Determining most efficient engines?

[strider3]

I am 2/3 of the way through my scientific exploration of Moho. What I have discovered is the tremendous amount of fuel I have expended to keep my science lander fueled and making landings on Moho. I originally sent a scanner, the lander and 3 refueling ships...and the amount of fuel actually available to the lander after the refuelers were in the working orbit was...depressing. I still have 4 landings at biomes to complete and I'm out of available fuel for my lander. I don't know if getting fuel to Moho orbit is always this costly or if it's the UT years I am trying to accomplish this in (11-13). Either way, I may have fallen into the "bigger is better" rabbit hole. I built large refuelers, refueled them from my Kerbin space station and sent them to Moho only to have (comparatively)  little fuel left after they were in my Moho working orbit of 30 km. I feel that part of the problem, besides the size of these refuelers, may be the efficiency of the engines I am using to get from Kerbin orbit to Moho orbit. How does one determine the efficiency of an engine (in vacuum)? What numbers do you look at as far as engine performance to determine which engine will be the most fuel efficient in getting to Moho?

My "off the cuff" calculations for fuel needed to complete 4 more landings, return to a close to equatorial Moho orbit (I'm at 54* inclination now) and return to Kerbin is about 30,000 fuel. I'm basing this number on my lander's Delta V when full of fuel as it will be the return to Kerbin vehicle (barely!). I realize there is way more to this than I am asking but, at this point, I just want to get my biome landings done and get Ferbin and Arbart back home.

[Fraktal]

41 minutes ago, strider3 said:

How does one determine the efficiency of an engine (in vacuum)? What numbers do you look at as far as engine performance to determine which engine will be the most fuel efficient in getting to Moho?

That's kinda what vacuum Isp is. What engines are you using in your craft?

There's also of course the matter that carrying more fuel for more delta-V means carrying more weight and thus less delta-V. Diminishing returns hit fast and hard, especially since the best vacuum engines are all weaklings in terms of thrust. Just yesterday I designed a medium-size refueling craft myself; it has two Rockomax 32 tanks of fuel for itself and another two tanks hauled for refueling purposes, meaning that it can supply any ship with over 3000 LF but can barely squeeze out 2000 m/s of delta-V for itself due to massing over 86 tons when fully loaded and having around 0.32 TWR due to running on two Poodles.

Simply put, improving fuel efficiency by selecting the right engine is one option, but it doesn't change the fact that fuel is heavy and you still need to lug that weight around. Hence why ISRU is a thing.

Edited June 29, 2020 by Fraktal

[strider3]

26 minutes ago, Fraktal said:

That's kinda what vacuum Isp is. What engines are you using in your craft?

Here's what I have been using to get fuel to various bodies. It's a triple asparagus setup and I do not eject the last 2 outer stacks. It's all Mammoth's. I leave them attached, drag their empty weight to my Kerbin refueling station (350 km orbit) and fill everything up. This has not really worked as it seems to result in very little fuel/Lox left once I have it at my working Moho orbit of 30km.

The vacuum ISP tells me thrust...but not really "efficiency"? It doesn't tell me "miles per gallon"?

Edited June 29, 2020 by strider3

[strider3]

36 minutes ago, Fraktal said:

Simply put, improving fuel efficiency by selecting the right engine is one option, but it doesn't change the fact that fuel is heavy and you still need to lug that weight around. Hence why ISRU is a thing.

ISRU?

[N_Danger]

19 minutes ago, strider3 said:

ISRU?

In-Situ Resource Utilization (ISRU)

https://www.nasa.gov/isru

Mining ore and converting it to fuel.

https://wiki.kerbalspaceprogram.com/wiki/Ore

[Fraktal]

27 minutes ago, strider3 said:

It's all Mammoth's.

That's the problem. Mammoths are not very efficient for vacuum (Isp 315). If you need to be able to land the refueling vehicle, use a Rhino (Isp 340), otherwise Poodles (Isp 345 but low thrust) or Nervs (Isp 800 but very low thrust for its weight and doesn't work well with other engines) will do the trick.

16 minutes ago, strider3 said:

ISRU?

In-Situ Resource Utilization, AKA cooking your own fuel on-site. That's what you brought the scanner for, I presume?

[strider3]

1 minute ago, N_Danger said:

In-Situ Resource Utilization (ISRU)

https://www.nasa.gov/isru

Mining ore and converting it to fuel.

https://wiki.kerbalspaceprogram.com/wiki/Ore

That I have done on Dres...but it doesn't really help on this situation? Or am I misunderstanding your point?

[Fraktal]

You haven't brought any drills and converters with you to Moho?

[strider3]

1 minute ago, Fraktal said:

That's the problem. Mammoths are not very efficient for vacuum (Isp 315). If you need to be able to land the refueling vehicle, use a Rhino (Isp 340), otherwise Poodles (Isp 345 but low thrust) or Nervs (Isp 800 but very low thrust for its weight and doesn't work well with other engines) will do the trick.

In-Situ Resource Utilization, AKA cooking your own fuel on-site. That's what you brought the scanner for, I presume?

Nope, the scanner was to find biomes only.

Just now, Fraktal said:

You haven't brought any drills and converters with you to Moho?

I am not looking to make fuel at Moho...only to land and get science. My mining operation will be on Dres.

[Fraktal]

Ah, I see. I'd say sending another ship with mining equipment over to Moho might be an option for now, if your guys don't mind tanning on Moho for another year or so.

[N_Danger]

6 minutes ago, strider3 said:

That I have done on Dres...but it doesn't really help on this situation? Or am I misunderstanding your point?

Just answering the question of what is ISRU?

[strider3]

What I need is to get out of the "bigger is better" hole and send some ships to Moho which will have fuel to continue the exploration and get all that info and my 2 Kerbals home.

Just now, N_Danger said:

Just answering the question of what is ISRU?

Thank you.

[strider3]

I'm still confused on the engine efficiency question? ISP seems to deal with thrust...not "miles per gallon". I would like to know which engines are the most efficient in their use of fuel and which numbers tell you that?

Edited June 30, 2020 by strider3

[Superfluous J]

13 minutes ago, strider3 said:

I'm still confused on the engine efficiency question? ISP seems to deal with thrust...not "miles per gallon". I would like to know which engines are the most efficient in their use of fuel and which numbers tell you that?

Isp has nothing to do with thrust. Well, not directly.

It's literally the change in impulse you get for spending a specific mass of fuel.

It's roughly how fast you throw stuff out the back, which will in turn throw your rocket forward faster than an engine with a lower specific impulse.

An Ion engine has a really high specific impulse, because it throws a small amount of mass VERY FAST. The probe with an ion engine can therefore have thousands of m/s of dV for very little mass, but takes forever to get going.

Eli Manning has a really low specific impulse, because he can throw a much larger mass, but it goes quite slow (at least in rocket terms). He would accelerate much quicker than the probe, but use up all his fuel (the football) very quickly and not actually achieve that much of a change in his velocity.

One of the puzzles of rocketry is actually the balance between Isp and Thrust. You want more of both but getting more of one almost always results in less of the other.

Edited June 30, 2020 by Superfluous J

[Grogs]

If you have Making History, the Wolfhound is a good long-haul vacuum engine for a fairly large craft. At 380 vacuum ISP, it's the best chemical engine I believe. I find it to be good for a transfer stage to move a relatively large payload a long way. I used 7 of them in an asparagus setup to move a 300-ton payload from low-Kerbin orbit to Jool. The stage gives about 4500 m/s dV.

 

For something mid-sized, the Poodle is a good engine. The ISP is a little lower, but it's half the mass which makes up for it. This was my Moho return craft and it had about 3500 m/s. That's not enough to make Kerbin orbit, but plenty to get back and make a direct re-entry.

 

If you need more dV, you can go with a Nerv setup. I think this one had about 5,000 fully fueled.

[Gargamel]

9 minutes ago, Superfluous J said:

Eli Manning has a really low specific impulse, because he can throw a much larger mass, but it goes quite slow (at least in rocket terms). He would accelerate much quicker than the probe, but use up all his fuel (the football) very quickly and not actually achieve that much of a change in his velocity

That, and he'd probably miss the target.....      Sorry, sorry!  Back on topic!

 

[Spricigo]

5 hours ago, strider3 said:

Either way, I may have fallen into the "bigger is better" rabbit hole.

 

A good refuelling tank is a combination of all the tanks you may need with the most efficient engine you can support the lack of thrust. Other then that is a matter of reducing the need for refuelling (optimizing the crafts that need it and the logistics of refuelling).

If for some reason you don't want to set up a mining base on Moho, consider Minmus instead (or Gilly from where the transfer to Moho will be cheaper) that way you at least cut the problematic launching from Kerbin.

Edited June 30, 2020 by Spricigo
spelling

[jimmymcgoochie]

ISP is fuel efficiency- it’s a measure of how long a specified quantity of fuel can provide a specified quantity of thrust (in Americanish, one pound of fuel sustaining one pound of thrust). A higher number means that you get more distance out of each unit of fuel so bigger is better.

Generally speaking, the more efficient engines have lower thrust and vacuum optimised engines have the best ISP in space but are much less effective in atmosphere because the air gets in the way. ISP is directly dependent on exhaust velocity so ion engines (which fire ionised gases like xenon at tremendous speeds) have huge ISP but really low thrust; nuclear engines like the NERV (which heat the fuel to very high temperatures) have very high ISP but still relatively low thrust; and chemical rockets provide differing ISP depending on whether they’re optimised for space (high vacuum ISP but low atmospheric ISP) or launches (moderate ISP all round and higher thrust in most cases); and solid boosters are the least efficient of all.

To get to Moho you could use either the Rhino (for high thrust) or Poodle (low thrust but a bit more fuel efficient), both are high efficiency engines. The Mammoth is a launch rocket and is heavy, less efficient in space and is also overkill for thrust when you’re already in orbit. If you have Making History then the Wolfhound is a possibility, but it’s pretty heavy so you might not see any real benefits over the Poodle.

 

[Streetwind]

7 hours ago, strider3 said:

I'm still confused on the engine efficiency question? ISP seems to deal with thrust...not "miles per gallon". I would like to know which engines are the most efficient in their use of fuel and which numbers tell you that?

Specific impulse is, first and foremost, a mathematical invention designed to bridge the gap between imperial and metric units. Still, there are a few different ways in which one can try to visualize this odd unit.

My favorite way is this: specific impulse, measured in seconds, is the amount of time a rocket engine can run while holding up a given amount of fuel against gravity, using only that amount of fuel.

In other words: if you have 1 kg of fuel, then in order to hold it up against standard gravity, you need to exert 9.80665 N worth of continuous force. Now imagine you have a rocket engine that is throttled to produce exactly this much thrust, and you feed it exactly 1 kg of fuel. How long does your engine burn before it runs out of fuel? This many seconds. This is specific impulse.

Obviously, the longer an engine runs with the same limited amount of fuel, the more fuel efficient it must be, right? Therefore, specific impulse must be the measure of fuel efficiency of rocket engines, and the higher, the better.

The reason why you may be confusing Isp with a metric that governs thrust is because they are sort of related in a not entirely intuitive way. Because there's a third figure that hasn't been mentioned yet: mass flow rate. In a rocket engine, you pipe in 11 parts oxidizer and 9 parts fuel per unit of time, and combustion happens, and out comes thrust. The mass flow rate is the absolute measure of throttle. If you throttle up to 50%, you're asking for half the maximum possible mass flow rate. If you're throttling up to 100%, you're asking for all of it. And so on. And in a perfect world, which we call the vacuum of space, 100% mass flow rate translates into 100% thrust.

But the world is not perfect. Specifically, the world that is called Kerbin has an imperfection that is called atmospheric pressure. This pressure pushes against the exhaust jet of the rocket engine. And therefore the engine must expend effort in order to push back against the atmosphere. This effort isn't free - it consumes some of the energy produced in the combustion. And energy that has been consumed cannot go on to become thrust that pushes the spacecraft. So the maximum possible mass flow rate in the atmosphere is still 100% at a throttle setting of 100% - but the thrust is not. It may only be 90%, or even less, if you're using an engine that is not designed for atmospheric operation.

Let's return to the example of the rocket engine that runs for this many seconds on 1 kg of fuel, which defines specific impulse. This experiment is only valid if the engine is throttled to exactly the required thrust to fight local gravity. But what if this experiment happens in an atmosphere? Then the engine's thrust output at a given throttle setting is reduced, and you need to throttle it higher to compensate. It needs a higher mass flow rate to produce the required amount of thrust. And if the mass flow rate is higher, then more fuel flows per unit of time. So your 1 kg of fuel is going to drain faster, lasting fewer seconds. Your specific impulse is lower inside an atmosphere by exactly the same factor that thrust at a given throttle setting is also lower inside an atmosphere. This is how the two are related.

 

Oh, and by the way - Moho is the single most dV-intensive destination in stock KSP. Even more so than Eeloo. You are struggling right now because you are attempting a hard challenge, not because you are bad at the game.   Sure, there are probably things you can do better, but Moho is not supposed to be, and will never be, fuel efficient to explore. The only solution that lets you avoid expending massive quantities of fuel to ship small amounts of fuel from Kerbin to Moho is to send a mining vessel and produce the fuel directly at Moho.

Edited June 30, 2020 by Streetwind

[Xurkitree]

Use Ion lander lmao. Moho is practically baked in sunlight, you barely need panels to run a full array of ion engines.

[strider3]

Let's see if I understand ( a big "if"  ). Say I replace the central mammoth with a GNR 2500 nuke engine and use it only once in Kerbin orbit for the transfer to Moho. I'd keep the 6 mammoth asparagus setup (minus the fuel ducts to the central stack) to get to Kerbin orbit. The GNR 2500 will take a much longer burn to get me on the way to Moho (lower vacuum thrust than a mammoth) but should, if I understand, use much less fuel per second (higher vacuum ISP). The question becomes...does the much longer burn required using the GNR 2500 negate the increased efficiency of it's higher vacuum ISP...making the whole thing "moot"?

[strider3]

I don't seem to have a "liquid hydrogen" tank available for the GNR??

[Streetwind]

7 minutes ago, strider3 said:

The question becomes...does the much longer burn required using the GNR 2500 negate the increased efficiency of it's higher vacuum ISP...making the whole thing "moot"?

There is indeed a point where acceleration becomes so miniscule that you start having to spend serious amounts of extra dV to get anywhere. For example, real life electric propulsion is so anemic that, in order to climb out of Earth's gravity well, they have to fly a spiral trajectory. Imagine setting "hold prograde", throttling to 100%, and then timewarping for days on end while looping around and around and around, slowly pushing the orbit outwards through constant application of miniscule thrust. Such a trajectory might cost more than twice as much dV as a normal Hohman transfer with near-instantaneous burns; the worst case scenario is 240-ish percent of the ideal Hohmann figure.

However, no engine in KSP is so weak as to force you onto a constant thrust trajectory. Not even the "Dawn" ion thruster. Well... assuming you don't stick a single one onto a gigantic mothership, anyway

You have an entirely different problem to contend with, though: your orbit around the planet. As your burn time increases, you must start to burn earlier before the node, and must continue burning longer past the node. But because you are in a circular orbit, that means the further away from the node you burn, the further off-prograde you burn. This does introduce some inefficiency, yes - but much worse is the way it pushes your orbit out of shape. If you have a really long burn and your periapsis is only a bit above the atmosphere, you may very well accidentally push it into the atmosphere by burning that early and that far off-prograde. And you introduce an error in your final trajectory that may mean that you now need a mid-course correction to hit your encounter despite executing the transfer node correctly.

Because KSP's planets are all fairly small, and your orbits quick, that means that this effect is much more pronounced in the game than it is in real life. Generally, burns that last more than a quarter of your orbit start to become seriously problematic. For instance, if you are on a 30-minute low Kerbin orbit, a quarter-orbit burn would mean burning for 7.5 minutes (3min45s each before and after the node). Personally, I try not to go much past an eighth of an orbit, even. Something like 4 minutes in low Kerbin orbit (2m each before and after the node). Such a burn length would cause no immediate problems.

In order to make use of low-thrust engines, people do what's known as "periapsis kicking". Each time you pass periapsis, you burn for a couple minutes, then stop, and loop around. Burn another couple minutes. And so on. Your primary goal is to get on an escape trajectory; once you have that, you can just keep your engines on, no matter how low your thrust. From low Kerbin orbit, you need to expend roughly 1000m/s in order to escape. Generally, you'd kick until you have expended the 800m/s or so required to stay just inside the Mun's orbit, then loop one last time and push for the rest of the transfer. Reason being that you don't want the Mun to come by and encounter your spacecraft as it loops around.

Additionally, if you raise the orbit much higher, looping around will take aaaaages. And that's an unfortunately tricky part of periapsis kicking. You'll have to point your apoapsis in the correct direction - the direction where it needs to be when you go for your final transfer burn, not where you start your periapsis kicking. So the less time you spend looping around, the smaller the time window between the start of your kicks and the actual transfer window, meaning it's easier to get the orbit just right.

Thankfully, with a nuclear engine, your thrust shouldn't be all that low. You can probably get away with splitting your burn into just two parts. Maybe three at most, if you're unlucky. Just remember to switch your fuel to all LF, no oxidizer for the nuclear engine. Otherwise your mass fraction is going to be garbage, and your rocket will be way heavier than it needs to be.

2 minutes ago, strider3 said:

I don't seem to have a "liquid hydrogen" tank available for the GNR??

...Oh right, mods. Not sure where the GNR-something engine comes from, but that mod probably expects you to install liquid hydrogen tanks along with it. You'll have to read up on the mod to figure that out (and whether there might be an option to run with just LF, like the stock nuclear engine does).

[strider3]

I'll be leaving Kerbin from 350 km orbit. More importantly, the GNR is apparently not going to work? It wants liquid fuel only tanks and I need both fuel and LoX to refuel my lander at Moho. At this point I guess I'll try the Wolfhound on the central stack and send 3 more refuelers to Moho to complete the Moho science mission.

Edited June 30, 2020 by strider3

[Streetwind]

Well, you'd obviously bring both - liquid fuel for your engine, and LF/Ox for the lander. Even if you brought only LF/Ox, the amount of fuel consumed by the engine doesn't go towards fuelling the lander. So does it really matter what kind of fuel and what kind of engine it is?