Identification Of Rock Types And Quantitative Analysis Of Major Elements Of Mars-like Mineral Based On LIBS

Deep space exploration is an important national development strategy of China.Mars became the next important exploration target after the success lunar landing mission.2020 is a critical time window for Mars exploration.At that time,China will put her first Mars landing mission Tianwen-1 into effect with the three engineering goals,i.e.orbiting,landing and patrol,at the same time.The Mars soil composition analyzer(MarsCoDe)carried by the Mars rover is the China's first LIBS system for planetary exploration.For its data interpretation,it is urgent to establish a LIBS spectral database and a quantitative element prediction model.To get this goal,one Mars chamber and one stand-off LIBS system were built and coupled each other in Shandong University,named SDU-Mars-LIBS,and aimed for Mars return LIBS data interpretation.During the period of system building,the author participated in the system design,optimization,construction,and the prediction model building of rock type identification and quantitative analysis of the major elements of Mars-like rocks.The author obtained the following research results:1)The main factors of physical matrix effect of geological samples were preliminary studied and the optimal system parameters were determined;2)more than 50 kinds of Mars-like geological standard samples were analyzed and their LIBS spectral were recorded both in Earth environment and in simulated Mars environment(7mbar,CO2).The original LIBS data were preprocessed and the LIBS spectral database of Mars-like rocks was preliminary established;3)A method using the Loading Space Distance of PCA to select the LIBS characteristic spectral lines with the largest contribution to sample classification was proposed,and 64 standard geological samples published by ChemCam team were identified and analyzed with an identification accuracy of 92.8%;4)two algorithms,support vector machine regression(SVR)and partial least squares regression(PLSR),were used to establish the quantitative prediction model for the major elements of mars-like rocks.There are six chapters in this thesis.The preface introduces the profile of Mars which mainly includes the Martian atmosphere,Martian rocks,and minerals.The first chapter mainly summarizes the human Mars exploration mission,including the current exploration results,important payloads of mineral element analysis and the overview of Chinese Mars mission.The second chapter introduces the working principle and instrument structure of LIBS,including ChemCam system,MarsCoDe,SDU-Mars-LIBS system and the corresponding LIBS data preprocessing.Third chapter is LIBS data qualitative analysis section,this chapter introduces a method using PCA principal component loading space distance as decision criteria for rock type identification was introduced.This can reduce the matrix effect to a certain extent and plays an important role in the identification of mineral types and the establishment of quantitative analysis models of elements.Chapter four introduces the recording and analysis of geological standard LIBS data,obtained by SDU-Mars-LIBS system,LIBS spectral library building and LIBS data quantitative analysis model building;Chapter five is text summary and work outlook.