Computational Intelligent Systems for Prototype Fast Breeder Reactor

Computational intelligent systems play a major role in enhancing the safe operation of the nuclear reactor. In Prototype Fast Breeder Reactor (PFBR) nearly 15000 process signals are digit ized by physically and functionally distributed embedded systems. Digitized signals are processed and relevant informat ion is displayed through large video display systems at control room. It is necessary that correct and reliable informat ion need to be provided to the plant operator for efficient plant monitoring and control. The paper explains the features of four such systems namely on-line validation of neutronic power channel through on-line thermal balance calculation, detection of anomalous reactivity addition through on-line react ivity balance computation, on-line computation of reactor power from fluctuations of core thermocouple signals and detection of flow blockage in fuel subassemblies.