Simulation of heat exchanger performance by artificial neural networks

The artificial neural network technique was applied to heat transfer throug h a series of problems of increasing complexity, For the simplest problem o f one-dimensional heat conduction with linear activation function, it is po ssible to give physical meaning to the synaptic weights of the network. A n etwork with sigmoid activation function was used for non-linear representat ion of convection problems where identification of the weights with physica l variables was not possible. Two cases of convective heat transfer with on e and two heat transfer coefficients and artificially, generated data were examined. Finally, the method was applied to the analysis of data obtained in the laboratory, for a single-row, fin-tube heat exchanger. It is shown t hat a better prediction with smaller scatter is obtained in comparison to a conventional power-law correlation for the heat transfer coefficients.