The Research Of Character Relationship Analysis And Visualization

In the development of automation and information in coal industry,automatic height adjustment of the shearer cutting drum on the coal working surface is a key technology,but the change in seam thickness caused by discontinuous geological changes posed a challenge for the automatic heightening.In order to improve the accuracy of the automatic height adjustment of the drum during coal mining,this thesis would analyse the drilling data and coal seam teaching data,and improved algorithms based on local correlation between the data,the three dimensional model of coal seam and the indirect identification method of coal-rock interface were realized,and the study will enable it to offer the basis of the data for the process of industrialization and mechanization of coal industry.According to the first theorem of geography and the principle of normal distribution,a Kriging algorithm based on triple variance distance and Radial Basis Function(TVD&RBF-Kriging-Kriging)was presented for solving the problem that is all known data processing and selection of the semi-variation function relying on experience.This algorithm improved the data computation ability during the interpolation process and selected radial basis function(RBF)to fit the semi-variation function.Compared with the Kriging interpolation method,this algorithm can handle unknown data accurately in the coal seam data,and offer the data support for coal-rock interface identification and prediction.Through reading and analysing relevant studies on coal-rock interface identification,it had been discovered that processed the coal-seam data in time series without considering spatial structure,processed the data with workface direction without considering the influence for the predicted points at other directions,and not considered the different effects of the equidistance point for the prediction point.Therefore,a Gaussian Process Regression algorithm based on structural element(3×3)transition probability(STTP-GPR)is presented to implement the coal seam thickness prediction.The experimental results show that the average absolute error in the prediction is less than 0.03m through our algorithm,and this algorithm improves the coal seam prediction accuracy,which is an important reference value for the coal-rock interface identification and coal mining,compared with BP neural network,support vector regression,Gaussian Process Regression(GPR)in the marching direction.