conditions, I sometimes have about a 20% climb vs. 80% cruise ratio, meaning the drag associated with elevator deflection is much higher during cruise. Drag goes up exponentially as speed increases, and since I spend more time running than circling, the drag penalty is much higher. I hope to be able to quantify these data this year to be able to determine an "optimum" CG point for weak, vs. strong conditions.

And then I will just forget what I learned and do the best I can to make some distance and avoid as much yardwork as possible.

I am of the opinion it much safer to launch with a full load of water then dump down to where you want. Partial tanks can have more slosh.