Indian Journal of Science and Technology

Abstract

Objectives: The aim of the present work is to study the application of ANN model on a complex system such as Nuclear Reactor and their adaptation for the early prediction of abnormal behaviour. Methods: Safe and reliable operation of reactor depends upon many parameters and many physical barriers at different layers. Application of ANN model on reactor control process monitoring requires many training samples and effective method for implementation. It requires the identification of possible occurrences of events and mapping of the each event the relevant parameters. Attention on the selection of parameters related to each of the IEs is paramount. Findings: Finding the cause for the transient events is important in every plant and the ANN model trained for all possible transient events is useful for incident analysis as well as avoid reoccurrence of the faults. Applications/Improvement: In order to provide heuristics as well as neural support for the similar NPP, collaborate entire NPP database under one roof with the help of cloud computing. History maintenance for future analysis and knowledge sharing will be feasible. However, the end result shall be confirmed with human verification for final decision. 

Keywords: Cloud Computing, IE, Pattern Recognition, Rule Based System, SVM, Transients