Modeling of Groundnut Shell Mercerization Process Using a Neuro-Fuzzy Technique

Natural fiber is growing relevant in composite processing due to its low cost, lightweight, and good mechanical properties; therefore, increased natural fiber composite development is desirable. This study predicted the mercerization effect on the moisture absorption properties of groundnut shell samples using neuro-fuzzy modeling. The groundnut shells were processed, dried, and treated with NaOH varying the time and concentration of the treatment. Sensitivity analysis using the adaptive neuro-fuzzy inference system) ANFIS’s exhaustive search showed that treatment time and concentration impacted the moisture absorption rate of groundnut shells. Parametric analysis via ANFIS surface plot indicated that an increase in treatment time and concentration decreased the moisture absorption rate of the samples. The characterization results from SEM(Scanning electron micrograph) and FT-IR (Fourier Transform Infrared Spectroscopy) showed that the groundnut shells were suitably mercerized. ANFIS optimum result showed that the moisture absorption rate of 1.23% was obtained at a treatment time of 4 hours and a concentration of 4 mol; pi membership function (mf) had the best coefficient of determination R2 (0.99364) and Mean Square Error (MSE, 0.011679) amongst other membership functions demonstrating a significant predictive behavior for the model. The observations from the study prove that the ANFIS technique is a practical approach for the prediction of the groundnut shell mercerization process.