Design of a meta-heuristic artificial intelligence (AI) model for an optimal photovoltaic module cooling system

The heat absorption and management methods of PV module are very diverse and constantly developing. The advantages and disadvantages of each method imply an efficient method for optimal selection. Therefore, this paper aims to design a multi-objective particle swarm optimization (MOPSO) model to search for a better configuration of cooled PV/T. The algorithm provides a more objective and comprehensive understanding of the trade-offs and advantages associated with different cooling methods. Seven objective functions have been implemented. The Cost of Energy (COE), Net Present Value (NPV), Internal Rate of Return (IRR), Ergonomic Factor (EF) and Payback Time (CPBT) revealed that photovoltaic /thermal (PV/T) systems with hybrid cooling (Passive/Active) with forced convection PCM/Air (Phase Change Materials) are better. Similarly, the evaluation of the total annual cost (TAC) shows that air cooling systems are more economical. On the other hand, the cost evaluation shows $${\text{CO}}_{2}$$ that hybrid cooling systems with NanoPCM/TNF (Ternary Nanofluids) are less polluting. However, the payback time of the cooled PV module unit is less than 4 years and the sensitivity of the savings is more than $20 in only 5 years of life. The MOPSO method deduces that the PV/T systems with hybrid cooling PCM/Air are optimal compared to the hybrid systems TEG/NF (thermoelectric/Nanofluids) which are the least efficient. The proposed algorithm offers a robust and accurate method to optimize PV/T systems, thus outperforming other approaches presented in the literature.