Integrating fuzzy clustering and regression models to optimize cost and performance of canola-derived activated carbon for diclofenac adsorption

• Fuzzy clustering identifies energy-efficient AC synthesis conditions (300°C carbonization, 500–800°C activation). • Fuzzy regression models predict BET surface area (R² = 0.998) and adsorption capacity (R² = 0.999) with high accuracy. • Canola-derived AC achieves maximum diclofenac removal at pH = 2 via hydrogen bonding, π-π interactions, electrostatic interactions and hydrophobic effects. • Cluster 2 delivers 33–40% lower production costs versus higher-energy clusters • AC retains 91.4% adsorption efficiency after five regeneration cycles, demonstrating industrial viability. This study explores the sustainable production of activated carbon (AC) from canola stalks and its application for diclofenac removal from aqueous solutions, employing advanced fuzzy clustering analysis (FCA) and fuzzy regression techniques. The research systematically investigates the optimization of AC synthesis parameters (carbonization temperature: 300–600°C, activation temperature: 500–800°C, carbonization time: 60–120 min) using FCA to identify energy-efficient production pathways. Fuzzy regression models elucidate the relationships between these parameters and key performance metrics, including BET surface area (up to 560 m² g⁻¹) and diclofenac adsorption capacity (up to 89 mg g⁻¹). The results reveal that Cluster 2, characterized by low carbonization temperatures (300°C) and shorter processing times, offers a cost-effective route with 33–40% energy savings, while Cluster 1 (higher temperatures) achieves marginally better adsorption at increased costs. The highest diclofenac adsorption capacity (89 mg g −1 ) was observed at pH 2, likely due to hydrogen bonding, π-π interactions, and hydrophobic effects, as supported by FTIR analysis. Kinetic data fit the pseudo-second-order model, and isotherm analysis follows the Langmuir-Freundlich model, suggesting chemisorption and heterogeneous adsorption sites. Thermodynamic analysis indicates that the adsorption process is spontaneous and exothermic (ΔH° = −52 kJ mol −1 ). The AC maintained 91.4% of its initial adsorption efficiency after five adsorption-desorption cycles, demonstrating excellent reusability. This study pioneers the integration of FCA and fuzzy regression in AC production, providing a robust framework for balancing performance, cost, and sustainability, with implications for wastewater treatment and circular economy strategies.