Samer Dahaweer
, September 21, 2023
Share this article

Critical systems are essential for society’s well-being. From power generation and distribution, to healthcare, telecommunications, and transportation, these systems are all vital to the health, safety, and welfare of the public. A failure in these systems could result in catastrophic consequences, ranging from loss of life to severe injuries and breaches of privacy and assets. As technology continues to advance, the use of artificial intelligence (AI) in critical applications is expected to revolutionize safety measures and enhance the efficiency and reliability of these systems. Keep reading to learn how.

The growing role of AI in critical applications

Over the next five years, we can expect to witness a substantial increase in the use of AI within critical infrastructure systems. This adoption is driven by the potential to improve safety, reduce human error, automate dangerous tasks, enhance employee monitoring, and detect crimes and fraud. AI offers a versatile set of tools that can be tailored to the unique needs of different critical sectors.

AI and machine learning in nuclear power

One of the most significant advancements in critical applications related to AI involves the integration of machine learning in the field of nuclear power. By combining digital simulations of real nuclear facilities with AI systems, the nuclear industry has the potential to revolutionize complex procedures, improve reactor design, enhance performance, and boost overall safety.

Potential uses for AI and machine learning in CANDU systems

AI and machine learning hold great promise for various applications within CANDU type nuclear power plants (NPPs), offering the potential to enhance safety, efficiency, and decision-making. Here are some potential applications:

  • Verification of refueling activities: AI and machine learning techniques can be employed for verifying refueling activities in CANDU NPPs, with direct applications in nuclear safeguards.
  • Fault prediction: AI-driven fault prediction systems for the primary heat transport system of CANDU-type pressurized heavy water reactors can improve safety and reliability.
  • Flaw detection: Deep learning can be used to automate the detection of flaws in nuclear fuel channel ultrasonic scans.
  • Radioactive substance classification: AI and machine learning can assist in classifying and visualizing radioactive substances for nuclear forensics.
  • Fuel defect localization: Improved online localization of CANDU fuel defects can be achieved using ancillary data sources and neural networks.
  • Ultrasonic inspection analysis: Development of an ultrasonic inspection analysis defect decision support tool can enhance safety measures for CANDU reactors.
  • Predictive maintenance: Predictive maintenance architectures driven by AI and machine learning can ensure the longevity and reliability of nuclear infrastructure.
  • Energy management: AI and machine learning-driven energy management is crucial for optimizing hybrid nuclear-wind-solar desalination plants.

Challenges in AI and machine learning for decision making

While AI and machine learning offer immense potential in enhancing safety and efficiency, there are several challenges associated with their integration into critical applications:

  • Historical data requirements: Machine learning for decision-making may rely on historical data to establish baseline operational decisions. However, this data may not account for unforeseen scenarios.
  • Data quality: High-quality data is essential for training machine learning algorithms. Imperfections in historical data can affect the algorithm's performance and decision-making.
  • Heuristic-based decisions: AI-driven decisions based on heuristics may not always be optimal, raising concerns in safety-critical scenarios.
  • Algorithm adjustments: While it's possible to adjust algorithms for specific scenarios, overreliance on such adjustments may limit the technology's potential.

Ensuring safety and reliability

The integration of AI into nuclear power plants raises a significant question: how can we ensure that these AI-based solutions meet the expected safety and reliability requirements? To address this challenge, it is essential to evaluate the existing software qualification standards in the context of emerging AI-based solutions.

Meeting stringent safety standards

The CSA N290.14-15 standard, developed by the Canadian Standards Association, serves as a guiding document for nuclear power plant design, construction, operation, and decommissioning in Canada. It encompasses various aspects of nuclear power plant safety, including seismic hazard assessments, environmental protection, emergency planning, and radiation protection control systems.  

What sets Alithya apart in the nuclear industry is our team of nuclear experts who played a critical role in developing the CSA standard. With over 30 years of experience, we have been the trusted provider of software qualification and compliance review services for some of Canada's largest power generators. Our extensive expertise has not only shaped the standards but also ensures that NPPs run safely and efficiently.

Embracing innovation

Alithya is dedicated to staying at the forefront of AI safety and compliance, collaborating closely with industry leaders and experts to ensure that new AI advancements are thoroughly vetted against standards such as the CSA N290.14-15 standard before implementation. We have a team focusing on open-source solutions, evaluating their relevance within the Nuclear Power Plant (NPP) context, and identifying any potential gaps in regulating software. Our commitment to industry safety is unwavering, and we take every step possible to maintain the highest standards.

We are here to assist others in their AI journey, offering our expertise to help our clients navigate the complex landscape of compliance. Alithya is your trusted advisor for software qualification and compliance, specializing in guiding our customers through the rigorous demands set forth by the CSA N290.14-15 standard. Whether you require support with software validation, safety assessments, or compliance audits, our solutions are tailored to meet your specific needs.

The way of the future  

The integration of AI and machine learning into critical applications, especially in the context of CANDU nuclear power plants, holds immense promise for enhancing safety, efficiency, and reliability. However, a cautious approach is crucial to identify suitable AI solutions and ensure they adhere to strict, proven standards and processes, such as the CSA N290.14-15 standard. At Alithya, we are excited about the positive impact this will have on the nuclear industry and are dedicated to supporting this transformative journey.

Share this article