A novel computational intelligence strategy has been introduced to optimize ternary hybrid nanofluids, aiming to significantly boost their thermal performance. This innovative approach minimizes dynamic viscosity and maximizes thermal conductivity by adjusting volume fractions, temperature, and mixing ratios of nanomaterials contained within the fluids. The study utilizes advanced computational intelligence methods, encompassing machine learning, multi-objective optimization, and multi-criteria decision-making techniques.
Nanofluids, which are fluids enhanced by the addition of nanoparticles, have garnered considerable interest for their ability to improve thermal properties. Specifically, when these nanoparticles are applied synergistically—such as graphene oxide (GO), iron oxide (Fe₃O₄), and titanium dioxide (TiO₂)—the potential for performance enhancement is significantly amplified. The research highlights how the interplay of various factors can be precisely tuned to yield optimal thermal efficiency.
The performance of these nanofluids is intrinsically linked to their thermophysical properties, which are the focus of the new optimization strategy. Researchers implemented three machine learning algorithms including the GMDH-type neural network, gene expression programming, and the combinatorial (COMBI) algorithm to create predictive models of dynamic viscosity and thermal conductivity based on temperature and input variables. The algorithms performed exceptionally, achieving R² values ranging from 0.99964 to 0.99993, showcasing their accuracy and reliability.
Utilizing multi-objective particle swarm optimization (MOPSO), researchers discovered a viable Pareto front—a range of optimal solutions informing decision-makers of the trade-offs involved between conflicting objectives like maximizing thermal conductivity and minimizing viscosity. The study's results revealed optimal conditions for ternary hybrid nanofluids, highlighting the importance of volume fractions. For example, around 93% of the optimal points for the 111 mixing ratio exhibited volume fractions below 0.5%.
Importantly, both temperature and volume fraction emerged as pivotal variables, with the temperature for optimal points consistently near the upper limit of 65 °C across all mixing scenarios. Such conditions not only exhibit improvements for thermal energy storage systems and electronic cooling applications but are also stated to hold promise within solar energy technologies, indicating broader industrial viability.
The research underlines the significance of adopting hybrid strategies using computational intelligence to address the limitations of traditional methods. The integrated approach allows the modeling of complex, non-linear relationships inherent to these systems, particularly beneficial for the configuration of hybrid nanofluids aimed at enhancing heat transfer capabilities.
"The MOO process indicated optimal points accept a wide range of volume fractions across all mixing ratios, with temperature maintained near maximum limits," observed the authors of the article. Their findings reinforce the robustness of the COMBI algorithm by demonstrating its superior accuracy over other algorithms studied. This predictive capability allows engineers and scientists to develop nanofluids customized for specific applications, enhancing stability and performance under varying operating conditions.
Future research avenues may involve the exploration of novel nanomaterial combinations or the refinement of machine learning algorithms to achieve even higher efficiencies. Integrative strategies could be developed to encompass the growing array of potential applications for ternary hybrid nanofluids, thereby accelerating advancements across energy-efficient technologies.