Influence of different clustering techniques of ANFIS on biomethane yield optimization and prediction from eutectic solvent pretreated groundnut shells

This study investigated the influence of different clustering techniques of the Adaptive Neuro-fuzzy Inference System (ANFIS) on optimizing and predicting biomethane yield from eutectic pretreated groundnut shells. Groundnut shells were pretreated at different conditions before anaerobic digestion. Three ANFIS clusters, Grid-Partitioning (ANFIS-GP), Subtractive-clustering (ANFIS-SC), and Fuzzy c-means (ANFIS-FCM), were used to predict the biomethane yield using the experimental data. Input process parameters of temperature, retention period, and pretreatment conditions with biomethane released as the outputs. It was discovered that clustering techniques significantly influence the performance of biomethane yield prediction from eutectic pretreated groundnut shells. All the cluster techniques considered have high accuracy in predicting biomethane yield with a correlation coefficient (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>) of more than 90%. The biomethane generated from feedstock pretreated at lower temperatures was predicted more accurately. ANFIS-GP and ANFIS-FCM predicted the biomethane yield of feedstock pretreated at higher temperatures more accurately. This study has shown the importance of clustering techniques in modeling biomethane from pretreated lignocellulose feedstocks.