Performance evaluation of shell and tube heat exchanger through ANN and ANFIS model for dye recovery from textile effluents

Abstract

ABSTRACT Performance of Shell and Tube Heat Exchanger for treatment of textile industry wastewater as a secondary treatment process is reported. An Adaptive Neuro-Fuzzy Inference System and Artificial Neural Network models were used for predicting the accuracy of the process. Condensate flow (m3/hr), feed inlet temperature (°C) and steam flow (kg/cm) for textile effluents were used to predict the condensate temperature (°C). The accuracy of the developed model was validated using Absolute Mean Relative Error, Root Mean Square, and Absolute fraction of variance. The different combinations of input parameters were determined with preceding parameters of the model. Comparison of the models indicated the advantage of ANFIS model over ANN model. Seventh configuration values of ANN reported 0.971%, 1.175%, and 0.9456 of the Absolute Mean Relative Error, Root Mean Square, and Absolute fraction of variance respectively. Similarly, the ANFIS reported 0.381%, 0.532%, and 0.9998 of the Absolute Mean Relative Error, Root Mean Square, and Absolute fraction of variance respectively. ANFIS model was found to be more effective in predicting the condensate temperature with a high Absolute fraction of variance R2 = 0.9998 the results were in line with the experimental data. In the present study the complication of nonlinear process has been resolved using ANFIS modeling technique. The treated water could be recycled using this process. Optimization of the process parameters helps in minimizing water consumption and offers economic benefits for wastewater treatment.