Palm Biodiesel as Flotation Collector for Carbon Recovery from Blast Furnace Sludge: Box–Behnken Optimization and Sensitivity Analysis Using Neural Network

Abstract

Blast furnace sludge (BFS), a carbon-rich waste generated in steelmaking, poses significant environmental challenges while containing valuable recoverable carbon. Motivated by the need for sustainable waste management and resource recovery, this study presents a flotation-based approach to extract carbon from BFS using a novel palm oil-based biodiesel (POB) as an eco-friendly collector. Spectroscopic characterization (Fourier transform infrared [FTIR] spectroscopy and gas chromatography–mass spectrometry [GC–MS]) confirmed the presence of surface-active functional groups that promote selective flotation of the carbonaceous fraction. Key process variables, including cell revolutions per minute (RPM), collector dosage, and frother dosage, were optimized using a Box–Behnken design (BBD). Under optimal conditions, the process increased carbon content in the concentrate from 24.92% to 35.60%, achieving total recovery of 54.79%. The recovered carbon could complement the carbon requirement in steel plant iron-making operations, reducing reliance on external carbon sources. Additionally, a feedforward backpropagation (FFBP) artificial neural network (ANN) model, trained with Levenberg–Marquardt and Bayesian regularization algorithms, was developed to predict flotation outcomes. Sensitivity analysis using Garson’s algorithm and the connection weight approach identified the most influential operational parameters. This work demonstrates a practical route for valorizing BFS, supporting circular economy practices, and enhancing sustainability in the metallurgical industry.