A new method of black-box fuzzy system identification optimized by genetic algorithm and its application to predict mixture thermal properties

Purpose This paper aims to present a black-box fuzzy system identification method coupled with genetic algorithm optimization approach to predict the mixture thermal conductivity at dissimilar temperatures and nanoparticle concentrations, in the examined domains. Design/methodology/approach WO3 nanoparticles are dispersed in the deionized water to produce a homogeneous mixture at various nanoparticles mass fractions of 0.1, 0.5, 1.0 and 5.0 Wt.%. Findings The results depicted that the models not only have satisfactory precision, but also have acceptable accuracy in dealing with non-trained input values. Originality/value The transmission electron microscopy is applied to measure the mean diameters, shape and morphology of the dry nanoparticles. Moreover, the stability of nanoparticles inside the water is evaluated by using zeta potential and dynamic light scattering (DLS) tests. Then, the prepared nanofluid thermal conductivity is presented at different values of temperatures and concentrations.