Sustainable water recovery from steel industry wastewater via direct contact membrane distillation

Direct Contact Membrane Distillation (DCMD) was investigated for water and heat reuse from steel industry wastewater, with a novel integrated framework combining experimental validation, thermodynamic simulations, and predictive modelling. Experimental tests with a hollow fiber polypropylene membrane demonstrated stable flux in the range of 1.4–4.8 kg/(m² h) over a feed temperature range of 40–60 °C, with salt rejection above 99.9% and no evidence of scaling or wetting, even up to 93% recovery. Then, solubility simulations with PHREEQC were performed, accurately reproducing solution properties and predicting undersaturation conditions even at high recoveries. The constant and high water activity calculated from these simulations explains the stable flux observed, while the predicted absence of solid phases matches the lack of scaling in experiments. The developed DCMD model was validated against experimental data and 14 independent datasets from the literature, achieving low errors (MAPE < 5%) and strong correlations (R² > 0.96) for flux and outlet temperatures. Parametric simulations revealed that higher feed temperatures and velocities enhance flux, productivity and specific thermal energy consumption (STEC), while reducing the membrane area required to produce 1 m³/day of water. These improvements result from increased driving force and reduced polarization effects but must be balanced against higher energy demands for heating and pumping. Overall, the study demonstrates that DCMD can achieve high recovery and excellent water quality for industrial wastewater treatment, and that the validated model provides a reliable tool for system design and scale-up in industrial contexts. DCMD treated steel wastewater with stable flux and > 99.9% salt rejection. PHREEQC simulations accurately predicted ion trends and saturation indices. Experiments confirmed no fouling or wetting up to 93% recovery. DCMD model showed high accuracy with MAPE < 5% and R² >0.96. Temperature and velocity identified as key levers to reduce membrane area.