Hybrid Deep Convolutional Neural Networks Combined with Autoencoders and Augmented Data to Predict the Look-Up Table 2006

Lookup tables derived from empirical research contain limited data for critical heat flux (CHF). Machine learning techniques can be employed to create CHF prediction models by training them with existing experimental-based CHF lookup tables. Consequently, these models can generate CHF look-up tables that provide expected CHF values for a wider variety of scenarios. This study presents a novel predictive method integrating auto-encoders with deep convolutional neural networks (DCNN). The purpose is to obtain precise and reliable predictions for CHF retrieval. The accuracy of these models was evaluated utilizing the coefficient of determination (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>), Nash-Sutcliffe efficiency (NSE), mean absolute error (MAE), and normalized root-mean-squared error (NRMSE). The findings demonstrate that the suggested hybrid model reveals an important degree of accuracy.