SPECIAL SQUAD TO SURVEY SALEMS SAGA

The Nuclear Regulatory Commission (NRC) announced on October 21, 2008, that it was

dispatching a special inspection team to look into “an apparent loss of reactor coolant system

inventory control” involving the Unit 1 reactor at the Salem Generating Station in southwestern

New Jersey the prior week. Operators had shut down the unit for refueling and were removing

water from the reactor’s cooling system to allow inspections and maintenance of equipment.

They did not realize that the instrumentation used to monitor the water level had malfunctioned

with the actual water level much lower than the level shown on the faulty instruments. Luckily,

the problem was discovered before it led to inadequate reactor core cooling and resulting fuel

damage. The operators added water back to the cooling system to restore the level to the proper

range.

Salem Unit 1 has a pressurized water reactor (PWR). When the unit is operating, the heat

generated by fissioning atoms in the reactor core warms water in the primary loop to nearly

550°F. Because this water is maintained at high pressure (about 2,200 pounds per square inch,

nearly 100 time the pressure of a car’s tire), it does not boil. Instead, this hot water flows inside

U-shaped tubes within the steam generator. Heat passes through the thin metal tube walls to boil

lower pressure water outside the tubes. The steam moves through the turbine, which is connected

to the generator producing the electricity that is the unit’s raison de etre. During reactor

operation, the pressurizer is partially filled with water. As implied by its name, the pressurizer

controls pressure. Electric coils in the bottom of the pressurizer can be turned on to raise the

pressure of the primary loop water. Sprays of cold water into the pressurizer’s upper region can

be turned on to lower the pressure.

October 22, 2008

Page 2 of 3

When the unit is shut down for refueling, water is drained from the pressurizer. Part of the

refueling outage steps include lowering the water level in the primary loop to what is termed

“mid-loop,” which is roughly when the piping from the reactor vessel’s nozzles partially empties

of water. This step is risky. Figure 2 is from NRC Information Notice 2000-13.* The horizontal

axis shows the day during a typical refueling outage. The vertical axis shows the risk of reactor

meltdown; lowest at the bottom and increasing towards the top. The two highest risk points

during refueling occur on days 6 and 20 when the primary loop (also called the reactor coolant

system or RCS) is drained to mid-loop. This NRC information notice also states that the risk

from 22 hours of mid-loop operation approximately equals the risk from operating the reactor at

full power for 33 days.

Why is the risk of reactor meltdown so high when the unit is shut down? Although the nuclear

chain reaction is terminated when the unit shuts down, the radioactive decay of fission

byproducts continues to generate large amounts of heat. In fact, right after the unit is shut down

and all control rods are inserted to stop the nuclear chain reaction, the heat from radioactive

decay equals nearly seven percent of the heat produced by the reactor operating at full power.

Two factors combine to make the mid-loop risk so high. First, the inventory of water in the

primary loop is intentionally lowered to its minimum amount. As a result, if something were to

happen, the margin to draining or boiling away this inventory and causing reactor core

overheating is very small. In addition, federal regulations require many safety systems to be

available when the reactor is operating. Should something happen, all of these safety system

must fail to cause a reactor meltdown. During mid-loop operation, the majority of safety systems

are disabled for testing, inspections, and maintenance. If something were to happen, it takes very

little else going wrong for reactor meltdown to occur.

* Available online at http://www.nrc. gov/reading- rm/doc-collectio ns/gen-comm/ info-notices/ 2000/in00013. html

October 22, 2008

Page 3 of 3

Salem’s saga is not the first water level instrumentation problem during refueling. For example:

Beaver Valley Unit 1 (PA): On March 26, 1981, leakage from tygon tubing used for

temporary water level monitoring instrumentation caused the indicated level to be

significantly higher than the actual level. As a result, the water level dropped low enough

to interrupt the flow through the reactor’s cooling system. In the hour it took the

operators to re-fill the system and restart the cooling system, the water temperature

heated up over 66°F.

Vogtle Unit 1 (GA): On October 26, 1991, the water level monitoring instrumentation

caused the indicated level to be significantly higher than the actual level. As a result, the

water level dropped low enough to cause the reactor cooling system pump to cavitate as it

tried moving water and entrained air.

Prairie Island Unit 2 (MN): On February 20, 1992, the water level monitoring

instrumentation caused the indicated level to be significantly higher than the actual level.

As a result, the water level dropped low enough to interrupt the flow through the reactor’s

cooling system. It took the operators 21 minutes to re-fill the system and restart the

cooling system.

Sequoyah Unit 1 (TN): On March 23, 1997, the water level monitoring instrumentation

caused the indicated level to be significantly higher than the actual level.

NRC Regional Administrator Sam Collins said the special inspection team will examine how this

near-miss at Salem occurred and “the actions the company has taken to prevent similar issues in

the future.” This last task will be ridiculously easy because this event has happened so many

times in the past. The true lesson from past events is in how to avoid them, not in how to repeat

them. Salem apparently learned the latter. Lightning may strike only once, but stupidity strikes

like a jackhammer. Whatever the NRC’s special inspection team finds, they need to exorcise

stupidity from all PWR sites before disaster occurs. It is foolish for the NRC to allow each site to

wait until it tempts fate before taking steps to prevent another mid-loop episode. The NRC must

act to have all PWR owners cease and desist the poor practices that cause these mid-loop nearmisses

to happen again and again.

Prepared by: David Lochbaum

Director, Nuclear Safety Project