Evaluation Of Oil From Oily Sludge With Ionic Liquid Assisted Solvent Extraction Based On Thermodynamic Analysis And Machine Learning

The oily sludge,produced in the process of oil exploitation and utilization,is a typical type of hazardous waste that needs to be addressed in the petrochemical industry.The oil content of oily sludge is generally between 15%-50%,which has a certain recovery value.However,the composition of oily sludge is complex,containing a large amount of aromatic hydrocarbons,asphaltenes and adhesives and other difficult-to-separate and difficult-todegrade organic pollutants.Resourceization and harmlessness have brought great challenges to oily sludge.Therefore,the development of high efficiency sludge treatment technology can not only solve the environmental pollution caused by sludge,but also can recover crude oil to solve the energy problem.In this study,green solvent ionic liquid was used to assist the extraction of oily sludge.Toluene was used as solvent,and various ionic liquids such as[EMIM][BF4],[EMIM][TA],[EMIM][N(CN)2],CHCl/U,CHCl/EG and CHCl/MA were used to assist solvent extraction to oily sludge to improve oil recovery.This study found that the order of the enhanced effect of different additives on the extraction system is[Emim][BF4]>[Emim][TA]>[Emim][N(CN)2]>ChCl/U>ChCl/EG>ChCl/MA.Among them[Emim][BF4]increased the oil recovery most effectively(about 20%),while ChCl/MA increased only 5%of oil recovery in the solvent extraction.The results demonstrated that at the same temperature,the oil recovery kept decreasing with the increase of soil particle concentration.As the temperature increased and the particle concentration increased,the activity coefficient continued to decrease,which means that the contacts and collisions between particles are more frequent,resulting in a decrease in effective adsorption sites and a decrease in activity coefficient.The fitting results showed that the Sips-SCA isotherm fitted best.The calculation result of the thermodynamic function showed that ?G<0 means the desorption of oil is a spontaneous process,?H>0 means the process of extraction and desorption is endothermic,?S>0 means the level of chaos in the extraction system has increased.The values of ?S were large in[Emim][BF4],[Emim][TA]and[Emim][N(CN)2]systems,and which were low in ChCl/U,ChCl/EG and ChCl/MA systems.This means in the interface between soil particles and petroleum,traditional ionic liquids are easier to dissociate and greatly reduce the interaction between the two,making the extraction system have a higher petroleum removal rate.Machine learning methods were applied to learn and analyze the experimental results.The results showed viscosity and surface tension is inversely proportional to oil recovery,while the pH and conductivity are directly proportional to oil recovery.Besides,algorithm analysis showed that the order of the influence of several physical properties on the extraction effect is as follows:conductivity>surface tension>acidity>viscosity.Finally,three machine learning algorithms(ridge regression,multi-layer perceptron,and support vector regression)were successfully used to predict the oil recovery rate of oily sludge using ionic liquid-assisted solvent extraction.