Nuclear plant operations: unlocking the human factor

At a time when nuclear energy could be playing a significantly larger role in decarbonising the world's energy mix, the nuclear industry finds itself with something of an image problem.

Beyond its immediate and ongoing consequences for the surrounding area, the 2011 meltdown at the Fukushima Daiichi plant sent ripples of distrust around the world, strengthening already-stiff pockets of public opposition to nuclear power and playing a part in the ramping down of civil nuclear programmes in several major markets. The sector now faces an uphill struggle to secure its place in the low-carbon future, despite its great potential as a method of low-carbon baseload generation to support the deployment of intermittent renewables.

Of course, a new generation of reactors that help solve nuclear's niggling issues will help to rehabilitate its image. But another important factor is deepening our understanding of the human stresses involved in nuclear plant operation; properly supporting nuclear engineers and control room operators could help improve the safety and efficiency in the sector as it transitions to new technologies, as well as making it feel like a safer bet for the public and policymakers.

Understanding and improving the performance of nuclear plant operators is exactly the focus of a new investigative project launched by a team at Vanderbilt University in Tennessee, US, funded by an $800,000 grant from the Department of Energy. With support from Areva nuclear specialists Paris Stringfellow and Mike Bonfiglio and led by professor of anesthesiology, biomedical informatics and medical education Dr. Matthew B. Weinger, the human factors research project will observe and interview nuclear plant operators in a simulated control room to find out how the transition from traditional analog instruments to new digital displays could affect their performance.

The team consists of participants from Vanderbilt's medical faculty (Dr. Dan France, Dr. Shilo Anders and Weinger himself) and the university's School of Engineering (Dr. Julie Adams and Dr. Sankaran Mahadevan), making it a truly multi-disciplinary effort. We sat down with Weinger and Adams to discuss the project, from its origins as a cross-industry conference in 2012 to the challenges of gathering data on human performance.

Chris Lo: Could you explain the basic idea behind human factors research?

Matthew Weinger: Human factors engineering (HFE) is the study of how systems, composed of humans, processes, and technology, do or do not support the goals of the humans who work in and/or rely on that system. Typically, HFE focuses on the design and evaluation of systems to achieve optimal effectiveness (system outputs like energy production) and efficiency (e.g. amount of work, cost of production) while maximizing safety and operator satisfaction.

CL: Why does the nuclear energy sector make a particularly good fit for this kind of study on human factors and safety?

MW: There is already a very long history of human factors in the nuclear power industry albeit with a predominant focus on safety. In fact, notwithstanding recent high-profile events, nuclear power is probably the safest high-hazard industry and most certainly far safer than healthcare. Thus, the industry is receptive to new HFE methods and approaches to enhance system performance as it modernizes its existing fleet of NPP [nuclear power plant] control rooms.

From a purely pragmatic standpoint, a huge advantage to conducting HFE research in nuclear power when compared to healthcare is the high fidelity and the routine use of NPP control room simulators. In healthcare, not only are our simulators crude but very few experienced clinicians train in our simulation facilities. The advantage of simulation is that one can reproducibly study a situation or intervention - for example new technology - without risk to the system or its operators.

CL: How did the joint healthcare and nuclear power workshop in 2012 and the subsequent monograph that you co-edited affect your thinking on the common ground between these two fields?

MW: Beforehand, I had envisioned that there would be little the nuclear industry could learn from healthcare. In contrast, I thought that the workshop would be a tremendous opportunity for healthcare to learn from the nuclear industry about making systems safer. Indeed, the workshop revealed many safety practices in the nuclear industry - creating a robust safety culture, developing resilient processes - that would benefit healthcare. What I did not appreciate until after the workshop was that nuclear's intense industry-wide collaborative focus on safety has had unintended consequences particularly with respect to slowing technology innovation, in both research and practice, something that is a real strength of the healthcare industry.

CL: The main aim of the investigation is to compare the effects of analog, digital and hybrid systems on power plant operator performance - how has the project progressed since your initial announcement in early November 2014?

MW: We chose to focus our project on activities that will help the industry 'upgrade' existing NPP analog CR [control room] instrumentation to digital technology. Based on decades of experience in other industries, including healthcare, we know that the transition from analog to digital instrumentation is neither simple nor straightforward. This may be particularly difficult in NPP CRs due to their complexity, size and operators' need for situation awareness across many subsystems. Hybrid (or mixes of analog and digital instrumentation) may actually be the worst situation because operators often need to use different strategies with the different technologies. Further, there are opportunities to improve how we evaluate the success of new CR subsystem designs using meaningful performance measures.

Because members of the project team are relatively new to the nuclear industry, aside from our team members from Areva, we are spending the first half of the first project year 'coming up to steam' on the domain.

CL: How do you plan to move the project forward in 2015? Is there a clear end date for the study?

MW: Project goals for this year include a synthesis of the relevant literature, not just in the nuclear industry, but in healthcare, aviation, military and so on, focused on, firstly, meta-level design guidance to support decision-making and situation awareness, and secondly, performance measures to assess the effects of new technologies. We will also be conducting observations and interviews at several NPPs to inform our initial studies of digital versus analog instrumentation and our choice of performance measures in those studies.

CL: Could you explain the process of gathering the necessary data from nuclear control room operators? Is it challenging to turn observations into actionable evidence?

Julie Adams: Data collection via observation and interviews is qualitative in nature and will occur in NPP control room simulators. Prior to observations, the team will identify elements of the activities that are most important to observe. Observers will collect information regarding how the control room operators (CROs) complete the activities, what information they access and use to complete the activity, and how they access and use the information along with how they proceed to complete the activity. Interviews will be conducted after the observations to clarify observed actions and procedures, identify difficult-to-access information or controls, and collect suggestions on potential interaction and information collection needs for completing the activities.

While it is challenging, many techniques exist for analysing and transforming such information into design guidance. Common techniques include cognitive task analysis and work domain analysis. These techniques permit the understanding of the tasks and how the tasks are impacted by the functional work environment. This information often informs additional data collection that provides deeper insights into potential human-system interaction designs.

A second, equally important aspect of our research is to identify methods to assess more objectively CROs' performance during simulation studies specifically designed to evaluate new digital technologies. For example, how does the substitution of a new digital control-display module for an existing analog module affect CROs' workload, workflow efficiency, heart rate variability or other performance measures?

CL: Is confusion over the use of newer or older technologies (digital and analog instrumentation in the nuclear power case) a common cause of impaired performance in high-risk, high-consequence working environments?

MW: Based on experience in other industries, there can be significant safety and efficiency effects of changing or mixing types of instrumentation. We aren't yet sure about the nuclear industry but suspect that there could be negative effects on efficiency, safety and operator satisfaction. That said, the devil is in the details of instrumentation design so there are opportunities for making the analog-to-digital transition easier and safer.

CL: Can the familiarity of older technologies in these environments sometimes outweigh the benefits of newer tech?

MW: Yes, and that's one of the reasons that older technologies persist in any system. However, in the nuclear industry, possibly bigger drivers of maintaining the status quo are safety questions about new untested technologies and the cost of change.

CL: Generally speaking, do you think there is more that could be usefully shared between the fields of medicine and engineering?

MW: Certainly. The National Academy of Sciences (which includes the Institute of Medicine and the National Academy of Engineering) has published several reports on the need for healthcare to utilize more engineering principles, practices and methods, and to engage engineering expertise in the healthcare delivery redesign necessary in the coming years to create the higher-quality, lower-cost system that our society requires. More generally, we believe that interdisciplinary learning and thinking are necessary to successfully address complex issues.

Follow Chris Lo on Google+