Dynamic viscosity modeling of nanofluids with MgO nanoparticles by utilizing intelligent methods

Nanofluids, as the fluids with modified thermal characteristics, are usable for improving the efficiency of renewable energy systems and enhancement of thermal management. Dynamic viscosity is one of the most significant features of fluids involved in their heat transfer and flow characteristics. This property of nanofluids is under effect of various elements and their modeling is somehow complex. Intelligent methods, regarding their outstanding performance in modeling complex problems, are attractive tool for this purpose. Due to the importance of dynamic viscosity, it is necessary to have accurate model for prediction of this property of nanofluids. Nanofluids with MgO nanoparticles have shown significant performance in thermal mediums and energy devices. In the present article, two intelligent approaches including Adaptive Neuro Fuzzy Inference (ANFIS) and Group Method of Data Handling (GMDH) are made use for modeling dynamic viscosity of nanofluids with MgO nanoparticles. Comparing the determined values by the intelligent methods and the actual quantities revealed significant function of the applied approaches. Among these two methods, using GMDH is preferred regarding its more exactness based on the statistical criteria. R2 of the mentioned methods based on GMDH and ANFIS were 0.9996 and 0.9616, respectively. Average Absolute Relative Deviation (AARD) is another criterion used for comparison of the proposed models that was around 5.13% and 19.41% for the mentioned models, respectively. According to these values, it is concluded that employment of GMDH is preferred in term of exactness.