Artificial neural network-Genetic algorithm optimized graded metal foam

The effective utilization of renewable energy directly correlates to efficient performance of integrated energy storage system. The efficiency of Thermal Energy Storage (TES) systems strongly correlates with the heat transfer rate during phase change processes. Heterogeneous metal foams which allow augmenting heat transfer without compromising the thermal capacity suffers from high manufacturing cost. A series of homogeneous metal foams stacked is employable to reap advantages associated with heterogeneous metal foam at a similar manufacturing cost as homogeneous metal foam. The present study focuses on optimizing the porosity distribution of graded metal foam to augment the performance of the TES system. The unidirectional phase change process, observed particularly during solidification, acts as a bottleneck. Therefore, the present study focuses on augmenting unidirectional phase-change processes. The present study discusses the influence of PCM and metal foam thermophysical properties and operational characteristics of the TES system on heat transfer augmentation in different graded metal foams. A numerical model is developed to quantify the influence of graded metal foam on heat transfer rate, which is used to train an Artificial Neural Network (ANN) model. The porosity distribution has been optimized using a Genetic algorithm, which employs the ANN model to ascertain heat transfer rate for different graded metal foams. The optimal distribution has been found to be 93.7% and 96.3% for two-layer graded metal foam and 92.8%, 95%, and 97.2% for three-layer graded metal foam, which augments heat transfer by 1160% and 1185%, respectively.