Peat-based sorbents for the removal of oil spills from water surface: Application of artificial neural network modeling

摘要

Removing of oil slicks from sea, rivers and lakes formed as a result of accidental oil spillage is of great concern. Such ecological accidents have created a great need to find more efficient and low-cost materials for oil spill cleanup. In this work three types of peat-based sorbents were compared in order to determine their potential for oil spill cleanup. The best peat sorbent PT-1 can sorb 12-16 times its weight from different oils. The retention profile and images captured by optical microscope have revealed that oil sorption process involves both capillary and adsorption phenomena. A feed-forward artificial neural network has been constructed to predict the removal efficiency of oil slick from water surface by peat sorbent PT-1 based on 45 experimental runs obtained in a laboratory study. The effect of input variables such as sorbent dosage, drainage time and initial thickness of oil slick has been studied to optimize the conditions for maximum removal efficiency. On the basis of confirmation run result, the optimal operating conditions involve the following values of input variables: sorbent dosage of 4.98g/dm~2, a drainage time of 5s and the initial oil-slick thickness of 3.3mm. For the optimal conditions the removal efficiency of 99.21% has been obtained experimentally being the maximum value of response in this study.