What Energy Requirements might you need for a Mecha?

[Crossroads Inc.]

Ok ok.. So first off before anyone says anything...
Yes I know this is all fictional in any case, so why think about "real numbers" at all?

Well.. I would like to consider the question from a different angle...

Ignoring things like 'flying', or cleaving a planet in two with some uber energy blade...
What would be the energy requirements for a large 'Humanoid Construct" to move about as a Human might?
Let us say we had a construct of, oh, 60ft at 200tones. If such a construct were 'alive' would there be a way to work out the caloric requirements for such a thing? And then working from that, could you work out what energy requirements would be needed?

Since most Mechs seem to have some sort of "magical" power source for crazy crap in any case. I am curious if there is a way to figure out "hard numbers" on normal things like, just moving and running and such.

[Lord Revan]

well first we might want to start from simply standing upright cause there's alot more involved in that then most people realize.

after that we might want deside will this mech be the Anime style of moving like a big metalic human in everyway or the more western style of being slow and ponderous.

[Crossroads Inc.]

Yeah this is why I wanted to approach it in such a way, people don't often think about how much Energy is required to move a Human around like a Human...

As far Anime vs Westren Mecha. I'd say Western since it will probably be easier to calculate and such. Really I am hoping for just a Very Raw estimation of "Energy" usage for a walking thing of such height and mass.

Just found this nifty website:
http://www.thecalculatorsite.com/conversions/energy.php

For an "average" human of 35 years at 160lbs, an 'Active' day would burn 2100 calories

That works out to 8,786.4 Joules or 8.7 KJ.
So... How would you scale that up to a 60ft Robot?

[madd0ct0r]

i've been thinking about it.

There's a lot of studies looking into the effeciny of different gaits - but unless your mecha has a bone and muscle system like humans it simply dosen't apply very well.

[Jub]

This is assuming that a mecha can built to be as energy efficient as a human including things like energy used while idol but powered up are equivalent to that of a human. I calculated that a 20 year old male, standing 5'9" and weighing 150lb. burns 0.084kj per second at rest, 0.3kj per second at a brisk walk*, and 1.08kj per second at a run**. If we scaled that man up to double his height while scaling all proportions equally, so 11'6" and 1,200lb, his energy usage should scale in the same way, and thus for every doubling of mass a human shaped machine should use 8x more energy per second. Now I'm rather bad at math, but somebody who's a bit smarter should be able to give you a rough energy usage for any given mass assuming the proportions stay the same.

*Walking, 2.8 To 3.2 Mph, Level, Moderate Pace, Firm Surface
**Running, 10 Mph (6 Min/Mile)

[Zeropoint]

Well, as a back-of-the-envelope calculation, let's take as our starting point a six-foot man with a 2000-Calorie per day energy requirement, and work from there to a 30-foot humanoid mech. That's a five-fold increase in linear size, for a volume of 5^3 or 125 times our starting size. Let's further assume that the average density of the mech is twice that of a human being (for comparison, average human density is about 1 g/cm^3 and solid aluminum would be about 2.7 g/cm^3) which gives us a relative mass of 250 times the human. Making the probably wrong assumption that energy consumption for activity scales directly with mass, and ignoring the fact that much of a human's energy goes to staying warm, and we end up at an energy budget of 250 * 2000 = 500,000 Calories per day. Remembering that a dietary capital-C Calorie is 1000 calories, this converts to 2,092,000,000 Joules per day. If one day is equal to 24 * 60 * 60 = 86400 seconds, then this is an average energy consumption of 2.092e9 J / 86400 s = 24213 Watts, or about 32.5 horsepower.

Keep in mind, however, that this is *average* power over a whole day. Wikipedia says that a well-trained cyclist can maintain an output of 400 watts for an hour; using our factor of 250 from above, that gets us an output of 400 * 250 = 100,000 Watts or 134 hp. Of course, this is output power and obviously, our mech cannot convert the output of the power plant to mechanical motion at 100% efficiency. Assuming a 25% efficiency and some sort of capacitor bank to handle momentary surges of higher power output, a 30-foot mech should be able to move more or less like a human on 536 hp. Assume a 75% efficient generator, and you'll need a 536 hp / 0.75 = 715 hp engine to run the mech.

Now, for comparison's sake, let's use Jub's numbers . . . he indicates 1080 watts for a running human, 2.7 times my value of 400 watts. Using that figure yields a requirement for a 1,931 hp engine.

For a 60-foot mech, the requirements would be eight times larger under my assumptions, for a requirement of 5,720 hp with my figures or 15,444 with Jub's. This puts us up into the high end of railroad locomotive power, or into marine engine power.

Also note that according to Caterpillar, an engine with an output in the teens of khp is nearly 40 feet long and weighs over 200 tons. I'm having a little trouble finding data for gas turbines, but one source indicates that a gas turbine in the 20 khp range weighs in at about 18 metric tons and maybe 25 feet long.

[Jub]

Thanks Zeropoint, I could have muddled through the math and unit conversions, but I was on shift at the time and wasn't able to look up as much as I'd have liked.

I'm actually somewhat surprised by how doable these numbers make that seem, but I suspect that things scale rather worse than we've spit balled in practice. Even so, a even a 25-30 khp turbine should be possible, if very fuel hungry, and keeping the weight lower than that of a solid block of aluminium could be done with exotic materials. So it might be possible, if very difficult, to make a mecha that can walk and maybe even jog.

[Zeropoint]

I'm actually somewhat surprised by how doable these numbers make that seem, but I suspect that things scale rather worse than we've spit balled in practice.

Yeah, I expected to find that it was completely out of the question, and was pleasantly surprised. The real killer problem is going to be the "bones", though: going by Wikipedia again, human bone has a tensile strength (yield) of a bit over 100 MPa. Scaled up ten times, we'd need bones 100 times stronger, and the only known materials with a strength in the 10 GPa range are various nanotubes that we don't know how to make into large structural elements.

Still, it's not physically impossible!

[Lord Revan]

you can probably make due with weaker "bones" if your mech can't do things like jump or run as those put the most stress to the human skeleton, basically make move like a human wearing heavy armor.

[Zeropoint]

Wait, I messed that up.

The cross-section of the bones would be 10 * 10 = 100 times greater, while the mass of the mech would be 10 * 10 * 10 = 1000 times greater (2000 with my assumption of higher density). We'd only need a material 20 times stronger than bone. This puts us in the range of high-end steel and existing composites.

you can probably make due with weaker "bones" if your mech can't do things like jump or run as those put the most stress to the human skeleton, basically make move like a human wearing heavy armor.

That's a valid point. However, it makes me think of modeling a giant robot on a human with a bad case of osteoporosis . . . which doesn't sound awesome.

[Crossroads Inc.]

Hey thanks for all the information thus far guys Been a big help.

And don't stress about things too much, after all what I am mostly after is Energy Usage.
Things like Materials and tensile strength are all assuming we are working with "Fantastic" materials... IE whatever is needed for said Robot to function without breaking apart. Same goes for the motor itself. Im working under the assumption of "Future tech" So motor size I am not worried about either. Although knowing it would be something around 6000HP to 16000HP is pretty awesome.

[madd0ct0r]

things like materials give you weights and densities of moving parts which has a very strong effect on energy usage.
so a swinging leg's energy usage might be considered in terms of KE = 1/2mv^2 where v is measured at that limbs center of mass - probably just above the kneecap for humans. the mass is another matter.

[Crossroads Inc.]

Aha, I had not thought of it like that, but it makes a lot of sense

[LaCroix]

You could also try to translate it to hydraulics (which is more reasonable than the comparison with a bulldozer). For every degree of freedom, you'd need at least 2 pistons, each capable of generating a multiple (probably a safety factor of 4+, so I'll go with 5) of the weigth it was supposed to move, in order to move it quick and with an eventual load.

So you'd need a couple of systems (at least 8 ) equal to 1000T presses for the legs (they hold all the weight of the mecha while the other leg is in the air, plus the impact from walking), and then a couple of 500T equivalent for torso and arms (easily 12). This will tell us the rough total power usage needed. A 1000T commercial system needs about 70KW electrical power to run ( a military one will just run faster and better, at the same power), so you'd be in the range of 8 of such systems, 12 half as strong, and then a few smaller for hands and head and stuff. You might note that each of that commercial 1000T presses will weight in at 50 tonnes, but this is the weight of the frame keeping the press together - in our case, it would be the skeleton itself to provide this. 30-40T for a leg is reasonable for a 200 t total weight.

You'll need some additional control units, cooling and stuff, so, rounding generously, we'd be at about 1GW electric power needed to drive the system (under the premise that every system has to be 'ON' at all times).

1 GW in distributed 50 or 100 kw units (You don't want one system for all things, and miles of hoses that could drain the whole system if just one gets severed.) means you need a total combustion engine power of about 1200 hp to run hydraulics. Add another GW for electric systems (computers for everything, limb position sensors, gyroskopes, life support, radar and other stuff you might want to add, and you are at 2500 hp, and you have awalking mecha..

A slow, lumbering, "you cannot stop this thing" walking mecha.

To 'run', you would probably have to massively upscale leg hydraulics in strengt and additional DOF, to the point they move at top speed without noticing the weight it carries. You could easily double or triple the energy requirements doing that.

So, anything from 2500 to 7500 hp would be reasonable, I'd say.

[Zeropoint]

Thanks, LaCroix! It's worth pointing out that you've used an entirely different line of reasoning to arrive at your figure, and yet it's in the same ballpark. This suggests that we're probably not too badly wrong in our collective estimates!

[biostem]

Below is a link to Boston Dynamic's "Wild Cat", which is pretty close to human size, (though 4-legged). It sounds like it uses a lawnmower or motorbike engine for power. I will also include a link to their earlier design, "Big Dog", which, though smaller, I think better demonstrates its ability to remain stable. These examples, I think, at least show that you don't need immense power requirements for a robot frame, (I think some sort of hybrid gas-electric system for power could easily meet the requirements). The issue, I think, is getting a design that is actually better in combat or for civilian purposes, (like construction), than any other specialized tools.



https://www.youtube.com/watch?v=wE3fmFTtP9g
https://www.youtube.com/watch?v=cNZPRsrwumQ

[Simon_Jester]

well first we might want to start from simply standing upright cause there's alot more involved in that then most people realize.

For a piece of machinery that might not actually be so much, because the machine's "reflexes" are fast enough that it doesn't need to worry about starting to overbalance and needing major muscular effort to overcorrect. Sure, a giant robot is dynamically unstable in a stiff breeze. But that's a problem closely related to, say, the problem that modern fighter jets are dynamically unstable and go into an out of control tumble quickly in the absence of control input.

Modern fighters would thus be 'unflyable' in that no human could keep control for long... but with a computer to note small shifts in the plane's balance and aerodynamics, and make the appropriate adjustments before they start to get out of hand, the plane becomes very flyable.

On the other hand, movement will be very energy-intensive because large masses of metal need to be raised and lowered with every step.

You could also try to translate it to hydraulics (which is more reasonable than the comparison with a bulldozer). For every degree of freedom, you'd need at least 2 pistons, each capable of generating a multiple (probably a safety factor of 4+, so I'll go with 5) of the weigth it was supposed to move, in order to move it quick and with an eventual load.

So you'd need a couple of systems (at least 8 ) equal to 1000T presses for the legs (they hold all the weight of the mecha while the other leg is in the air, plus the impact from walking), and then a couple of 500T equivalent for torso and arms (easily 12). This will tell us the rough total power usage needed. A 1000T commercial system needs about 70KW electrical power to run...

You'll need some additional control units, cooling and stuff, so, rounding generously, we'd be at about 1GW electric power needed to drive the system (under the premise that every system has to be 'ON' at all times).

Uh, I think you jumped from "one megawatt" (enough electricity to power a large neighborhood) to "one gigawatt" (enough electricity to power a respectable city). Your horsepower figures are reasonable, though- and scale roughly to "on the order of one megawatt," since one horsepower is less than one kilowatt.

[LaCroix]

]Uh, I think you jumped from "one megawatt" (enough electricity to power a large neighborhood) to "one gigawatt" (enough electricity to power a respectable city). Your horsepower figures are reasonable, though- and scale roughly to "on the order of one megawatt," since one horsepower is less than one kilowatt.

Ups, my bad - yes it's MW - I don't know why I wrote GW in both places, because I obviously calculated the hp requirements for MW magnitude...


@biostem - the wild cat is using a "2stroke go cart motor" - which means it will be in the 5 to 20 hp ballpark, if you think "standard" engine. (It could have anything up to 60 or 70 hp... but unlikely for such a contraption )

It's a improved "Big Dog", so it will also be in the 100kg ballpark, and about 3 ft tall.

Scale it up to 60 feet - 20 times larger, we have 8000 times the volume and weight - ~800 tons. Scaling it to 200 tons - the thing needs to be 12.5 times larger to get there (~2000 times the volume). This means 2000 times the engine power - 10.000 to 40.000 hp. I'd go with the lower values, since hydraulics do scale very benevolent in respect to hp requirements, as you only need to drive a pump. Also, the 4 legged design is much easier on the drive, as it's inherently stable and the weight is much better distributed between multiple legs, which allows for a less power-hungry system.

Same ballpark - low tens of khp...

[madd0ct0r]

a kangaroo based mech might be more efficient. Kangaroos have an exteremyl efficient gait for their calorie intake.

[Borgholio]

a kangaroo based mech might be more efficient. Kangaroos have an exteremyl efficient gait for their calorie intake.

And they can kick like a motherfucker.

[LaCroix]

And their 'gait' is probably the most complex to emulate for a robot, as it leaves little room for error. (High needs in regards to balance, speed, constant high g spikes, and an absolute need for an ability to identify and aim for a right landing spot, constantly.)

That's why we usually stick with wheels, 4 legs, and when we feel lucky, try a bipedal.

[biostem]

Below is the world's largest bulldozer. Wikipedia lists its power output as 1150 HP or 860KW. It doesn't list its endurance, but you figure it has to be able to work at least a full day without refueling, (though said fuel capacity isn't listed). The bulldozer is about 40' long, and I'd imagine it's design would be "denser" than a similar mecha. After all, a mech would have just its legs as a majority of its height, and not the treads.

http://en.wikipedia.org/wiki/Komatsu_D575A

Something like this would probably be a more practical design, IMO:




EDIT: This site says it can go through "440 gallons of diesel during a day", but doesn't quantify how much run time is considered a day, (I'm going to assume they mean 8-10 hours). Link: http://www.vincelewis.net/bulldozer.html

[Zeropoint]

Below is the world's largest bulldozer. Wikipedia lists its power output as 1150 HP or 860KW.

For a more military point of comparison, the M1A2 Abrams is 68 tons, 26 feet long excluding the gun, and has a 1500 horsepower engine.

[Sea Skimmer]

Below is the world's largest bulldozer. Wikipedia lists its power output as 1150 HP or 860KW. It doesn't list its endurance, but you figure it has to be able to work at least a full day without refueling, (though said fuel capacity isn't listed). The bulldozer is about 40' long, and I'd imagine it's design would be "denser" than a similar mecha. After all, a mech would have just its legs as a majority of its height, and not the treads.

Largest series production dozer FYI. Below is the biggest one period, and it looks like the awesome kludge of Italian and American parts it is.
http://en.wikipedia.org/wiki/Acco_super_bulldozer

Anyway the Japanese one has a top speed of less then 9mph and I do believe unsprung suspension as well. Which is barely acceptable at 9mph, but the point being for higher speed you need springs or spring like function and that costs some performance in realistic terms. Human running speed is already high enough to need this, we get shock absorption from our bones being able to bend slightly, and certain connective tissue. These tissues are so good at what they do major effort is going into trying to synthetically replicate them. A mech would probably have hydraulic buffers just above the ankle joints, which might be able to recoup some energy if so equipped, but overall legged locomotive still costs you serious energy in a way the suspension on a tracked or wheeled vehicle does not on reasonably flat ground.

So anyway bulldozer not good comparison for this. But actually a very simple rule of thumb is that military tracked vehicles need 20hp per ton for reasonable agility, and 25hp for really good agility. Though in reality its the torque curve that counts in most situations, and this is why the US military pays out the butt for a gas turbine engine in the M1 tank. The torque curve is more of a vertical line then a curve on the turbine, any speed it runs it has nearly maximum torque.

For a mech any form of direct drive is out of the question and so the power plant would probably actually be designed to run at just a narrow range of speeds powering either a generator or a hydraulic pump. That will help efficiency some. Though I would note even in the electrical course of action you'd still probably have local hydraulics as on the 787 airliner, but just powered by motors locally. That way a single hydraulic leak isn't completely crippling and numerous flexible hydraulic couplings can be avoided.

So basically the mech could gain back some fuel efficiency by using an engine designed for only a narrow speed range, and capacitors or hydraulic accumulators providing a power buffer to make up for how insanely variable on a millisecond basis actual power demands would be while running. However your still stuck with the problem that if its power/speed curve is similar to a human then the power to run 20mph will be about twenty times higher then the power to walk at 3mph.

Which means either you have a vehicle which is grossly overpowered in most situations. Or you have a mech which is just like a human, running cannot be sustained and relies on stored power replenished by the engine at a much lower rate. Adapting the latter course is well worth considering.

Of course speed-power curves affect all vehicles, its just a matter of the scaling effect being much worse for human locomotion, and less linked to air resistance then just how bad legs are at speed.

As for energy consumption we can apply math to this fairly easily. A simple gas turbine these days has SFC of about .35lb per HP per hour. The best current diesels (which are huge and slow reving, that part is important) are about .26lb per hp per hour.

So lets say 20,000shp is required from a turbine engine output shaft (which is different then the effective power at the joints of course) then that means at peak power its burning ~7000lb an hour. Diesel is about 7lb a gallon, so that is 1000 gallons an hour

Note gas turbines can run on diesel if adapted to such. Jet-A is about 6.8lb a gallon with similar energy content. Gasoline is more like 6lb a gallon. Exact values depend on the exact fuel grade and of course, temperature.

In more direct terms 1000 tons is that's a large 8x8 2,500 gallon tactical fuel truck drained every 2.5 hours.
http://olive-drab.com/images/id_hemtt_m978_700_06.jpg
One of these.

In comparison the M1 tank holds 500 gallons, so the same truck can refuel five of those tanks, and the fuel will typically last 8-12 hours of operations or provide a 300 mile road march. As is infamously known by many, the M1's turbine basically burns the same fuel to idle as to move slowly, and actually gains fuel efficiency with speed (to an uncertain limit).

Basically your probably going to end up with the fuel truck fleet needed to resupply an heavy tank battalion to support a couple of these mechs engaged in mobile operations.

You also need a place to store a tactically useful amount of fuel on the mech, and even allowing for the fact that it could have lots of small fuel tanks (which are also radically more likely to explode when hit, bigger fuel tanks can absorb more energy and avoid bursting and explosions), and exploit fuel tanks as armor, that's bound to be a problem even in a giant mech. Though probably not as big of one as finding space for an air filter and exhaust suppressor able to cope with an engine the size of an LM2500. Even if you don't care about noise tactically, or the huge dust cloud being blown around, your probably going to need the suppressor to avoid physical damage to the vehicle from the vibration. It's that kind of loud you get out of combustion engines this big.

Now if we need 40,000hp from a .26lb/shp/hr diesel our 1 hour max power consumption would be, skip to the end, 1,485 gallons an hour. Five freaking tons of fuel.

Yeah, I expected to find that it was completely out of the question, and was pleasantly surprised. The real killer problem is going to be the "bones", though: going by Wikipedia again, human bone has a tensile strength (yield) of a bit over 100 MPa. Scaled up ten times, we'd need bones 100 times stronger, and the only known materials with a strength in the 10 GPa range are various nanotubes that we don't know how to make into large structural elements.

Still, it's not physically impossible!

Remember a human bone is basically a single tube surrounded by muscle which does not support weight, though it does absorb a bit of impact shock. On a mech you could out of hand have a tube which is inside of a stressed skin, and hell, maybe has multiple honeycombed tubes inside of it. You don't have to make all the same compromises the human body did. Though you will need new ones.