Читать книгу Meltdown - Yoichi Funabashi - Страница 10

2:52 P.M., MARCH 11. Let us turn back the clock an hour or so. Units 1 and 2 Central Control Room.

A loud voice reverberated.

“The IC is online.”

Unit 1 was equipped with two isolation condenser systems (IC) as an emergency reactor cooling mechanism. This was an appendix-like object attached only to the early types of light-water boiling water reactors (BWR), but in Japan it had been left in Unit 1.

High temperature steam generated in the reactor was cooled by passing through a pipe coil in the IC’s cooling water tank located on the fourth floor of the reactor building and condensed into a liquid, flowing back as water into the reactor. The fact that the IC had automatically started meant that steam could now be returned to the reactor as cool water.

After SCRAM took place, the cooling of the reactor core went smoothly. All the operators had to do now was keep switching the IC on and off to gradually lower the temperature of the reactor. The reason it was not left on was to avoid cooling the reactor too rapidly.

Keeping a close eye on the instrument indicating the temperature of the reactor, the operators kept opening and closing the valve of the IC using a lever. The reactor temperature began to fall slowly. Forty minutes after the reactor shutdown, the reactor temperature, which had been close to 300 degrees Celsius when operating, had dropped to about 180 degrees.

The operators were relieved.¹⁷

However, flooding by the tsunami changed everything. This was because it robbed the plant of both its AC and DC power.

The lever on the IC operation panel was designed to always return to a central position when released after operation. If the valve was open, a red light came on; if it was closed, a green light. Since they had to use the operation lever time and again, they determined its status by the different colored lights.

Those lights had disappeared.

Was the IC valve open or closed, which was it?

The operators could no longer remember. Was the IC working? What about it? Nobody knew.

We can’t operate the valve without the display light.

The operators were completely shaken up.¹⁸

4:44 P.M. Izawa, duty manager in the Central Control Room, was contacted from the Anti-Seismic Building.

“There’s steam coming out of the Pig’s Nose.”

“How much?”

“A puff.”

“A puff?”

The Pig’s Nose was the two exhaust pipes twenty meters high on the west wall of the Unit 1 Reactor Building. When the IC was working, part of its job was to discharge the steam generated from the IC outside.

Uncertain if the IC was working, Izawa had asked the ERC in the Anti-Seismic Building to check whether steam was coming out of the Pig’s Nose. He had heard stories from older operators that when the IC was activated, white steam would gush out of the Pig’s Nose.

A worker from the ERC’s power generation team had gone out into the parking lot of the Anti-Seismic Building and confirmed steam was coming out of the Pig’s Nose.

“The steam’s coming out hazily.” This was what he had reported.

The ERC interpreted this as “steam is coming out.” But Izawa was adamant that the IC was not working, because he could not hear the release of steam, which, he had learned in training, was a crucial marker of a working IC. Regardless of this insight, Yoshida and personnel in the ERC were so engrossed in starting the water injection that the IC stoppage failed to catch their attention and remained uncommunicated to senior managers throughout the evening.¹⁹

The ERC continued to be at the center of communication problems that further strained onsite relations. Tasked with micromanaging onsite conditions and operations, the ERC demanded that all information be passed on to them. Izawa was simply overwhelmed with the sheer volume of data he now had to communicate to the ERC while managing reactor operations at the same time. At one point, an ERC worker questioned Izawa over the phone on how much battery power was left in the control room. “I f****** told you already that we don’t have any batteries! I told you that we’ve lost electrical power! What’s wrong with you?” Izawa shouted.²⁰

4:56 P.M. The water level in the reactor fell down to 1.9 meters above the top of active fuel (TAF). Top of active fuel serves as the reference point for water-level readings in a reactor; it is the uppermost point in a fuel rod that contains uranium. A few minutes later, the water level gauge again was no longer visible. In the fifteen minutes that the water level gauge had been visible, this worked out to a sixty-centimeter drop.

At 5:15 p.m., in a videoconference connecting TEPCO’s Tokyo Head Office and the Anti-Seismic Building, the voice of the person in charge of the technical crew rang out.

“If the Unit 1 drawdown continues at this pace, one hour to TAF!”

It was a shocking prediction. The exposed active fuel would melt, releasing radioactive material.

Is the information about steam coming out of the Pig’s Nose a mistake?

Perhaps the IC isn’t functioning.

It was only natural to harbor such doubts. However, it had only just been reported by workers from the power generation team that steam was coming out of the Pig’s Nose. If steam was coming out, the IC was probably working. But they could not be sure. It was just a guess; it might be just barely working. That was the general conception. Yoshida shared this as well.²¹

No questions were asked about the status of the IC in the videoconference, and no links were made between the falling water level and the IC. Someone took the microphone and shouted, “Entering the Office Building is prohibited!”

Other officials interrupted in rapid succession.

“We can’t go to help because seawater has flooded in as far as the seaside bus stop!”

“Suspected fire behind Unit 4 in the light water tank. Smoke has risen about five meters!”

Information on the drop in the water level at Unit 1 had been buried in the avalanche of other reports flooding in.²²

Izawa sent operators to the field to check if the IC was working. Although he had received a report that steam was coming out of the Pig’s Nose, he still was not convinced. It was at this time that the information on the decrease in the Unit 1 water level came in. The only way they could confirm whether the IC was working or not was to verify it directly.

The two ICs, A System and B System, were lined up together on the fourth floor of the reactor building. They decided to examine the water gauge attached to the side of the condensate tank to check the amount of cooling water to see whether it was sufficient. The door of the reactor building was a double door.

5:50 P.M. When the operators tried to open the door, the needle on their dosimeter went off the scale. It was over the 2.5 microsieverts (0.0025 millisieverts) maximum value. They had never seen this high a dose measured outside the double doors. Giving up on going inside, they returned to the Central Control Room.²³

6:18 P.M. The operators had gathered in front of the control panel in the Central Control Room. The light indicating the status of the IC valve in Unit 1 was dimly lit up. For some reason, some of the batteries flooded by seawater had recovered and some of the instrumentation and lights were once again visible.

“Green.”

“Closed.”

The IC light was green. The valve was closed. If the valve in the IC piping was closed, that meant that the steam had stopped flowing and the IC was not moving.

“It’s not moving. Can we start it up?” Izawa instructed the operator in charge to open the valve using the control panel lever.

“The valve is open. IsoCon startup confirmed.”

“Roger that. The time is 18:18.”

The operators here referred to the isolation condenser (IC) as IsoCon.

The light had changed from green to red. The IC to cool the Unit 1 reactor had at last started. Approximately two and a half hours had passed since the total power outage at 3:37 p.m.

Izawa ordered the operators to go outside and confirm whether steam was coming out of the Pig’s Nose. The operators who had gone out to check came back in a hurry.

“It was billowing out at first, but started to peter out, then disappeared.”

Apparently, the cooling water level in the IC tank had dropped and was not generating much steam. This was Izawa’s assessment. When the cooling water in the tank ran out, it would fall into a boil-dry state, incurring a risk of damaging the IC piping and releasing radioactive material outside.

“Do you want to continue running the IC?”

After thinking a little, Izawa said:

“It can’t be helped. Let’s stop the IC. For now, let’s close the 3A valve.”

6:25 P.M. Izawa ordered the IC valve closed. The control panel light changed from red to green. After a mere seven minutes, the IC had been stopped again.²⁴

However, the fact that they had closed the IC valve was not precisely transmitted to the ERC in the Anti-Seismic Building. From Site Superintendent Masao Yoshida down to senior managers in the ERC, even after this, they still believed that the IC was working and continued to respond to the emergency on that basis.

Why was there such a miscalculation? Why were mistakes made? What was happening onsite?

According to an investigation conducted by Niigata Prefecture, the operators and deputy director in charge of operating the IC in the Central Control Room understood that “the IC was not working.” After all, electrical supply had been lost; they firmly believed they had closed the valves themselves; and with data up until that point lost from the power outage, their memory was the only gauge they could use.

In fact, the IC had not been working at that time, and their understanding was correct. However, when Izawa asked the onsite director about the condition of the IC, the answer was, “I don’t know.” The operator and the deputy director’s understanding that the IC was not working had not been conveyed to the onsite director. Had Izawa asked the operator of the main reactor directly, he would have been aware of their assumptions on the condition of the IC. This was the first misunderstanding.

Later on, a second misunderstanding emerged as a result of this communication breakdown between the Central Control Room, the Anti-Seismic Building, and the Emergency Response Headquarters. Although Izawa made sure to constantly report to the power generation team leader in the Anti-Seismic Building, information on the condition of the IC had failed to reach Izawa himself, and therefore the Anti-Seismic Building, too. The power generation team leader continued to believe that “since we lost all power supply while the IC was running, it should still be working.” In a chain of miscommunication, the fact that the IC had stopped working never reached Yoshida.

By 6:18 p.m., some of the power in the Central Control Room had been restored. The light showing the condition of the IC valves now indicated “closed.” Izawa promptly reported this to the Anti-Seismic Building and began to operate the IC while the onsite director opened the valves. However, even this turn of events had been miscommunicated. The report was vague and simply indicated that “the IC is working” rather than communicating that it was working because “a valve had been opened that was closed before.”

At 6:25 p.m., the onsite director judged that the effect of the IC had been limited and shut down the IC, but even this operation failed to reach the Anti-Seismic Building. By eleven p.m., however, the radiation dose in the turbine building of Unit 1 had risen, and the pressure on the storage container now exceeded the maximum that it was designed to endure. It was at this point that Yoshida became suspicious and realized that the IC might not actually be working. “The IC is working, the water level has a surplus, but something is strange. I started to realize that the situation inside is probably terrible,” Yoshida recalled to the Government Investigation Commission.

The misunderstanding on whether the IC was working or not also appeared in TEPCO’s official announcement. During the late evening hours of March 11 and early morning of March 12, TEPCO Head Office announced that they were “cooling the steam inside the nuclear reactor using the IC.” But it was only after four a.m. on March 12 that TEPCO revealed that the IC had stopped working.

In domino-style, miscommunication had occurred at three different levels. First, between the operator in the Central Control Room and the onsite director; second, between the Central Control Room and the Anti-Seismic Building; and third, between the team leader and the director of the Anti-Seismic Building. This produced “a chain of misunderstanding.” On this point, Yoshida later testified the following:

“As far as the IC is concerned, there was a certain water level being shown, so I was under the impression that the IC was still working.… That was because no mechanism had been set up for the shift supervisor to call me at that time. I know I should have made sure many times at that point whether the IC was really functional or not.… I’ve been intensely reflecting upon the fact that I didn’t question that then, but at that time an SOS signal from the site had not reached me.”²⁵

To make matters worse, those at the Anti-Seismic Building, including Yoshida, and operators in the Central Control Room, including Izawa, did not accurately understand the IC itself. In theory, when power supply is lost, the system to detect failure in devices stops functioning and the IC automatically closes the valves. This is what is called “fail-safe function.” If a common understanding of this function had existed, the misunderstanding that “the IC is still working” could have been avoided. Yoshida testified to the Government Investigation Commission on this point that “no one besides the onsite operators at Units 1 and 2 knew about it … the IC is an extremely unusual system, so, to be honest, I don’t know myself.”²⁶

There was not a single operator onsite who had experience in actually operating the IC, including the operators at Units 1 and 2. According to a 2015 report that TEPCO submitted to Niigata Prefecture, the last time the IC had been operated at Fukushima Daiichi was on June 29, 1992.

However, it was not only the status of the IC that the ERC was wrong about. The water level gauge also fooled Yoshida. He was concerned that the water in the IC tank, which was the only means of cooling, would be gone, but it turned out that even after ten hours, it still contained water. After being stopped for three hours, the IC was restarted. However, at this point, the meltdown had already progressed to a state where the IC could no longer be cooled.²⁷