QUICK AND RELIABLE ESTIMATION OF POWER DISTRIBUTION IN A PHWR BY ANN

Knowledge of the distribution of power in all the channels of a Pressu rised Heavy Water Reactor (PHWR) as a result of a perturbation caused by one or more of the regulating devices is very important from the op eration and maintenance point of view of the reactor. Theoretical desi gn codes available for this purpose take several minutes to calculate the channel power distribution on modern PCs. Artificial Neural networ ks (ANNs) have been employed in predicting channel power distribution of Indian PHWRs for any given configuration of regulating devices of t he reactor. ANNs produce the result much faster and with good accuracy . This paper describes the methodology of ANN, its reliability, the va lidation range, and scope for its possible on-line use in the actual r eactor. (C) 1998 Published by Elsevier Science Ltd.