Specific heat capacity prediction of hybrid nanofluid using artificial neural network and its heat transfer application
Sahatsawat Seawram, Prathana Nimmanterdwong, Teerawat Sema, Ratchanon Piemjaiswang, Benjapon Chalermsinsuwan
Abstract
Currently, many industries aim to improve the heat transfer system to reduce cost and carbon emission. For improving heat exchanger apparatus, the convectional working fluid such as water can be replaced by hybrid nanofluid. The hybrid nanofluid is a suspension of nanoparticles into conventional base fluids to improve its properties. Generally, different combinations of nanoparticle and base fluid can variably enhance the specific heat capacity for removing heat from the system. In this study, the artificial neural network was thus employed to develop the prediction model for predicting specific heat capacity using feedforward and cascade forward propagation networks with Levenberg–Marquardt learning algorithm. The best artificial neural network (ANN) model topology was selected to predict specific heat capacity of hybrid nanofluid with R value of 0.9919, 0.9473 and 0.9673 for training, validating and testing, respectively. Then, the specific heat capacity value from ANN model was applied to evaluate and analyze the heat transfer rates. The hybrid nanofluid with Al2O3 and CuO in water exhibited the best heat removing medium in heat exchanger application.