Great New Physics Thread!

[NathanKell]

OhioBob: By extracting the curves used:

1. Kerbin uses 80% values of Earth's Standard Atmospheric Model for pretty much everything, however:

2. While density varies based on temperature (i.e. day vs night, equator vs temperate vs poles), pressure is purely based on altitude

3. The temperature offset (from the baseline altitude->temperature curve, which is indeed an 80% Standard Atmospheric Model) varies based on height. So in the stratosphere, it doesn't much matter if it's day or night or where on the planet you are, but in other atmospheric regions the sun-based offset is more considerable.

4. There is no seasonal temperature variation because without tilt (either real tilt, or inclination-induced tilt) there are no seasons, so looks like nobody bothered to make those curves.

[OhioBob]

Now, somebody riddle me this - what's an engine that has a reasonable enough ISP to get off the surface at those pressures?

GoSlah27 investigated this and concluded that the only two viable options are the Aerospike and the Mammoth.

- - - Updated - - -

OhioBob: By extracting the curves used:

1. Kerbin uses 80% values of Earth's Standard Atmospheric Model for pretty much everything, however:

2. While density varies based on temperature (i.e. day vs night, equator vs temperate vs poles), pressure is purely based on altitude

3. The temperature offset (from the baseline altitude->temperature curve, which is indeed an 80% Standard Atmospheric Model) varies based on height. So in the stratosphere, it doesn't much matter if it's day or night or where on the planet you are, but in other atmospheric regions the sun-based offset is more considerable.

4. There is no seasonal temperature variation because without tilt (either real tilt, or inclination-induced tilt) there are no seasons, so looks like nobody bothered to make those curves.

Thank, Nathan. That pretty much confirms what I already thought.

[barfing_skull]

GoSlah27 investigated this and concluded that the only two viable options are the Aerospike and the Mammoth.

Huh. I'm sure this has been requested before, but I'd love to be able to run some simulations without actually having to send a craft to Eve.

Cheers,

-BS

[jwenting]

Gravity Turns are much more accurately modeled! If you take any rocket with tail fins, you can do a control free gravity turn.

Step 1 - launch as normal

Step 2 - pitch over to 10 degrees off horizon at about 1000m

Step 3 - Disable SAS, let the fins and gravity do the work - no control input needed, other than throttle adjustments!

tried that. But once I reach about 7-8km anything I build that has a meaningful payload starts oscillating wildly, flips over, and nosedives into the ground.

[OhioBob]

Huh. I'm sure this has been requested before, but I'd love to be able to run some simulations without actually having to send a craft to Eve.

What Slashy did was to study the ISP curves as defined in the engine config files to estimated what the ISP would be at 5 atm pressure. The only two engines that he thought would have adequate ISP and thrust to launch from deep in Eve's atmosphere were the Aerospike and the Mammoth. I haven't studied it myself to either confirm or deny.

[barfing_skull]

What Slashy did was to study the ISP curves as defined in the engine config files to estimated what the ISP would be at 5 atm pressure. The only two engines that he thought would have adequate ISP and thrust to launch from deep in Eve's atmosphere were the Aerospike and the Mammoth. I haven't studied it myself to either confirm or deny.

I was thinking more of something along the lines of the equivalent of a wind tunnel test. You simulate the atmospheric conditions, load, and drag (and probably a few other things) on Kerbin without launching a craft or gaining any science or funds. I have to imagine actual space agencies do this under perhaps idealized conditions. Perhaps something similar for aerobraking. Or would that make design too easy?

Cheers,

-BS

[Chaos_Klaus]

Oh, I would love to measure the drag coefficient of my rockets in a wind tunnel.

[Martinoss]

Kerbin Atmospheric Temperatures

(Moderators, feel free to split my posts off into a separate thread if you don’t think this is the appropriate place for them.)

I’m starting to get a handle on the atmospheric model. I’ve obtained some good temperature measurements and a logical pattern is starting to emerge.

I obtained atmospheric temperatures using two methods. First, I sent up a probe and took direct temperature measures. To do this I used the debug tool’s “external temperatureâ€Â, which is a better tool than the 2HOT thermometer for obtaining instantaneous ambient air temperatures. This method only worked below 25,000 meters when I could suspend the probe on a parachute. Above that altitude the probe was experiencing aerodynamic heating, either from powered flight or reentry, which elevated the temperature reading and made it useless for my purposes.

For the second method I sent up a probe with a barometer and took over 100 pressure measurements as the probe descended through the atmosphere. By analyzing the pressure rate of change I could calculate the scale height. From scale height, temperature could be calculated using T = Hg/R. For R the value 213.75 J/kg-K was used, which I computed from temperature and pressure measurements obtained at the KSC launch pad (described in my previous post).

From the direct measurements it became obvious that the temperatures are based on the U.S. Standard Atmosphere (USSA). Major changes in temperature gradient were observed at 8.8 km and 16 km. The USSA has the same gradient changes occurring at 11 km and 20 km. This tells me that the vertical height scale of the USSA has been compressed by a factor of 0.8 to make it better fit the smaller scale of Kerbin. Between the altitudes of 16 km and 25 km, the observed temperatures correspond exactly to the USSA (accounting for the compressed height scale). Below 16 km the measured temperatures were greater than the USSA, though this is likely to be due to a latitudinal modifier. The USSA is based on cooler mid-latitudes while KCS is located at the equator. More measurements are necessary to know for sure, but my guess is that Squad applies a temperature modifier based on latitude and time of day. This modifier is added to the USSA temperature. The adder is greatest at sea level , follows a curved function, and eventually tappers off to zero at 16 km.

My second method of temperature determination is less accurate and displays considerable fluctuation. However, the results clearly show that the Kerbin temperature profile follows the general shape of the USSA temperature profile. Once above an altitude of about 35 km, the pressure-derived temperatures are consistently lower than the USSA. Because of my inability to take direct temperature measurements at these altitudes, I don’t know if this is a real observation or an artifact resulting from some inaccuracy in my method. I would not be surprised if Squad has also applied a latitudinal-diurnal modifier to the upper atmosphere temperatures as well. As a general rule, when lower atmosphere temperatures are greater than the norm, upper atmosphere temperatures are lower than the norm, and vice versa. This pattern seems to hold true in my pressure-derived temperatures.

All of the above is summarized in the graph below. The USSA temperature profile includes the 20% reduction in vertical scale.

http://www.braeunig.us/pics/KSP/Kerbin_TempZ.gif

I’m encouraged enough by the pressure-derived temperatures that I think this method should work reasonably well to obtain temperature-height data for other planets. So far I’ve focused entirely on Kerbin, though I hope to also reconstruct other atmosphere models.

One final note regarding the specific gas constant. Although R = 213.75 J/kg-K produces some unrealistic values for molecular weight and air density, it seems to be the value that you want to use for anything involving pressure and scale height calculations. A more Earth-like value of 287 J/kg-K seems to be what you want to use for any calculation involving density, speed of sound, drag, etc. This is an internal inconsistency that exists in KSP that wouldn’t in real life. In KSP, air pressure decreases with increasing height more rapidly than it should according to the values of temperature and density.

Next up, a discussion of drag coefficients. I’ll post some experimental results latter today.

Would your data points be available to the public? And what was the scale height you found? I once wrote a script to determine the trajectory of a re-entry, and now that the drag is overhauled, it looks like I am gonna need it. Is this sort of data also available for the other planets?

I was also wondering if there are still parts physicsless like in KSP 0.90??

Edited May 7, 2015 by Martinoss

[John FX]

Oh, I would love to measure the drag coefficient of my rockets in a wind tunnel.

You may want to try Nu-FAR. It shows you a graph of your transonic drag in the VAB. Also calculates drag based on the surface shape of your craft...

[OhioBob]

Oh, I would love to measure the drag coefficient of my rockets in a wind tunnel.

There are ways to determine it experimentally in the game. If you use the AeroGUI mod you have a display giving dynamic pressure, drag force, and Mach number. Dividing drag force by dynamic pressure gives you AC_d. You then just have to divide by the area to give you the drag coefficient. The exact value of the drag coefficient is depended on the reference area that you use, therefore C_d can vary. You just have the state what the reference area is and be consistent. I like to use A equal to the frontal area of the rocket. Once you've computed Cd you can plot it as a function of Mach number.

You can also compute Cd from terminal velocity, but the above method is easier.

[barfing_skull]

Google works well on the forums, since the built-in search, uh, leaves a little to be desired. Anyway, I couldn't find if this had been addressed elsewhere, but how do you keep things like ladders and mobility enhancers which couldn't be put in a service bay (could they?) from overheating and exploding on re-entry?

For the final re-entry to Kerbin, I don't particularly care because I'm just going to recover the craft anyway, but I've got a lander I'd like to be able to land on other bodies with atmospheres, then bring back. I obviously can slow down the craft a little with retro-rocketry and chutes, but I'd like the atmosphere to do more of the work to preserve delta-v. Sadly, my tests on Kerbin result in the ladders and mobility enhancers kicking the bucket.

Put a larger (than the rest of the stack) heat shield just below the ladder on the stack? Hmm...possible but unwieldy. I suppose a service bay is possible, but again, a bit unwieldy.

Thoughts?

Cheers,

-BS

[OhioBob]

Would your data points be available to the public? And what was the scale height you found? I once wrote a script to determine the trajectory of a re-entry, and now that the drag is overhauled, it looks like I am gonna need it. Is this sort of data also available for the other planets?

I've learned quite a bit about the Kerbin atmospheric model since my last post on this topic. I will now share with you what I've learned.

First, just about everything you need to know about atmospheric models can be found here: Rocket & Space Technology, Atmospheric Models. If at any time I reference a table or formula, it can be found in that web page.

The U.S. Standard Atmosphere (USSA) is summarized in Table 4. As has been noted, Kerbin's atmospheric model is based on the USSA, though with the vertical scale compressed to 80% normal height. However, it's not a simple as just taking the height values in Table 4 and multiplying them by 0.8. Altitudes in the USSA are in units of geopotential meters, while the altitudes measured in KSP are in units geometric meters.

What Squad has done is to take the heights in Table 4, convert them to geometric heights using equation 15 (where r_o= 6371 km), and then multiply the result by 0.8. For example, the first temperature transition point is,

z = 6371*11/(6371-11) * 0.8 = 8.81522 km

Once these conversions are completed, computing the base temperature is easy because the temperature-height profile is just a series of line segments. However, to this base temperature are added latitudinal and diurnal modifiers. The latitudinal modifier at sea level varies from +17 K at the equator, to 0 K at 30^o latitude, to -50 K at the poles. The diurnal modifier varies according to latitude and solar incidence angle. At sea level at the equator, the diurnal modifier varies from +9 K at the halfway point between noon and sunset, to 0 K at the halfway point between midnight and sunrise. The diurnal variation is greater at mid-latitudes and less at the poles.

The base temperature adjustment also varies as a function of altitude. It is greatest at sea level, decreases to zero at approximately 16 km, and then begins to increase above approximately 26 km.

At sea level at the equator (approximately the location of KSC), the maximum and minimum air temperatures are,

T_max = 288.15 + 17 + 9 = 314.15 K

T_min = 288.15 + 17 + 0 = 305.15 K

Atmospheric pressure is also based on the USSA. As with temperature, the vertical scale of the pressure-height profile is reduced to 80%. Although this makes sense in terms of gameplay, it creates an internal conflict. Many calculations related to Kerbin's atmosphere, such as air density and speed of sound, use a gas molecular weight of 28.9644 kg/kmol (same as Earth). However, since the pressure-height profile is reduced to 80%, this means that the pressure drop that would be expected to take place over 1000 meters in altitude must now take place over 800 meters. This means that Kerbin air must be heavier than Earth air, i.e. M = 28.9644/0.8 = 36.2055 kg/kmol. The rate of pressure decrease is internally inconsistent with the atmospheric properties used in other computations.

Also be advised that atmospheric pressure is not effected by temperature. The same pressure-height profile is used globally regardless of local temperature.

In KSP, atmospheric pressure is calculated using a floatCurve. As such, pressure doesn't follow a pure exponential function. Furthermore, the normal real-world equations used to compute pressure (equations 20 and 21) won't work correctly when the input is geometric height rather than geopotential height.

To derive equations for pressure as a function of altitude, I found it easiest to simply plot LN(P) versus z and then fit the data points with a trendline. The equations provide quite good accuracy (particularly at lower altitudes); in almost all cases the error is just a fraction of a percent. Above 62.3 km the pressure curve turns abruptly as the pressure is forced to suddenly go to zero at 70 km.

Below is a summary of all the equations. These equations are not always an exact recreation of numbers produced by KSP, but they are a very close approximation. Note that the temperature given is the "base temperature". It does not include any of the latitudinal or diurnal adjustments.

[TABLE=class: grid, width: 1000, align: center]

[TR]

[TD=colspan: 3, align: center]Kerbin Standard Atmosphere, 0 to 70 km[/TD]

[/TR]

[TR]

[TD=align: center]Geometric

Altitude, z

(km)[/TD]

[TD=align: center]Kinetic

Temperature, T

(K)[/TD]

[TD=align: center]Pressure, P

(Pa)[/TD]

[/TR]

[TR]

[TD=align: center]0-8.815[/TD]

[TD=align: center]288.15 – 8.11117 × z[/TD]

[TD=align: center]EXP[ –5.6684193E-05 × z³ – 1.9580663E-03 × z² – 0.14836114 × z + 11.5260885 ][/TD]

[/TR]

[TR]

[TD=align: center]8.815-16.050[/TD]

[TD=align: center]216.65[/TD]

[TD=align: center]EXP[ 3.6215104E-06 × z³ – 1.1135486E-04 × z² – 0.19514262 × z + 11.753405 ][/TD]

[/TR]

[TR]

[TD=align: center]16.050-25.729[/TD]

[TD=align: center]196.7513 + 1.23980 × z[/TD]

[TD=align: center]EXP[ 1.8765033E-06 × z³ + 4.0682342E-04 × z² – 0.21018261 × z + 11.868445 ][/TD]

[/TR]

[TR]

[TD=align: center]25.729-37.879[/TD]

[TD=align: center]139.7102 + 3.45679 × z[/TD]

[TD=align: center]EXP[ –4.1690244E-05 × z³ + 5.0761226E-03 × z² – 0.36509550 × z + 13.508117 ][/TD]

[/TR]

[TR]

[TD=align: center]37.879-41.129[/TD]

[TD=align: center]270.65[/TD]

[TD=align: center]EXP[ –3.6083972E-04 × z³ + 0.044146606 × z² – 1.9496301 × z + 34.813685 ][/TD]

[/TR]

[TR]

[TD=align: center]41.129-57.440[/TD]

[TD=align: center]411.8568 – 3.43327 × z[/TD]

[TD=align: center]EXP[ –8.6910710E-07 × z³ – 1.0609374E-03 × z² – 0.062235305 × z + 8.6153226 ][/TD]

[/TR]

[TR]

[TD=align: center]57.440-68.798[/TD]

[TD=align: center]354.7555 – 2.43916 × z[/TD]

[TD=align: center]See below[/TD]

[/TR]

[TR]

[TD=align: center]68.798-70.000[/TD]

[TD=align: center]186.946[/TD]

[TD=align: center]See below[/TD]

[/TR]

[TR]

[TD=align: center]57.440-62.300[/TD]

[TD=align: center]See above[/TD]

[TD=align: center]EXP[ 1.5412356E-05 × z³ – 3.5432535E-03 × z² + 0.058616347 × z + 6.7777398 ][/TD]

[/TR]

[TR]

[TD=align: center]62.300-70.000[/TD]

[TD=align: center]See above[/TD]

[TD=align: center]EXP[ –9.37298921E-04 × z⁵ + 0.304103311 × z⁴ – 39.4630615 × z³

+ 2560.291576 × z² – 83044.59921 × z + 1077314.96395 ][/TD]

[/TR]

[/TABLE]

From the data in the table above, air density and speed of sound are calculated using the following:

Density, ÃÂ (kg/m³) = P/(RT)

Speed of sound, C (m/s) = (γRT)^1/2

where,

Specific gas constant, R = 287.058 J/kg-K

Specific heat ratio, γ = 1.400

Edited May 18, 2015 by OhioBob

[Wakaleo]

By the living Harry, there's a lot to learn in this game. I wish I'd paid more attention during maths class! But that was sooooo long ago in a galaxy far, far away. This forum is a great resource for saving Kerbalkind from reckless rocketeers (hmmm, could that be the byline of the builder/s of an enhanced re-entry capsule).

My parachute was destroyed during re-entry over the north polar region - Jebediah was doomed but looked happy as a pig in you-know-what until I reverted the game just before impact to save him. Could this self-immolating chute be an effect of higher atmospheric density over the poles and therefore higher re-entry temperatures?

[johnsonwax]

I've learned quite a bit about the Kerbin atmospheric model since my last post on this topic. I will now share with you what I've learned.

This is excellent information. I wonder if we know anything about how it could be adapted to the atmospheres of other bodies? Clearly there's a scaling of altitude and a scaling of sea level pressure, and a different temperature model. Any thoughts on how we could work out a path from here to there?

My guess is that if they've adapted the earth atmospheric model, they haven't developed a completely different base model for the other bodies - instead they just put a reasonable set of coefficients and scaling factors on the same model.

Edited June 8, 2015 by johnsonwax

[OhioBob]

This is excellent information. I wonder if we know anything about how it could be adapted to the atmospheres of other bodies? Clearly there's a scaling of altitude and a scaling of sea level pressure, and a different temperature model. Any thoughts on how we could work out a path from here to there?

My guess is that if they've adapted the earth atmospheric model, they haven't developed a completely different base model for the other bodies - instead they just put a reasonable set of coefficients and scaling factors on the same model.

I have some information on the other planets as well, though I don't yet have it in a presentable form. I'll probably post something eventually.