Energetic thermo-physical analysis of MLP-RBF feed-forward neural network compared with RLS Fuzzy to predict CuO/liquid paraffin mixture properties

Dynamic viscosity of novel generated Copper Oxide (CuO)/Liquid Paraffin nanofluids is obtained experimentally for various temperatures and concentrations. To optimize the empirical process cost-efficiency, Feed-Forward Neural Networks (FFNNs) were modeled compared with Recursive Least Squares (RLS) Fuzzy model. prepare CuO/ liquid paraffin nanofluids, CuO nanoparticles are dispersed within paraffin. Then an input-target dataset containing 30 pairs available T=25,35,40,50,55,70(∘C), φ=0.1,0.5,1.0,3.0,5.0(%). Based on results, two types FFNNs examined RLSF model to predict CuO/liquid nanofluids. evaluate best optimization methods nanofluid viscosity, Multi-Layer Feed forward (MLF), Radial Basis Function (RBF), discussed. The MLF network provides a global approximation while RBF acts more locally, further, better fit. On contrary, has properties from generalization noise rejection points view. Also, networks can be applied in online manner. Further, three curves RLS by Parabola2D, ExtremeCum, Poly2D models fitted data compared. ExtremeCum showed least margin error employed data.