Back in the good old days every electrical engineering student was required to study linear control theory in all its canonical glory. Identify the n system state variables (usually node voltages or mesh currents), and write out the the resulting nth order system in terms of n first order equations. Package everything up in matrix form and you were ready to either tediously solve by hand or (if you were lucky) by computer.
Then if you wanted to some way optimize the system you could figure out what variables to optimize, form the appropriate scalar function, set the derivative(s) to zero and voila! Your homework was almost done. Too bad your mechanical engineering buddies were stuck with that nasty tensor analysis stuff.
I loved control theory, particularly later on as I explored the tidy theoretical world of a grad student being funded by a NASA grant.
But then I went to work in the research department of a very large electric and gas utility, and suddenly it was Katie bar the door! This was the real world of the grid.
Sometimes it was almost surreal – for example, my friends in the transmission planning department had the latest load flow analytics that were expensive and elegant but not very accurate when compared to what the substation meters were actually reading. We found that much of the error, but not all, was due to about 30 percent of the substation VAr transducers being installed backwards and these outputs were used to calibrate the load flow state estimator. Accurate voltages and phase angles couldn’t be obtained to allow the load flow software to work. The theory was fine, but the execution in the utility environment was iffy at best.
The dilemma was that in order to optimize a system you needed to control it. But you can’t control what you can’t observe.
Not that we didn’t try.
I’m not sure, but I think I might have designed the first applied ‘synchrophasor’ during stability testing of the new west coast 500kV intertie and by inference, the entire western transmission region. We built zero crossing measurement circuits out of (gasp) bipolar and unijunction transistors and other discrete components and installed them in the intertie substations from Washington to southern California. We didn’t have time-stamped digital communications then so we had to rely on analog microwave with all the attendant signal jitter which we had to statistically (by guess and by-golly and plenty of hand waving) average out.
It all seemed to work and we planned a power step function test for a weekend. However, an earthquake knocked out the intertie on Thursday and our circuits caught the data. The power swings were what we predicted and the big cheeses in the general office we’re happy. A report was written and the circuitry was put in storage in our R&D warehouse (affectionately known as the home of dead dreams). The engineers on the research team got a free fast food lunch and the technicians got massive overtime checks. A job well done.
But, optimize the grid? HA HA. Whaddya, nuts? Even a decade ago it couldn’t be done. The best you could do was provide incremental improvements where technology and the usually benevolent commission allowed.
Now, years later the limitations are falling away. Synchrophasors and other sensor technology and fast, broadband communications are available. It’s possible to collect those esoteric voltages and phase angles as well as monitor tower structural strains and incipient failure during a storm. We can accurately isolate outages, reduce truck rollout and reduce outage times.
But we can do more than observe – we can control. Storage technologies, demand management, maybe even better control over dispersed solar. We’re not just talking transmission either – look at all the smart switches and other distribution/microgrid-type products.
Optimization of what the National Academy of Engineering called the greatest achievement of the twentieth century – the electric grid- is becoming possible.
Now the key question is what weighted combinations or qualities do we want to optimize? Cost? Reliability? Safety/security? Everyone has their own idea and agenda so you can bet we’ll be increasingly chewing on these issues for years to come.
Let’s chew together - share your comments below.
Interested in finding out how you can be of even more help, perhaps in an advisory role? Send me an email: [email protected]
Then head on over to the Grid Optimization reader poll Top Grid Changing Technologies? and vote on which technologies you think will contribute the most to grid optimization.
Catch the spirit!