USS Clueless - Un-blacking out
     
     
 

Stardate 20030815.0131

(On Screen): So why is it going to take so long to bring everything back up again? If the power could go out in a wide area in 3 seconds, why will it take days to bring it all back up?

One reason is that they need to check everything to make sure it wasn't damaged when it went down. If a transmission line tripped its breakers because it was approaching overload, then the breakers are supposed to throw before anything is damaged – but you never know for sure, and it's better to check. (Expensive components are legendary for burning out first in order to protect fuses and breakers.)

Another reason is that this kind of failure takes place when the demand for power comes close to the supply. As you get close, the system becomes less robust, and a local surge in demand or an unexpected loss of generation capacity can set off what's known as a cascading failure where each time something overloads and shuts down it causes the next piece in turn to overload, sort of like a series of dominoes falling over.

Unfortunately, if the power's been off for a long time (fifteen minutes or more), then when the grid comes back on again it actually draws much more than normal for quite a while.

For the first few seconds there is a huge spike, caused by all the electric lights and everything with a motor turning on. Both incandescent and fluorescent lights draw more power as they come on than they do steady state. Electric motors also draw more power as they're spinning up. But that only takes a few seconds. However, power draw will remain a lot higher than normal for anything from half an hour to several hours.

The other effect is caused by everything which has a thermostat. In summer that means air conditioning; in winter it means electric heating. And year round it means freezers, refrigerators and electric water heaters. All of those are high-power devices, and every one of those has been drifting towards ambient temperature all the time the power was off. Once the power comes back on, every one of those sucks power continuously until it can get back to its temperature set point. And UPS's trying to recharge their batteries contribute to this, but that's a very small effect compared to refrigerators, which are usually the big villain since they're the most common of these devices. (Although this time of year in New England, air conditioning might well be bigger. It's tough to say: air conditioners are controlling larger volumes, but refrigerators have a lower temperature set point. And though air conditioners are common, every home has a refrigerator.)

Ordinarily a refrigerator doesn't run more than 10-30% of the time, and different ones run at different times. It's a random thing, but when you're talking about thousands of them it evens out statistically and the power needed by all of them collectively is pretty constant. But the steady-state power consumption is far below what is needed if they're all running at once, which is what happens when the grid first comes back up. Up to a certain point, the longer the power's been off, the longer it takes for everything with a thermostat to get happy again. (If the power's been off for two days, everything is pretty much at ambient, which is the worst case in terms of bringing the power back on.)

If, once they've checked everything, they were to turn a huge area (e.g. Manhattan) back on all at once, they'd overload their generating capacity and the whole system would shut itself back down again in self-defense. What they have to do is to bring the grid back up gradually, a piece at a time. After each piece comes back on, they wait until the load drops back to something near normal before bring the next lot back online. And since the sections they turn on later were off longer, it takes them longer to settle back to normal loading. (And that's also when they're again nearing the power generation capacity.) Which means that the last few sections take a lot longer.

Of course, if they had a lot of excess generation and power transmission capacity, then they could bring it all back up a lot faster. But if they had that, the system wouldn't have gone down in the first place.

Update: Benjamin writes:

As someone that plans part of the transmission grid in three states (CO, KS, & MO), I felt it was needed to point out items about power generation that you have left out. These items help explain why power takes a while to be restored to blackout areas.

First off is that most base load power plants are steam (either nuclear, coal, or gas). Once these plants have been knocked out of service, the fuel for providing the steam is shut off. This cools the plant down.

Most large base load plants (steam), also do not have the ability to start on their own (aka black start capability). These plants rely on the grid to excite their generators to get them to start. Consequently when the grid goes down, they can not start. The only way to get these generators excited is to provide power from a small generator (usually less that 20MW and diesel or gas).

While the fuel has been shut off the water providing steam has cooled. Once fuel is turned back on it takes time for the temperature to rise to provide steam to turn the turbine. To make sure enough steam is available to the turbine while enough power is available to excite the generator is complicated.

All of the major power companies in an area work together from a p

Captured by MemoWeb from http://denbeste.nu/cd_log_entries/2003/08/Un-blackingout.shtml on 9/16/2004