Application of RSM and ANFIS in the optimal parameter evaluation for crude oil degradation in contaminated water amended with PES

Abstract

In this work, palm-bunch enhanced stimulant (PES) was synthesized from palm bunch ashes (PBA), and applied in the optimized bioremediation of crude-oil laden water. Characterization of PBA revealed that the elemental composition composed sufficient constituents capable of propagating microbial driven biodegradation process. Gas column chromatography analysis of the raw and treated samples, confirmed that the residual sample contained only 82 lightweight hydrocarbon (HC) components as against 197 heavyweight hydrocarbon (HC) components contained in the untreated wastewater. Experimental design using response surface methodology (RSM) illustrated that PES concentration (F-value = 106.03) was the most significant process variable that influenced the bio-elimination of total petroleum hydrocarbon (TPH), while treatment time having an F-value of 4.56 exerted the least influence on the process. Mode appraisal analysis using error functions demonstrated that adaptive neuro-fuzzy inference system (ANFIS: R2 = 0.9725) modeling technique performed better than response surface methodology (RSM: R2 = 0.9480) in capturing the quadratic nature of the system. Particle swarm optimization tool (PSO) predicted an optimum TPH removal of 85.8% at PES conc. of 532 mg/L, 150 mg/L crude oil conc., pH value of 8.2, and treatment time of 3.4 days (81 hr., 36 min.). This optimum predicted value was validated using a set of triplicate experiments, and an average value of 85.2% TPH removal was obtained at the stipulated optimum process variables.

Keywords

• ANFIS
• Bio-remediation
• Bio-stimulant
• Oil-polluted water
• Optimization