Launch optimal TWR

[chd]

I don't think I have a good idea about what the best TWR should be at launch; I've done a fair bit of Googling, and I'm finding conflicting answers (maybe because things changed back when the aerodynamics were changed?).

On one hand, "high TWR is best, because drag is largely irrelavent compared with gravity losses."

On the the other, "too high a TWR ratio -- more than somewhere in the 1.5 - 2.0-ish range -- is bad because you get too fast before you get high enough and the drag kills you."

So, what's the real story? :-)

[bewing]

High TWR engines are heavy (and expensive), so they waste fuel and money. And once they get into the upper atmosphere or space, they have a lower Isp than the low-thrust engines -- so they are not efficient, either.

So, a 1.5 to 2.5 TWR tends to be a reasonable compromise. Especially if you stage away your launch engine when you are halfway to space.

Edited August 29, 2020 by bewing

[VoidSquid]

Opinions will differ, @chd, some folks like me go for a lower TWR (usually 1.3-1.4), others for higher TWR for launching. I think a reasonable launch TWR is between 1.3 and below 2, as @bewing already said, a high TWR becomes inefficient in terms of money, weight. Plus, the higher the TWR, the more difficult to steer the rocket during gravity turn, and the more friction in the atmosphere.

In the end, I suggest you play KSP the way YOU like it.  

[Grogs]

I was playing around with a rocket design yesterday, and I had two choices:

1. Fire just the outer SRBs at launch, and then stage the inner core liquid engines when the SRBs burned out. Pad TWR ~ 1.3

2. Fire both the SRBs and the core LFO engines at the same time. Pad TWR ~ 2.0.

I tried both a couple of times using Mechjeb (so no pilot skill, or lack thereof to cloudy the results).

For option 1, it almost seemed like the ship struggled to get out of the atmosphere at the end, and the engines kept firing until almost 60,000 meters. However, the flight was extremely smooth, and once the Ap was up to 80 km, the arc was very wide. There was about a 5-minute coast before the final circularization burn, and it was only 300 m/s more.

Option 2 seemed to be a bit more efficient in terms of overall dV, but it was quite a bit more hectic. Final Ap (and SRB sep) was achieved when the ship was at around 35,000 meters. Following the standard gravity turn profile, this meant that the ship was on a high, narrow arc. The circularization burn was something like 1100 m/s. The biggest issue here was that the ship had to clear the atmosphere, and then very quickly turn about 30 degrees to the maneuver node to make the burn. On a couple of tests, I didn't get the ship turned in time and ended up with a higher Ap than I wanted.

Ultimately, I decided I liked Option 1 (the lower TWR) better. It's much more forgiving to small errors in piloting, and the long coast at the end means that I have plenty of time to set up the circularization burn and make it accurate.

Most of the time with other rockets, I shoot for something in the 1.6-1.7 range. I feel like that's a good range to avoid either the feeling of struggling to get out of the atmosphere or rushing to make a huge circularization burn in time. That said, rocket design is a compromise of a lot of competing factors. If I have a ship that can make it to orbit with sufficient dV remaining using a launch TWR of 1.2, I don't see a reason to strap on an extra ring of boosters just to have extra fuel in orbit that will probably go to waste anyway.

[Gargamel]

7 hours ago, VoidSquid said:

In the end, I suggest you play KSP the way YOU like it.  

I like to build a set launcher, with SRB's, and then adjust the thrust of the SRB's to get the TWR down to 1.8.  I then use MJ to control the throttle to keep the TWR at 1.8-2 during the launch.  If I built it right, eventually the Main engine is almost at 0 throttle as the SRB's burn out and stage off.  Allows me to carry a decently strong engine on the core stage that also works decently as a second stage insertion engine.

[Okhin]

It depends if you have control over the throttle during flight or not. If I have Solid Boosters (I do have them 70% of my launch), I tend to opt for a low TWR because it will rise fast as I ascend, and under 12000m in Kerbin atmosphere, I try to not push too hard through the atmosphere, especially with wide fairings (like when launching rotating rigs for instance, or super wide bases). Usually after the 12 000 m mark, I can usually go crazy and blast full throttle on my non SRB engines.

However, too low of a TWR can lead to unstability, especially when you try to flip over the rocket. So when I'm borderline (under 1.2), I need to start the gravity turn way later than what I usually do, or the rocket will just flip around too much and crash.

If I can manage without SRB, then I'm going crazy, and I can go to 2 TWR on the launchpad (if I'm above 2, I generally consider that my launcher is too heavy for the jo I need, and I scale things down). But again, under the 12 000 m altitude, I tend to stay between 1.4 and 1.8 TWR. It seems aerodynamics fights my rocket under this altitude. Then I gradually throttle up and once I reach 20 000, I'm going full throttle (and sometime with a TWR of 3 or 4) and I embrace the plasma, aiming for the shortest possible circulation burn (those atmospheric engine I have at this stages under-perform when they lack atmosphere, so I might as well use them in atmospheric situation).

So it depends on your use of SRB and the wideness of your payload, according to my experience.

[Wobbly Av8r]

I've never really looked at TWR being "the" factor in deciding to use all engines or just the SRB's to get going off of Kerbin - if you're talking about efficiency, i.e. least amount of fuel used to establish LKO, the decision is influenced by how efficient your LFO engine(s) are vs. the increased amount of drag you will have to overcome by accelerating too quickly in the lower, denser atmosphere.

A different way to think about it would be that, say, if there were no atmosphere on Kerbin, it would probably take ~2400 dV (wild stab to illustrate the principle) to establish a 70-75 km orbit, so the atmospheric drag is what causes you to require more dV (on average 3100 - 3400 dV). That atmospheric drag requirement increases if you accelerate too quickly in the denser, lower atmosphere, and most (if not all) LFO engines burn their fuel less efficiently down low as well.

Beyond these quantities are some interesting aerodynamics, like the Twin-Boar, whose drag rises quickly up to Mach 1... and then quickly drops (the drag being produced not from 'headwinds' but from the 'suction' of the base of the booster...). Again, the joy of KSP is discovering the strengths and weaknesses of any number of parts and strategies, each somewhat unique in their application, which in turn perform better at various TWR's, not just one ideal value.

Edited August 29, 2020 by Wobbly Av8r
average burn for Kerbin LKO

[jimmymcgoochie]

The higher your TWR, the higher you’ll get the same rocket with the same mass as it can fight against gravity better. On airless bodies, this is all you need to worry about- for example on Tylo, which has around 0.8g gravity- and on lower gravity bodies it’s easy to get more than enough thrust even with smaller vacuum engines.

However, in atmospheres you also have to deal with aerodynamic drag. Drag is highest at low altitudes and builds up a lot when you go supersonic, so too much thrust can cost you more fuel by going too fast too soon- especially on Eve- and so a more moderate TWR is probably the most efficient compromise between gravity and aero losses.

I personally try to get a TWR between 1.35 and 1.5 on the launch pad, any lower and it doesn’t climb fast enough but any higher and the drag can build up too much. TWR rises as you climb due to a combination of decreasing mass (as fuel gets burnt) and increasing ISP (less atmospheric pressure) so if necessary throttle back a little to keep your TWR below 2 until you reach around 30km, at which point air drag is minimal and you should focus on gaining orbital velocity. Just try to avoid cooking your payload on the way up!

[OhioBob]

There's really no "best" answer to the TWR question.  But I've found that it generally comes down to what it is you are trying to optimize.

If we want to minimize the amount a delta-v it requires to reach orbit, then higher TWR is better.  This is because when we use a high TWR a larger fraction of our rocket's thrust is accelerating the rocket upward rather than simply counteracting gravity, so we reduce gravity losses.  Drag losses will be higher with a high TWR rocket, but drag losses are typically small compared to gravity losses, so it shouldn't matter too much.  I would say that anything greater than a TWR of about 2 is probably too much, but that's just my opinion.

If our goal is to optimize for cost, i.e. the least cost per ton of payload put into orbit, then lower TWR is better.  This is because of the high cost of the engines in comparison to everything else.  It's less expensive to use smaller engines and lower TWR.  Using low TWR means that our gravity losses will be greater, hence it will require more delta-v to reach orbit.  We attain the extra delta-v by increasing the mass ratio, or the fuel fraction of our rocket.  But fuel tanks and fuel is cheap in comparison to large powerful engines, so our launch cost goes down.  The minimum useful TWR is about 1.2.

I prefer the later method.  I typically target a launch TWR in the range of 1.3 to 1.4.  That's being said, however, if the only engine available to me produces a high TWR, then I use the high TWR.  I don't purposely throttle an engine back just to get it within some preferred TWR range.
 

Edited August 29, 2020 by OhioBob

[GuessingEveryDay]

I just get it as high as I can, and then take it easy.

[Aelipse]

Good question. In an environment devoid of atmosphere, the highest possible TWR would be the best for the reasons mentioned in the original post.

Once you are launching from an atmospheric body, the main factor is aerodynamics. If you're launching a large rover and you didn't bother to put it in an aerodynamic fairing, you want to go as slow as possible through the atmosphere and a TWR of 1.2 - 1.4 will be optimal. If on the other hand you are obsessed with aerodynamic shapes like me, to the point of putting solar panels and antennas into cargo bays, you might wanna aim for a TWR of 2 and more.

What I am trying to say is that what you are asking is dependent on too many variables: gravity, atmosphere density and its height, not to mention the rate of change of pressure and density with altitude; then you need to take into consideration the efficiency of your engines (with engines of incredible efficiency it would be probably best to get through the whole ascent in subsonic speeds to keep the drag at its minimum), weight of your craft, staging and last but not least the drag coefficient of your craft.

I imagine there must be some insane calculations going into planning of the flight path of real world space crafts, and even then I would be not surprised if they only aimed for 'good enough' instead of 'optimal'.

And that would be my advice to you. Try a few ways of launching your vehicle and stick to whatever is 'good enough'.

[CBase]

29 minutes ago, Aelipse said:

the main factor is aerodynamics

Actually I don't think so. For the launch profile challegene I did quite some test if limiting TWR on high drag phases helps to get more efficient. It didn't, the gravitional loss was bigger than aerodynamic gain. My guess after this challegene is that dry weigth impact from number of engines is main factor. At least if you do SSTO for recovery. With expandable booster very high TWR for short duration might pay off.

[Aelipse]

18 minutes ago, CBase said:

Actually I don't think so. For the launch profile challegene I did quite some test if limiting TWR on high drag phases helps to get more efficient. It didn't, the gravitional loss was bigger than aerodynamic gain. My guess after this challegene is that dry weigth impact from number of engines is main factor. At least if you do SSTO for recovery. With expandable booster very high TWR for short duration might pay off.

That really depends on the range of mass / thrust / engine efficiency and drag coefficient within which you were performing the tests. I haven't tested this myself and I'll be happy to take your word for it.

However, if aerodynamic drag was not significant, everybody would go straight for insane TWR and tilt their rockets directly to the east right after lift off, which is not something I have ever seen from any experienced players. Besides, I have recently launched a large rover attached to a sky crane, which means I wasn't able to put it inside a fairing, and I can tell you, even though the starting TWR was below 1.5, I soon found myself throttling down just to avoid the thing breaking apart, not to mention the drag force losses. So, in some instances, the drag does play a noticeable role.

 

EDIT: After reading through my earlier post and your reaction again, I realised there has been a poorly worded claim on my part. What I meant by 'aerodynamics being the main factor' was that aerodynamics is the main factor in favour of the steep climb. The other factor being gravity only favours flat, horizontal ascents. In other words, if you are deciding how steeply you should climb, the "dragginess" of your craft is the main thing you should be concerned with. I hope we understand each other now.

Edited August 29, 2020 by Aelipse

[CBase]

4 minutes ago, Aelipse said:

everybody would go straight for insane TWR

No, since higher TWR => more engines => more dry weight => more overall weight => less efficient

And on very shallow profiles skin overheating is another issue. Actually I even modified MechJeb to better throttle down to avoid overheating for very aggressive profiles. But without shallow ascent profile the high TWR looses advantages as you simply coast earlier.

On real scale earth escape velocity is a lot higher, ascent profile very different and therefore aerodynamic losses bigger, so real rockets do throttle down and do not benefit from a high TWR. But for Kerbin it hardly matters.

[Aelipse]

6 minutes ago, CBase said:

No, since higher TWR => more engines => more dry weight => more overall weight => less efficient

For SSTOs perhaps. Otherwise you just drop the extra weight of the engines and their respective tanks. Boosters never make any craft less efficient since you just leave them behind.

7 minutes ago, CBase said:

And on very shallow profiles skin overheating is another issue. Actually I even modified MechJeb to better throttle down to avoid overheating for very aggressive profiles. But without shallow ascent profile the high TWR looses advantages as you simply coast earlier.

It's true I didn't specifically mention aerodynamic heating. Isn't aerodynamic heating part of the aerodynamics? I imagine the heat transfer must be lower for crafts with smaller cross-section area.

[Wobbly Av8r]

Quote

For the launch profile challegene I did quite some test if limiting TWR on high drag phases helps to get more efficient. It didn't, the gravitional loss was bigger than aerodynamic gain.

While I can appreciate your point, the challenge addressed the same aerodynamic shape on different profiles (as far as I can tell). What I think your challenge would confirm is that certain designs are best blasted into the heavens at the speed of heat, but I've got designs that definitely are more efficient with a 'slower' profile, whether that comes from launching more vertically or throttling down in the lower atmosphere and using the increasing Isp to get a few extra miles per gallon - in fact I discovered that by running out of fuel before establishing orbit and then changing to the slower profile and getting into orbit with fuel to spare!

Edited August 30, 2020 by Wobbly Av8r

[Spricigo]

 

 I consistently launch rocket with more launchpad TWR than what people usually call "reasonable" in those forums.  TWR as high as 3 or more can be fine if you know what are you doing. 

Yes you may have more engines than needed and getting more aerodynamic dag than you need, but if you can afford it, why not? Put those engines to good use instead of being afraid of being "inefficient".

 

[CBase]

12 hours ago, Wobbly Av8r said:

changing to the slower profile and getting into orbit with fuel to spare

Interesting, could you share the craft and a rough description of the slower profile ? If it isn't a wardrope without a fairing that you try to put in space, I am curious to see it.

 

15 hours ago, Aelipse said:

Boosters never make any craft less efficient since you just leave them behind

Absolutely true, I am only flying SSTO rockets, so my thoughts on optimal TWR are focused on these. If you consider booster an option, "Moar Booster" for short duration is definitely improving performance until you hit ~ mach 1 when it is aerodynamic drag that is main factor.

SSTO and other liquid fuel rockets do waste a lot of fuel in first seconds when vertically fighting gravity. But just like Elon Musk recently twittered: even loosing something like 40% payload pays off after 3 launches if you recycle.

[Wobbly Av8r]

23 hours ago, CBase said:

Interesting, could you share the craft and a rough description of the slower profile ? If it isn't a wardrope without a fairing that you try to put in space, I am curious to see it.

Yes - not sure what the preferred method of making a craft file available here would be, but here is an image - the boosters are intended to lift a 43.789t payload into LKO. I have, indeed, found a narrow set of solutions solely based on launch profile, none of which occur at max possible TWR from liftoff.

While there are virtually infinite ways to go about accomplishing a task in KSP, my intent of using this booster set was due to the fact that the Twin-Boar and attached Jumbo 64 tank can be recovered near KSC for > 20,000 kerbucks, which reduces my overall cost. And as it applies to our conversation, it is just fortunate happenstance that the only way I could get this vessel into LKO was a very narrow window of profiles (with 3 m/s dV left in the booster...) [ Edit: I was actually able to eke the left-over fuel to 5 m/s of dV ! ]

Edited August 31, 2020 by Wobbly Av8r
Tweaked profile to achieve 5 m/s dV leftover...

[Wobbly Av8r]

Thanks for the info link, CBase; hope this link works...

https://kerbalx.com/Wobbly_Av8r/Get-Me-To-LKO

Give her a rip at max TWR and see if you can find a profile that will achieve LKO. If you can, let me know what profile you used, and if you are unable to, I'll let you know what I found to work.

The intent of the design is to enable the booster to achieve LKO so that I can deorbit at the time of my choosing that enables me to get close to KSC for max credit for recovery.

Edited August 30, 2020 by Wobbly Av8r

[Wobbly Av8r]

Pardon for brushing aside the tumbleweeds of this thread, but wanted to share what I found out about the Get Me To LKO vessel profile. To minimize deviations, I performed the profiles using MechJeb to compare one profile to another)

At the most optimal angle for this specific design (~25% profile shape) and starting the turn at 30.5m altitude, the closest I could come to establishing an orbit at full thrust (max TWR) was short by ~64 m/s dV and as a result the booster never raised the Pe above the atmosphere initiating a premature re-entry. By limiting the throttle to 65% all the way up, I was able to get the booster into LKO with ~5 m/s dV left, which when the payload is staged becomes about 40 m/s dV for de-orbiting the booster, which works well because that will occur 10-20 minutes (or another orbit or two) later.

As to the TWR question of the OP, while the amount of dV here is actually not so significant, it definitely shows that Kerbin's atmosphere has an influence on efficiency - a relatively simple change of profile in this case decreases the overall cost of the mission at least 10%, in practice even more.

[RizzoTheRat]

On 8/29/2020 at 10:47 PM, CBase said:

Actually I don't think so. For the launch profile challegene I did quite some test if limiting TWR on high drag phases helps to get more efficient. It didn't, the gravitional loss was bigger than aerodynamic gain. My guess after this challegene is that dry weigth impact from number of engines is main factor. At least if you do SSTO for recovery. With expandable booster very high TWR for short duration might pay off.

I keep meaning to revisit that challenge, it's a bit limited in that it's a single design, I had plans for rounds 2 and 3.

My standard approach for a multi stage rocket, which I did a bit of experimentation to produce and is what triggered that challenge, is to aim for about 1.6-1.8 TWR at launch, with a first stage powerful enough to get the Ap out to 60 seconds, then 1.2-1.4ish TWR second stage to maintain that 60 seconds.  Throttling to keep the Ap 60 seconds then improves efficiency by reducing the gravity loss is the initial pitch over for the gravity turn wasn't aggressive enough.

I used to limit the ascent speed based on Q but in the current aerodynamics model I don't think it really makes a lot of difference unless you have a very high dV

Edited September 3, 2020 by RizzoTheRat

[CBase]

On 9/3/2020 at 3:50 PM, Wobbly Av8r said:

At the most optimal angle for this specific design (~25% profile shape) and starting the turn at 30.5m altitude, the closest I could come to establishing an orbit at full thrust (max TWR) was short by ~64 m/s dV

Unfortunately I do not have Making History to check this specific design. I did rebuild it closely with some SpaceY boosters and a weight at top, but at this small dV differences even small build changes make too much difference.

But on the topic of this thread about optimal TWR it is a nice example that the answer is not purely to look at TWR and indeed the added hammer booster only barely help to achieve orbit: TWR ~2 is high enough so you are not loosing a lot of efficiency to gravity and going more shallow with the extra thrust is eaten by increased drag, therefore your solution to throttle the main stage early on makes sense.

However it might be more sensible to look at your vacuum Isp during ascent beyond 15km altitude: The twin boar engine is sub optimal compared to a main sail or pure vacuum engines. The major dV proportion for gaining orbit is mostly at low pressure. So the second stage or SSTO stage should always have an engine with okay to good vacuum Isp. Then you add either a first stage or boosters to get high enough TWR at launch pad. For SSTO you add engines until you reach at least 1.3 TWR. A true two staged designed will propably aim for higher TWR at launchpad and lower TWR with the vacuum stage, like RizzoTheRat posted.

[Wobbly Av8r]

2 hours ago, CBase said:

But on the topic of this thread about optimal TWR it is a nice example that the answer is not purely to look at TWR [...]

However it might be more sensible to look at [ ...]

No disagreement that there are a LOT of different ways to get things done and I will look into what you described - I'm sure it will help my designs down the, uh, 'road'.

But all said and done, your quote about TWR is essentially what I was trying to convey to @chd and anyone else following these types of debates - the TWR is more of an indirect result than a primal factor. In this case the TWR was preempted by considerations of efficiency, allowing optimal recovery, ultimately making the cost of lifting 43+ tons into LKO cost  a meager ~23,000 due to the "recovery friendly" properties of the Twin-Boar booster!

All said and done, I appreciate your feedback and the conversation.

[Edit: OH! And I HIGHLY recommend the DLC Breaking Ground - the whole franchise has been worth every penny!]

 

Edited September 5, 2020 by Wobbly Av8r
Breaking Ground... highly recommended!

[CBase]

3 hours ago, Wobbly Av8r said:

OH! And I HIGHLY recommend the DLC Breaking Ground

I do have Breaking Ground DLC for the robotic parts, just not Making History. It didn't appeal me somehow.

 

3 hours ago, Wobbly Av8r said:

"recovery friendly" properties of the Twin-Boar booster

Actually what is recovery friendly on Twin-Boar ? I would always consider a Main Sail engine with 2 Jumbo tanks and small control wings more recovery friendly as the performance across all altitudes is better. Personally I added fixed and recovered 1.25m tank and engines to scale performance until I unlocked bigger diameters.