Modelling vital as nuclear regulators turn to performance-based licensing
Modelling nuclear power stations is essential to reduce costs and streamline development, especially as regulation focuses more on a plant’s performance.
Related Articles
A nuclear plant must be seen as both safe from meltdown and secure from an attack by bad actors looking to provoke a system collapse.
To meet strict safety and security requirements, operation and maintenance (O&M) costs for a reactor in the United States cause many utilities to struggle to keep plants functioning at rates competitive with natural gas power stations.
O&M costs to protect a nuclear power station account for approximately 7% of the total cost of power generation, with labor making up half of that cost, according to the Department of Energy’s (DOE’s) ‘Economic Analysis of Physical Security at Nuclear Power Plants’.
The analysis forms part of the DOE’s 2020 Light Water Reactor Sustainability Program which was established to support research to facilitate lowered O&M costs through direct research into modeling and simulation.
According to the study, physical security forces account for nearly 20% of the entire workforce at several nuclear power plants.
For exclusive nuclear insights, sign up to our newsletter.
Labor costs make up more than 60% of the total physical security budget, and this contribution has been rising steadily since 2008, suggesting the physical security system has shifted toward a more labor-intensive approach, the analysis says.
Evolution from 1990 to 2019 of percentage of total cost for the four types of physical security costs
(Click to enlarge)
Note: (a) Single unit nuclear power plant (b) Dual unit nuclear power plant
Source: DOE's Light Water Reactor Sustainability Program's 'Economic Analysis of Physical Security at Nuclear Power Plants'
A new approach
The approach to security and safety is changing for next-generation reactors.
Current regulations are mostly prescriptive – a plant must have a set number of barriers, all at a specific height, and a set number of armed responders on site . However, the proposed regulations focus more on performance-based requirements.
The new regulations would examine each plant’s plans individually to determine whether, for example, the developer’s proposed wall, in concert with other aspects of the physical protection system, is sufficient to defend against an adversary.
Current requirements also aim to prevent sabotage, defined as core damage with localized fuel melting, though the new rulemaking aims to instead focus on radiation doses expected at the exclusionary boundary as a consequence of an attack.
This new approach places greater emphasis on accurately modelling reactor designs.
“This is a performance-based requirement, which completely changes the paradigm for advanced and small modular reactors, hence, us moving to more modeling and simulation,” says Christopher Chwasz, Regulatory Development and Licensing Engineer at Idaho National Laboratory (INL).
“Instead of the security program preventing the core from melting, we're now looking at the performance of the reactor in response to the security event.”
The lab’s work is pioneering ‘safety and security by design’ and have developed software, the ‘Event modelling Risk Assessment using Linked Diagrams’ (EMRALD), to integrate tools that simulate scenarios at a power plant.
Combining EMRALD with other software tools, such as a thermal hydraulics tool and timelines associated with adversary and response force action, forms the Modeling and Analysis for Safety and Security using Dynamic EMRALD Framework (MASS-DEF).
One example of the use of the MASS-DEF framework would be to model a water makeup system designed to give a reactor core 24 hours of extra time in the case of a safety incident.
From a safety perspective, 24 hours may be sufficient, but from a security perspective, if the the simulated armed response time for this reactor is longer than 24 hours, then a larger tank may be necessary.
“We're entering a new world for physical security, and so modeling and simulation is vital, because we're talking about doing all of this strictly during the design phase,” says Chwasz.
With the right, performance-based modeling, it is possible to reduce the number of armed responders needed to protect a plant.
By taking just one guard off the payroll, a utility can save tens of millions of dollars over a plant’s 40-year license period, Chwasz says.
Passive safety
At the DOE’s Argonne National Laboratory (Argonne), researchers are working on a half-scale model of a reactor system, the Natural Convection Shutdown Heat Removal Test Facility (NSTF), which allows them to study passive safety systems that can cool a reactor without humans or power.
The data gathered is qualified at the national standard for quality assurance in the industry, or the National Quality Assurance-1 (NQA-1), and is shared with developers, vendors, and regulators to validate computational models and guide licensing of new reactors and components.
While a passive safety system doesn’t aim to replace human operators, it does add an extra layer of security in worst-case scenarios.
“Your options are going to run out if you lose your power and some of your systems are relying on valves closing or switches flipping that are based on electricity. You don't have that concern with a passive safety system,” says Principal Nuclear Engineer at Argonne Matthew Jeffrey Jasica.
“If you lose power and you lose your switch or valve control, it doesn't matter, because you have this extra layer of safety that is still going to keep coolant flowing no matter what’s happening.”
Modelling has become more accurate as high-tech sensor systems have been developed, but challenges to predicting systems, especially at edge cases, can arise.
An example is when water passing through a system starts to boil, a complex chaotic event that can be difficult to model due to changing pressure waves and fluid loss, Jasica says.
Argonne is currently developing a suite of datasets that examine different responses to boiling water, allowing modelers to use the data to validate their models.
Standardized, shareable data means certain security testing doesn’t need to be done individually by developers, reducing costs for everyone.
“One of our key contributions to reducing present day costs is, we've got a facility in place, we have data we're generating now, and we’re collaborating with them to make sure that what we do next is working with their needs, so they don't have to do it, and they don't have to spend the money,” Jasica says.
By Paul Day