On September 8, 2011 the electrical grid in and around San Diego, California experienced a blackout that lasted for more than 12 hours. By some accounts more than 5 million people were effected. The initiating event was a human error that caused a large transmission line from Arizona to turn off unexpectedly. I recently discussed why a single failure as occurred that day should not have caused such a widespread grid failure, and how New York City will be much more susceptible to similar events if Indian Point Nuclear Plant is shutdown prematurely.
As it was designed to do, the San Onofre nuclear plant automatically disconnected itself from the grid and shut down then the blackout occurred. This was done as part of the plant’s protective scheme to shield the plant from unintended consequences from the falling grid voltage and frequency. A similar thing happened to nine nuclear plants in the eastern USA during the blackout of 2003.
Why do nuclear plants trip off line when a blackout happens?
While this is a somewhat simplified answer, it covers the fundamentals. Please be aware my experience is with pressurized water reactors, but the same basic principles should apply to boiling water reactors.
The nuclear plant’s generator, like that of any electrical generator supplying the grid, is electrically locked to the voltage and frequency of the grid. As grid voltage drops, so does the voltage sensed inside the plant. Most large electric loads inside nuclear plants are electric motors on pumps, valves, fans, and other such equipment. To drive a fixed mechanical load connected to the shaft, a motor must draw a fixed amount of power from the power line. The amount of power the motor draws is roughly related to the voltage times current (amps). Thus, when voltage gets low, the current must get higher to provide the same amount of power. Thus, as voltage drops, current inside the motors rises. This increase in current can cause overheating and short circuits.
Note: the paragraph above was revised to correct an oversimplification & error in my original post. The results are the same, my explanation was lacking.
Also, normally the alternating current on the grid operates at 60 cycles per second (60 hertz). As the grid collapses, the frequency begins to drop. If allowed to continue this would cause the nuclear plant’s reactor coolant pumps to run slower, thus moving less water through the reactor. Less cooling water could potentially lead to higher than normal fuel temperatures. To protect against the reactor operating with degraded cooling water flow, nuclear plants have various means of sensing low grid frequency or coolant flow. When electrical frequency or reactor cooling flow drops below a defined threshold it triggers an automatic shut down. Some of these protection schemes are anticipatory in nature – they happen predicatively before the grid situation has a chance to deteriorate to the point of causing a challenge to the reactor or plant equipment.
Why can’t nuclear plants stay on line when a black out happens?
While it’s possible to design a nuclear plant to be able to stay online during a loss of off-site power, it would require some large and expensive equipment, and a redesign of the reactor protection system.
The loss of electrical power to equipment inside the plant is not the only aspect of a loss of off-site power (LOOP) that designers have to consider. Another significant challenge is designing mechanical and control systems to withstand an instantaneous loss of load from 100% power to around 10% power. The reactor is putting out 100% power one instant, and the next instant the “grid” is gone and the only load on the rector is in-house loads. Since reactors can not change load that quickly, the reactor will be generating excess heat until reactor power can drop to balance with the new load. While reactor power is greater than the load there is excess heat being generated. That heat has to go somewhere; it causes the water in the reactor coolant system to heat up and to expand. Thus, to accommodate a 100% loss of load a nuclear plant needs a reactor coolant system with a large surge volume to accept that expanding water, and a large heat dump system to reject the extra heat. Both of these attributes can be designed into a reactor system – I personally operated a prototype naval reactor that was designed to accommodate a near instantaneous 100% load rejection. However, in a land based power plant the extra system hardware would be costly. Since base load power plants are not expected to withstand a loss of grid transient often, it is tough to justify the extra expense.
It’s possible that some of the new small modular reactors could be designed to stay on line during a LOOP. Perhaps some of my SMR friends will add some comments to this post below?