The Use of Artificial Neural Network for Modeling Coagulation of Reactive Dye Wastewater Using Cassia fistula Linn. Gum

Abstract

Natural seed gum extracted from Cassia fistula Linn. (CF) was experimentally evaluated to treat reactive dye (Red 195) in an aqueous solution, whose color and Chemical Oxygen Demand (COD) were to measure the treatment efficiency. To investigate five parameters i.e. pH, reaction time, agitation speeds, dye concentration and CF gum concentration were used to implement a one-factor-at-a-time experiment with Jar-test apparatus. Carried out under weak basic condition (pH 10) for 30 min, the COD and decolorization efficiency of the dye stuff wastewater was observed at 42.4% and 57.8%, respectively. A single-layer Artificial Neural Network (ANN) model was also developed to predict the removal efficiency of the dye by using the determination coefficient (R2) and the root mean square error (RMSE). The observed and predicted outputs were found to be 0.924 and 3.759, respectively. Furthermore, the ANN model was analysed using Garson’s algorithm, connection weight method, and neural interpretation diagram to understand the influence of each operation factor on the treatment process.