الفهرس | Only 14 pages are availabe for public view |
Abstract The present study aims at confirming the developing role of neural netweoks towards faster response and a broader contribution to our upgrading life. This study depends upon the use of the back propagation network algorithm for the one step DNBR calculation, instead of the three step procedure conventionally used to calculate DNBR. The first step using a system analysis code, secondly a heat flux calculation code and finally a detailed DNBR calculation code should be used. The back propagation network is trained with the DNBE training data using the ’ZEBRA’ computer code. Extension of the code has been done to accept variation in different parameters in the same run so that the effect of these parameters on DNBR can be examined, at the same time, a comparative evaluation of the neural network can be conducted. Results confirm that the back propagation network can reasonably and quickly calculate the DNBR. |