Modeling and Prediction of Thermal Properties of Formulated SMAW Coating Flux Using ANFIS Model
摘要
Adaptive neuro fuzzy inference system (ANFIS) models for estimating the thermophysical and physicochemical characteristics of SMAW coating flux have been generated. The coating fluxes were formulated using an extreme vertices design approach. Thermogravimetric analysis and hot disk were used to calculate weight loss and thermal diffusivity. The developed model has been tested for its prediction capability and compared with conventional regression analysis predictions. The ANFIS model of weight loss and thermal diffusivity has shown improvements in root mean square error of 19.62% and 30.72%, respectively. Statistics such as Pearson’s r and Adj R-squared have also shown significant improvement. Slag viscosity was theoretically estimated for developed fluxes since it controls the chemical reaction kinetics during weld. The new optical basicity of the flux was calculated and employed for the estimation of slag viscosity. The slag viscosity of the developed fluxes ranges from 0.1027 Pas to 0.1998 Pas at 1673 K.