How do nuclear power plants make use of large data?

Thanks to the development of information technology, more and more machine parts ' health and performance can be monitored during operation. The large amount of data generated by such sensors requires not only complex computations but also extensive industry experience. The American Institute of Electric Power (EPRI) has implemented a new demonstration project aimed at combining the two.

It takes a lot of time and resources to diagnose the potential problems of nuclear power plants. While a large amount of data can now be obtained through online monitoring systems and other performance assessment devices, it is not easy to distinguish between pros and cons and to identify critical or emergent problems.

EPRI is developing a predictive and health management software tool to support nuclear power plants to achieve these goals. For example, an abnormally high temperature reading of a pump or motor bearing may be the first indication of a device problem, but extensive damage can be avoided if the problem is corrected in a timely manner. The software automatically identifies the high temperature readings and alerts the user to a list of possible problems that trigger the condition. The software then obtains additional information from the user to determine which issues need to be corrected.

Omni-Directional predictive and health management (FW-PHM) tools

A software tool called omni-directional Predictive and health management (FW-PHM) that integrates a wide range of technologies to enhance monitoring, including online advanced recognition patterns, automated diagnostics, and residual service life (Rul) algorithms. The software tool was tested in several nuclear power plants in 2013, proving it could shorten the time to diagnose problems and allow system engineers to focus on solving problems. The full package will be used more widely in the first half of this year.

Monitoring techniques, such as FW-PHM software, are at the forefront of advanced industrial management information systems. Based on the so-called "Big data" analysis, industrial monitoring technology automatically provides information for nuclear power plant personnel to improve the reliability, efficiency and productivity of nuclear power plants.

The increase in industrial and economic productivity over the past 30 years has been based on the development and application of digital information and control technologies. With the ageing of nuclear power plants, the nuclear power industry is slow to adopt these technologies. Its current business model and hardware facilities are in the 80 's, and if the industry does not adopt an efficient business model, it is difficult to keep pace with competitors. Monitoring technology has been highly cost-effective, delaying the adoption of a potential increase in operating costs and limiting the long-term economic viability of nuclear power plants.

The FW-PHM software suite makes it more sensible for the plant personnel to operate and maintain the equipment. It can:

Reduce equipment failure

Factory personnel can anticipate and avoid problems that may limit the production or shortening of plant life.

Improve reliability and productivity

This tool helps the nuclear power plant develop a fail-safe plan that can be used until the failure can be remedied. As it helps guide the troubleshooting process, the tool continues to ask questions as new information emerges, helping plant engineers decide what is the next most important information to collect.

Create a more formal industry-wide diagnostic experience record

When systemic causes are identified, nuclear power plants can change their response so that problems do not recur.

Expand your knowledge

The system is "learning" through continuous database development each time it is used. Although each nuclear power plant or utility company has its own monitoring center, EPRI will collect new troubleshooting information from the experience of the nuclear power plant, compile it and provide it to all users.

Promote the health of nuclear power plants

By acquiring the knowledge of experienced engineers-many of whom have been running nuclear power plants safely and reliably for decades-and may be considering leaving the workforce, tools can integrate the knowledge and experience of plant employees to help provide ongoing equipment health management.

Limitations of existing technologies

In many industries, there are already good precedents: the provision of timely corrective or preventive maintenance of monitoring information to avoid production losses or catastrophic accidents. The application of nuclear power plants, engineering efforts to assess anomalies and provide a useful diagnosis of real aging conditions, has become an obstacle in the expansion of technology applications. Epri admits that an automated diagnostic and prognostic advisor is needed to capture anomalies and provide diagnostic information immediately before the project enters the process.

In addition, most nuclear power plants do not have an automated learning structure to capture the acquired experience while overhauling the fault and then solving the problem. As a result, some industry problems have gone through the same lengthy diagnostic process.

Providing a risk-telling and maintaining a useful life expectancy will allow the plant to prioritize maintenance activities. If automation is properly implemented, this capability will fill the gap between the need for timely and accurate performance analysis and the availability of plant engineering resources. The project focus will shift from identifying potential problems, diagnosing asset failures, and providing maintenance guidance to accurate solutions. Valuable operating assets can avoid premature or untimely failure, or avoid life-limiting mode of operation.

(Responsible editor: Mengyishan)