Three-dimensional Fluorescence Combined With Multi-way Calibration And Pattern Recognition Algorithm For Detection Of Petroleum Type Oil

As an indispensable source of energy and chemical raw materials,petroleum promotes the rapid development of society,and frequent oil spill accidents also bring serious property loss and ecological environment problems.Therefore,how to achieve rapid and accurate detection of petroleum type oil has become a hot topic of concern.Aiming at the problems of complicated operation process,high material cost and long analysis time faced by traditional analysis methods,a new analysis strategy of three-dimensional fluorescence spectroscopy combined with multi-way calibration and pattern recognition algorithm for detection of petroleum type oil in this paper.The specific research contents are as follows:(1)In order to clarity the theoretical basis for the detection of petroleum type oil,the specific process of photoluminescence is explored,and the relationship between the concentration of fluorescent substance and fluorescence intensity is revealed,and then the main factors affecting fluorescence intensity are obtained.The basic structure and working process of fluorescence detection system are systematically explained to provide a theoretical basis for obtaining fluorescence spectrum data.And the comparison of different implementation methods for the analysis of multi-component complex systems is introduced to lay the foundation for the subsequent detection of petroleum type oil.(2)Taking jet fuel and lube as the research objects,an analysis strategy based on three-dimensional fluorescence spectroscopy combined with second-order calibration algorithm is established to achieve the detection of oil species.By designing specific experiments,after preprocessing and estimation of component number of the obtained fluorescence spectrum data,the analytical effects of different second-order calibration algorithms are compared.The experimental results show that the detection of oil species in petroleum type oil can be achieved effectively by this analysis strategy,and it also provide technical support for judging the nature of source of oil spill.(3)Different experimental temperatures are introduced as an extra dimension,and an analysis strategy based on three-dimensional fluorescence spectroscopy combined with third-order calibration algorithm is established to achieve qualitative and quantitative detection of petroleum type oil.The obtained fluorescence spectrum data is stacked to construct a four-dimensional response arrays,after preprocessing and estimation of component number of the obtained fluorescence spectrum data,and the analytical effects of third-order calibration and second-order calibration algorithms are compared.The experimental results show that this analysis strategy provides a "mathematical separation" method to achieve qualitative and quantitative detection of petroleum type oil,and further enriches the ways of petroleum type oil detection.(4)An analysis strategy based on three-dimensional fluorescence spectroscopy combined with second-order calibration and pattern recognition algorithm is established to further achieve the characterization and classification of petroleum type oil.The second-order calibration algorithm is used to decompose three-dimensional response arrays for achieving components characterization,and then several classification models are established and their classification effects are compared.The experimental results show that the main components characterization and sample classification in petroleum type oil can be achieved effectively by this analysis strategy,and it provides a new analysis strategy to detect petroleum type oil.