Wednesday, December 29, 2010

family stuff and Tron Legacy

Here are the three of us at Christmas dinner
at the house of my ex-wife in Maryland.

Chris persuaded us all to see Tron Legacy (in IMAX 3-D) a couple of days ago. At first, it just seemed like entertainment, not really addressing the kinds of global issues I usually think about, or even state of the art computing. But then I realized -- it really reflects a lot of the basic feelings here in our house in Virginia, overlooking the wilderness, but ten minutes from some tricky complex systems. Even Zeus is a neat reflection of Chris's head principle.

But little games like conventional energy technology and economics are still a bit better than sudoku or doing nothing at all, in the off-grid world...

This is not the whole family -- just the three living in this house.

Sunday, December 5, 2010

exciting new energy options from the world of electrical engineering

Learned a lot from what may be the most advanced cross-cutting meeting
on electrical engineering
anywhere, last week. The new work changes a lot of stories. Among them:


Background (the previous story):

Several years ago, the best available numbers (SAE) said that the
Toyota Prius hybrid
cost $3000 more to make than the comparable conventional car -- $1000
due to new power electronics,
$2000 due to battery. The plug-ins and fuel cell cars out there
probably increase both costs, maybe doubling
them, but reduce gasoline use a lot more. The mileage of the Prius
hybrid itself is improved 15%
through a no-cost upgrade in control, using the new control system
developed by Danil Prokhorov (whose
earlier work I funded from NSF).

The Volt and many other PHEVs can be recharged in your home, at 110 volts
(8 hours) or 240 volts (3 hours), but true electric cars (like Nissan
Leaf) require "gas stations" which can recharge
your car in 5-15 minutes. Fast recharge stations would also enhance
the value of PHEVs to the owner and to
the power grid, and enhance our ability to survive any sudden cutoff
or price shock of oil.
Therefore, many US states are spending lots of money on building
$100,000-$200,000 fast recharging stations,
using Japan's "Chademo" standard for fast DC charging. Large vested
interests have begun to spring up.
I have the impression that this is like "$150,000 per parking place."
SAE may be on the verge of blessing Chademo
as part of new standards.

The new options/story:

Alireza Khaligh and of Illinois and Hui (Helen) Li of Florida State
presented important and exciting new results on the power electronics
part of it.
They have developed integrated, more universal systems which could
replace the power electronics of cars
ranging from Prius to Volt to Leaf. The new technology is expected to
cut the overall net cost of power electronics
by 1/3 to 1/2, AND AT THE SAME TIME make it just as easy and safe to
recharge the car using 480 volts AC as
with the Chademo standard. THE IMPLICATIONS FOR RECHARGE ARE HUGE: the
$100,000 to $200,000 I was citing
is for a recharge station that takes 480 volts AC as INPUT. If
recharge stations can be used which are "free" (starting from
the same 480 volts AC input, and not accounting for billing system
add-ons in either case), it would be a whole lot easier and cheaper to
quickly build up a national network of fast recharge stations. What's
more, if the costs on the car are LESS than what
GM is already paying for on the Volt... this is huge. Friends in the
auto industry have explained to me why this kind of
breakthrough really often has to start in the university research
community (especially when major companies are escaping from recent

I have done some due diligence on this. The poster paper giving
details of Khaligh's work was co-authored with his colleague Emadi,
one of the authors of the most definitive recent book out there on the
kind of motors used in cars. They have actually retrofitted Fords
to prove that the technology works. We are either AT or VERY CLOSE to
having good justification for some kind of $10 million demo,
aimed at deploying the maximum number of cars and recharge stations to
really prove the power of this new lower-cost technology,
more precisely combining:

1. Fast recharge stations at least as fast as the Chademo stations
being set up for Nissan Leaf, without compromising 8-10 year battery
2. A network including the most competent university centers to feed
technology to the project (Illinois, Georgia Tech, FSU, U. Michigan,
U. Hawaii at minimum);
3. The new power electronics, Switched Reluctance Motors (which do
pass due diligence now, and free us form need to use
rare earths), control as advanced as Prokhorov's or more, and
batteries at least as good (in kwh/$ meeting other minimum specs)
as the "present big five" -- Excellatron rechargeable lithium/air,
ReVolt rechargeable zinc/air, Chinese best lithium-ion (Thunder Sky,
BYD or RJE).

NSF does not fund that kind of larger-scale project. But I still have
questions about a crucial technical area, i.e.:
Can we really reconcile regular use of fast charge (as fast as what
Oregon/Leaf are installing) with 8-10 year battery lifetime?
This may be just a matter of my own ignorance. For example, the folks
deploying that right now probably have a story on battery lifetime.
Maybe NSF might have some role in improving the public, open
literature understanding of this key requirement.
Or maybe people putting together demos could make some arrangements
with folks who have that proprietary knowledge, to allow a demo to get

Bottom line: It would be of enormous value to world economic and
war-and-peace security to somehow get such a demo going as soon as
possible, somehow, dependent on the two obvious milestones: (1) having
a more complete story on the battery lifetime fast charge issue;
and (2) working out the exact plans to produce the maximum number of
PHEVs and recharge stations possible within the $10 million

Of course, folks developing standards should not enforce the Chademo
standard so rigidly that it gets in the way of something better and
Maybe it's just as well that some of the present day clean cities
programs have not gone faster than they have, because there is a risk
of wasting unnecessary sums of money creating something entrenched and



At times, I have heard salesmen trying to tap into smart grid money
say: "Just think how well we could
use renewables and prevent blackouts if only we have the
communications, meters and sensors in place..
a vast new flood of data, enabling better decisions." Good new
communications and sensors are
important, but a flood of data is not the same as a better decision.
To make better decisions, to better accommodate
and use renewable energy, we need two more essential ingredients: (1)
more real intelligence, to learn from
that stream of data and translate it into action implications
(requiring MUCH more than the antiquated stuff the salesmen
usually offer and dress up in fancy buzzwords); (2) CONTROL AUTHORITY
-- something to decide upon.

For many years, people actually trying to stabilize grid flows (as
oppose to tapping public money) has stressed the
need for new "FACTS" digital switching technology, which can be used
to change power flows, and actually give us something to decide
upon. But these have presented problems with cost and reliability,
which have limited deployment in the US.

The New Story

Two researchers at Georgia Tech (one funded from my area at NSF) have
come up with new technology for "thin
AC controllers" which does what FACTS does, but at lower costs with
safe failure modes. The key people are Ron Harley and Deepakraj Divan.
I expect to hear a whole lot more on this in coming years, as they
work directly with utilities and tell us more. But it turns out --
a good place to insert such systems is in front of big transformers.
They say this would allow total protection against big DC power
such as people expect with solar flares (at least with some warning,
which need not be too far ahead of time). Dual use technology --
more grid capacity and ability to exploit renewables, ALONG WITH
insurance on the solar flare side.

There is also a story on wind which I will probably say more about later.


People worried about the PHYSICAL security of the power grid, versus
terrorists, have generally pointed towards transformers as
the number one vulnerability to worry about. Big expensive oil drum
things, all one of a kind, sometimes
imported over the ocean, hard to stockpile and build up, etc., etc.

The interagency current wisdom seems to be: "Solid state US-made
transformers are not only less efficient (like 97% versus 98%) than
the big old oil drum kind.... they are also far more expensive.
Because of cost, they are not suitable even as backups to
be held in readiness in case of emergency. There are certain physical
laws at work... at 60 hertz, calculate the size.."

Prof. Sudip Mazumder -- an NSF grantee who also serves on one of the
frontline interagency committees on all this -- reported how
the new story is quite different.

The underlying solid state technology is closely related to what Li
and Khaligh have been developing. People have worked out how to work
electricity at frequencies like 20,000 to 40,000 cycles per second,
which allows deep reductions in mass and cost.
(They have worked out safety and shielding aspects, etc. This is
working stuff, not just theory.)

The technology is THERE TODAY and competitive enough to replace
traditional transformers in distribution systems
and in renewable systems (like big solar farms or wind farms), for
multimegawatt applications. His feeling is -- let's
start with revolutionizing the distribution system first, where it's
easier, and work out way to the transmission system from there.
Why not wait, say, 5 years, for the gigawatt applications, which
really want stuff like the new 20kv chips which will be coming in a
few years?


There was also a radically new design for wave power presented,
CycWEC, by Stefan Siegel of the Air Force Academy. It still needs to
prove a lot
of things, but it uses more basic principles in design than earlier
wave energy efforts. This allows something like three times
to efficiency in extracting wave energy, and, more important,
something like a 50-fold reduction in mass, according to Siegel.
Tests so far have been positive, but he wants a mid-scale demonstrator
before moving to the kind of full-scale ocean demonstrator which would
make it possible to design full-scale power plants and estimate what
they would cost. (As with other wave energy schemes, I do wonder
where the wires will go to the power plants, etc.)


All these should be followed up, by the energy community.

These are just a small selection of things I heard/learned at the 2010
NSF ECCS Division Grantees' Conference,
in Hawaii, this past Wednesday to Friday.

In actuality -- I do not know of any other meeting which brings
together some of the top (peer-reviewed, front line)
creative people across ALL disciplines of electrical engineering in
one place, in a way which encourages cross-cutting conversations.
Beyond the stovepipe. Everyone present either worked for NSF or for an
NSF-funded research project, except for one, so far as I know.
That one was from the NSF of China, interested in improving US-China

Other discussions I saw included things which will help change our
understanding of time and mind, electronics beyond Moore's Law,
and various themes in advanced photonics. (There was also a magnetic
bottle fusion talk which tells me that there are ways those folks
could move a lot faster, but not my top priority...). There was a huge
amount of all kinds of nanotechnology and communications technology
and microelectromechanical systems. Folks were complaining that there
are only 5 real experts left in the US in high voltage motor design...
three of whom I happen to know...
all past retirement age and worried what happens when they DO retire...


Best of luck,



I forgot to mention that the folks with the thin AC converters estimate that their
technology would let us hook up more wind and solar to the power grid, cutting the cost of new transmission requirements in half or more.

Some folks on the global energy list asked for more... and I said:

EMP is yet another important topic. We got very deep into that
in a recent IEEE dialogue. Living life "at 300,000" feet, it would be reasonable to conclude from the most authoritative reports now available that there might be a 25% probability of trillions of dollars
worth of damage if we do not insert new protections before the next big solar flare event, which could
come as early as 2012. I would want to help.

A Congressional office on the other list asked about this item too. My reply:

Looking on the web, I immediately find two references to the thin AC converters:

and two to the key people I mentioned:

These references basically reflect the talks they gave at the grantees' conference.
When I saw the diagram of a switch right in front of a big transformer, I ASKED...
in the open question period... the obvious question... whether this would be able to isolate the transformer,
in case of a solar flare. They answered that of course it would ... if it would be a high voltage DC surge.
(And that's what it would be. Hopkins APL and Iridium have given detailed presentations
on solar flares to NSF recently, in connection with the NSF-funded "Ampere" program they are working on.)

I have asked Harley for more. They have not been doing this for the PURPOSE of EMP
protection. It's "just a side benefit." And of course, it's worth double-checking anything at this early stage.
The method itself they have vetted in great detail, but they didn't even think about the side benefit until I raised
the question this week.

They have stressed how much cheaper this is than conventional FACTS technology, but
I do not know what it would take to have it deployed, say, over half the country's big transformers by 2012,
as part of smart grid efforts.


By the way, this meeting was in Hawaii. As I post this on Sunday...

I got back from Hawaii Saturday morning, still in pain from a fall Wednesday morning,
at the start of the meeting. Bruised or cracked rib, it seems. Not one glimpse
of the beach (except as I waited for the shuttle bus to the airport to go home),
every minute in the hotel (except for two breakfasts at the Wailana cafe across the street), but still a whole lot of fun talking to really interesting people.