Centre for Research in Computer Science and Applications
Application of artificial intelligence to improve the safety of nuclear power plants
Improving nuclear power plants' availability, reducing the cost of operation, helping the decision-making in the nuclear control room, and assisting the preparedness of different accident levels
It is very challenging to use artificial intelligence techniques to identify accidents that happen very quickly, such as large breaks of coolant pipes. It is even more challenging to identify the minimum length of the data and use such a minimum data to find an accident model. Our research activities have included close collaboration with both expert academics and Nuclear Power Plants (NPPs).
Safety is an important aspect to be considered in a nuclear reactor. Maximum care is exercised to keep the likelihood of potential risks to a very low value. However, in the event of such an unlikely occurrence, operators must take necessary actions relatively faster, which involves complex judgements, making trade-offs between partly incompatible demands and requires expert opinion.
Timely and correct decisions in these situations could either prevent an incident from developing into a severe accident or to mitigate the undesired consequences of an accident.
Our work is to develop a system, which assists operators in identifying an accident quickly using artificial intelligence that diagnoses the accidents based on reactor process parameters, and continuously displays the status of the nuclear reactor.
BARC’s diagnostic system is an operator support system for accident management. It performs well for small breaks of inlet headers of nuclear reactors, but has difficulty to predict the large breaks. Our work helps improve the detection for large breaks.
Our work has produced predictive models for mitigating the risk of loss of coolant accidents within nuclear power plants, which can have the potential for critical consequences. New predictive techniques have been developed based on the analysis of data. We developed an online monitoring system for loss of coolant accidents to enhance the safety of nuclear power plants, to reduce risk through early failure prediction.
The project exploits the use of information (through monitoring), artificial intelligence and signal processing to enhance the safety of nuclear power plants. The systems act as early warning devices to facilitate emergency preparedness, prevent accidents from occurring and predict potential loss of coolant accidents. They also monitor accident progression at nuclear power plants, predict the onset and evolution of an accident, and support operators in their decision-making process. The developed support system will maintain plant availability and reduce accident-handling costs in nuclear power plants.
As part of dissemination activities, we ran an end SMART project workshop to articulate the importance and outputs of the research. We have attended conferences in the US, UK, France and China, which attracted the major players in the field, from both the academic and industrial world, and also those with safety-related interests.
Outputs and recognition
- D. Tian, J. Deng, G. Vinoid, T. V. Santhosh, H. Tawfik, A constraint-based genetic algorithm for optimizing neural network architectures for detection of loss of coolant accidents of nuclear power plants, Neurocomputing, Vol. 322, pp. 102-119, 2018
- D. Tian, J. Deng, G. Vinoid, T. V. Santhosh, H. Tawfik, A neural networks design methodology, for detecting loss of coolant accident in nuclear power plants, Applications of big data analytics: trends, issues, and challenges, pp.43-61, ISBN 978-3-319-76471-9, Springer, 2018.
- T. V. Santosh, G. Vinod, P. K. Vijayan, J. Deng, PCA based neural network model for identification of loss of coolant accidents in nuclear power plants, Technology for Smart Future, pp. 345-354, ISBN 978-3-319-60136-6, Springer, Sep., 2017
- D. Tian, J. Deng, G. Vinoid, T. V. Santhosh, A constraint-based random search algorithm for optimizing neural network architecture and ensemble construction in detecting loss of coolant accidents in nuclear power plants,11th International Conference on Developments in e-Systems Engineering, Cambridge, 2-5 September, 2018, Cambridge, UK.
- D. Tian, J. Deng, G. Vinoid, T. V. Santhosh, Selecting a minimum training set for neural networks using short time Fourier transform in detecting loss of coolant accidents in nuclear power plants, The 2018 World Congress in Computer Science, Computer Engineering, & Applied Computing, 30 July-02 Aug. 2018, Las Vegas, USA
- D. Tian, J. Deng, E. Zio, F. Maio, F.Liao, Failure modes detection of nuclear systems using machine learning, The fifth International Conference on Dependable Systems and Their Applications, 22-23 September 2018 , Dalian, China
- D. Tian, J. Deng, G. Vinoid, Santhosh, et al., Identification of loss of coolant accidents of nuclear power plants using artificial neural networks, IAEA Fourth International Conference on Nuclear Power Plant Life Management, Lyon, France, Oct 2017
- G. Colantuono, J. Deng, G. Vinoid, Santhosh, et al., Principal-component-based detection algorithm for fault detection, IAEA Fourth International Conference on Nuclear Power Plant Life Management, Lyon, France, Oct 2017
- T. V. Santosh, G. Vinod, P. K. Vijayan, J. Deng, PCA based neural network model for identification of loss of coolant accidents in nuclear power plants. The international SEEDS Conference (Sustainable, Ecological, Engineering and Design for Society), Leeds, Sep. 2016
Study With us
Study for a research degree at Leeds Beckett and you'll join a thriving academic community in an inspiring and supportive environment. The Graduate School supports our increasingly active postgraduate research community and encourages students to make a difference to the university’s research culture and environment.
research with us
Leeds Beckett University can conduct research on your behalf to help you to implement change and realise your business potential. Validating your ideas with academic evidence can be an essential part of winning contracts and fuelling business growth.