Wednesday, May 6, 2015

The Need for a Better Alpha

What the heck is Alpha?

What Alpha am I talking about? A power source's ratio of mass to power. The first time I ran into this quantity was in a NASASpaceFlight discussion of VASIMR's 39 day trip to Mars.

Kirk Sorensen worked for NASA 10 years. He has Master's degrees in aerospace as well as nuclear engineering. He is co-founder of Flibe Energy, a company that hopes to build thorium nuclear reactors. I don't always agree with him but I believe he has some expertise in this field. For the time being I am talking his word for it.

So this Magic Alpha, .50 kg/kWe, what is that? Here's an attempt to portray it:

This Ford Focus has a 160 horsepower engine. One horsepower = ~750 watts, so the engine is about a 120,000 watt power source. Not only the engine but also gasoline and oxygen. Pictured with the Focus is Dominique who masses about 60 kilograms. If she were a power source replacing the engine, gasoline and oxygen, we'd have an alpha of .5 kg/kW.

An electric car like the Tesla uses a battery. But just as a gas engine must make periodic stops to gas up, a Tesla must be frequently recharged. Enroute to Mars there are no gas stations and no electric outlets.

Later in the same NASA Space Flight thread, Sorensen says:
I don't really care if Samim Anghaie is crook but his 0.5kg/kWe number is a fantasy. Why FCD builds his VASIMR sales case on that number when all other reputable electric propulsion researchers have rejected it (even though makes their thrusters look incredible too) is beyond my understanding.

Indeed. Such a power source would make Hall thrusters look great. So far as I know Franklin Chang Diaz and Samim Anghaie are the only folks whose schemes rely on such an alpha.

Thermal Watts vs Electric Watts

39 day VASIMR trips to Mars are mentioned on Page 42 of The Plundering of NASA by Rick Boozer. Boozer argues that SLS and Orion are pork barrel make work programs and that money could be better spent on SpaceX and other programs. In general I agree with Boozer but was disappointed to see his endorsement of VASIMR.

So I asked Boozer about the Magic Alpha. Boozer came back with Project Nerva, a nuclear thermal rocket. He wrote:
Project NERVA claimed up to 5 GW possible with total mass of 38,600 kg. That works out to .00008 kg per W or .008 per kW. According to that Sorenson is incorrect.

NERVA's output is thermal watts. Thermal and electric watts are two very different things.

A nuclear electric power plant must first convert thermal watts to electric watts. But that's not the only problem.

The plant must dump waste heat. Massive cooling towers have become an icon for nuclear energy. From  Wikimedia:

Earthly nuclear power plants can use water to carry off waste heat. In space there are no neighboring streams a nuclear power plant can use. In fact vacuum is a great insulator. A nuclear electric power source would need massive radiators.

I mentioned to Boozer that thermal and electric power sources were very different things.

He replied:
David, I just don't know where you are coming from. There you went earlier lecturing me about the difference between thermal and electrical energy which is something that I teach physics students all the time. If I wasn't competent in physics my students wouldn't be making the high grades they are and I couldn't have got my Master's in astrophysics.
I was hoping Boozer would demonstrate Sorensen was wrong. Sadly, pointing to his students' good grades and his degree did absolutely nothing to demonstrate the plausibility of Diaz' Magic Alpha. I was convinced of one thing though: Rick Boozer isn't credible.

I did not bother reading past page 42 of The Plundering of NASA.

What's the best plausible Alpha?

I search space forums for discussions of low mass power sources. So far as I can tell, thin film photovoltaics show the best promise. Roll Out and Passively Deployed Array (RAPDAR) might deliver 250 watts/kg. RAPDAR's thin film solar cells use an Elastic Memory Composite (EMC) for support and structure. Rolled up and cooled, the EMC will fit in a small volume and thus can fit under a fairing. When the sun warms it, the EMC will expand to the shape it needs to be.

On a Nasa Space Flight thread space entrepreneur Jeff Greason opined:

While I won't speak to this specific design, more generally I am quite convinced that thin film solar approaches 1 kW/kg are definitely possible near term. However there is very little serious work going on, and packaging such systems for launch and deploying them without spoiling the mass is not at all trivial. 
But do keep thinking -- it is not crazy, at least 1 kW/kg rather than two. 
Thin film solar is extremely fragile, however, so the packaging is really challenging.
That's the rub, packaging. How useful are acres of Saran Wrap® with no structure? There needs to be a supporting frame to keep the film spread. It also needs to be kept pointing towards the sun so the supporting frame needs to be attached to gimbals and motors. What is the Alpha including supporting structure, gimbals and motors?

How will we deploy acres of Saran Wrap® from a small volume that fits within a fairing?

However Greason's optimism is somewhat reassuring. Being a bonafide space-cadet, I cling to optimistic opinions as long as I can.

Why is Alpha such a big deal?

As mentioned at the beginning of this post, a great alpha would make ion thrusters a more formidable tool. Presently ion thrusters have great ISP but very slow acceleration. A big cut to parasitic mass would give ion thrusters better acceleration.

Good Alpha would also make ISRU more plausible. Readers of my blog know I'm gung-ho on use of extra-terrestrial propellent, either near earth carbonaceous asteroids or frozen volatiles in the lunar cold traps.

Let's say we do mine water in the moon's neighborhood and we want to crack it to hydrogen/oxygen bi-propellent. Cracking a mole of water (18 grams) takes 287000 joules. A tonne of water is 55555 moles. 55555 moles*287000 joules/mole =13166666667 joules. If we wanted to crack 10 tonnes of water per day, we'd need a 1.5 mega-watt power source. And that doesn't include refrigerating the cryogens.

Cracking water isn't the only ISRU electricity hog. Just about all extra-terrestrial mining and industry will need lots of juice.

In the 50's and 60's NASA and the military provided big incentives to miniaturize electronics as low mass and small volume circuitry is a pre-requisite for rockets and missiles. I believe miniaturizing a power source should be a top goal for NASA. If we hope to settle space, a better Alpha should be given a higher priority than Apollo redux.


Melcon37 said...

For the Ford, there is a misprint. 160 HP * 750W/Hp is 120,000W, not 12,000W.
At the start the units are kg/kW, later they are kW/kg. Hwich defines alpha?

Hollister David said...

Thanks, Melcon. Typo corrected.

I've usually seen alpha given with the kg in the numerator. However a .5 kg/kWe power source would also be a power source that deliver 2 kilowatts per kilogram.

Anonymous said...

Cutting down parasitic mass will not change the ISP of an electric propulsion system, I think you meant to say it would increase acceleration rate.

Hollister David said...

Anonymous, when you post could you give your name or a nick name? I usually try to make it clear who I'm replying to in the first sentence. Replying to "anonymous" can be ambiguous if more than one person posts anonymously.

Any way, thanks. I made an edit.

Nilof said...

Regarding the whole thing about units, engineers like using units of kg/kW because if you keep the power fixed, it allows you to get the total mass by simply adding the alpha numbers of different components such as arrays, power processing units, thrusters ect ect.

Regarding anonymous' comment above me, reducing the mass of a solar panel does not by itself increase isp, but it does increase the optimal isp, since if you have lighter panels you can have more of them for the same mass budget, and running your thrusters at a higher power but the same mass flow increases both Isp and thrust.

Hollister David said...

Thanks Nilof. Now the kg/kW makes more sense.

I hadn't meant to say cutting parasitic mass increases ISP, that was sloppy writing on my part. I'm happy anonymous pointed it out.

But maybe I was right after all. Does more juice result in high exhaust velocity? I badly need a primer on ion engines.

Anonymous said...

This is Impaler the earlier Anonymous, sorry was in a hurry last time. Nilof is correct that a redesign of the craft towards higher ISP should follow an improvement in the alpha of the propulsion system. I just wanted to point out it is not a direct effect.

For Electric propulsion (ion is generally used to refer to only one particular type) Exhaust velocity is proportional to the square root of power, so quadrupling power will double exhaust velocity. Thrusters with variable ISP generally operate by varying mass flow rate while keeping power constant, low flow for high ISP and low thrust, high flow for low ISP and high thrust.

Jim Baerg said...

What I wonder is why some people talk as though such short travel times as 39 days to Mars are 'must have' rather than 'nice to have'?

In previous centuries people would spend many months sailing around Africa for the spice trade or paddling from Montreal to Western Canada for the fur trade. I don't see why 6 months to Mars should be considered an unreasonable trip time.

Also if humanity is to become a space faring species that means learning how to live for years or decades or lifetimes off earth. We might as well learn that capability & use it for reaching Mars.

Chris said...

How about separating the power source and propulsion? Ablative Laser Propulsion has potential to have the performance of Orion (the nuclear one), but without the pesky nuclear fission and radiation.

Suppose we want to launch a ship to Mars. We put a solar powered laser in LEO, and blast the ablative end of the Mars-bound ship, injecting it into a near escape orbit (that takes hefty 3 km/s). The remaining deltaV (0.4 km/s) needed to reach the hyperbolic orbit speed the spacecraft can do itself with at a small expense. Moreover, the power source, now stationary in LEO, can be relatively hefty and can be reused over and over and over again. Keep in mind, that it would be also a high Isp, high thrust propulsion system allowing for near instantaneous maneuvers, as opposed to the slow spirals of ion drives.

The problem with it is that it is not tested in space, however it works on well established physics and engineering. Capital cost is not a problem, because whether we need a 10 MW power source on a spacecraft, or 10 MW power source positioned permanently in Earth orbit, it doesn't really matter. The stationary one may actually be cheaper and it will not eat into the payload. The added cost of laser is the only extra, and the whole system can be amortized over multiple launches. Separation of the power source and the spacecraft is the most reasonable thing that can be done. Of course, beamed propulsion will work at relatively small distances, probably not more than 20 km, so the spacecraft and the power plant will have to maneuver quite closely in LEO. The LEO cannot be too low because of the atmospheric drag, but 2000 km altitude should do. The ISS generates nearly 100 kW of power. Making power station 100 times the power output is not unthinkable - it will be only 10 times longer and wider. The ISS solar power arrays are quite ligthweight, so 10 MW PV array is not unthinkable.

Robert Clark said...

Thanks for that. Robert Zubrin wrote a critique of Vasimr here:

The VASIMR Hoax.
By Robert Zubrin | Jul. 13, 2011

The problem is that it needs a 1,000 watt per kg power source, assumed nuclear, and the best that had been done for space nuclear power sources was only 10 watts per kg.

However, space solar power sources are getting close to that. As you say Hop, this should be a major focus of NASA and the military and industry, yet not nearly enough is been done to develop this capability.

One way this is very important to the satellite industry is that satellite developers such as Boeing want to use solar electric propulsion to make the final push to GEO to save on the weight and launch costs.
But the SEP thrusters now put out so little thrust it takes months to get the satellites into position. That's months of revenue not being made. But by using high power, lightweight solar power sources we can get the satellites into position in a few days, comparable to the time with chemical propulsion.

There are two solar power techniques I know of that are close to getting the 1,000 watt per kg threshold for getting plasma propulsion engines such as Vasimr to allow weeks travel times to Mars.

Interestingly they have both been successfully tested in space. One method is by solar concentrators. This uses mirrors or lenses, which are lighter than solar cells, to concentrate light collected over a large area onto a smaller area covered by solar cells. This gives a larger amount of power being produce by a smaller number of solar cells, thus saving weight. This has been successfully tested on the now famous Deep Space 1 spacecraft:

Deep Space 1.
1.2 SCARLET concentrating solar array.

From memory I believe the solar power system used here was able to 300 watts per kg. But solar concentrators in use on Earth have been able to get hundreds to thousands of times solar concentration. So increasing the concentration level for the SCARLET system should allow it to get the need power per weight for the plasma propulsion.

A second method that could work is thin-film technology. This has been successfully tested on the TacSat satellite:


I don't know what the specific power was here but I've seen references on thin-film solar cells in the lab able to get 2,000 watts per kg.

Bob Clark

Doug Plata said...

What's the relationship between specific power and alpha? I usually see specific power with units listed as W/kg.

Hollister David said...

Doug, one is just the reciprocal of the other (I believe). They give the same info. I'm more comfortable with specific power but see Nilof's explanation above why the ratio is given with mass in the numerator.

cryogen said...

I'm not convinced about the "magic alpha". I can Google up literature talking about space-nuclear alphas that high, for example:

In figure 1.2, the estimated alpha for the closed-cycle system is 1-2 kWe/kg, at the 100 MWe scale.

I get the suspicion that both sides have credible figures, except they're talking about systems at vastly different scales -- three orders of magnitude apart. The JIMO reactor would have been only 20-30 We/kg, but that's at the 100 kWe scale, that's more appropriate to small probes than human missions.

What I think is happening is (?) the reactor mass grows as a strongly sub-linear function of power, so alpha increases with size. At the 100 MWe scale, most of the mass is in the radiators, which are even lighter (per kW) than photovoltaics.

kim holder said...

I saw a presentation at the ISDC on a possible new electric drive that uses principles similar to an arc welder. The Neumann drive, a "wire-triggered pulsed cathodic arc system". Website at It doesn't have nearly the detail that was in the presentation, i'm going to email Paddy to see if he will pass along some of the slides he used.

In the video on the website he states exhaust speeds of 60 to 70 km/s, and i recall a figure of 6 N of thrust for the drive model he was presenting.