May-June 2019 NPJ

Nuclear Plant Journal, May-June 2019 NuclearPlantJournal.com 21 With these digital systems, plants will have many more sensors for data collection and analysis because these systems have more rich streams of data built into them. This is where there are opportunities to apply artificial intelligence and other advanced data techniques for new learnings about equipment strengths and potential vulnerabilities. 3. Will the data which has already been stored at the plants for the last so many years be of use or it will be all new data? There are plant databases which already have thousands of data points going back years. That’s a rich source of data that we’re looking at right now using advanced techniques to see if we can predict potential failures earlier than it would have been predicted using traditional techniques. 4. What would be the applications for virtual reality in the plants? One application is training. We already have an application for maintenance training on a Terry Turbine. Terry Turbines are used in the emergency core cooling system of many plants. Hands-on Terry Turbine training is costly and time-consuming. With virtual reality, there’s no need to purchase and maintain spare plant components, and it’s easier to set up and administer than hands- on training. You can also do things like color code the parts and see which part is which, because when you take apart the turbine in reality, it looks like a bunch of stainless steel parts. Another example for virtual reality is using it to visualize inspection data. An important part of inspecting plants is to make sure there’s no aging going on in pipe welds, for example. Instead of having someone looking at a two dimensional scanner and interpreting that, you could take the data from the inspection, overlay it onto the actual weld in augmented reality and you can literally see the flaws inside. 5. Analytics examine large data sets, for patterns and correlation. How does it happen that you get data from another plant and you bring it to your own plant and customize it to help you with the operation of the plant? One of the great things about the nuclear industry is that it is a sharing industry. Sharing data is an extension of sharing of operating experience. Many plants are sharing their work order data, for example, with EPRI and we’re looking across it for patterns and trends, and using that to share best practices and new learnings. One trend for example might be this: If you had six months of pump data and you know that the pump failed towards the end of that period, you could look in the data for patterns that provide correlations to the failure, which would be an early warning of the failure. 6. How do you plan and equip sensors needed for thousands of valves, motors and switches needed for the new data application? We’ve done several things in that area. First we’ve defined what are the critical sensors to measure equipment health in our Continuous On-Line Monitoring Quick Guides (see EPRI Products 3002014054 and 3002014055 and a YouTube video at https://youtu.be/ MGfY2T38-pI). We’ve also developed a common design package for installation of wireless sensors and a wireless network so plants can more efficiently install wireless systems, which are typically less expensive than wired sensors systems (see EPRI Products 3002011820 and 3002011818). 7. Describe some of the work you are doing around automation tasks as part of plant modernization. One repeated theme in modernization is replacing manual time-based tasks with automatic condition-based tasks. One of our research projects is exploring conditioned-based maintenance, but it also applies to water chemistry sampling as well. Many plants are doing water chemistry sampling manually – going out and filling up a beaker, carrying it back to the lab and then doing tests on it. EPRI has constructed several “skids” – or a collection of instruments – which can automatically sample the water and do many of the tests previously done in the lab. The instruments then report the water chemistry condition and send it to the plants databases. This results in obvious efficiencies and savings for the plant. Those skids have been demonstrated at one set of sites already. They’re on their way to a second set of sites that are slightly different reactor designs. In the end, we hope this research will be able to demonstrate that automatically shared data that is transmitted leads to better problem prediction and solutions. There’s a similar theme in radiation monitoring. A lot of radiation monitoring is plant staff going out and doing swipes or carrying around portable meters. The questions we’re asking and researching are whether plants install permanently mounted equipment, and eliminate those tasks. For example, can you replace people taking radiation samples with drones or robots? 8. Concluding remarks. Modernization can equip nuclear plants with updated technologies and improved processes to deliver greater economic efficiency and reliability – helping the future of carbon-free energy generation. To this end, EPRI’s Nuclear Plant Modernization effort through 2021 will provide technology demonstrations, common information models, standardized implementation methods, and tools to inform decisions on plant capital investments. Our final product will be a handbook with case studies that outlines technology business cases, full modernization conceptual design, and an implementation business case. Contact: Christine Rivers, Electric Power Research Institute, 1300 West WT Harris Blvd., Charlotte, NC 28262; telephone: (980) 495-7478, email: crivers@epri.com .