Reimagined and Revamped. Fighting the spread of nonsense often feels like a Sisyphean task. However, the joy is in making the information available, not the hope of conversion.

micronuke vs solar: some quick math

Hyperion Power Generation has been making some waves recently. They, like Toshiba, have designed a micro-nuclear plant. When I say 'micro' think "powers a small town", not "fits in your back yard.

I first heard about Hyperions device on SGU, where one of the group said something like: "well if it can power a town for 7 years, then it can power my bunker for 100 years!".

Sorry, it doesnt work like that. The reaction takes place whether you use the power or not. The fuel is hot, and as it gets expended, it cools, which is why the life time is limited.

Edit: Rod Adams corrected my original contention that this was not how these reactors worked. They in fact should last longer if they are used less. This is not due to a control mechanism within the nuclear reactor. The fuel itself is designed such that if the temperature starts getting too hot, then the reaction dies down. It self regulates. It is the act of extracting heat (either by diverting it and using it, or by boiling water into steam for a turbine) that controls the reaction. Its a very safe way of doing the reaction, and further the fuel should last longer with less use. However, I'll note that Hyperion still has not claimed any life longer than 10 years, I'm not sure if this is due to the expectation that the load will always be there, or if there is another decay mechanism, related or unrelated to the actual nuclear reaction. More info and thoughts here.

No where do they say exactly what the fuel is, but I wonder if they are not taking waste from a normal nuclear plant and instead of planting it in a cooling pool, they are simply encasing it in concrete and providing a path for heat to get out.

Edit: In fact they do say what the fuel is. Its uranium hydride.

Anyway, this comment and the one on their website got me thinking. Here is their claim:

"Hyperion’s innovative energy technology is even more affordable than many developing “alternative” energy technologies."
Well on a $/watt basis this is 100% true. Checking current solar module prices, we have 3-4 dollars per watt. Non-utility installation will double this price (these prices are the #1 reason why solar is not more ubiquitous). If you convert the heat to power you get about 25 megawatts, they claim. So we are really talking about 1 dollar per watt. This is supposed to include security, installation and so forth. I dont know if it includes the turbine and water system required to get the power, but lets assume not.

But what about the energy capable of being created over the life of the product?

Well the lifespan of the hyperion device is about 5 to 7 years (lets give it the upper end). 25mW*7 years*365 days/year*24hours/day
*.001 kW/W = 1.5x10e9 kWh= $0.016/kWh.

Wow, that is pretty good. Industrial electricity costs about 5 cents/kWh, and residential can be 2x to 3x higher than that!

What about solar? For 25 million dollars at 4$/W, you can get about 6.25 MW. But the life of a solar panel is very very long. The general thining is that solar panels last 20-25 years. Some companies are claiming 35 years. For for our 25 million dollar investment, we get:

6.25 MWe *25 years*365 days/year*24 hours/day*.001 kW/W = 1.37x10e9KWh= $0.018/kWh

Yeah that is a bit more expensive, but that looks like a wash to me, because we have not gone over the fine details of each of these, mostly because this data is not available from Hyperion. these details include:

  • Installation: Does the price tag include the turbine and new water system? Does it include the water routing? Does include the tie in to the grid?
  • Maintenance: Does this price include the security guard? The upkeep of the turbine? The upkeep of the power switching and load ballancing hardware (less lectricity is used at night).
  • End of Life: Does it include waste disposal? Digging the unit back up?
Solar has similar questions associated with it. So once again, the Hyperion system is not a clear winner and to claim that it is less expensive than "alternative" energy solutions is really pushing it. Further, photovoltaic solar panels may not be the best comparison to make.

Solar thermal, which provides energy storage, and therefore does not suffer from the cyclic power generation of photovoltaic solar systems, is a far more likely substitute. Costs for a ST system show that a kWh can be had for as litle as $0.10, in a single year, not as analyzed above over the life of the plant! Geothermal is another system that may be more applicable for comparison.

All that being said, Im am totally for the deployment of the Hyperion systems for the following reasons:
  • They do not emit greenhouse gasses
  • They are inherently safe and can not melt dowm
  • They are a simple device with few mechanics involved for its operation (besides the turbine)
  • They are clearly cost effective
  • They provide a decentralized power system. this is beneficial with regard to loading on the power grid, i.e. a blackout need not effect large areas. But also they make terrible targets for enemies as damage from an attack would be very minimal.
  • They are economically feasable now, they do not need government sponsorship (something the republicans call socialism) to make them financially effective. Solar, wind, geothermal, biodiesel and other forms of alternaive energy do not share this feature right now.
Nuclear fuel is cheap and most of it comes from people who are friendly with us (Canada and Australia). At least they are friendly now.

So I say, if your community wants to invest in this system, Go For it! You could provide cheap power to your residents which may in turn boost your local economy.

But as a national plan, we need to get to the far harder to do "alternatives". We need to get a carbon capturing system in place. Algae to oil methods are the only viable method, but we are suffering from the current inability to convert the algae into oil in a financially viable (read: energy positive) way. We know how to grow it and what species to grow, we know what methods work to convert the oil to fuel. But we are stuck and the actual oil extraction part.

Similar problems abound with solar, right now the energy it takes to make a panel is only 1/7th of the energy is produces over its lifetime. This is a crappy ratio and is one of the things that leads to the high cost of the PV modules.

My moral of the story is here, endorse the local installation of decentralized, passive nuclear power, but ask your politicians to fund the alternatives, nuclear is not a long term solution. Oil and natural gas are not a long term solutions. Solar (including solar thermal and wind), geothermal, and algae are the long term solutions we need to figure out how to make financially competitive to excel forward.


On 11/22/08, 10:35 AM , Rod Adams said...

techskeptic - Your interpretation of the fuel use for nuclear fission plants is not accurate. The amount of fuel consumed per unit time and thus the length that the plant can last is completely controllable.

Reactors are devices that are designed to reach and achieve a stable temperature, but the amount of fuel that needs to be consumed to maintain that temperature is a function of the amount of heat that is removed from the reactor to produce power.

Because of the natural laws of thermodynamics, a Hyperion Power Module that produces 70 MW of thermal energy can provide enough heat to produce about 25 MW of electricity. In order to produce 70 MW of thermal energy, the reactor will be fissioning about 70 grams of fuel (U-235 or Pu-239 in this case) each day.

If the load on the electrical system is only 1 MW, for example, then the heat production (thermal power) in the system will drop by a factor of 25 so it will only be about 2.8 MW. Producing that heat output only requires the consumption of about 2.8 grams of fuel each day.

If the Hyperion Power Module is loaded with enough fuel to provide 70 MWth for 5-7 years, it would produce 2.8 MWth for 125-175 years. Of course, 1 MWe is far larger than any individual home - the normal American home uses an average of about 1-2 kw. (1/1000 to 1/500th as much, but I thought the math would be simpler to follow with a 1 MWe example.)

On 11/22/08, 4:37 PM , carnifex said...

But, Rod, isn't that assuming the Hyperion system is a fission reactor? From what I see on their site, that doesn't seem to be the case, thought I'm no expert and can't pretend that it's very clear to me. As it stands I have this idea of a small nuclear pile decaying at a rate that is thermally significant but non-fissioning, effectively giving it a fixed output. A full on reactor wouldn't be so black boxable, I wouldn't think.

On 11/22/08, 10:44 PM , Techskeptic said...


I wrote a long response, and now deleted it because I found this.

There are a couple of goodies in here:

" We're leveraging the design of a very common reactor, called a TRIGA reactor."

More on that in a sec.

"How long will each reactor last?

John R. "Grizz" Deal: Probably eight to 10 years, but that depends on your use because it is load following. The more you use it, the faster it is used up, just like a battery. Again, there's that battery analogy! "

So I stand corrected, they claim that the reaction self governs based on the load.

I understand now, the reaction self regulates based on temperature . If the heat is not taken away, the fuel will want to heat up and moderate its activity back down. So there is control, but it happens up at the place where the heat is removed, not at the nuclear reaction.

But to be clear, it is still consuming fuel even if the electrical load goes to zero. It will still maintain a hot temperature. I think this is why they have fuzzy numbers about the life of a system, I'm now reading 7-10 years depending on load. It does not look like the life to load graph is linear.

The activity will be reduced with less electrical load, I just don't see any metrics of how much.

On 11/23/08, 2:53 AM , Rod Adams said...


When the heat demand for power production goes to zero, the only heat required to keep the reactor at its elevated temperature is the amount required to overcome natural heat transfer on the surface area of the reactor.

That should be a very small amount, probably on the order of a few kilowatts. As you can imagine, a reactor that is well shielded an buried underground is also well insulated from a thermal perspective.

In addition, like all fission devices, there will be a certain natural heat production that comes from the radioactive decay of fission products. Immediately after heat demand for power production stops, the fission power level drops down, but the decay heat is initially (for the first few minutes) as high as 6% of the initial power level.

That drops down to less than 1% within an hour and to about 0.5% within a day.

I have tried to convince John Deal and his marketing team that applying a battery analogy to their fission based heat source is stretching way too far. I like to think of it more like a furnace or boiler that does not need any new fuel or produce any emissions for a very long time.

If you want heat from that furnace, it is a pretty simple matter of the right kind of piping. If you want electricity, you have to build and operate a power plant. That is a bit more complicated that hooking up to a battery, but it is still a huge advantage over the competition.

However, I have to take my hat off to Hyperion - they have actually been funded to a substantial level to develop their product, something that I have not yet been able to achieve.

Rod Adams
Founder, Adams Atomic Engines, Inc.
Publisher, Atomic Insights
Host and producer, The Atomic Show Podcast

On 11/23/08, 6:40 AM , Techskeptic said...

yeah I think i commented that I understood that the control mechanism for the power plant lay with the heat extraction part, as opposed to active control over the nuclear reaction.

I literally can not discuss the magnitude of heat leakage from these things. They have not published data on it and I am not willing to become expert in nuclear systems to know the ins and outs. I'll take your word on it, even though dense materials (like concrete and steel) are generally not great thermal insulators, but there are good engineering solutions to counter their poor thermal performance.) But yeah, I am sure that if all heat removal is stopped, that it would not take more than a couple of tens of kilowatts to keep it hot.

Regardless, the point of the article was to compare the claim of being cheaper than alternatives, and in particular I compared this claim to solar on a kWh to kWh basis over the life of the product. Its doesnt matter if is it used minimally at low power or drained at its maximum rate for that comparison (unless the efficiency significantly changes over the time interval).

As for their marketing, I gotta agree with them. Their goal to make this product viable is to make it seem as little like a normal nuke as possible. While I think your boiler analogy is fine and perhaps more accurate, they have only one or two sentences to communicate how this thing works, before eyes glaze over.

A battery loses charge directly depending on the load. It loses charge while sitting on the shelf due to internal resistance and electricity can be used to power electrical devices and heaters.

Its not 1:1, but for people with little atomic knowledge (i.e. clearly not you), I think its good enough.

My hat is off to them also. I hope they sell thousands of these things, while the country works to replace them in years to come with longer lasting 'alternative' solutions.

On 11/23/08, 10:27 AM , Rod Adams said...

techskeptic - I am not a marketer by trade, training or even inclination. I prefer to under promise, over deliver, and sometimes even sought to fire customers who were too impatient to listen to my cautions about their purchases.

Anyway, one of my challenges with people wanting to compare life cycle "costs" of nuclear generated electricity to solar generated electricity is that such a computation ignores the "value" side of the equation. Power that is there on demand 8760 (or some number very close to that) hours per year is far more valuable than power that is only available on its own schedule.

It is the difference between groceries and garden grown food. Nothing wrong with the garden food and it can be very inexpensive per unit of produce, but it does not fill the belly in the winter, spring and much of the summer.


On 11/23/08, 6:56 PM , Techskeptic said...

hmm.. perhaps I should repeat the analysis using geothermal oreven solar thermal. It would certainly be more apples to apples.

On 11/27/08, 1:28 AM , Rod Adams said...


I would be interested in your comparison of Hyperion to geothermal or solar thermal.

One aspect to remember is that the Hyperion concept is a device that can be built in a factory and delivered to a site that is potentially far removed from the grid. That device will provide power 24 x 7 on demand.

Geothermal may be able to do the "on-demand" part, but except in some very special areas of the world, geothermal requires a lot of site work for drilling deep into the earth to a depth of approximately 10,000 feet or more.

Solar thermal will require a vast quantity of collector area to produce the same amount of power as a Hyperion Power Module, and it will require a great deal of storage volume. Remember, once the sun goes down, the power production using stored thermal energy means that the storage volume gives up its heat at a rate high enough to support the steam production. It is not hard to figure out the relationship between specific heat capacity of the storage fluid, the range of acceptable temperatures, and the volume of fluid needed for various power demand profiles.

Looking forward to the computation.

On 1/23/09, 6:11 AM , Cyril R. said...

Two major issues with this simple analysis:

1. The time value of money - financing.
2. Variable cost (O&M plus fuel)

With respect to number 1 the micro nuke would be favorable. With respect to number 2 the PV wins quite definately.

However, Hyperion won't licence the design before 2013.