Quantitative Analysis Of Hydrocarbon Substances In Drilling Fluid Based On Raman Spectroscopy And PLS

In the actual process of petroleum exploration,the detection while drilling technology has been widely used.However,what follows is how to make qualitative analysis of the mixed materials detected while drilling in the current underground environment,to determine whether there is crude oil,and then whether to adjust the direction of drilling forward trajectory.Combined with the development of actual mining,this paper chooses laser Raman instrument to obtain the laser Raman spectrum of mixtures,which is taken as the basis of qualitative analysis.The material information on Raman spectrum is obtained and converted into data information by a certain method,so that the mixture of oil-bearing drilling fluid is qualitatively analyzed,and on this basis,the quantitative calculation is completed.According to the characteristics of Raman spectrum,on the premise of fluorescence filtering of the original data,this paper eliminated the interference of influencing factors through the way of "preprocessing" of the spectral data."Polynomial moving average filtering method " is selected as the filtering processing means,and the polynomial fitting method is used to "baseline correction" of the filtering data.Finally,"normalization " is carried out on the basis of the same variable selection.After the above processing,based on the high-dimensional situation of spectral data,"principal component analysis" and "singular value decomposition" methods are adopted to reduce the dimension of data while retaining the original real information and reduce the impact of data collinearity on data information.At the same time,according to the differences between the spectral spectra obtained under different experimental times,the data of each concentration were grouped to ensure that the bond strength of the data in each group could change with the mixture concentration.In the part of qualitative analysis and quantitative calculation,the partial least squares regression analysis method is adopted to group the processed data into "modeling data" and "verification data",based on the information of the modeling data.The mathematical model was established by PLS(Partial least squares)algorithm,and then the concentration values of the "validation data" which were not involved in the modeling were predicted based on this.The simulation results of MATLAB prove that the current mathematical method can realize the basis of qualitative analysis of the mixture of oil-bearing drilling fluid,that is,to determine the existence of crude oil substances in the mixture.Secondly,when the current mathematical model predicts the data of "unknown concentration",the deviation between the calculated value and the real value can be controlled within a small range,which is relatively ideal.In order to ensure the reliability of the model obtained by current processing,two groups of different types of data are used for processing,and the applicability of the current model is finally determined.