Study On Quantitative Identification Of Fuel Oil By Multidimensional Chemically Fingerprinting

In recent years,the frequent occurrence of oil spill accidents at sea has a great impact on the marine environment and organisms.It is urgent to deal with oil spill in time and quickly,and oil spill identification as a preliminary preparation for accident handling,has also become crucial.By consulting a large number of domestic and foreign literature and research records,the identification technology of crude oil has been very mature,while the identification of fuel oil is relatively less.Because of the complex composition of fuel oil,a single method can not identify it accurately,so this study adopts Fisher discriminant model of multi-dimensional chemical fingerprint method to quantitatively identify fuel oil,which improves the accuracy of fuel oil identification.In this paper,eight kinds of fuel oils are taken as the research object,and the three characteristic indexes of diagnostic ratio,fluorescence characteristic and Ni/V of oil are analyzed by GC-FID,fluorescence scanning and F-AAS,and the influence of short-term weathering on them is explored.The results showed that short-term weathering had little effect on C17/Pr and the fluorescence intensity at I380nm,except C18/Ph,Pr/Ph,(C19+C20)/(C19?C22),fluorescence intensity at I280nm?I302nm and I332nm,Ni/V,which could be used as characteristic variables for oil spill identification.The results of similarity analysis and cluster analysis of 8 oil samples show that the accuracy of single method is only 75%.The results of co-linear analysis of C17/Pr,C18/Ph,Pr/Ph,(C19+C20)/(C19?C22),fluorescence intensity at I280nm?I302nm?I332nm and I380nm show that C17/Pr and C18/Ph are co-linear,and the fluorescence intensity of I380nm is co-linear with other variables,After eliminating the fluorescence intensity at C17/Pr and the fluorescence intensity of I380nm,there was no collinearity among the variables in each group.The results of one-way ANOVA show that the P values of all variables are less than 0.05 except the fluorescence intensity at I332nm.It is considered that they have significant differences in light fuel oil,heavy fuel oil and residual fuel oil,and are suitable for modeling variables.Because the fluorescence intensity at I332nm and I380nm and C17/Pr,Pr/Ph is greatly affected by weathering,and the fluorescence intensity at I380nm nm and C17/Pr is in line with other variables,there is no significant difference between the fluorescence intensity at I332nm and other variables,so it is not suitable to be used as a modelling variable.Finally,the Fisher discriminant model was established by choosing C18/Ph?(C19+C20)/(C19?C22),fluorescence intensity at I280nm?I302nm?Ni/V as variables.The discriminant results showed that the discriminant accuracy rate reached 100%.The discriminant accuracy of the validation model is 100%.It shows that the Fisher discriminant model of multi-dimensional chemical fingerprint method can improve the accuracy of fuel oil identification and is suitable for the identification of weathered fuel oil.It has a certain practical significance for the study of oil spill identification.