Author Topic: MCT Speculation and Discussion Thread 4  (Read 621614 times)

Offline Lar

  • Fan boy at large
  • Global Moderator
  • Senior Member
  • *****
  • Posts: 10649
  • Saw Gemini live on TV
  • A large LEGO storage facility ... in Michigan
  • Liked: 7512
  • Likes Given: 5288
Re: MCT Speculation and Discussion Thread 4
« Reply #2160 on: 06/06/2016 04:09 PM »
The point about driving AI is not driving per se, although that's a useful part of the overall problem... it's that they implemented something that learns.

...

with two years of learning time the units might be better than when they arrived...

The way that most successful AI learning works is that they use large amounts of data for training. For instance Tesla collections 1 million miles of driving data every 10 hours. This data is used to train the algorithms, but also used to evaluate how the AI does on real world data, it is also used to detect real world corner cases and allows humans to tweek the AI to handle these unusual situations better. Tesla also runs autopilot in shadow mode on customers cars, comparing what new algorithms do with what the human driver does.

For rover driving on Mars we have none of that, no large data sets, no real world data around the landing site, no ability to look at corner cases and no ability to compare AI results with what a human would do. This applies not only to driving but also to many other tasks the mining system would have to perform.

Read up on simulated annealing, please.

Also, I realise that there won't be large datasets,  but every time an excavator moves its bucket, there is a chance for improvement in routines for bucket motion by evaluating how the motion went.
"I think it would be great to be born on Earth and to die on Mars. Just hopefully not at the point of impact." -Elon Musk
"We're a little bit like the dog who caught the bus" - Musk after CRS-8 S1 successfully landed on ASDS OCISLY

Offline Impaler

  • Full Member
  • ****
  • Posts: 1283
  • South Hill, Virgina
  • Liked: 363
  • Likes Given: 0
Re: MCT Speculation and Discussion Thread 4
« Reply #2161 on: 06/07/2016 01:29 AM »
Robotic deployment of thin-solar panels to power an atmospheric water adsorption system is the most practical means to refuel an initial landing vehicle, it requires the least knowledge about the martian subsurface and is the most reliable due to minimal moving parts and minimal contact with abrasive regolith.

Offline Lar

  • Fan boy at large
  • Global Moderator
  • Senior Member
  • *****
  • Posts: 10649
  • Saw Gemini live on TV
  • A large LEGO storage facility ... in Michigan
  • Liked: 7512
  • Likes Given: 5288
Re: MCT Speculation and Discussion Thread 4
« Reply #2162 on: 06/07/2016 02:01 AM »
Robotic deployment of thin-solar panels to power an atmospheric water adsorption system is the most practical means to refuel an initial landing vehicle, it requires the least knowledge about the martian subsurface and is the most reliable due to minimal moving parts and minimal contact with abrasive regolith.
Absolutely, given what we know now. But a Red Dragon or two could inform more efficient techniques, if conditions supported it.
"I think it would be great to be born on Earth and to die on Mars. Just hopefully not at the point of impact." -Elon Musk
"We're a little bit like the dog who caught the bus" - Musk after CRS-8 S1 successfully landed on ASDS OCISLY

Offline Paul451

  • Full Member
  • ****
  • Posts: 1469
  • Australia
  • Liked: 736
  • Likes Given: 598
Re: MCT Speculation and Discussion Thread 4
« Reply #2163 on: 06/07/2016 05:12 AM »
If they simply install a large (but very lightweight) crane
A crane isn't mining equipment.
...on Earth.
[...]
I never said you'd use a bucket-line

Okay, in all seriousness, how did you envision using a crane to do mining, without a bucket or scoop?

Offline Robotbeat

  • Senior Member
  • *****
  • Posts: 28478
  • Minnesota
  • Liked: 8352
  • Likes Given: 5483
Re: MCT Speculation and Discussion Thread 4
« Reply #2164 on: 06/07/2016 05:11 PM »
If they simply install a large (but very lightweight) crane
A crane isn't mining equipment.
...on Earth.
[...]
I never said you'd use a bucket-line

Okay, in all seriousness, how did you envision using a crane to do mining, without a bucket or scoop?
A dragline is a specific type of machine. You can have a bucket or a scoop without having a dragline. (I accidentally said "bucket line," but we were talking about a dragline... sometimes it's better NOT to trim quotes.)
« Last Edit: 06/07/2016 05:12 PM by Robotbeat »
Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

Offline Impaler

  • Full Member
  • ****
  • Posts: 1283
  • South Hill, Virgina
  • Liked: 363
  • Likes Given: 0
Re: MCT Speculation and Discussion Thread 4
« Reply #2165 on: 06/08/2016 12:42 AM »
Robotic deployment of thin-solar panels to power an atmospheric water adsorption system is the most practical means to refuel an initial landing vehicle, it requires the least knowledge about the martian subsurface and is the most reliable due to minimal moving parts and minimal contact with abrasive regolith.
Absolutely, given what we know now. But a Red Dragon or two could inform more efficient techniques, if conditions supported it.

It is unlikely that even insitu observation alone can validate a complex mining system, at best it can tell us that their is something worth mining (prospecting) and that certain techniques will NOT work. 

A mining system would need to actually RUN and be closely observed by engineers to see what is and isn't working and how it will need to be refined or scaled up. 

Atmospheric water on the other hand can be fully validated by small scale demonstrators and then scaled up with far fewer issues due to the non-contact and uniformity of input principles.

Offline Paul451

  • Full Member
  • ****
  • Posts: 1469
  • Australia
  • Liked: 736
  • Likes Given: 598
Re: MCT Speculation and Discussion Thread 4
« Reply #2166 on: 06/08/2016 04:49 AM »

You still haven't said how you can use a crane to do mining. (Let alone how this could be automated.)

Offline philw1776

  • Full Member
  • ****
  • Posts: 1088
  • Seacoast NH
  • Liked: 728
  • Likes Given: 309
Re: MCT Speculation and Discussion Thread 4
« Reply #2167 on: 06/08/2016 02:16 PM »
From the FH discussion speculating on Raptor upper stage for FH


Raptor with a 4 m nozzle loses about 1% of ISP compared to a 4.8 m nozzle: 376 s vs 380 s.

This is based on sim in RPA lite using: Methane/LOX at:
9.7 MPa chamber pressure (same as Merlin)
2.8 O/F ratio (optimum for methalox at 9.7 MPa)
165 expansion ratio for the 4.8 m nozzle (same as Merlin Vac)
115 expansion ratio for the 4 m nozzle (assuming same throat diameter as the 4.8 m nozzle)

IF Rvac does have a 4.8m diameter, what does this say about max # of Rvac engines for the hypothetical 2nd stage BFS @ a given stage diameter?  Makes the case for >10m with even 15m having issues with # of Rvacs.

10m is out for > 3 engines
12m only fits 4 engines
15m seems OK for a ring of 6 engines, or 5 engines and a center engine.  Maybe too little swivel clearance for CE.

My guess is 15m BFS with slightly under 4.8m Rvac diameter bell, say ~4.5m such that a center engine is feasible.
4.5m also allows a 12m BFS to hold 5 engines in a ring.

“When it looks more like an alien dreadnought, that’s when you know you’ve won.”

Offline envy887

  • Senior Member
  • *****
  • Posts: 4624
  • Liked: 2544
  • Likes Given: 1395
Re: MCT Speculation and Discussion Thread 4
« Reply #2168 on: 06/08/2016 04:02 PM »
From the FH discussion speculating on Raptor upper stage for FH


Raptor with a 4 m nozzle loses about 1% of ISP compared to a 4.8 m nozzle: 376 s vs 380 s.

This is based on sim in RPA lite using: Methane/LOX at:
9.7 MPa chamber pressure (same as Merlin)
2.8 O/F ratio (optimum for methalox at 9.7 MPa)
165 expansion ratio for the 4.8 m nozzle (same as Merlin Vac)
115 expansion ratio for the 4 m nozzle (assuming same throat diameter as the 4.8 m nozzle)

IF Rvac does have a 4.8m diameter, what does this say about max # of Rvac engines for the hypothetical 2nd stage BFS @ a given stage diameter?  Makes the case for >10m with even 15m having issues with # of Rvacs.

10m is out for > 3 engines
12m only fits 4 engines
15m seems OK for a ring of 6 engines, or 5 engines and a center engine.  Maybe too little swivel clearance for CE.

My guess is 15m BFS with slightly under 4.8m Rvac diameter bell, say ~4.5m such that a center engine is feasible.
4.5m also allows a 12m BFS to hold 5 engines in a ring.

4 engines matches the expected values best IMO. BFR thrust levels (15mlbf) indicate a 10x scaling from F9, which would put the total US thrust at 900t. Musk said optimization was indicating 230t per engine which gives 920t thrust for 4 engines.

4 engines precludes a center engine landing, but I don't think a center engine landing is optimal anyway for orbital or interplanetary reentry velocities. Side engines firing at an outward angle is a requirement for increasing drag during Mars retropropulsive reentry, and they can be far enough forward to gimbal behind a heatshield for Earth entry (or at least far enough to stay inside the bow shock and keep heat load manageable).

With four 4.8m engines, the interstage needs to be 12m at it's widest point for clearance... but it doesn't necessarily have to be cylindrical since it's not pressurized and supports mostly axial and bending loads. If a 10m vehicle is desirable the interstage could have engine cowlings like the Falcon 9 v1.0 did for the tic-tac-toe engine arrangement.

Offline philw1776

  • Full Member
  • ****
  • Posts: 1088
  • Seacoast NH
  • Liked: 728
  • Likes Given: 309
Re: MCT Speculation and Discussion Thread 4
« Reply #2169 on: 06/08/2016 04:11 PM »
Good points but I don't see the BFR/BFS as a direct scale up extrapolation.  Stage 2 will be heavier because (1) it's fully recoverable/re-useable and (2) from models I've run I think it's also going to need to be heavier for propellant needed for delta V for escape LEO to Mars xfer & Mars to Earth return. 
So, I get 5 engines estimated and even with that the ignition T/W < 1 for LEO launch, which is OK.  Plus I assume Rvac thrust > Raptor sea level thrust.

Of course we're assuming that raptors still have the actual thrust levels similar to Musk's 18 months ago statement.  I doubt this will be so as actual design always makes changes from early design calculations.
« Last Edit: 06/08/2016 04:14 PM by philw1776 »
“When it looks more like an alien dreadnought, that’s when you know you’ve won.”

Offline envy887

  • Senior Member
  • *****
  • Posts: 4624
  • Liked: 2544
  • Likes Given: 1395
Re: MCT Speculation and Discussion Thread 4
« Reply #2170 on: 06/08/2016 04:35 PM »
You're right, it's not a direct extrapolation. However, Falcon can put over 26 metric tons in LEO (22,500 kg of payload plus a 4,000+ kg upper stage). If BFR/BFS is scaled ~10x it would put over 250!!!! metric tons in orbit. It only needs to take 100t of payload to Mars, so they can dedicate ~100 tons to a reuseable US. Even with booster RTLS.

It really doesn't need more dV than the Falcon US, because it will be refueled in LEO and at Mars surface. Even at really poor mass fractions (about 10:1 vs Falcon's 27:1) it can still to a 3-month Mars injection almost every synod, and a 6-7 month direct return from Mars surface with a 25t payload.

So the thrust requirement is pretty much fixed, and the only reason to add engines is if they are a little smaller. 5 engines with a 4 to 4.5m nozzle and 180-200t thrust would make sense.

Offline philw1776

  • Full Member
  • ****
  • Posts: 1088
  • Seacoast NH
  • Liked: 728
  • Likes Given: 309
Re: MCT Speculation and Discussion Thread 4
« Reply #2171 on: 06/08/2016 04:59 PM »
By more US DV I mean (1) less performance needed from 1st stage BFR making the big boy's re-use environ slightly more benign and lowering cost {My unsubstantiated assumption}, (2) when you re-fuel in LEO you still need the tankage structure mass for propellant for the BFS's DV needed for fast transfer to Mars, say 90-120 days whatever, possible or maybe not propulsive breaking, aerobraking and powered landing.

Using today's F9 as a model, the Rvac would have ~ 110% the thrust of the assumed 230mT Elon referred to.
« Last Edit: 06/08/2016 05:01 PM by philw1776 »
“When it looks more like an alien dreadnought, that’s when you know you’ve won.”

Offline envy887

  • Senior Member
  • *****
  • Posts: 4624
  • Liked: 2544
  • Likes Given: 1395
Re: MCT Speculation and Discussion Thread 4
« Reply #2172 on: 06/08/2016 06:22 PM »
By more US DV I mean (1) less performance needed from 1st stage BFR making the big boy's re-use environ slightly more benign and lowering cost {My unsubstantiated assumption}, (2) when you re-fuel in LEO you still need the tankage structure mass for propellant for the BFS's DV needed for fast transfer to Mars, say 90-120 days whatever, possible or maybe not propulsive breaking, aerobraking and powered landing.

Using today's F9 as a model, the Rvac would have ~ 110% the thrust of the assumed 230mT Elon referred to.

Falcon with 10-15t of LEO payload stages quite low and slow. A 15m lbf BFR should have little problem putting 100t and a reusable US up while still doing a comfy RTLS.

Transfers as fast as 80 days and never longer than 120 days aren't all that difficult if you have LEO refueling capability; they requires 4.5 to 4.9 km/s from LEO. The EDL maneuver burns another ~1.5 km/s, for a total of ~6.0 to ~6.4 km/s so there's no need for tanks that are much bigger than Falcon's (relative to payload, of course).


Offline Robotbeat

  • Senior Member
  • *****
  • Posts: 28478
  • Minnesota
  • Liked: 8352
  • Likes Given: 5483
Re: MCT Speculation and Discussion Thread 4
« Reply #2173 on: 06/08/2016 06:26 PM »

You still haven't said how you can use a crane to do mining. (Let alone how this could be automated.)
Honestly, you can't figure it out? Different requirements from Earth mining. Don't need nearly the throughput. You're talking like a factor of 1000-100,000x less than large mining equipment on Earth, like a few tons of soil per day to fuel up an MCT in a synod (even less for a subscale MCT). So if you can pick up and drop some weight, you can mine with it. If it takes 15 minutes to move 100kg of soil, that's fine. That would, of course, be a joke on Earth. And yeah, you could outperform it with a hand shovel. But that's irrelevant.

But fine, I'll posit something /intentionally/ non-optimal just to show you can do it.
On the end of the crane, you'd have a clamshell bucket or orange-peel grabs or something like a blend between the two. It'd be awkward and slow (and could be enhanced by dragging it with another cable), but it still could work. Randomly grabbing scoops of soil within reach would be effective enough to get a few tons of material per day.
« Last Edit: 06/08/2016 07:04 PM by Robotbeat »
Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

Offline Robotbeat

  • Senior Member
  • *****
  • Posts: 28478
  • Minnesota
  • Liked: 8352
  • Likes Given: 5483
Re: MCT Speculation and Discussion Thread 4
« Reply #2174 on: 06/08/2016 06:42 PM »
...
4 engines precludes a center engine landing...
Not true. An axi-symmetrical "Y" shape with 4 engines would allow it.
« Last Edit: 06/08/2016 06:42 PM by Robotbeat »
Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

Offline envy887

  • Senior Member
  • *****
  • Posts: 4624
  • Liked: 2544
  • Likes Given: 1395
Re: MCT Speculation and Discussion Thread 4
« Reply #2175 on: 06/08/2016 07:38 PM »
...
4 engines precludes a center engine landing...
Not true. An axi-symmetrical "Y" shape with 4 engines would allow it.
Sure, but that takes up as much space as 7 engines. I suppose if you're pointing a central engine in the direction of reentry, you want a very wide, very light stage anyway. Optimize for reentry ballistic coefficient, and just live with the drag on the way up.

Besides aerodynamics, which factors really optimize stage diameter? Road transport isn't a consideration.

Offline Impaler

  • Full Member
  • ****
  • Posts: 1283
  • South Hill, Virgina
  • Liked: 363
  • Likes Given: 0
Re: MCT Speculation and Discussion Thread 4
« Reply #2176 on: 06/09/2016 03:39 AM »
A central landing engine wouldn't need a vacuum bell nozzle on it, it would be simple to use 4 vac engines and a central sea-level engine tucked in between them, at landing just gimbal the large nozzles outward to give the central engine plume as much clearance as possible.

On Mars your going to be firing all 4 vac engines in a hover-slam and only firing the central engine in an emergency that requires a shutdown of a pair of outer engines.

I'm betting on 12.5 m diameter 4v1s (4 vacuum, 1 sea-level) for the second stage.  And ~30 engines first stage.
« Last Edit: 06/09/2016 03:40 AM by Impaler »

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6769
  • Germany
  • Liked: 1797
  • Likes Given: 1786
Re: MCT Speculation and Discussion Thread 4
« Reply #2177 on: 06/09/2016 05:50 AM »
The nozzle extensions of vac engines will need protection to survive reentry at least on earth. So they will probably be retractable for that reason. Can the remaining nozzle part formed like first stage engines and work as such in retracted state?

Offline envy887

  • Senior Member
  • *****
  • Posts: 4624
  • Liked: 2544
  • Likes Given: 1395
Re: MCT Speculation and Discussion Thread 4
« Reply #2178 on: 06/09/2016 01:28 PM »
The nozzle extensions of vac engines will need protection to survive reentry at least on earth. So they will probably be retractable for that reason. Can the remaining nozzle part formed like first stage engines and work as such in retracted state?

It won't be formed like a SL nozzle, because to get optimal vacuum expansion it needs a rapid increase in area immediately after the throat. In this area, the gas is expanding radially, so the velocity component is reduced in the thrust direction. Just look up a picture of the Mvac compared to the SL Merlin, and see how much fatter it is at the same distance from the throat. In theory still work, just at a lower efficiency and thrust. At sea level the flow would separate about 1/3 the way down the nozzle anyway, so it's not really less efficient than running a vac engine at SL.

Based on some more simulations with varying nozzle efficiency, the unextended nozzle should get 70 to 90% of the impulse and thrust of a SL optimized engine at SL, depending on how overexpanded it is. But it will separate a lot earlier in the throttle band, which might cause control issues.

Also of interest: in a vacuum it gets 92 to 96 % of the performance fully expanded vac nozzle. If the mission requires that last 4-8% of performance, then an extension failure would cause LoM.

Offline Robotbeat

  • Senior Member
  • *****
  • Posts: 28478
  • Minnesota
  • Liked: 8352
  • Likes Given: 5483
Re: MCT Speculation and Discussion Thread 4
« Reply #2179 on: 06/09/2016 03:36 PM »
...

Also of interest: in a vacuum it gets 92 to 96 % of the performance fully expanded vac nozzle. If the mission requires that last 4-8% of performance, then an extension failure would cause LoM.
...for a LOT of missions, success isn't entirely binary. I can point to several times when RL-10-based upper stages had a significantly early shutdown for one reason or another but the performance was still made up for, perhaps by the payload. This may lead to lower on-orbit fuel, which is annoying and can shorten the life of the payload, but this isn't LoM.
Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

Tags: