Response surface methodology and improved neural network coupled with optimized pulse electrochemical oxidation for rhodamine B degradation

Abstract

To address the challenges of dye pollutant removal and high energy consumption in industrial wastewater treatment, a novel boron-doped diamond (BDD) electrode bidirectional pulsed alternating electrochemical oxidation (BDD-PAEO) system was developed. Key operational parameters (current density, initial pH, and reaction time) were optimized via response surface methodology (RSM) to explore their combined influence on Rhodamine B (RhB) removal efficiency (Re) and Electrical energy consumption (EEC). To enable accurate process prediction and control, an enhanced intelligent model based on improved particle swarm optimization coupled with a backpropagation neural network (improved particle swarm optimization - back propagation (IPSO-BP)) was developed. Under optimized conditions (current density: 13 A/m², pH: 1, reaction time: 66 min), the system achieved a 98.870% removal of RhB with a minimal energy input of 0.124 kWh/m³. The IPSO-BP model demonstrated strong predictive performance, with a root mean square error (RMSE) of 6.26 and a determination coefficient (R²) of 0.96, outperforming traditional models in both nonlinear fitting accuracy and generalization capacity. Compared with conventional direct current systems, the bidirectional pulsed alternating mode approach lowered EEC by 44.69% and enhanced current efficiency by 90.30%. Mechanistic studies confirmed that RhB degradation primarily proceeds via hydroxyl radical (·OH)-driven heterogeneous oxidation and direct electrochemical combustion at the electrode surface. This advanced electrochemical strategy offers a promising and energy-efficient pathway for treating refractory industrial wastewater, with practical implications for improving water quality and supporting sustainable development.