VCU news

By Brian McNeill

As the United States increasingly eyes nuclear power as part of its strategy to minimize CO2 and greenhouse gas emissions, a new emphasis is being placed on modernizing the nation’s aging nuclear power plants, most of which were built between 1970 and 1978.

When you go into a nuclear power plant, the infrastructure you see looks like it came out of the 1960s.

“When you go into a nuclear power plant, the infrastructure you see looks like it came out of the 1960s. And, in some cases, it may actually be from the ‘60s, or it’s a replica of technology from the ‘60s. You can’t just go down to your local electronics supply warehouse and buy an analog computer,” said Carl Elks, Ph.D., assistant professor in the Department of Electrical and Computer Engineering in Virginia Commonwealth University’s School of Engineering. “Obviously, that’s a problem in the 21st century.”

For one thing, Elks said, the old technology is becoming increasingly difficult to replace with similar technology. For another, it is becoming harder to find skilled people who understand the obsolete technology.

“While this older technology may be very safe to rely on for the next 20 years, we need to have technology upgrade options available,” he said. “So the nuclear industry is looking at ways to incrementally replace analog technology from a bygone era with modern digital or computer-based technology.”

Yet updating the technology in nuclear power plants is far more complicated than simply installing new computer-based systems. If those systems fail, it might lead to loss of power generation or an accidental loss of reactor protection and safety.

“One way these digital systems can fail, particularly those that are software-driven, is that they might have bugs in the software that you haven’t foreseen,” Elks said. “And these bugs, if they weren’t found during development and testing, could become active when an anomalous condition triggers the software – then you can have what’s called multiple co-incident failures or common cause failures. Failures of this type could lead to tripping the plant off the grid or temporary loss of safety protection – all of which are a bad day in a nuclear power plant.”

Elks is part of two teams that have received recent grants from the Nuclear Energy University Program of the Office of Nuclear Energy in the U.S. Department of Energy to develop formal model-based testing and “design for verification” methods to minimize the chances of common cause failures with digital and computer based technology.

“If you could list the worst things that can occur in digital-based instrumentation and control system at a nuclear power plant, common cause failures are up there toward the top. That’s one of the major concerns the [the Nuclear Regulatory Commission] is really worried about,” Elks said. “The goal of our research is to provide methods and tools to significantly reduce the possibility of common cause failures.”

The first grant, “Development and Demonstration of a Model-Based Assessment Approach for Qualification of Embedded Digital Devices in Nuclear Power Applications,” is a three-year project totaling $254,490. It is being conducted in partnership with researchers at the University of Tennessee and Ohio State University.

The grant is aimed at developing new ways to test the power plants’ digital systems. “We’re developing a testing framework, which we call model-based testing, which is a rigorous new method for testing software,” Elks said.

The second award, “Realizing Verifiable Embedded Digital Devices,” is a three-year grant totaling $542,000 and is being done in partnership with the Electric Power Research Institute.

For that project, two unique and different approaches are being researched. The first approach is called SymPLE, which is a concept where verification requirements are “baked in” the design and development process from the very beginning, thereby allowing designers and testers to constrain the design so it is easier to verify and eliminate the occurrence of unforeseen software errors.

The second approach is being led by Gary Atkinson, Ph.D., associate professor in the Department of Electrical and Computer Engineering, and it aims to explore the potential of non-digital, miniaturized technology called Micro-Electrical Mechanical Systems, or MEMS, as a possible failsafe alternative to digital technology.

“If everything else fails, then you have MEMS technology to provide defense in depth. That’s something that the nuclear industry has always liked, this notion of defense in depth and diversity,” Elks said. “Nobody has ever really looked at MEMS technology as one of the defense mechanisms.”

If successful, both grants could be game-changing, disruptive new technology for modernizing the nation’s nuclear power plants, Elks said.

“Software common cause failures are one of the things we have to get a handle on before digital computer-based technology is accepted and commonplace in today’s plants,” he said. “That’s what both of these grants are doing. They’re looking at ways to really rigorously design and test these systems so that a very high level of trust, assurance and dependability can be justifiably placed on them.” 

 

Subscribe for free to the weekly VCU News email newsletter at http://newsletter.news.vcu.edu/ and receive a selection of stories, videos, photos, news clips and event listings in your inbox every Thursday.