[PLUG-TALK] The unstable "smart" grid

Keith Lofstrom keithl at kl-ic.com
Thu Oct 22 15:29:30 UTC 2009


I attended the TiE "smart grid" presentation at Intel Ronler last
night.  Interesting.  Frightening.  The idea is to use computers to
regulate usage and make electrical demand sensitive to availability
through a rapidly adjusting price mechanism.  There were so many
claimed benefits, it is clear that the smart grid will bring about
the Second Coming!

Let me see - large interconnected systems, delay, and gain.  Where
have I seen that before?  I know - oscillators!  

Anybody remember October 19, 1987, Black Monday, when the S&P and
the DJIA dropped 20%?  They blamed that on computerized trading.
The California energy crisis of 2000 and 2001, helped along by 
our buddies at Enron?  Adding computers doesn't automatically make
a system more stable, especially if the system designers aren't well
versed in stability math, and don't run lots of worst case simulations,
and don't understand what happens when self-interested players game
the system.

When you mix in large scale financial incentives, you get into "model
wars", where companies with well tuned and rapidly reactive system
models are able to rattle the system and draw money out of the
variations.  Often at the expense of their competitors, with their
own models.  The financial markets now are driven by millisecond
trades, and woe unto the Linux kernel designer that improves bulk
processing at the expense of millisecond latency.  Arbitrage is
normally a good way to damp out instability in an economic system,
but the system has to be dampable in the first place, and the
arbitrageurs must be smaller than the intermodulation products.

When I mentioned large systems of coupled, nonlinear differential
equations,  and how difficult such systems were to understand, 
much less stabilize, the room full of policy wonks, computer geeks,
and CEO types had no clue how all that esoteric math could possibly
relate to their visionary nirvana.  I did not even get into the fact
that the communication system is powered by the electrical system;
making the electrical system dependent on sophisticated communication
is not a formula for robustness.

I hope all the wonderful things the smart gridders dream about will
someday come to pass.  I wonder how many lethal lessons will be learned,
and how many power grids and economies will collapse, on the way?

BTW, the computer slide presentation system failed partway through ...

Keith

-- 
Keith Lofstrom          keithl at keithl.com         Voice (503)-520-1993
KLIC --- Keith Lofstrom Integrated Circuits --- "Your Ideas in Silicon"
Design Contracting in Bipolar and CMOS - Analog, Digital, and Scan ICs



More information about the PLUG-talk mailing list