DOI: 10.1615/ICHMT.2009.CONV
ISBN Print: 978-1-56700-261-4
ISSN Online: 2642-3499
ISSN Flash Drive: 2642-3502
CRITICAL HEAT FLUX PREDICTION IN SUBCOOLED BOILING REGION USING ARTIFICIAL NEURAL NETWORK
ABSTRACT
The critical Heat Flux (CHF) is an important parameter for the design of nuclear reactors. Many theoretical and experimental methods of prediction were performed in the last decades. The aim of this study is the implementation of a technique based on artificial neural networks (ANN) of multi-layer type (MLP) for prediction of the critical flux heat in vertical uniformly heated tube using experimental databases. In this way, we have developed one neural model. A set of 2520 databases for subcooled boiling region was selected to train the ANN. 42 hidden layers neurons are used to computation, learning and linearization. MLP predicts CHF with absolute average error (AAE) of 0.30%. Finally, we have made a comparative study between neural networks and model proposed by the literature. This study have revealed in one part the precision of the results obtained and in the other part the efficiency of the conceptual model for the calculating of critical heat flux.