Numerical investigation of MXene-based ultrawideband solar absorber with behaviour prediction using machine learning

We have designed a multilayered solar absorber composed of several materials, including tungsten, magnesium fluoride, MXene, silicon, and silver. The structural frame is composed of a resonator constructed from silver (Ag) in the form of Jerusalem. The absorption results are compared with the traditional AM 1.5 data, and the finite element method computational tools are used to analyse the structure of the absorber to investigate the many physical elements that contribute to changes in absorption. Except for the resonator height, this design demonstrates the least amount of variation in absorption when considering all physical characteristics. Hence, the resonator's height is one of the most critical aspects in deciding the final output. The solar absorber achieves constant absorption values for a large incidence angle range (80 degrees). We have also used a prediction model based on the findings of our computational analysis. We used an ANN prediction model for the suggested prediction design, and the results for the 2500 epoch data set were impressive with values of different parameters RMSE = 0.029, MAE = 0.020, MSE = 0.0013, and R2 = 0.96. The accurate model developed in this work may be used to predict the absorption levels for various values of the structure's physical properties. Solar absorber structure design for infrared, visible, and UV light absorption is made easier with the help of the provided findings and predicted model.