Neural Network Model for an Event Detection System in Prototype Fast Breeder Reactor
Subhra Rani Patra1, R. Jehadeesan2, T.V. Santosh3, T.Jayanthi4, S.Rajeswari5, S.A.V. Satyamurty6, M. Sai Baba7
1Dr. Subhra Rani Patra, Department of Computer Division, Electronics and Instrumentation Group, Indira Gandhi Centre for Atomic Research Kalpakkam 603102, (Tamil Nadu) India.
2Sri R. Jehadeesan, Department of Computer Division, Electronics and Instrumentation Group, Indira Gandhi Centre for Atomic Research Kalpakkam 603102, (Tamil Nadu) India.
3Sri. T. V Santosh, Reactor Safety Division, Bhabha Atomic Research Centre, Mumbai (Maharashtra), India.
4S.Rajeswari, Department of Electronics and Instrumentation Group, Indira Gandhi Centre for Atomic Research, Kalpakkam (Tamil Nadu), India.
5S.A.V. Satyamurty, Department of Electronics and Instrumentation Group, Indira Gandhi Centre for Atomic Research, Kalpakkam (Tamil Nadu), India.
6M. Sai Baba, Department of Electronics and Instrumentation Group, Indira Gandhi Centre for Atomic Research, Kalpakkam (Tamil Nadu), India.
7T.Jayanthi, Department of Electronics and Instrumentation Group, Indira Gandhi Centre for Atomic Research, Kalpakkam (Tamil Nadu), India.
Manuscript received on 20 January 2014 | Revised Manuscript received on 25 January 2014 | Manuscript published on 30 January 2014 | PP: 105-111 | Volume-2 Issue-6, January 2014 | Retrieval Number: F0932012614/2014©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: System failures are identified and quantified by modeling artificial intelligent systems using the required process parameters that cause the failure. In this paper, an artificial neural network (ANN) model has been implemented for detection of various events in Prototype Fast Breeder Reactor (PFBR). Using the conventional, in-house developed thermal-hydraulics model of PFBR operator training simulator, input data has been generated to train the ANN model for various events associated with PFBR subsystems. The subsystems considered here are Primary Sodium Circuit and Neutronics system of PFBR. Operators have to take immediate actions in order to tackle the unsought occurrence of events due to mechanical and electrical failures, thereby ensuring the safe operations of the power plant. In those scenarios, neural network serves as a useful tool in identifying the events at the early stage of their occurrence. The artificial neural network (ANN) models developed here are able to identify the events quickly as compared to the conventional methods. Various learning algorithms based on back propagation network has been successfully applied to the neural network model and the network has been fine tuned towards detecting the events accurately. The resilient back propagation algorithm shows better results compared to other variants.
Keywords: Nuclear Power Plant, Event Detection, Prototype Fast Breeder Reactor, Neural Network, Back Propagation Network.
Scope of the Article: Neural Information Processing