As mentioned in Chapter 3, the Nuclear Safety Commission, in May 1992, set out “Accident Management: Measures against Severe Accidents at Light Water Nuclear Reactor Facilities for Power Generation” [1] and recommended that utility companies promote their own accident management. MITI also guided the utility companies to promote accident management as their own project.

Given these directions, TEPCO took about 10 years to establish its accident management for Fukushima-1, Fukushima-2, and Kashiwazaki-Kariwa and submitted their “Accident Management Development Report” [2] and “Accident Management Development Effectiveness Evaluation Report” [3] in May 2002 to METI (MITI up to January 5, 2001).

The accident management development by TEPCO at this time included the following four arrangements: (i) developing accident management measures like water injection functions to the reactor core and Containment Vessel (CV), (ii) assigning groups and organizational structures for actual accident management, (iii) building procedural documents like operational procedures at times of accidents and an accident management guide, and (iv) education of employees assigned to accident management groups.

As noted in the previous chapter, causes of severe accidents were mainly limited to internal factors, and thus, accident management against external factors like natural disasters was not sufficiently covered in the scope of the accident management described here.

TEPCO settled with the accident management completed in 2002, and from then on, it concentrated on “lateral dissemination” of countermeasures learned from accidents and findings at domestic and overseas nuclear reactor facilities. For example, after the July 2007 Chuetsu Offshore Earthquake, Kashiwazaki-Kariwa suffered damage to the office building and a fire in the transformer area, and the “lateral dissemination” led to the stationing of two chemical fire engines and a tank-mounted fire truck at Fukushima-1 by February 2008. Also in 2010, several new fire reservoirs were placed, water inlets were newly installed on the turbine building for each unit to reach the cooling systems, and the emergency response headquarters were transferred to the seismic isolated building from the administration area of the main building.

The seismic isolated building was equipped with emergency power generators and protection against radiation. That building played an important role in this emergency response at Fukushima-1. If the seismic isolated building had not been available, accident response at Fukushima-1 would have been much more difficult, and this point was commendable.

As seen earlier, TEPCO carried out lateral dissemination from 2003 on, but did not push for stronger accident management, including measures against external factors.

Japan is an earthquake-prone country. Natural disasters like earthquakes and tsunamis have to be considered as potential causes of severe accidents. TEPCO, however, did not consider ways of preventing damage to the reactor core in the case of severe accidents caused by natural disasters that exceed the design assumptions. TEPCO’s concept of preparation against natural disasters was to assume certain levels of natural disasters like earthquakes or tsunami and that it was sufficient to design the nuclear reactor facilities in accordance with the Guidelines for Safety Design Reviews and Guidelines for Aseismic Design Review by the Nuclear Safety Commission. TEPCO (as well as other utility companies in Japan) did not prepare adequate countermeasures against severe accidents with its existing nuclear reactor facilities. Instead, they ran a series of aseismic back-checks (evaluation of aseismic safety) to see if the facilities could withstand natural disasters of the set magnitudes, and if locations with insufficient resistance were found, they were modified as countermeasures.

The Fukushima NPP accident suffered severe damage to three nuclear reactor cores at one time. Submergence of the facilities led to station blackout (SBO), for which the station was unprepared. There was an underestimation of risk with regard to exceptional external events. TEPCO thus left the preparations against severe accidents in an incomplete state.

4.2.2 TEPCO's measures for accident management

What accident management measures were actually in place at Fukushima-1? We will review its (i) measures against loss of power, (ii) preparation of fresh- and seawater injection from fire engines, and (iii) communication tools in emergency situations before the accident took place.

TEPCO, following Article 10, Paragraph 1 of the “Act on Special Measures Concerning Nuclear Emergency Preparedness,” established a rule that when an event specified in the paragraph occurred, and the nuclear emergency preparedness manager or the plant manager declared the first degree state of emergency, applicable nuclear power plants and the central headquarters should establish their emergency response headquarters.

At 15:42 on March 11th, Fukushima-1 Plant Manager Yoshida judged that an event specified in the “Act on Special Measures Concerning Nuclear Emergency Preparedness (Nuclear Emergency Preparedness Act),” Article 10, Paragraph 1 had taken place and notified applicable ministries and agents, local governments, and the TEPCO main office. The notification was responded to by settings of emergency response headquarters in the TEPCO main office and at Fukushima-1. At that time, the emergency disaster response headquarters had already been placed an hour earlier in the seismic isolation building with the Great East Japan Earthquake attack (Figure 4.1). The emergency response headquarters in Fukushima-1 merged the one for earthquake response within to lead the overall disaster response management.

Figure 4.2 shows the relationship between the local plant headquarters and the main office headquarters. The local Plant Emergency Response Headquarters were at the frontline of emergency responses, and the Main Office Emergency Response Headquarters supported the local headquarters by confirming and giving authorization for important procedures.

Emergency responses in Fukushima-1 NPP were carried out while all lights were lost and the rubble of equipment and facilities were scattered everywhere except in the seismic isolated building. In the late afternoon on March 11th, to worsen the situation, damage to the nuclear reactor had started to release radiation near the reactor. Field workers of TEPCO and its contractors had to work in pitch darkness with high levels of radiation. The hardship they had to tolerate is beyond what we can imagine.

Although the circumstances of the work area were difficult, TEPCO’s response to the emergency showed weaknesses. In the following paragraphs we will discuss the failure to correctly assess the operational status of the Unit 1 Isolation Condenser (IC) and weaknesses in handling Unit 3 alternative water injection.

4.4.2 Misjudgment of Unit 1 isolation condenser (IC) status

IC, as discussed in detail in Chapter 2, is an emergency cooling system that cools the Reactor Pressure Vessel (RPV) steam in the condenser tank and feeds the condensed water back to the RPV by gravity without the use of a pump. Its continuous operation cools the reactor core. Fukushima-1 had two IC systems.

If these ICs had operated normally, as we pointed out in Chapter 2, the core damage process of Unit 1 would have been different. The ICs, in reality, were not operating properly, and the Emergency Response Headquarters had mistakenly judged they were in proper operation.

The biggest reason for the misjudgment was, including the main office, TEPCO engineers’ lack of knowledge about the basic operation of ICs. These IC systems were designed so their logic closes the IC isolation valves immediately after loss of DC power for a fail-safe purpose. None of the staff at the plant or the main office emergency response headquarters realized this.

Between 16:42 and about 16:56 on March 11th, there were symptoms of trouble with the ICs such as lowered reactor water level or high levels of radiation near the Unit 1 reactor building. The engineers should have realized the high likelihood of IC trouble with these phenomena. Hardly anyone at the response headquarters realized the trouble, and thus proper instructions were not given and the right actions were not taken in the field.

The operator in charge at 18:18 on the same day began to be concerned that the IC isolation valves might not be responding when he tried to open valves 2A and 3A after the control panel display was recovered. He then consulted with the plant response headquarters, but because they did not understand the IC basics, headquarters did not change their view on the IC operation.

The hearing by our investigation team of the Fukushima-1 engineers found that no one in the plant had recent experience with IC operation, and the operators had only heard of such experience by word of mouth. Testimony during the hearing claimed training had been given about functions and operations of the IC, but none in the case of loss of DC power.

The plant response headquarters was at the front end of accident response, and the main office response headquarters' role was to support the plant. No one in these two headquarters had a complete understanding of the IC mechanism and function. The operators in charge were not familiar with its operation, either.

Cooling the reactor to prevent core damage must have the highest priority at the time of an accident. In the authors' opinion a company with such poor understanding of the functions and handling of IC, that was most relied on at a time of an accident, lacked the qualifications to run a nuclear power plant.

The functional failure of IC called for a quick alternative water injection. This injection could not have been made without lowering the RPV pressure. At about 0:00 on March 12th, instructions were given to prepare to drop the pressure; however, the work did not start until about 14:00 on the same day.

A long delay in dropping the RPV pressure and injecting water meant a delay in the core cooling. The misjudgment about the IC operation was the biggest cause for this delay.

SBO is a seriously severe situation, and no matter what the circumstances were, measures to cool the core had the highest priority, However, the plant and main office response headquarters failed to recognize the IC operation status and thus, did not rush in creating a method of alternative water injection. In addition, instructions to vent the CV were also delayed. After all, misjudgment about the IC operation triggered the chain of delays in responding to the Unit 1 accident.

Core cooling of Unit 1 failed in the way described above, and in the afternoon of March 12th, a hydrogen explosion blew up the top floor of the reactor building. The explosion broke the temporary cables that had been laid out for electric recovery of Units 2 and 3, further delaying the work on these two other reactors. The misjudgment of IC operation, thus, not only affected the Unit 1 accident response, but also caused a negative impact on the responses for Units 2 and 3.

4.4.3 Mishandling of Unit 3 alternative water injection

After the Unit 1 reactor building exploded at about 15:36 on March 12th, continuous core cooling of each unit was recognized as the most critical process. The mishandling of alternative water injection for Unit 3 happened under such circumstances.

When water injection with a certain method fails, it is absolutely necessary to immediately switch to another method without any delay. At about 02:42 on March 13th, without an alternative method prepared, operators and others in the field manually turned off the high pressure core injection (HPCI).

Furthermore, a measure against discharged batteries had not been in place, the drained batteries could not restart the HPCI, and the pressure reduction for alternative water injection failed as well. Water injection to the reactor core was, thus, suspended for over 6 hours. To make matters worse, the trouble was not immediately reported to the senior staff. Alternative water injection did not start until about 09:25 on the same day, and the Unit 3 core was damaged.

The decision to stop the HPCI only with the Unit 3 staff at the time and the plant response headquarters power generation team without requesting authorization from a senior member of staff was questionable in terms of a response in an emergency situation. In addition, if the information about manual stoppage of HPCI had been fully shared within the plant response headquarters, there might had been a quicker remedy to the operator’s misjudgment of manually stopping the HPCI without an alternative method for water injection.

4.4.4 Consequences of the inaccurate assessment of Units 1 and 3

As Chapter 2 discussed in detail, hydrogen explosions took place in the reactor building of Unit 1 at about 15:36 on March 12th, and in Unit 3 at about 11:01 on March 14th. The two buildings suffered great damage. These explosions were caused by damage with the fuel rods inside the reactor core. The Zircaloy that covered the nuclear fuel reacted with water to produce hydrogen. The hydrogen then made its way out from the RPV, through the CV to accumulate inside the reactor building.

The question is if the response headquarters had properly acknowledged the status of Units 1 and 3 and had dropped the pressure earlier to inject water, could the core damage have been prevented to avoid hydrogen generation and thus the explosions?

To precisely answer this extremely important question, further detailed investigation is needed about the core and possible water injection, and thus, finding an answer at this point is difficult. We can say at least that if earlier depressurization had been made for alternative water injection from fire engines, the core damage would have proceeded more slowly, and less radioactivity would have been released.

4.4.5 Faults by the local response headquarters and main office response headquarters

The May 2002 “Accident Management Development Report [2]” describes the basics of the Fukushima-1 emergency response. This report states, “For complex events, technical evaluation is highly important and vast information is required for understanding the accident status and deciding which accident management measure to take. The supporting organizations, therefore, perform these technical evaluations to support the decision making.” In other words, the teams for information, engineering, safety, recovery, and power generation were expected to gather all necessary information to carry out the technical evaluation and provide suggestions and instructions to the chief on duty.

Let us take the event we explained earlier to describe this idea. When the supporting organization received information about Unit 1 IC and intended to cool the core in isolation, it would properly evaluate the IC status and in case such information was not available would call the chief on duty to actively collect the information. Unfortunately, the reality proceeded differently, with the headquarters failing to perform such a role and without correcting the misunderstanding of the IC status.

The teams of the main office response headquarters were also expected to gather important information from teleconferences and so to evaluate the information from viewpoints different from those of the plant response headquarters and support the plant’s decision making. Each team in the main office did not provide effective advice to the plant response headquarters. Inability to provide effective advice and instructions when faced with the serious problem of delay in core cooling showed weaknesses in accident management for Fukushima-1.

When all AC power was lost, power supply from batteries was also headed toward drainage. When over a day had passed from the loss of all AC power before dawn on March 13th, engineers working on Fukushima-1 should have been concerned with the battery drainage for operating HPCI and the reactor core isolation cooling system (RCIC) of Unit 3. If they had, the plant response headquarters could have worked earlier on securing alternative water injection from fire engines instead of relying on HPCI running at the time but at risk of losing the function once it was stopped.

Also, before dawn on March 12th, with the rubble already removed, the fire engines that had been parked near Units 5 and 6 were available, and there was time for collecting batteries for activating SRVs for RPV depressurization.

The plant response headquarters at the time, however, had only discussed and prepared alternative water injection from the standby liquid control system, which is a mid- to long-term response, and did not start planning for alternative water injection using fire engines until hearing about the trouble of manually stopping the Unit 3 HPCI from the operator on duty. For Unit 2, Plant Manager Yoshida had given instructions to prepare alternative water injection around noon on March 13th before its RCIC stoppage; thus, if the reactor status was well understood, similar responses could have been taken for Unit 3 as well. The plant response headquarters did not recognize the urgent need to prepare alternative water injection for Unit 3.

It is understandable that the plant response headquarters had difficulty in facing an unprecedented emergency situation. The response headquarters in the main office was not effective in supporting the plant response headquarters.

TEPCO had been educating and training those in nuclear power generation to the level required by law before the Fukushima-1 NPP accident. The investigation committee interviewed a number of TEPCO employees, and those in the nuclear power department had ample knowledge of nuclear power technology at a level equivalent to plant manufacturers.

However, when we reviewed the responses by the employees to this Fukushima NPP accident, in many cases their knowledge was not used to its full extent. A typical example was the misjudging of the Unit 1 IC operational status, and a similar confusion can be noted about the reactor water level gauge.

During the accident, when the reactor water level gauge indication stood still for a long time, no one in the main office or on-site at Fukushima-1 voiced the possibility that the actual water level could be lower than the reactor side piping tap. Records show, however, that there was one worker who pointed out the possibility of a higher indication of the reactor water level but a lowered water level in the reference condensate pot. Nevertheless, no evaluation or discussion was made about such a possibility that could keep the reactor water level indication unchanged.

Also, the employees had plenty of knowledge about the containment atmospheric monitoring system (CAMS) and how to apply its readings to accident management; however, during and after the accident, no one tried to evaluate the plant status by estimating the RPV and CV integrity from the CAMS readings, and they simply followed the manual to calculate the core damage rate and reported the results to NISA.

It is understandable that the workers were under high tension in the middle of a difficult accident response; however, it can be argued that TEPCO showed real weakness in its capacity to respond to an emergency situation. This fact is due to the company not providing education and training to enhance such skills in coping with an emergency. The real cause was not the ability of each employee in the field.

When the authors further traced this problem of weakness in dealing with emergency situations, we concluded, finally, this was because companies such as TEPCO and even government officials underestimated the risk that a severe accident like a meltdown could happen with a Japanese nuclear power plant.

Employee qualities and competence in responding to severe accidents cannot be built in a matter of days, nor through classroom lectures. Ability to respond to accidents is gained not just from textbook knowledge, but also through the power of thinking of various possibilities from information at hand, making choices, judging what the best move is at the time, and executing the move.

TEPCO, as the first company in nuclear safety, should seriously review its education and training to enhance the quality and competence of every employee and contractor so they are ready to respond to accidents.

4.5.2 Problems with specialty-based sectionalism

TEPCO’s disaster prevention plans and accident management guide called for teams such as power generation, recovery, engineering, and so on in the emergency disaster response headquarters. The scheme aimed at a systematic and integrated response to nuclear disasters. Each of the teams, however, tried to cover its own responsibility, but lacked managerial skills to see the total situation, find its role within this overall picture, and carry out the necessary supporting actions.

The corporate culture, similar to other electricity companies, was that TEPCO employees would refer to each other according to functions associated with their job, such as “operating guy,” “safety guy,” “electrical guy,” or “mechanical guy.” A few could gain key experience in a wide range of areas to ascend to a position at the director level; however, most employees in the nuclear power department were “some guy” that stayed in that area for years. Such workers have ample knowledge in their own fields, but do not have sufficient information to function in other fields, even those closely related to their own.

Once an organization is formed with such resources, the field of view for each employee is narrowed, and even when the organization appears trouble free at normal times, once an emergency situation breaks out, the weakness will be exposed. The organization cannot get a total and cross-sectional view of the situation, fails to decide what is important in prioritizing its actions, and thus ends up with a delayed start in important actions.

For example, Plant Manager Yoshida gave the order to evaluate water injection with fire engines soon after the accident on March 11th; however, it was not a documented procedure. Each functional team or group did not recognize its responsibility, and no actual evaluation took place until before dawn on March 12th. This was a typical example of the weakness described here.

Another case was the opening of SRVs. When electricity is available, it simply requires turning the switch on the control panel in the central control room. However, with the power lost, the recovery team had to connect the terminals in the back of the control panel to batteries with a total power of DC 120 V. There was confusion about whether the operator on duty or the recovery team was opening the Unit 2 SRVs on the evening of March 14th. This is another example of the weakness of sectionalism.

4.5.3 Lack of education or training for severe situations

As mentioned earlier, one of the reasons for the functional teams in the plant or main office response headquarters not performing their roles was sectionalism. In addition, poor performance of the functional teams was due to TEPCO not providing sufficient education and training to deal with total loss of AC power with multiple reactors.

TEPCO’s accident manual does not discuss multiple-unit SCRAM followed by days of SBO. These manuals assumed that AC power would recover within hours or, at most, a day. There was no description of the process of AC power recovery either. These manuals, at a glance, appear fairly detailed; however, they were incomplete with critical oversight.

The report TEPCO wrote in 2002, “Accident Management Development Report” described the loss of AC power with the phrase, “In an event that all AC power is lost, the phenomena will progress slowly with plenty of time margin”; however, no clarification was provided on why the phenomena would progress slowly. Preparing such incomplete procedures and training the employees with them can only cope with smaller events such as a local power loss.

Training was also insufficient. For example Fukushima-1 had, in late February 2011, a simulated training exercise assuming Nuclear Emergency Preparedness Act Article 10 notification. The postulation was loss of external power at a reactor caused by an earthquake, followed by the transformer break and failure to start the emergency generator. It was a state of loss of AC power. An assumption with this training, however, was that the emergency diesel generators would recover, and the simulation was on how to prevent catastrophic events from happening before the recovery of the generators. Nothing in the training had the severity of this accident, and no consideration was given to submergence of the switchboards or loss of internal power.

It has been claimed that loss of almost all power in Fukushima-1 caused by the earthquake and tsunami was beyond reasonable expectation. However, effective emergency planning should have taken more account of extreme events, as well as testing more effectively employees' ability to respond to emergency situations.

4.5.4 Excessive reliance on contractors

So far, in Fukushima-1 NPP, subcontractors had operated fire engines and heavy machinery; however, there had been no agreement that they would be handling them at times of emergency or abnormal situations. Their contract did not even mention a situation with this level of accident, where there is a high possibility of exposure to radiation (i.e., even the agreement with contractors ignored the possibility of severe accidents).

In the evening of March 11th, after the tsunami had subsided, the Fukushima-1 plant area was covered with rubble and wreckage of facilities that blocked people and vehicles. The plant wanted to remove these obstacles with a backhoe; however, there were no operators on site, and TEPCO had to make an urgent request to its contractor to dispatch operators. The contractor could have refused the request saying that it was beyond the scope of the agreement, but following their past practice, they did send operators to the site.

The next episode also shows TEPCO’s heavy reliance on contractors: when the time came to inject water from fire engines, contractors like Nanmei Kosan had to run all the fire engines. None of the TEPCO employees knew how to operate them, and the water injection was delayed for that reason. Even though the hardware was ready, without knowing how to operate it, a quick initial response did not take place.

In the authors' opinion, TEPCO's excessive reliance on contractors had led to a situation where TEPCO employees lacked the hands-on skills to carry out the tasks for maintenance and emergency responses. The over-reliance of TEPCO on contractors and subsequent lack of experience and training of TEPCO employees has been highlighted in the investigation reports [12, 13].

4.5.5 Insufficient safety culture within TEPCO

The Fundamental Safety Principles by IAEA [11] states in its first principle that “The prime responsibility for safety must rest with the person or organization responsible for facilities and activities that give rise to radiation risks.” In other words, the prime responsibility for nuclear safety is assigned to individuals and companies in the nuclear business.

And principle 3 states, “Effective leadership and management for safety must be established and sustained in organizations concerned with, and facilities and activities that give rise to, radiation risks” to emphasize the importance of stationing the safety culture within the organization.

Principle 3 lists the following three points as part of safety culture:

– Individual and collective commitment to safety on the part of the leadership, the management and personnel at all levels

– Accountability of organizations and of individuals at all levels for safety

– Measures to encourage a questioning and learning attitude and to discourage complacency with regard to safety

TEPCO's preparation against nuclear disasters that could cause serious damage to the reactor core was insufficient, and it had not made satisfactory preparation against the risk of tsunami attack over the design standards. Its capacity in responding to emergency situations was weak, education and training for severe accidents were insufficient, and their accident plan had flaws.

A portable radiation dosimeter with an alarm is indispensable for dosage management for nuclear plant workers. Fukushima-1 had about 5,000 of them; however, most of them were broken by the tsunami. As a result, not all of the workers had dosimeters, and TEPCO had many workers work without them for about a month. On March 12th to early 13th, 500 dosimeters were shipped to Fukushima-1 from Kashiwazaki-Kariwa NPP. However, TEPCO employees did not correctly identify the shipment and make use of the dosimeters.

Our investigation committee invited five experts from overseas for an international peer review on February 24 to 25, 2012. Among the five guests, Mr. Lars-Erik Holm, Board Director, Swedish National Board of Health and Welfare, highlighted this problem. Mr. Holm criticised TEPCO for (1) having contractor employees work for several weeks after the accident without dosimeters, and (2) the fact that dosimeters had been on-site.

TEPCO also had problems in terms of communicating information to the public. For example, during the press conference at 19:00 on March 11th, it repeated the wrong information that both Units 1 and 2 were under cooling and gave out only limited information about detection and leakage of radioactivity, which was of most interest to the residents. TEPCO only publically admitted that there was core damage from the accident on May 12th, two months after the accident. These problems in communication undermined confidence in TEPCO and its response to the accident.