Research On Measurement Method Of Laser Induced Fluorescence Spectroscopy For Oil Spill On Water

With the continuous and rapid development of the global economy,the development and utilization of petroleum resources have become more frequent.The exploitation and transportation of a large amount of petroleum resources has led to aggravation of marine oil pollution.Although oil has brought huge economic benefits to mankind,oil spill pollution accidents have seriously damaged the marine environment.The emergency treatment of oil spills has also become a hot spot in the research of global environmental issues.Therefore,establishing a set of fast and real-time oil spill on-site measurement and analysis methods to obtain timely and accurate information on the source and quantity of oil spill is of great significance for the rapid formulation of effective pollution control measures.Laser induced fluorescence spectroscopy is an active optical detection technology,which has the characteristics of fast measurement speed and high measurement accuracy.It is possible to obtain information on the type of oil spill,the amount of oil spill and the spread of oil at the same time.It can be mounted on ships or airplanes for fast and efficient measurement of oil spill related parameters in a large area.Therefore,it is one of the oil spill detection methods with great development potential.However,the current use of laser-induced fluorescence technology for rapid oil spill monitoring still faces many problems:How to accurately construct the oil film thickness and spectral signal model,and perform accurate quantitative inversion of the oil spilled oil film;How to solve the problem of low spectral recognition rate of similar oil sources,etc.Therefore,in response to the problems faced in the on-site monitoring of oil spills,this paper has carried out research on the quantitative and qualitative analysis methods of oil spills based on laser induced fluorescence spectroscopy.The main results of the thesis are as follows:(1)This work designs and builds a laser-induced fluorescence detection system.The Nd:YAG laser with an excitation wavelength of 266nm was selected as the laser transmitter,and 10Hz was selected as the best operating frequency.A fiber-coupled optical path structure is designed to improve the efficiency of fluorescence coupling,and a miniature spectrometer is used as a fluorescence detection device to realize rapid and sensitive detection of oil fluorescence.The spectral denoising and scattering correction methods are studied to realize the accurate extraction of the oil spill fluorescence spectrum signal.(2)The method of laser-induced fluorescence spectroscopy for oil species identification is studied.This work proposes an oil spill classification and identification method based on principal component analysis combined with K nearest neighbor classification algorithm.The results show that:by using PCA to extract the characteristics of the spectral matrix,and using KNN to classify and identify the oil fluorescence emission spectrum,the average accuracy of the six types of oil spill identification is 73.3%.Among them,the recognition accuracy of 92#gasoline is the highest,which can reach 90%.Oils with high fluorescence spectrum similarity,such as Lifeguard fluid AG6 and Shell Helix 10w-40,have the lowest recognition rates,which are 70%and 30%respectively.(3)In order to further improve the accuracy of oil species identification,a concentration-emission fluorescence spectrum matrix was established to increase the amount of spectral information.Wavelet analysis extracts the spectral characteristics of the concentration-emission matrix,and combines the SVM algorithm to realize the identification of different types of oil spills.The results show that the wavelet analysis method combined with the GA-SVM algorithm can accurately classify and identify six oils,and the recognition accuracy can reach 100%,and the running time of the algorithm is only 75.5s,which can meet the needs of rapid oil spill monitoring.(4)This work studies the quantitative inversion method of oil film thickness based on the water Raman suppression method.Using the Gaussian function fitting method to simulate the characteristics of the fluorescence spectrum and subtract them to realize the correction of the Raman spectrum.The Gaussian function fitting method is used to correct the interference of the fluorescence spectrum on the Raman spectrum.The non-linear least squares algorithm is used to establish the water Raman spectrum-oil film thickness inversion model,and the accuracy of the model is evaluated by the relative error and the average relative error.The results show that the water Raman spectroscopy-oil film thickness inversion model is used to analyze the six oils:92#gasoline,0#diesel,Mobil motor oil 20w-40,Shell Helix 10w-40,Lifeguard Fluid AG6 and crude oil.The relative error of oil film thickness prediction is 1.06%-30.26%,and the average relative error is 8.14%-15.81%.The relative error of oil film thickness prediction decreases with the increase of oil film thickness(5)This work studies the quantitative inversion method of oil film thickness by laser induced fluorescence spectroscopy.The non-linear least squares algorithm was used to establish the oil film fluorescence intensity-oil film thickness inversion model,and the accuracy of the model was evaluated by the relative error and the average relative error.The results show that the fluorescence intensity-oil film thickness inversion model is used to predict the film thickness of 92#gasoline,0#diesel,Mobil motor oil 20w-40,Shell Helix 10w-40,Lifeguard Fluid AG6 and crude oil.The relative error is 0.32%-21.36%,and the average relative error is 6.65%-8.90%.Fluorescence spectroscopy is more accurate in predicting oil film thickness than Raman suppression.The reliability of different types of oil film thickness measurement is verified with actual seawater as the background.The average relative error of the measurement is between 6.82%and 8.86%,which can realize quantitative measurement of oil film thickness.(6)This work proposes a quantitative analysis method of oil concentration using bicubic interpolation algorithm and least squares spectral matching algorithm.The results show that the recovery rate of oil concentration prediction by this method is 80.80%-117.07%,and the relative error of prediction is between 0.58%-19.2%.This method can realize quantitative analysis of oil spill concentration.