Research On Analysis Method Of Potassium Salt Mine Based On LIBS Spectroscopy And Its Data Fusion Technology

With the increase of the global population and the development of industry and agriculture,the demand for potassium has continued to surge.China is not only a big agricultural country,but also a major potassium deficiency country.According to incomplete statistics,China’s potassium-deficient soil has reached 3,000 hectares.When potassium is lacking in the soil,the yield and quality of crops will be directly affected.In order to meet the urgent needs of potash fertilizer in agricultural production and ensure the stable and safe supply of potash,potash geological exploration has received more and more attention in China.Therefore,the rapid and accurate analysis of potassium in potash ore is of great significance to whether the mine has mining value.Based on LIBS and Raman spectroscopy combined with stoichiometry,this paper analyzes the potassium content in potash ore samples quickly and accurately.This article is mainly containing the following:The first chapter establishes a rapid analysis method for potassium content of potash ore combined with LIBS technology combined with Partial least squares（PLS）.Firstly,the LIBS spectra of 30 potassium samples were collected to construct a PLS model based on the original libs spectrum,followed by a five-fold cross-validation to optimize the different pretreatment methods,input variables and model parameters,and finally,the potassium content in the potash mine was quantitatively analyzed based on the optimized PLS model.The results show that the optimal R_p²,RMSEP and MRE are 0.9575,0.1546%and 0.1632%,respectively.This paper shows that LIBS technology combined with PLS is a feasible method that can be used for rapid and accurate analysis of potassium elements in potash ore,which can provide a new idea and method for the quantitative analysis of potassium elements in the process of potash mining.The second chapter establishes a rapid analysis method for potassium content of potash ore combined with Raman technology combined with PLS.Firstly,potassium salt ore samples and sodium nitrate were decomposed to generate potassium nitrate,and the Raman spectra of 30KNO₃samples were collected to construct a PLS model based on Raman’s original spectrum;secondly,the quintifold cross-validation was used to optimize different pretreatment methods,input variables and model parameters;finally,the potassium content in KNO₃samples was quantitatively analyzed based on the optimized PLS model.The results show that the optimal R_p²,RMSEP and MRE are 0.9177,0.1647%,and 0.1622%,respectively.It shows that Raman technology combined with PLS is a feasible method that can be used for rapid and accurate analysis of potassium in potash ore,which can provide a new idea and method for the quantitative analysis of potassium in the process of potash mine mining.The third establishes a rapid quantitative analysis method of potassium content in potash mines using LIBS-Raman data fusion technology combined with PLS was established.The spectral information collected in chapters 2 and 3 adopts data fusion technology.First,the primary data fusion technology is used,that is,the two single spectra are normalized and preprocessed,and then the fusion is performed end-to-end;secondly,the primary data fusion technology is used.On the basis,the VIP feature variable extraction is performed on the two normalized single spectra,that is,the intermediate data fusion technology.Then,five-fold cross-validation was used to optimize different preprocessing methods,input variables and model parameters.Finally,based on the optimized PLS model,the potassium content in potash mine samples was quantitatively analyzed.By comparing the prediction performance of the calibration models established by two different data fusion techniques,the research results show that the R_p²,RMSEP and MRE of the calibration model established by intermediate data fusion are 0.9621,0.1546%,and 0.1441%under the optimal parameters,respectively.It shows that LIBS-Raman data fusion technology combined with PLS is a feasible method for rapid and accurate analysis of potassium in potash mines,and it can provide a new idea and new method for quantitative analysis of potassium in potash mines.