Design Goals
In the previous post I was talking in system-wide terms. That plan envisions a significant fraction of homes generating their own power. But only an insignificant fraction of homeowners can afford to install multi-kW photovoltaic systems or build dams to drive expensive microturbines.
I don’t have a river, an ocean or a volcano in my back yard, so my house is pretty much limited to solar power. There was a relevant quote in the most recent issue of Make Magazine.
You can divide most solar power research into two camps: increasing efficiency or reducing cost.
I would like to add to that quote: “And one of these camps is barking up the wrong tree.”
Efficiency is a red herring when fuel is free and space plentiful. The sky is raining fuel, don’t break your back trying to squeeze every last watt out, just set out more buckets.
A typical home uses about 2-3kW averaged over a day. A square 6+ meters (20 feet) on a side, if converted at 10% efficiency, can power your home. That’s much smaller than your roof, let alone a typical yard. Not enough or want some backup? Just build another square. With a good, cheap design this should be easy.
So what’s needed is a good, cheap design for turning solar power into electricity. 1 The usual solution, since the chemical processes of PV are beyond DIY, is some kind of heat engine, usually a Stirling because of the simplicity.
Making a Stirling engine that runs on solar power is not that hard. Making a big one that could conceivably power a home is not much harder, provided you have access to pretty sizable machine tools. Stirlings are very uncomplicated, but any machine with moving parts and a requirement to be air-tight is going to need some special skills or tools. 2
And I think I’m going to leave this on that note. I have what I think is a good, simple, cheap design but I have to build and tweak it to make it work and have pictures/video.
I’m far from a font nerd, but man is that capital D ugly. Looks like the punchcut theme doesn’t let me control that easily though.
“Efficiency is a red herring when fuel is free and space plentiful” is /almost/ (but not strictly) true. Moreover, efficiency/cost is a false dichotomy – what we strive for in a good design is to produce the greatest efficiency we can manage at the lowest cost.
In order to meet that goal, we need to begin by discovering how to produce the greatest efficiency, whether that efficiency meets minimum requirements, and then establish some kind of relationship between efficiency and cost.
If maximum efficiency doesn’t satisfy requirements, we can conclude that the particular design approach doesn’t provide a viable solution to the problem we’re trying to solve – and we can then pursue other approaches.
If we see that the design does provide a viable solution, then (and only then) can we meaningfully begin looking for ways to reduce the cost without degrading the design to the point that it no longer meets our needs/requirements.
Right, but I think that’s almost backwards from what I’m trying to do. I don’t want to approach it from the mindset of “how cheaply can I achieve efficiency X”. I want to approach it from the mindset of “how efficient can I be with cost C”.
It’s all the same calculation, I’m just weighting cost much more heavily than efficiency. I’m doing it that way because fuel costs are low and in any case efficiency is an indirect measure of that. Am I better off efficiently producing 1W or inefficiently producing 1kW?
So I’m optimizing for things I actually care about: power and cost. Efficiency is tunable.
This is a conversation I’ve already had with myself. :-)
If we’re pumping, we care about how much water we can move – and if we’re generating electricity, we care about how many kW we can produce. Before we begin designing our pump/generator *system*, we need to know something about the efficiency in order to determine the amount of input power required. For example, if we require 1kW of output power and have two design approaches, one with an efficiency of 10%, and another that costs twice as much with an efficiency of 50%, then we can calculate that the 10% system will require a minimum of 10kW input to meet the design requirement and that the 50% system will require a minimum 2kW input to meet the same requirement.
Now, the photon stream coming from the sun is available without cost, as you pointed out, but the collectors to capture that energy are not, and the 5:1 cost differential in collectors is likely to be greater than the 2:1 cost differential in engines.
It’s at this point in the process that one might pause to reconsider the possible system cost advantages of the more efficient engine that costs twice as much…
More about Carnot cycle engines (of which the Stirling cycle is just one type) – the maximum possible efficiency can be calculated from the hot head temperature and the cold head temperature (both in Kelvins) according to the last formula on the web page at http://www.iedu.com/DeSoto/Projects/Stirling/StirlingCycle.html – and if you play with that formula and different hot head temperatures, you’ll discover that the cost relationship becomes still more complex (but working through those complexities is an essential part of the problem-solving process!)
/And/ there’ll be a need to work through everything again on the output (pump/generator) part of the project…
All of which is why it seemed to me to be a good idea to stop trying to do it all myself and enlist the help of volunteer R&D teams – and why my favorite pun has become “Many hands make light work.” :-)
For example, if we require 1kW of output power…
I don’t have an output requirement. This isn’t a module that needs to interface with a well-defined input requirement somewhere. This is an attempt to get as much as I can as easily as I can.
…you’ll discover that the cost relationship becomes still more complex (but working through those complexities is an essential part of the problem-solving process!)
Exactly. The costs are multidimensional and non-linear. Time, money, access to materials, access to tools, etc are the dimensions. Non-linearity comes in when you consider that a dimension like “access to tools” doesn’t vary smoothly but may end abruptly when you pass “handsaw”.
So I can’t just say “a 5x more efficient engine costs 2x as much” because “2x” implies a non-existent linearity (not to mention single dimensionality). What if that more efficient engine requires a lathe I don’t have? I’m stuck with the lower efficiency, lower cost engine and I have to make the best of it.
Sure, I could factor in the cost of buying the lathe and call it 3x more expensive instead. But I’ve also raised the barrier to someone else following my same design. How do I I factor *that* in?
Or maybe this is a better way to think of it: I’m still exploring the design space. Because of the multidimensional non-linearities, I have to sample around to figure out how cost and efficiency depend on each other. Then I can choose among them based on the logic you so clearly outlined.
“Because of the multidimensional non-linearities, I have to sample around to figure out how cost and efficiency depend on each other.”
Exactly what I was trying to say. ;-)