Integrating artificial neural networks, multi-objective metaheuristic optimization, and multi-criteria decision-making for improving MXene-based ionanofluids applicable in PV/T solar systems

• Published in: Scientific reports 2024-11, Vol.14 (1), p.29524-21
• Main Authors: Hai, Tao, Basem, Ali, Alizadeh, As’ad, Sharma, Kamal, jasim, Dheyaa J., Rajab, Husam, Mabrouk, Abdelkader, Kolsi, Lioua, Rajhi, Wajdi, Maleki, Hamid, Sawaran Singh, Narinderjit Singh
• Format: Article
• Language: English
• Subjects: 639/166; 639/301; 639/4077; Artificial neural network; Decision making; Humanities and Social Sciences; Multi-criteria decision-making; Multi-objective optimization; multidisciplinary; Multiple criteria decision making; MXene-based nanofluid; Neural networks; Optimization; Photovoltaics; PV/T system; Science; Science (multidisciplinary); Solar energy; Specific heat; Thermal conductivity
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-024-81044-3

Optimization of thermophysical properties (TPPs) of MXene-based nanofluids is essential to increase the performance of hybrid solar photovoltaic and thermal (PV/T) systems. This study proposes a hybrid approach to optimize the TPPs of MXene-based Ionanofluids. The input variables are the MXene mass fraction (MF) and temperature. The optimization objectives include three TPPs: specific heat capacity (SHC), dynamic viscosity (DV), and thermal conductivity (TC). In the proposed hybrid approach, the powerful group method of data handling (GMDH)-type ANN technique is used to model TPPs in terms of input variables. The obtained models are integrated into the multi-objective particle swarm optimization (MOPSO) and multi-objective thermal exchange optimization (MOTEO) algorithms, forming a three-objective optimization problem. In the final step, the TOPSIS technique, one of the well-known multi-criteria decision-making (MCDM) approaches, is employed to identify the desirable Pareto points. Modeling results showed that the developed models for TC, DV, and SHC demonstrate a strong performance by R-values of 0.9984, 0.9985, and 0.9987, respectively. The outputs of MOPSO revealed that the Pareto points dispersed a broad range of MXene MFs (0-0.4%). However, the temperature of these optimal points was found to be constrained within a narrow range near the maximum value (75 °C). In scenarios where TC precedes other objectives, the TOPSIS method recommended utilizing an MF of over 0.2%. Alternatively, when DV holds greater importance, decision-makers can opt for an MF ranging from 0.15 to 0.17%. Also, when SHC becomes the primary concern, TOPSIS advised utilizing the base fluid without any MXene additive.