Выбрать главу
The TMI-1 reactor system was typical of pressurized-water reactors used all over the world, but some oddities of the B&W design contributed to a disastrous breakdown. The complete lack of water-level instrumentation in the reactor vessel was a big problem that the Nuclear Regulatory Commission would make illegal after the accident.

The TMI-2 reactor had been originally contracted for the Oyster Creek Nuclear Generating Station in New Jersey, supplementing a BWR brought online by General Electric back in 1969, but the Jersey craft-labor corruption was starting to get out of hand. Jim Neely, the negotiator for Jersey Central Power and Light, had been dealing with the mob ever since a worker made a point by dropping a wrench into the gearbox of the crane hoist while they were lifting the 700-ton reactor vessel into place for Unit 1 at the plant. That was alarmingly close to a disaster. Now a mob representative wanted one percent of the construction budget for the new B&W unit to ensure peace among the workers. That would be $7 million for one individual, and he was only the first in line. It was just not worth it to build a nuclear plant in New Jersey anymore, and Neely gave up. He amended the license application, and gladly transferred the contract to Met Ed, Pennsylvania. Maybe they would have more luck with it. TMI-2 was constructed without incident and began delivering power on December 30, 1978.

On March 16, 1979, a movie, The China Syndrome, opened in theaters nationwide. It was a fanciful cautionary tale about a potential nuclear power accident that could spread deadly radiation covering an area “the size of Pennsylvania.”[232]

It was March 28, 1979, heading toward 4:00 a.m., and the Shift Supervisor on duty was Bill Zewe. Zewe was 33 years old and had learned the nuclear business in the Navy. The previous shift had left him a problem. The steam that runs the turbine is turned back into water by the condenser beneath the turbine deck, dumping the excess heat to the twin, iconic cooling towers out back. The towers are each 30 stories tall, and they cool a million gallons of water per minute, making white, fluffy clouds rise into the air above.

The water out of the condenser must be “polished” before it returns to the delicate, expensive steam generators, removing anything that may have dissolved in it as it cycled through the pipes, valves, steam generator, pumps, and condenser. A bit of rust, for example, could have been picked up along the way, but the water is made sparkling pure as it is pumped through a line of eight 2,500-gallon tanks in the basement of the turbine building en route to the steam generators.

Each tank is filled with tiny balls of purifying resin, and they must be flushed out and replaced as they become contaminated or loaded with gunk out of the condensed water. Unfortunately, the resin beads tend to mash down and stick together, reminiscent of the problem that blew up on the Atomic Man back at Hanford, and the back-flushing system installed by B&W was underperforming. In Tank 7, the resin was stuck tight and was not moving. Zewe left two men working on the problem and climbed the eight flights of stairs to the control room. He asked Fred Schiemann, the foreman, a Navy man, to go down there and encourage them. As he left the control room, Schiemann gently reminded Zewe that the PORV was leaking, which was nothing new. In the 177FA design, B&W had replaced the troublesome Crosby PORV with a Dresser 31533VX30. In terms of reliability, it was proving no better than the Crosby. There was nothing particularly fatal about a little bit of leakage out the top of the pressurizer, but it was a pain, having to readjust everything as steam slowly escaped the primary coolant system and blew off into the containment building. It was just an irritant, and nothing more. It would be on the list of things to be corrected during the first refueling shutdown, which was not scheduled for two more years.

The men who had built this plant were an industrious, creative lot, and when they found that the resin beads could not be flushed using the factory-designed system, they added a compressed air line from the general-purpose compressed-air system in the plant. The air pipes were about the size of a garden hose. You could just open a valve, and the air bubbles would stir up the beads in the tanks and break them loose from sticking. There was not quite enough air in the system, so they cross-connected it to the instrument compressed-air system, which could then be used to open and close valves remotely, using switches in the control room. But ceasing to manually check those valves would be a problem, as somebody on the previous shift had air-flushed the tanks, but had forgotten to close the air valve. The one-way check valve in the air line was leaking, so for the past 10 hours, pressure from the 5,000 tons of water per hour running through the tanks had forced water up the instrument air line, almost to the point where it would cut off the air going to the valves on top of the eight tanks, which would slam them all shut at once and stop the flow through the steam system. There was an electrical backup system that would prevent such an improbable, almost impossible catastrophe, but it had not been wired up. The valves were supposed to be left open while the steam was running, but you could call up the control room and ask an operator to close the inlet valve on one of the eight tanks. Tanks could thus be cleaned out one at a time as the plant ran at full power.

Schiemann, down at the resin tanks, tried to assess the situation. They had not been able to dislodge the beads, and they had tried everything. A water flush, compressed air turned up all the way, and even steam had been unleashed on Tank 7. Schiemann climbed on top of the enormous water pipe so he could watch the sight glass and see the level of water in the tank. It was hard to see in the dim light. It was 3:58 a.m., and suddenly there was an awful quiet in the normally rumbling water pipe. Uh-oh. The water had backed up in the air pipe just enough to close all the valves atop the resin tanks.

He could feel it under his feet, a water hammer, caused by the sudden perturbation in the steam system, coming down the pipe, hot and fast, like a ballistic missile. He leaped free, just as the pipe jumped out of its mounts and ripped out the valve controls along the walls. Scalding hot water blew out into the room as the pump at the end of the pipe flew apart.

Back in the control room, every alarm tile on panel number 15 came lit up at once, and the warble-horns started going off. The turbine, sensing that it was not going to get any more steam, threw itself off line, and the reactor followed eight seconds later with an automatic scram, ramming all the neutron-absorbing controls deep into the core. The main safety valves in the secondary loop opened and blew the excess steam skyward. It sounded like the building was coming apart. The floor in the control room trembled, as the four main feed-water pumps shut down. Pressure in the reactor vessel, now denied its two primary cooling loops, rose sharply, and in three seconds the PORV opened automatically, blowing extremely hot water and steam into the drain tank on the containment building floor. On the control console a red light came on, indicating that the PORV had received an OPEN signal. Ten seconds later, a green light came on, indicating that the PORV had received the CLOSE signal. The sharp spike in the primary loop pressure had quickly dropped below 1,800 pounds per square inch, so there was no longer a need for an opening in the normally closed cooling system.

The senior men, Zewe, Faust, and the operator Ed Frederick, had seen this before, and knew it was a turbine trip. Regardless of the blinking lights and the pulsating horn blowing in their ears, it was nothing to get panicky about, and all the systems were acting correctly.

This feeling of tense calm lasted about two minutes, when the two high-pressure injection (HPI) pumps, a main part of the ECCS systems, switched on automatically. Now this was something new to Zewe, Faust, and Frederick. Why did the reactor think it needed emergency cooling? The temperature in the reactor was too high for this lockdown situation, and the pressure was too low. A minute later, Schiemann made it to the control room, gasping for breath after having sprinted up the staircase.

вернуться

232

The film starred Jane Fonda as a television reporter, Michael Douglas as her camera operator, and Jack Lemmon as the shift supervisor at a fictitious nuclear power plant in California. It was inspired by the Browns Ferry fire, which had occurred two years earlier, and it was a fair assessment of the growing public angst over nuclear power. Although there was a great deal of built-up worry and end-of-the-world dread in the movie, nothing melted, there was no radiation escape, and the reactor safety systems basically acted as they were supposed to. The only harm to come to any of the characters in the movie was by sniper rifle and motor vehicle. It was thus a parody of a typical commercial nuclear plant accident.