If enough power plants in grid fail, then that grid will not make it. How many have to fail for this to occur? Is this 5%, or 10% or 50%? Does anyone even have an estimate of the number of plant failures it will take to bring down a grid?

If we know the answer to this, we can then read the SEC 10Qs (starting in January 99) and start making our own estimates of plant and grid failure.

-- Anonymous, November 02, 1998

Answers

Steve, I don't believe it's that simple, which is a pity. There have been failures in ONE place that have caused a grid to fail, and there have been failures which have not brought down a grid. A simple failure in one transmission line in Utah (1996) triggered a power outage extending from Calgary Alberta to California and Texas. And we all know of times when a regional area has been without power and the overall grid holds up. It seems to depend on what fails where, and when, and what the load is on the grid at the time, and how the other generating facilities react, and on and on...

Just too many variables. The only certain thing is that Y2K will present the biggest challenge the power industry has ever faced: the possibility of synchronous multiple failures across the nation.

-- Anonymous, November 04, 1998

The answer totally depends on the load on the system at the time. The system is a complex arrangement of interconnects that attempt to balance load, power factors, etc. to maintain a stable system. If enough capacity drops off line to effect this balance, the system is in peril of failure.

Three examples: (1) Approximately two years ago, the entire panhandle of Texas lost power for 8-12 hours. Toke Station is a 1000MW coal fired generating station. A workman was cleaning dust from the big transmission transformers with isotonic water. Too much static electricity built up and a 1000MW boom followed. The 1000MW that was taken out of the pool caused the remaining generators to be fully loaded. Under-frequency fail-safes tripped and took 6+ more generation units off line in a matter of minutes. After the second one went down, the rest followed quickly. (An under-frequency or under-rotation condition, if allowed to continue, can shed a generator.) Approximately 400,000 people were without power for 12 hours.

(2) A city has its own utility, but can continuously produce only about 75M, but with a few hours warning can produce up to 150MW for a short time. It buys 75MW to supply its normal needs. One fine day its 75MW source failed. The city started its supplemental generators and had most of the city (145MW) back on-line in 2 hours. At noon, a small 10MW co-generation unit failed. The city was running very close to capacity and now only had 140MW to supply 145MW capacity and the entire city experience brown-out voltages and then went down for hours.

(3) I personally know of several instances where a regional grid was stressed due to high energy usage, generator maintainance, etc. and the grid came dangerously close to "melt down". The system controllers purposely opened the breakers in sub-stations for 5 minutes or so and then re-closed then. That small action turned the power off for 5 minutes to customers, but more importantly shut down inductive load (motors). Most large motors are started with magnetic contactors and do not automatically come back up when power is restored. Thus load can be shed for short periods of time. That action repeated in different localities during the peak period is the only thing that kept the entire regional grid from overload and automatic shutdown. (If their customers knew that this action was on purpose, I suppose many lawsuits would follow in order to pay for damaged goods, car wrecks, ...)

Our electrical grids are marvelously complex and interrelated. I'm very certain that a bump in one place can be rapidly transmitted to vast regions. Under normal circumstances, the grid can be brought back up slowly, but in a matte of hours. Under Y2K conditions, I could certainly see days, maybe even weeks of on and off power while the offending components are located and work-arounds are created. One thing seems certain, the cost of spot power will be high. If a supplier can deliver metered amounts of power reliably, they should make some serious money.

-- Anonymous, November 04, 1998

On a normal day economics will drive the generation and transmission mix in a region. Least cost units will be on-line and power transfer will maximized between utilities and control areas to move this "cheap" power from seller to buyer.

At Y2k rollover, contingency plans will dictate that: units are not scheduled to be down for annual maintenance(we have rescheduled our annual outage), transmission lines will not be scheduled down for annual maintenance, scheduled power transfers between control areas are kept to a minimum(expect MW costs to be higher while transfers are restricted), and key staff vacations are not scheduled during this period. All of this buys operating flexibility to react to the unscheduled outages that occur.

Security dispatch will come before economic dispatch during this period.

Jim

-- Anonymous, November 04, 1998

So does it make any sense at all to attempt to reduce our overall 'demand' going into this? My layperson thinking here was that by reducing demand we might increase 'reserve generating capacity', thus giving ourselves the ability to withstand more problems without widespread catastrophic failures -- i.e we give the grid(s) a bit more resilience.

I realize this is not THE solution, but could it play an important role? If so, what would it take to reduce demand by an amount sufficient to give us such a 'cushioning'?

If this is not a viable or practical approach, why not?

I understand that residential use is not the biggest portion of the 'customer pie' so to speak, but I could easily (and fairly quickly) reduce my demand for electricity in my own household by 50-75% or even more if it would help.

-Arnie

-- Anonymous, November 06, 1998