Nuclear power sector: also-ran or star turn in the 21st century?
Most people associate the word “nuclear” with two main other things — meltdowns and weapons. Neither of them is going to be the topic of this brief. Despite the intensive research programs funded widely by both the public and private sectors, nuclear incidents still occur. The nuclear power sector has made a strong progress in designing safety in the last few years. The last one happened a few years ago in Japan called Great Tōhoku Earthquake — at the Fukushima nuclear plant, after a magnitude 9.1 earthquake off the coast of Japan. The earthquake generated a tsunami wave at speeds that approached about 800km per hour and 3.5m heights which damaged the construction of steel and concrete storage tanks full of radioactive water. Thanks to the advanced technologies the range of the destruction area was strongly limited. So what makes our power plant safer?
To properly give an answer to this question, we should start from the time when nuclear energy was the future. Currently, there are 186 nuclear plants in Europe (most of them are in France — 58 and Russia — 36) generating 163.68MWe. The development of nuclear weapons had a notable impact on the entire field of nuclear science. The most fascinating and dangerous period of the time was called the “Atomic Age” in the 1950s and 60s. In the aftermath of World War II, many countries inspired by the potential of atomic bombs started developing this science intensively including the United States Army and its Manhattan Project. Since the atomic bombing of Hiroshima (“Little Boy” August 6, 1945) and three days later Nagasaki (“Fat Man” August 9, 1945) nuclear weapons and nuclear science have evolved a lot. Even though the Frisch-Peierls Memorandum (the first technical documentation of possible usage of nuclear power written by German physicists Otto Frisch and Rudolf Peierls in 1940) did not emphasize generating electricity using nuclear energy, in 1951 the first nuclear reactor was built. It was a small experimental breeder reactor called EBR-1 designed and developed by American Argonne National Laboratory based in Idaho, US. What is more, the year 1953 brought a new beginning for the use of atomic power as President Eisenhower proposed the “Atoms for Peace” national atomic program which reoriented significant research effort towards electricity generation and civil usage of this energy. A year after, the Soviet Union (succeeded as the Russian Federation) started using the world’s first nuclear power plant (Atomic Power Station 1 Obninsk) in Obninsk (100km far from Moscow). After more than 48 years, the power plant was closed (2002).
Photograph of the Shippingport Atomic Power Station in Shippingport, Pennsylvania, the first full-scale nuclear power generating station in the United States which began operating in 1957.
The main parts of every power plant, either in the 1950s or the 21st century, are automation and IT support systems. Automation is something we come across every day. From a coffee express button to advanced software products, automation has been giving you better performance on a daily basis. The need to improve performance and safety has led to increased use of automation. Moreover, the digital revolution is leading power plant designers, through economic and performance initiatives, to increase the range of automation. At the same time, business wants to continually increase general operational efficiency. More automation is not the complete answer to this business challenge however it can improve the process of improving safety within nuclear power plants and other complex industrial facilities.
IT and automation at nuclear power plants 700
Although nuclear power plants are designed to exacting standards including international regulations approved by the International Atomic Energy Agency (IAEA) using thorough quality assurance systems, power plants cannot be made one hundred percent perfect. In reality, the extent to which a power plant can be analyzed is limited. It can give an impression that there is a continuing role for an operator and human existence within a plant of the future. As automation can take over most repetitive tasks, the role of the human becomes that of an automation manager — a person who manages the plant. Of course, automation has led to decreasing costs, improving safety and increasing the control, efficiency, and scale of work within a nuclear power plant. Though a man has to initiate the entire system, once the system is being operated, the next steps are almost automated in order to keep the quality at the highest level. To meet this challenge, the global trends are to simply add software products and automation systems. The fact is that the use of computers in nuclear power plants is increasing rapidly. But currently, according to global statistics, the industry has lagged behind other industries in the application of new technologies.
Generally, it’s important to consider also the fact that automation is related to both machines and technology in this case. Despite this, the main point is how to achieve the optimum balance between man-machine interaction. Unfortunately, software products can be encumbered with bugs and glitches which make them among the least reliable. A nuclear power plant has many critical components which, for known reasons, have to work stably and continually. That’s why there are many backup systems behind the general solution. Once it stops working, an alternative platform can be turned on. In other words, such systems must work when needed; that’s why in the most critical parts of the system, an operator is in charge of monitoring the status of machines, not an automation system.
The control room of Dresden Station — The nation’s first full-scale, privately financed atomic power facility, the Dresden Nuclear Power Station, is owned by the Commonwealth Edison Company and was built by the General Electric Company.
Modern computers help automate and optimize processes throughout a nuclear power plant, ensuring increased efficiency. However, the ultimate goal is to maximize safety, without compromising quality and reliability. When the first versions of nuclear systems were written some 25 years ago, no quality standards were available. Nowadays, because of cyber attacks and other threats, the systems have to meet numerous requirements if they are going to be used at a power plant. The certification of a software product is a standard procedure which is used by all IT companies globally. Most known certification models are defined by ISO itself (ISO — International Organization for Standardization body which works in 162 countries). However, because the process is very rigorous and really expensive, it is needed mainly for the most critical system components. What is more, IT companies have to provide high-quality technical maintenance services which meet both regulation and customer requirements. The software products’ failures are completely different from the hardware solutions. Hardware technologies can be verified by different online testing mechanisms against any failures. Software errors and defects are not so easy to be found through testing processes. The digital systems are vast and complex. Failures of those systems are difficult to intuitively check and correct, which may affect the general performance. The source codes of those technologies, consist of hundreds of thousands of lines developed a long time ago. Even if the software management systems have been verified for decades, they still have new bugs.
Data science in the nuclear energy sector
With the rapid growth of information and digital systems (e.g. ICT), data has been increasing rapidly. It’s the most important element of the big data model which derives values from data. Although data science has been applied in many industries according to statistics published by the company Gartner, it is difficult to find relevant information and applications of big data usage in the nuclear energy sector. Especially, the process of equipment digitalization and usage of information systems are not widely used in power plants. The main challenge is the perception of the adoption of big data analytics by the energy sector, particularly power plants still have an indefinite point of view on the big data concept. Basically, it is a good approach for new startups to figure out how to apply a big data model effectively in nuclear power plants.
A big opportunity for data science is within health and performance monitoring management systems. Diagnosing potential problems at nuclear plants can take significant time and plant resources. While a big amount of data is available now through online monitoring platforms, separating the relevant data is not an easy task. Not many companies are able to do it correctly, and even less can generate an additional value for a customer. For instance, there are a few companies which offer a wide range of services within predicting failures and checking asset health status. The main point of those software products is to prevent critical asset failures while optimizing asset lifecycle costs. In other words, the software is able to prioritize and resolve risks (by efficient maintenance service planning) before they materialize. By reducing the ineffective time and expensive shutdowns a company is able to reduce the general maintenance costs and create more efficient compliance management through lean workflows. Many of the nuclear reactors are going to turn 20 years old and because of this, the costs of maintaining inventory are significant. Being able to predict correctly when a component requires maintenance, gives opportunities to cut down on the downtime of a reactor and help with other expenses.
Apart from this, the nuclear industry is suffering from a massive skills gap as much of the workforce retires. New technologies, both automation and software, can be an alternative way to handle this issue. Deploying wireless sensors and automated monitoring, nuclear plants can reduce not only labor costs but also cope with the labor shortage. Another example of a software product is an interactive application with 3D graphics technology allowing workers or operators to complete maintenance and inspection tasks more effectively, improving both plant safety and reliability.
Over the past, 10 or even more years automation platforms have progressed from primary hardware and software designs to systems that maximize the use of industry-standard hardware and software. The typical power plant automation system is populated with an enormous number of modules that meet strict standards for isolation and environmental specifications. One of many advantages of automation implementation is the possibility to move from control processes to supervision. Currently, more and more operation teams are moving from control processes which put more pressure on automation systems to not only control them but also provide real-time information to all parties concerned with the power plant condition. What is more, technology can enable personal interaction with third-party experts in remote locations who have real-time access to information from a plant process.
The plan for the future is not to control a simple process but supervise it. The plant business processes such as operation, maintenance and management will be integrated across all plant functional models and data will be available from each part of a plant. The system will provide the latest data and multiple communication channels. Flexible and interactive interfaces will be the core of the automation system. Additionally, collected information has a significant contribution to the process control, supporting cooperative works. This study can give some suggestions on how to design future systems or reorganize the current business processes with the practice of users. In fact, integrating design thinking and data science can impact business initiatives and can give a list of amazing ideas on how to improve the efficiency of a power plant. The future automation system will not be only able to collect current data but also store past information which could be used to recheck the reasons why some incidents happened.
Although automation is widely used in nuclear power plants, it has many limits. It should be also mentioned that the nuclear power sector has downsides itself as well but it will not be the consideration of this article. Despite the social benefits that may result from deploying automation technologies, in many cases a worker whose job has been taken over by a machine undergoes a period of emotional stress (retraining processes also take time). Moreover, automated equipment includes a high capital expenditure required to invest in automation (automated systems can cost millions of dollars to design, pilot and install). These dangers aside, automation technology if used wisely and effectively, can relieve humans from repetitive, hazardous and unpleasant workplaces in all forms.
Summary
Definitely, the nuclear power sector has the future. About 11% of the world’s electricity is generated by about 450 nuclear power plants with about 50 reactors under construction now, equivalent to 16% of existing capacity, while an additional 160 are being planned. There is a clear need for new generating capacity around the world, both to replace old fossil fuel units and increase the usage of clean and reliable energy technologies reducing carbon dioxide air pollution. Many countries still invest a lot of money in their nuclear power sectors. Adding data science solutions including machine learning and artificial intelligence, the lifecycle of nuclear reactors can be extended easily from 30–50 years to even more.
Proactive and predictive maintenance services have become common practices in many industry sectors. In the nuclear power sector, the maintenance strategies have also continuously advanced — evolving from a reactive to a predictive maintenance model. Condition monitoring can be a great opportunity for the nuclear power industry to identify the root cause problem. What is more, deploying data science technologies can also improve the safety and performance of a nuclear power plant. Big data management can enable better decision-making processes based on data. Using advanced analytics tools and machine learning to develop forecasts for asset availability and behavior and design different parameters and indicators can be one of many answers to current market challenges. As long as commercial and operation optimization is concerned and recognized as one of the biggest challenges, data collection can deliver knowledge and practice from the fields of operations, developing new frameworks for deterministic optimization.
In general, automation and software technologies can bring significant change in the nuclear power plants, helping to mitigate risks in some points and increasing it in others. A particular risk to consider is a cyber risk. According to international regulations, the most critical components of a power plant system have to be separated from not only other systems but also from the Internet completely. It can give almost one hundred percent confidence that no one is able to enter and damage the system from outside. Because many of the plants are still operating with analog control and safety systems, they are even less vulnerable to cyber attacks. Responding to this growing threat is not easy. Today’s strategy isn’t sufficient with the current nuclear power business models. International organizations including business partners have to accelerate efforts to develop new approaches that can scale to the threats of the future.
The future role of nuclear energy is attracting new attention. Once, the innovation is welcome widely in the industry, a new beginning will start. Unfortunately, nobody can say for sure which technologies are best suited to solve internal industry challenges addressed in this brief. Indeed, nobody knows whether there will be a single technology which can optimize and protect nuclear power plants effectively. Business has a clear choice to make — whether to stay on its current path with analog solutions becoming an also-ran on the global market or being a star turn delivering world-class value.
Written by Lukasz Kudlak
