Leonard J Bond
In February 2002, during a routine inspection at Ohio’s Davis-Besse nuclear power station, inspectors found three cracks in the lid of the reactor’s pressure vessel, the mighty steel cylinder that encloses the radioactive core. One crack was in the housing of a mechanism that drives control rods into the reactor core to manage the nuclear reaction. The flaws needed to be repaired, but there was no sense of urgency—that is, until workers began fixing the crack in the control rod mechanism and they felt a wiggle. A wiggle that was all wrong.
The control-rod housing moved slightly, which should have been impossible, as it was supposed to be surrounded on all sides by the 15-centimeter-thick steel of the reactor vessel. When workers investigated, they found a cavity roughly the size of an American football in the steel next to the housing. This void left less than one centimeter of metal protecting the pressurized interior of the reactor vessel, with its radioactive core. If the vessel had ruptured while the reactor was in operation and at pressure, the water that cooled the core would have gushed out through the hole. Such a serious “loss of coolant” accident might have led to serious core damage. To fix the vessel, the plant’s owners installed a new lid at an estimated cost of US $600 million.
Investigations by the U.S. Nuclear Regulatory Commission (NRC) and the plant’s owner determined that a tiny fissure probably appeared in the control rod mechanism as early as 1990. By around 1995, acidic water from inside the reactor was leaking through the crack and eroding the surrounding steel of the pressure vessel; it ate away at the steel for seven years before workers discovered the metal loss. Nuclear researchers are acutely aware that this kind of slow, steady degradation becomes more likely as nuclear power stations age. Every day of operation, the rugged steel and concrete that make up a reactor’s containment structures are bombarded with radiation and stressed by both high temperature and high pressure. Given enough time, these forces can potentially weaken even the toughest materials.
In the aftermath of Japan’s Fukushima Dai-ichi nuclear accident, governments all over the world are reevaluating the safety of their nuclear power plants. In the United States, where nuclear power supplies 20 percent of the country’s electricity, attention is focused on the aging of the country’s 104 active nuclear reactors, which are 32 years old on average. (Four Westinghouse AP-1000 reactors now being built in the United States are the country’s first new nuclear construction projects in decades.) When these reactors went online, regulators granted them licenses to operate for 40 years, a conservative estimate of their life span. Now these plants are being awarded license extensions; 73 reactors have already received approval to operate until they’re 60 years old, and 10 of those reactors have already entered this new era of extended operation.
But that’s not the end of the story. Operators are performing major “midlife” refurbishments that can cost $1 billion per plant. Meanwhile, regulators and nuclear researchers are studying these aging plants to find an answer to one of the most important questions now facing the industry: Would it be safe and economically sound to keep these plants running until they reach 80 or more years of age?
That question, on which billions of dollars will depend in coming years, is also being asked in Europe, Asia, and former Soviet states. Although license periods and practices vary across countries, it’s in the United States, with its large concentration of aging plants, that regulatory and industrial decisions will establish guidelines for reactors around the planet. If the U.S. nuclear industry demonstrates that refurbished nuclear power stations can operate until they’re octogenarians or even longer, other countries will likely follow its example.
Current plans for managing aging reactors include periodic inspections of the components that are most difficult to replace: the pressure vessel, the concrete containment structure that surrounds it, and the main pipes and cables that connect to it. Over the past 15 years at Pacific Northwest National Laboratory (PNNL), in Richland, Wash., my colleagues and I have sought new types of online monitoring and nondestructive testing technologies that can provide early warnings of degrading materials. Our goal has been to transition from the current “find and fix” approach to one we call “model and predict.”
Inside a nuclear power station, fierce forces are at work. In the pressurized water reactors (PWRs) and boiling water reactors (BWRs) that generate power in the United States, the nuclear cores consist of rods of uranium dioxide. Inside this radioactive material, a nuclear fission reaction produces energy and many forms of radiation, including gamma rays and neutrons. The extremely high radiation levels are reduced by about a factor of 20 by the steel walls of a reactor pressure vessel, and then to safe levels by the massive reinforced concrete containment structure that jackets the vessel.
Both reactor types use water as the coolant. In a PWR, water enters the reactor core at about 275 °C and is heated as it flows upward through the core to a temperature of about 315 °C. The water remains liquid due to high pressure, usually around 15.5 megapascals (about 150 times the atmospheric pressure at sea level). In a BWR, the cooling water is maintained at about 7.6 MPa so that it boils in the core at about 285 °C. In both cases steam is produced to drive turbines that generate electricity.
High temperature, high pressure, and radiation all stress a reactor’s components. Inside a reactor, neutrons bombard the pressure vessel’s steel walls; over a period of years, that bombardment can cause reactions that displace atoms in the material and produce impurities and tiny voids. These microscopic phenomena can reduce the metal’s toughness and its ability to resist cracking.
The NRC and the nuclear industry, working with the Electric Power Research Institute, are now determining how to measure and monitor the aging of a reactor’s key components. The major concerns are embrittlement and cracking in the reactor pressure vessel and its piping; degradation of the concrete containment; aging cables; and corrosion in buried water pipes. At the moment we just don’t know which of these problems will be the most critical in any given plant. After all, no one has ever before operated a commercial-scale nuclear reactor for six or seven decades. We have entered a new era in the atomic age.
During the past 30 years, many parts of plants have been replaced or refurbished, including turbines, some major piping, and pressure vessel lids. But the central components of a nuclear plant—the pressure vessel itself and its reinforced concrete-and-steel containment—were never designed for replacement. The pressure vessel of a typical 1-gigawatt power plant weighs about 300 metric tons and is more than 12 meters tall. Most analysts believe that it would be easier to build a new plant than to cut into the containment to extract and replace a pressure vessel.
So how do you determine whether a vessel or another major component is robust enough to last another 20 years?
If you want to know what’s happening in an aging reactor, to really understand how its thick steel and tough concrete are faring after years of relentless bombardment, the best thing to do may be to listen to it. Nuclear researchers are now testing acoustic and ultrasonic monitoring techniques drawn from the civil and aerospace engineering communities. The same methods used to monitor the structural integrity of a bridge or an airplane may work for a nuclear pressure vessel as well.
One promising technique was demonstrated decades ago in an operational nuclear plant. In 1989, inspectors at the Limerick Generating Station, in Pennsylvania, found a tiny crack in the welding around a pressure vessel pipe that brought cooling water into the bottom of the reactor. The operators concluded that the flaw didn’t pose a threat, but they wanted to see if it was possible to monitor crack growth in an operating plant. They turned to a technique called acoustic emission monitoring, which is used to check on metallic structures like pipelines and wind-turbine blades. This method relies on the fact that when a crack grows, acoustic energy is released in tiny pulses—much the same way an earthquake sends out seismic waves. Once the acoustic system was installed, operators could listen for the ultrasonic waves that would indicate a growing fracture.
The acoustic system was kept in place for three years, during which time researchers listened in as one part of the crack grew to a depth of 12 millimeters. The system also detected the growth of minuscule cracks that wouldn’t have been noted by traditional monitoring methods, and researchers deemed the technology demonstration a success. In the decades since, fossil-fueled power plants and petrochemical facilities have installed acoustic emission systems to monitor vessels and pipes. However, nuclear power stations in the United States have been slow to adopt this proven technology.
With advances in both computer hardware and processing software, acoustic emission systems are now little larger than a laptop and are capable of displaying data nearly in real time. At PNNL, my colleagues and I recently tested acoustic emission monitoring along with another technique for metal monitoring that makes use of “guided waves.” In this technique, transducers generate ultrasonic waves with specific frequencies, which propagate through a structure such as a metal pipe or the walls of a pressure vessel. Because the ultrasonic waves are scattered and reflected by discontinuities in a material, they can provide clear indications of cracks or corrosion. This technique could be particularly useful because it wouldn’t require inspectors to strip off insulation to inspect pipes, like the all-important cooling pipes that circulate water through the reactor core.
In our recent laboratory study, we tested these two monitoring methods on a fatigued stainless steel pipe. The acoustic emission monitoring detected signals caused by the formation of a crack before we visually confirmed that tiny fissure. After we knew the crack was there, we monitored it with the guided-wave technique. When the waves encountered a crack, they bounced back to the sensor; by monitoring those received signals we were able to follow the growth of the defect from a starter notch that was 2.45 mm deep and 47.7 mm long to a fissure that measured 68 mm long. This may not seem like dramatic growth, but such a crack would be a serious cause for concern at an operating nuclear power plant.
Guided-wave technology, which is rapidly maturing, is now regularly used for pipe testing in the oil and gas industry. In the nuclear industry, regulators are working to standardize the monitoring procedures. To use the technology inside an active plant, however, operators must overcome challenges like high temperatures—it can hit 200 °C inside a light-water reactor’s primary piping. That’s far too hot for the most common type of transducers, which use piezoelectric materials to convert electricity into ultrasonic waves in the transmitters (and vice versa in the receivers). To get around this problem, some researchers are testing more rugged piezoelectric materials. Others are experimenting with different ways to generate the waves—for example, a laser pulse that heats and expands a pipe’s surface to create waves that ripple outward.
Two other ultrasonic techniques show potential for long-term deployment. A kind of phased array, which is commonly used as a diagnostic tool in medicine, uses a grid of elements to generate many small ultrasonic pulses. By using electronics to control the timing and interaction of the individual pulses, operators can create a single wave front and control the direction of the wave. Phased-array technology is now routinely used in periodic inspections of nuclear power plants, but the technology has the potential for continuous monitoring, where a single transducer is fixed in place and electronic beam steering is used to scan critical structures. This technique can check for degradation in coarse-grain materials like cast stainless steels and can also look for flaws in welded areas.
Finally, an approach drawn from seismology could be useful to monitor the formidable concrete structures in a nuclear power station. In this “diffuse field” technique, an ultrasonic pulse is introduced into a coarse-grained material such as rock, concrete, or cast stainless steel. As the ultrasonic wave propagates through the substance, the grains interfere with the initial pulse of energy and send echoes back to the transducer. The resulting signal, showing all the interactions from within the textured material, provides a distinct signature for that material. This signature changes if the substance’s elastic properties vary or if a crack or other degradation is introduced. So far, diffuse ultrasonic tools are being used only for research in the nuclear industry, but their potential for inspections and long-term monitoring has been clearly demonstrated.
If the United States wants to continue relying on nuclear power to keep one out of five lightbulbs lit, the NRC needs to be assured that a sound technical basis exists to support a second round of reactor life extension. By about 2020, the NRC must decide if it needs to establish additional rules and standards for subsequent licenses to allow for operation to go from 60 to 80 years. These additional rules would give operators a clear framework for the crucial and expensive decisions before them. No decision from the NRC would, in effect, preclude extended operations because it takes many years to plan for component replacements, refurbishments, and upgrades. If utilities and other nuclear plant operators don’t have an explicit framework from the NRC by 2020 that enables them to schedule their capital investments, they’ll have no choice but to start planning the decommissioning of the country’s nuclear reactors.
Retrofitting and upgrading nuclear reactors will not be cheap. Some plants have already reported that they will spend up to $1 billion per plant to support the 40- to 60-year license extension. It may be that economics, rather than technology, will in the end determine if it’s feasible to extend the life of a given plant beyond 60 years. But there’s also a strong financial case for keeping aging reactors running: The loss of the existing plants after 60 years of operations would be a crippling economic blow. In the United States, the annual electricity demand is projected to increase about 21 percent by 2030 to roughly 5000 billion kilowatt-hours. It’s difficult to imagine meeting that demand without the help of most of the country’s 104 existing reactors.
If the United States decides against further license extensions, massive investments of trillions of dollars will be needed to replace the more than 100 GW of base-load generating capacity represented by the country’s aging nuclear reactors. Whether the investment would be in new nuclear plants, cheap natural gas plants, or renewable energy facilities, it would be a monumental national project to replace the power we’d lose. Keeping a careful eye—and ear—on our aging nuclear infrastructure may be the more attractive option.
This article originally appeared in print as “Old Reactors, New Tricks.”