Deep neural network model for predicting thermal-hydraulic performance of a solar air heater with artificial roughness: Sensitivity, generalization capacity, and computational efficiency

• DOI: 10.1016/j.psep.2024.05.133
• Publication Date: 2024-06-02T14:13:11Z
• Authors: Siddharth Suman; Sarvapriya Singh; Santanu Mitra; Manish Kumar
• Keywords: Softmax function

A novel coupled computational fluid dynamics-deep neural network approach is proposed to accurately predict the performance of an artificially roughened solar air heater. Computational dynamics data sets are used develop optimized model for prediction thermal-hydraulic factor. The with softmax transfer function in hidden layers has architecture 3-37-37-1 and it predicts factor heater maximum, mean, minimum error 2.54%, 0.245%, 0.0009%, respectively. concept Shapley values adopted evaluate global sensitivity analysis deep found that rib height most influential design parameter generalization capacity evaluated by comparing its predictions experimental data, demonstrating superior predictive over traditional simulations. artificial achieves a time gain approximately 100% against dynamics.