A neural network-based prediction of oscillatory heat transfer coefficient in a thermo-acoustic device heat exchanger

The growing electricity demand in the world has brought about significant challenges in her economic development. This concern has prompted the need for electricity generation from different technologies, such as the thermo-acoustic engines. These engines are low-cost electrical power generators. They are alternative and sustainable solutions for electricity generation in developing countries because they generate clean energy. Although the engines have good thermal efficiencies, their oscillatory heat transfer coefficient (O H T C) estimation is often a challenging task. This study, therefore, considers the evaluation of thermo-acoustic engines O H T C using aranciai neural network (ANN) model. The input parameters considered are frequency and mean pressure. Experimental data from literature were used to evaluate different hidden-layer architectures of the network configuration. It was concluded that the best solution was obtained with a root mean square error of 0.64 from a model with 4-10-2 architecture.