No single source of low carbon energy will be sufficient, but without a major expansion of nuclear energy around the world it simply doesn’t seem credible that the world will be able to achieve deep reductions in CO2 emissions while simultaneously having access to enough energy for even modest economic growth.

The IEA recently made a projection of what would be needed to prevent an increase of more than 2°C in the global average surface temperature by the end of the century. There are many uncertainties in these kinds of calculations, but the bottom line of the IEA projection is that the global nuclear fleet would have to be meeting about a quarter of the world’s electricity needs by 2050 to achieve this goal. That would require an installed capacity of about 1200 GW, compared with 370GW today. In other words, we would have to build more than twice as many nuclear plants in the next forty years as we built in the last forty. And since much of the current nuclear fleet will likely have been retired before the year 2050, the requirement for new-build would be larger still.

This would mean completing new nuclear plants at an annual rate at least as large as during the period of peak nuclear construction in the mid-1980s, and doing that every year for the next four decades. Not an impossible task, but surely a very formidable one.


What would it take to achieve such a goal? Here I want to mention three key challenges: strengthening nuclear governance worldwide after Fukushima, developing new nuclear technologies, and rapidly increasing the supply of well-trained people in every branch of the nuclear workforce.

First, it hardly needs to be emphasized how important it is for the international nuclear community to absorb and apply the lessons of Fukushima. Even today, there are still things that we don’t know about what happened. But it’s already clear that, just as at TMI and Chernobyl decades earlier, there was a breakdown of nuclear governance: of systems of management, control, and regulation.

The proximate cause of the accident was of course a physical event, but as with previous accidents the consequences of that physical event were magnified by breakdowns in the performance of human beings and the organizations they populated, the procedures they followed, and the institutional structures they had created. Those breakdowns aggravated what otherwise might have been a very serious but manageable accident.

Another lesson from Fukushima is that failures of nuclear governance can occur anywhere. It’s worth pointing out that the three most serious breakdowns of this type in history occurred in vastly different political and institutional settings: the United States, the former Soviet Union, and now Japan. It would be unwise for nuclear leaders in any country to assume that it could never happen to them. Indeed, we might reasonably conclude from the evidence that there is no national system within which major failure cannot occur.

There are some universal principles of effective nuclear governance that apply everywhere, regardless of national context. Principles like the transparency of decision-making, independence of regulation, and the importance of maintaining a safety culture in nuclear organizations.

The first challenge, then, is that both existing nuclear countries and new entrants into the nuclear field must fully embrace these principles and implement them with the highest standards of dedication and quality.

On this question of nuclear governance it’s too early to say what progress the international nuclear community is making. National governments are paying attention, but it’s clear that stronger international efforts will also be needed to drive the needed reforms. And, the fact is that our existing international institutions are limited in what they can do.

The second and equally important challenge for the international nuclear community concerns technological innovation. If nuclear power is to expand around the world, an already-safe technology must be made demonstrably safer, as well as less expensive, more secure against the threats of nuclear terrorism and proliferation, and more compatible with the capabilities and the limitations of electric power systems and the utilities that operate them.

To these familiar targets of innovation, we must also add the problem of cycle times. Everything — facility planning, siting, design, regulatory development and licensing, construction, R&D, demonstration, waste disposal — now takes so long in so much of the world that it is unlikely that the global nuclear power industry could grow by a factor of three or more between now and 2050 even if nuclear power plants were cost-competitive everywhere. There simply aren’t enough cycles to make that possible.
As just one example, some experts recently estimated that it would take 20 years to bring a new ceramic fuel cladding technology for light water reactors to the marketplace. The fact that this is a relatively modest technical change underscores the larger point. Of course, long lead times for physical infrastructure of all kinds are now common around the world. But nuclear lead-times are the longest of the long. Since it is now routine, even in advanced societies, to be able to build a few hundred megawatts of gas-fired capacity in three years or less, the much longer lead times for nuclear projects are a serious competitive disadvantage.

So innovation to improve safety, to reduce costs, and to reduce cycle times is the second challenge facing the international community. That is a different agenda from what has historically been the primary goal for nuclear innovation: uranium conservation through breeding.

Future directions

On the question of nuclear innovation, much of the effort today is focused on developing advances in light water reactor technology, and the next generation of large LWRs offers many safety advantages relative to earlier designs. That said, the high costs, delays and cost overruns that are being reported at some current LWR construction projects are a cause for serious concern.

We are moving away from fuel breeding as the dominant objective of innovation. That was the driver of innovation in the nuclear industry for decades. I think people are recognizing today that uranium conservation is only one objective of nuclear innovation. We have much more pressing problems to solve: escalating capital costs, concerns over safety and, crucially, the ability to build plants faster. Another important consideration is to have nuclear plants that can coexist easily with the growing amounts of non-dispatchable renewable resources on power grids. These are really the innovation imperatives for nuclear. Whether we achieve them through light water reactors or other kinds of reactors and fuel cycles is another question.

A number of innovative reactor and fuel cycle technologies are under development around the world. The challenges that face the developers of new nuclear technologies are daunting. There are many who say that today’s light water reactor technology is safe enough, that instead of searching for something better we should be concentrating on bringing down the cost of what we already have, through standardization and learning by building a series of these plants. Certainly we should be doing more of those things. But to argue that that’s all we should be doing seems, to me at least, to be short-sighted. Indeed, I believe we are quite likely still just in the early stages of developing nuclear energy technology.

It’s still only 75 years since the nuclear fission reaction was first demonstrated. In chronological terms, that puts the field of nuclear engineering today roughly where the field of electrical engineering was in the year 1900, eighty years after the great discoveries in electrochemistry and electromagnetism by Faraday and others. Just think about what the electrical engineers then achieved: the creation of the electric power grid, radio, television, the revolutions in micro-electronics, in computation, in telecommunications, the internet, and much, much more. A truly astonishing series of advances, and I doubt that there was a single electrical engineer in 1900 who anticipated any of these things.

Likewise, no one today can foresee the range of applications of nuclear science and technology at the end of the 21st century. The most we can say is that the nuclear power plants of the year 2100 are likely to have about as much resemblance to today’s workhorse light water reactors as a modern automobile has to a 1912 Model T Ford. Perhaps we will see greater reliance in the future on passive heat removal mechanisms. Certainly the new generation of LWRs is moving in this direction, but more advanced designs go much further towards the goal of ‘walk-away safety’. Is it so unlikely that this goal will become a requirement for all nuclear power reactors 50 or 100 years from now?

Other longer-term possibilities include reactor cores with a single charge of fuel over their lifetime, and integrated nuclear power plant-waste disposal systems, with spent fuel never leaving the power plant site and waste disposal in modular deep boreholes several miles below the earth’s surface in the stable, dry bedrock that is available in abundance in most countries. Over this kind of timeframe there are opportunities to make truly radical advances.

In the nearer term, there are hopes that much smaller nuclear power reactors will expedite the application of advanced manufacturing techniques, modular construction technologies, and ‘dynamic’ economies of scale or learning processes. Reductions in capital-at-risk, shorter lead times, better matching with smaller power grids, and increased flexibility may make small reactors especially well suited to new nuclear countries, as well as to many operators in mature nuclear states.

Small modular light water reactors are a relatively near term development, and I think the verdict is still out on their economics. Unless we can figure out how to reduce the operating costs of SMRs they may have real difficulties competing. The operating costs are going to be determined in part by regulatory requirements. Until we get a little further down the road we just don’t know how the economics will play out, particularly because one of the principal economic arguments for these SMRs is the ability to control the manufacturing environment in a way that you can’t with field construction, and to exploit economies of learning in a more controlled manufacturing environment. That’s easy to say, but it’s going to be hard to demonstrate without investing in the entire system, not just a single reactor module.

Role of technology

Tremendous gains in computing power already enable more advanced simulations of nuclear reactor behaviour than ever before. And the role of advanced modeling, simulation, and visualization technologies will be increasingly central. The need for training simulators and associated training materials and services will continue to grow as the nuclear workforce turns over and expands and as the shift to digital instrumentation and control continues.

There will also be a growing need for engineering simulators that can be used for the development, design, integration and validation of next-generation nuclear power reactors. For example, the neutronic and thermal-hydraulic design of earlier generations of nuclear reactors depended heavily on physical experiments. At the time, these were relatively inexpensive, whereas computational resources were prohibitively costly. That ratio has shifted radically. It’s now very time- and resource-intensive to build and operate nuclear test facilities, while the cost of accurately simulating the performance of nuclear reactors has declined dramatically along with the cost of computing.

Physical experiments will always be important, but advanced computation and simulation has become an essential tool for nuclear designers. In some areas simulation will be even more important. For a high-level waste repository, regulatory authorities and the general public will demand assurances of effective waste isolation over periods measured in the tens of thousands of years. The ability of physical experiments to contribute to this will be very limited indeed. Computer simulations of engineering materials and natural geo-hydrological structures and processes will be an indispensable tool for nuclear regulators.

In my own department of nuclear science and engineering at MIT, we’ve responded to these developments in a number of ways. We have brought advanced simulation and computation into the core of our curriculum, as these methods and technologies will become increasingly important for the 21st century engineering leaders in our field."