Study On Fast Detection Method Of Ash Based On Spectral Features

Ash is one of the important indicators to measure the quality of coal.Rapid and accurate detection of coal ash is a crucial issue for each coal preparation plant.The traditional ash detection methods are cumbersome and the analysis cycle is too long.However,most real-time ash detection methods have different levels of radioactive materials and are expensive.Based on spectroscopy and colorimetry,this paper explores the influence law between coal ash,reflectivity and chromaticity tristimulus values,and attempts to establish a fast and safe detection of coal ash.method.The main content includes the following aspects:A single factor test was designed to explore the effect of coal ash on the reflectivity and tristimulus values.Finally,the law of increase in ash content,reflectivity and tristimulus value was found,which laid a foundation for subsequent research.Prepare-0.075 mm,-0.25 mm and-0.5mm coal samples with different ash content at different particle sizes,collect spectral information,and explore the effects of coal particle size on reflectance and tristimulus values.When the ash content is the same,the reflectance of the wavelength between 780 and 1000 nm increases with the decrease of the particle size,and there is no obvious relationship between the tristimulus values X,Y,and Z.At a particle size of-0.25 mm,the linear correlation between the average reflectivity and ash and the tristimulus values X,Y,Z,and ash are the most linear.Therefore,in the subsequent study,coal samples with a grain size of-0.25 mm were used for testing.X-ray diffraction and X-ray fluorescence spectrometry were used to qualitatively and quantitatively analyze the mineral composition and elemental composition of the sample coal samples.The X-ray diffraction(XRD)was used to analyze the mineral composition of the coal samples.The X-ray fluorescence spectrometer(XRF)was used to analyze the elemental composition of the coal samples.The K value method was used in combination with the elemental composition of the coal samples on the gangue coal.Samples were subjected to quantitative analysis of minerals.Four kinds of inorganic minerals such as quartz,kaolinite,dolomite,and pyrite were selected to prepare coal samples,and the effects of single inorganic minerals on reflectivity and tristimulus values were explored.Studies have shown that for each mineral blended coal,the average reflectivity and tristimulus value increase in varying degrees with the increase in the proportion of minerals,and there are differences in the types of minerals,and there are also differences in the average reflectivity and tristimulus values.Among them,quartz,kaolinite,and dolomite three kinds of minerals increase the coal blending trend is basically the same,and pyrite blending coal has the smallest increase.Collect reflectance spectra and tristimulus values for 100 different ash coal samples,select 85 of them as correction sets,adopt partial least squares method to establish reflectance-based ash detection model,and use the remaining 15 as prediction set test model prediction accuracy.Then,the spectral data was preprocessed by using data smoothing and MSC multiple scatter correction,and the basic ash detection model was optimized.It was found that the regression model with 25 points smoothing of the original data and MSC multiple scatter correction had the best effect.The same method was used to establish coal ash detection model based on tristimulus value and coal ash detection model based on reflectivity and tristimulus value.It was concluded that the coal ash detection model based on tristimulus value had the best effect and the correlation coefficient reached 0.945.The RMSEC and RMSEP values are 4.9503 and 5.0453,respectively.The reflectivity-based regression model has the worst effect,and the two methods combine the established model in an intermediate effect.