Study On Coal-rock Classification Method Based On Terahertz Time-domain Spectroscopy And CA-PCA Model

Coal-rock identification has always been one of the key problems restricting the unmanned mining of coal mines.Coal industry is an important basic industry related to the lifeblood of national economy and energy security.Coal is China's basic energy and important raw material,accounting for 90% of China's fossil energy resources.In recent years,the proportion of coal in primary energy consumption has gradually decreased,but the main energy status of coal will not change.Because the working environment of traditional artificial coal mining is extremely complex,it is difficult to accurately find the interface between coal and rock,which is easy to cause under cutting or over cutting.At the same time,a large amount of solid waste-coal gangue will be discharged in the process of coal mining and washing.It is a dark gray rock that is harder than coal and contains less carbon.The output of coal gangue is high,and its emission accounts for about 15%-20% of the raw coal output.Therefore,the separation of coal gangue is also the key problem restricting the output and quality of coal.As a non-destructive detection technology,terahertz spectroscopy can reflect the physical and chemical information of the measured object.It can become an effective method to study the identification of coal,coal gangue and rock.Therefore,based on the background of coal rock medium and coal rock classification and recognition,this paper studies the terahertz spectral feature extraction and recognition model of coal rock samples.The main work contents are as follows:(1)Obtain terahertz spectrum data of the sample.The terahertz signals of different coal and rock media can be obtained by terahertz time domain spectroscopy system.The terahertz time domain signals of the samples are analyzed from the perspective of time domain and frequency domain,and the optical parameters of the samples in the frequency band of 0.5-3.5THz are retained.There are differences in terahertz data of various samples in this frequency band,which verifies the feasibility of using terahertz data to identify coal and rock media.(2)The influencing factors of terahertz physical properties of different kinds of coal and rock in terahertz frequency band are studied.The terahertz signals of four coal media are different in time domain and frequency domain.The principal component analysis method is used to process the optical parameters of various coal media in terahertz frequency band,and the PC1 score of the optical parameters of each sample is extracted.Combining the PC1 score with the carbon content and ash content of different coals,it is verified that the carbon content and ash content of coal medium are the main factors affecting its terahertz physical properties in terahertz frequency band.(3)A recognition model based on terahertz time domain spectroscopy is established.PCA method is used to extract the dimensionality reduction features of refractive index spectrum,absorption coefficient spectrum,dielectric constant spectrum and power spectrum of different coal and rock samples in terahertz frequency band.After dimensionality reduction,the contribution rates of the first three principal components of various optical parameters are higher than 99%.CA algorithm can classify samples by the similarity between various types of coal and the difference between coal,gangue and rock.Among the classification models of THz-CA-PCA,the model based on Terahertz Absorption coefficient spectrum has the best recognition effect on six coal rock media,and can realize the accurate classification of coal,coal gangue and quartz sandstone;The model based on terahertz power spectrum can accurately distinguish rock media,but it cannot distinguish coal gangue from coal samples.In the classification model based on refractive index spectrum and dielectric constant spectrum,coal,coal gangue and rock media are not completely distinguished.The experimental results show that the application of terahertz time domain spectroscopy in coal classification and recognition has certain research significance.