CAES (Compressed Air Energy Storage) is being promoted as a way to smooth the delivery of intermittent supplies of power from e.g. wind.  This would increase its ability to displace other supplies of electricity and reduce carbon emissions from the same.

Of course, pumping lots of air around is going to have secondary effects (you cannot do just one thing).  A question I have not seen addressed yet:  how much CO2 would be handled directly by CAES in the process of compressing air?  What if some fraction of this CO2 was chemically bound and not released back to the atmosphere?  Could this make a significant difference in atmospheric CO2 levels?

I don't know, and I don't have the energy at this moment to come up with a ballpark estimate.  But given the large volumes of air involved in CAES, widespread use might just be a two-fer.

¶ 5/31/2008 01:20:00 AM 19 comments

Pablo Picasso once said, "Computers are useless, they can only give you answers."  He was right.  Correct answers to the wrong questions get you nowhere.

The crisis in commodities is feeding back into the cost of energy.  The amount of raw materials (both steel and fuel) required to bring new oil fields into production is growing rapidly.

Sooner or later, you'll get more energy return in less time by using the oil to make composite blades and the steel to make pylon towers than to drill in really out-of-the-way places.  The question is, how close are we to that day?  Could some places be there already?

Maybe that's not the best question, but it looks like a good one.

Labels: commodities, energy, peak oil, resources

¶ 5/28/2008 11:30:00 PM 0 comments

I've been suffering with the products of a company called Sauer-Danfoss.  This company makes industrial controllers and displays, programmed with a proprietary graphical system resembling ladder logic.  Here is my litany of complaints (addressed as an open letter, because I would like the world to know what a pain this stuff is; maybe S-D will take bad press as a good incentive to improve their product, rather than mealy-mouthing explanations for why The Product Is Perfectly Designed For Its Intended Users).

Dear Sauer-Danfoss,

I have some serious complaints with your Plus+1 Guide software.  The poor packaging and lack of proper documentation (plus counter-intuitive behavior of the system versus the description in the help) have caused me to waste a large fraction of a week thus far.  Your "tutorials" lack the most basic amenities for the aspiring learner.  Several of the extended sequences can neither be paused nor backed up to repeat sections.  Even YouTube does better than that, and their stuff is free.

The graphical system is painfully cumbersome.  Take, for example, a common calculation:

degrees_F = (degrees_C * 18 + 5)/10 + 32

This takes about ten seconds to type in C code.  It takes no fewer than four calculation blocks, four typed constant blocks and drawing a number of dataflow "wires" to do it in Plus+1 Guide, and the product is inherently obfuscated by the complexity and harder to document.  The fact that the graphical mess is translated into roughly the above C code before being compiled adds insult to injury.

Packaging needs to be radically improved.  If you download the Plus+1 Guide software, you have neither the device templates nor the function libraries.  The rest of the world has figured out how to create interactive installers, why haven't you?  These things should either be included and installed by default, or prompted during the install process.  If the user should fail to install them, they are left with blank tabs in Plus+1 Guide.  Nothing mentions a missing element, no pop-up tells the user what else is required.  How useless is that?

Or take the "context-sensitive" help... please!  It claims that you can click on something and press F1 to get help on that block.  I opened an existing application (trying to hit the ground running), clicked on a block with a triangular symbol in it (looked like an op-amp) and pressed F1, hoping that I could get some help to tell me what this unknown block was.  I got the standard help screen.  Try again, same result.  Apparently, telling me what this function block was didn't rate in Sauer-Danfoss's priorities.  I found this not the least bit endearing.

The "component" menu was no help either.  It's full of little icons for various categories, but the icons bear no resemblance to the symbols actually used in the program itself.  You have to click through list after list after list until you find the item which bears a similar symbol to the unknown.  THEN, once you have gone through the laborious search, you can click on the item in the component MENU and use F1 to get the details.  You cannot use this to get details on a component already in a design and save time and effort.

This is the sort of design detail which leads me to suspect that the product has been made deliberately obscure, to sell expensive training classes.

The "service tool", which downloads compiled programs to the units, is equally ill-designed (if not worse).  It appears to require an ECU list before it will do anything (even with hardware connected), but the navigator entry for the ECU list won't open when clicked (perhaps because the company firewall blocked a download - but there is no diagnostic for this!) and there is no help entry for ECU lists.  How do I get an ECU list?  Do I need to download something?  The "help", doesn't.

In conclusion, you have the sorriest, lousiest, most cumbersome, most time-wasting excuse for a programming interface that I have seen in more than 20 years of work on embedded systems.  I have accomplished more useful work in less time with assembly code.  I am forced to use them in my job, but I will be making a personal recommendation that design engineers avoid your products at all costs.

¶ 5/06/2008 12:24:00 AM 3 comments