Review of pool boiling improvement with additives and nanofluids utilizing artificial intelligence

In recent years, there has been a lot of interest in improving nuclear pool boiling by changing fluid properties. The methods for improving surfactants, polymer additives, and nanofluids (NFs) were extensively reviewed in this article. A thorough analysis is conducted on each method, considering its effects on the critical heat flux (CHF) and nucleation heat transfer coefficient (NTTC), as well as the proposed methods for enhancing heat transfer, and the prediction models. One advantage of heat transfer is the ability to incorporate polymers designed for specific uses, but adding a surfactant to the base fluid changes the nucleation area of the boiling curve to have lower surface superheats, which means boiling starts earlier and the nucleation heat transfer coefficient is better. There are conflicting findings on the effect of NFs on the nucleation heat transfer coefficient, even though CHF increases significantly with most NFs. The diverse results are a result of the complex interplay between the starting surface roughness, nanoparticles (NPs), and the base fluid. Notwithstanding the prospective advantages of NF heat transfer, several significant practical issues must be meticulously evaluated before the use of NFs in real-world cooling applications. This paper reviews works on boiling pool heat transfer modeling using machine learning techniques and presents their conclusions. Based on the research that was considered, intelligent methods have the potential to provide accurate predictions of pool boiling heat transfer, with an R2 value of about 0.99 in some cases. In addition, situations involving NFs or porous media may have their heat transfer predicted using these approaches. Several aspects, such as the inputs used, methodology employed, and functions used, impact the accuracy and usefulness of these models. Employing the suitable technique with ideal parameter values for the respective intelligent approach results in enhanced accuracy in modeling.