When Discussing Safety, Science Should Be at the Epicenter


Woody Epstein, Manager of Risk Consulting for Scandpower

On June 19, Japan’s nuclear regulatory authority (NRA) released new safety requirements for the restart of 48 idled nuclear reactors. The new safety regulations oblige nuclear operators to protect against more severe accidents than ever before, using the most extreme seismic occurrences as standards for earthquake damage prevention. Forum on Energy recently caught up with Woody Epstein, Manager of Risk Consulting for Scandpower in Japan, to get his thoughts on nuclear safety from a seismic risk point of view.

Forum on Energy: We understand that the Nuclear Regulatory Authority (NRA) is setting national safety regulations, including earthquake safety guidelines. Given the geologic and other variations at each power plant, how does this process work?

Woody Epstein: The NRA is trying to lay the conditions for restart. This is useful if the process is scientific and involves communication between the regulator and the utility.

Each utility files an application to restart the nuclear power plant (NPP). Just like in the United States, each submission is considered on its own merits. A utility can argue why some criteria should be “screened out,” but the reasoning must be included in the application. To be successful, I believe that cogent scientific arguments must be the focus of the conversation.

Forum on Energy:There is some discussion of different ways to assess risk, with some pushing for deterministic risk assessment and others advocating for a change to probabilistic risk assessment. Could you describe the difference between the two methods?

Epstein: Let us look at seismic risk, for example. Deterministic risk assessment involves looking at historical data and the ground motion which a structure must be able to withstand. That measurement is multiplied by a factor based on engineering judgment. So if the largest tsunami in that location’s history was nine meters tall, a sea wall might be constructed to be 18 meters tall. The problem with the deterministic method is that it disregards uncertainty. To disregard this is nonsense.

Probabilistic risk assessment uses the same types of data, but evaluates the probability of events recurring. We use probability to measure our uncertainty — the likelihood of a greater event happening in the future. In the case of the nine meter tsunami, the sea wall is built according to the likelihood and its uncertainty of a future 18 meter tsunami.

Forum on Energy: We heard a rumor that the NRA considered regulating that all NPPs must be built to withstand an earthquake with a ground motion speed of 1,000 galileos, the same as the 2004 earthquake in Rumoi. Is this true? If so, what would the consequences be? [Editor’s note: Peak Ground Acceleration or Design Basis Earthquake Ground Motion is measured in Galileo units – Gal (cm/sec2) or g – the force of gravity, one g being 980 Gal.]  

Epstein: You are talking about “floating earthquakes” — earthquakes that occur along a previously unknown or still undetermined fault line.

The NRA asked utilities to consider safety measures that address floating earthquakes and presented a list of 16 floating earthquakes between the years 1996 and 2013. But the design basis has not been increased to the 1,000 gal that you mentioned. The design basis ground motions do not seem to have changed much since 2006 in the current submittals for restart. 

There are many ways to measure the strength of an earthquake, and the data is available online. I took a look at the 16 floating earthquakes and discovered that neither peak ground acceleration, nor magnitude, nor cumulative absolute velocity alone are good measures of possible damage, which is the point of risk assessment.

For example, a recent earthquake in Nikko, Tochigi Prefecture on February 25, 2013 had strong ground motion which lasted about five seconds and had a peak ground acceleration (PGA) of 1,214 gal. But there was no damage to the buildings. You cannot predict damage in theory or on a lab’s shake table. You need to look at similar sites.

Magnitude is the strength of the earthquake at the epicenter. But it has no meaning elsewhere because it does not consider constraints such as local geology; soil structure and interactions; and tectonics. Geology is important. You need to study tectonically similar earthquakes around the world to understand what may happen at a particular site.

Another factor is length of time. A ten-second earthquake does not cause as much damage as a 300-second earthquake. The 3/11 Tohoku earthquake lasted 300 seconds.

The point is, you can’t just use one number, you must use many different measures to understand the potential for damage.  

Forum on Energy: You stated that scientific metrics should be used to guide the safety features at each NPP, and regulations should set a goal to prevent damage rather than to prescribe resistance to maximum levels of galileos or magnitude. In this spirit, let’s look at the next step: how would you describe the process of developing and implementing these regulations with sometimes competing interests and agendas on the table?

Epstein: Science needs to be at the forefront of a collegial discussion between the regulator and the utilities. At this point, I do not see a transparent scientific discussion on whether NPPs are safe. Rather, the conversation is focused on understanding how the NRA can be independent without being isolated. The utilities view the NRA as against them from the outset. The NRA believes it must be isolated from the utilities. Both groups must drop this adversarial stance and work for the safety of the Japanese people and the good of Japan.

In the United States, by comparison, the Nuclear Regulatory Commission (NRC) talks to utilities all the time and yet maintains its independence. The Long Term Seismic Program between the Diablo Canyon NPS and the NRC should be used as a model for utility/NRA relations in Japan. Independence is the foundation of American culture. This makes it possible to have a strong peer review system. On an individual level I take my peer review duties seriously, even if the work in question is by a good friend or classmate in school.

Forum on Energy: So how can one break the deadlock on communication between regulators and the utilities?

Epstein: Make science the centerpiece of the conversation. Japan has few natural energy sources. Geothermal, wind, solar…these are not going to cut it. The choice is between using fossil fuels alone or also using fossil fuels and nuclear energy. Right now, Japan needs good science, good faith and good communication.