Modeling refrigeration cooling power of R2TMO6 oxides for green magnetic cooling applications using intelligent learning computational methods

Double perovskite R2TMO6 oxide belongs to the class of ABO6 perovskite in which T-site consists of transition metal, M-site consists of another metal while the R-site consists of rare earth metals. Rare earth-based oxides lately received significant attention due to their demonstrated interesting physical properties such as spin- phonon coupling, ferroelectricity, multi-ferocity, dielectric, and huge magnetocaloric effect. In addition to the huge magnetocaloric effect, large refrigeration cooling power (RCP) is highly essential in determining and tailoring the candidature of double perovskite R2TMO6 oxide refrigerant for green magnetic cooling applications. This green technology is more efficient, compact, and environmentally friendly as compared to the conventional gas compression/expansion system of refrigeration. This work models RCP of double perovskite R2TMO6 oxide refrigerant through random forest regression (RR) algorithm and extreme learning machine (ELM) with sine (S) and sigmoid (G) activation functions using crystallography (cryst) features of the oxides and ionic radii (radi) of oxide elemental constituents as descriptors. For correlation coefficient (CC), root mean square error (RMSE) and mean absolute error (MAE) performance evaluation metrics, the developed SELM-Cryst model outperforms GELM-radi, RR-radi, and RR-Cryst models with an improvement of 2.11 %, 5.62 % and 12.97 %, respectively for CC yardstick, 184.00 %, 241.24 % and 178.06 %, for RMSE yardstick and 264.19 %, 384.18 % and 240.75 % respectively, for MAE performance yardstick. The developed model further investigates the significance of applied magnetic field on the RCP of different classes of R2TMO6 oxide. The precision of the developed models coupled with the ease of the descriptors would facilitate the exploration of double perovskite R2TMO6 oxide for green cooling applications and ultimately address the current global energy crisis.