Predictive Modeling of pH and Viscosity in Emulsion Paint Production Using Adaptive Neuro-Fuzzy Inference System (ANFIS)

Description

Predictive Modeling of pH and Viscosity in Emulsion Paint Production Using Adaptive Neuro-Fuzzy Inference System (ANFIS)

By Akpan Ima-Obong Joseph, Aririguzo Julian Chika and Amaghionyeodiwe Cyril Anosike

This study investigates the predictive capability of an Adaptive Neuro-Fuzzy Inference System (ANFIS) in modeling the pH and viscosity of emulsion paints. Eighteen paint samples were produced using standard formulation protocols, with
three levels of mixer speed (200, 300, and 400 rpm), ambient temperatures (18°C, 23°C, and 27°C), and storage durations before use (12 and 24 hours). Analysis of Variance (ANOVA) was employed to assess the impact of these parameters on pH
and viscosity. Results indicated an average pH of 8.0 and a mean viscosity of 3.22 poise, consistent with industry benchmarks. While the combination of variables significantly influenced viscosity, it had a limited effect on pH prediction. The
ANFIS model demonstrated superior performance over linear regression, achieving root mean square errors (RMSE) of 7.47 × 10⁻⁶ for pH and 3.81 × 10⁻⁶ for viscosity.

These findings underscore the model’s effectiveness in capturing nonlinearrelationships in paint production.