Research On Three-Dimensional Reconstruction Of Scraper Conveyor Running Trajectory In Intelligent Working Face

Intelligent mining is the only way for the coal industry to move towards highquality development,and "three straight and two flat" is the inevitable requirement for fully mechanized mining face to realize intelligent mining.Therefore,the accurate perception of the scraper conveyor running trajectory can not only provide a reasonable cutting path for the shearer to reduce the running resistance,but also provide data support for the hydraulic support electro-hydraulic control system to be precisely pushed.In this thesis,taking the running trajectory curve reconstruction of the scraper conveyor as the research objective,the curve reconstruction algorithm based on different local activity frames is carried out with the FBG 3D curvature sensor as the sensing means.The feasibility and accuracy of the algorithm are verified by numerical simulation,laboratory experiments and field application practice.The main research work of this thesis is as follows:(1)Based on the structural composition and working principle of scraper conveyor,this thesis sums up the change law of its trajectory curve under three cutting feed methods,studies the formation process of "S" shaped bending section of scraper conveyor during the process of being pushed,and deduces the calculation formula of the length and the number of middle chutes of bending section.At the same time,the mathematical relationship between the FBG center wavelength change,strain and curvature are studied based on the basic principle of FBG sensing,and the real-time sensing system of the scraper conveyor trajectory is constructed.(2)To address the problem of discrete curvature during curve reconstruction,interpolation and fitting methods such as Lagrange,spline interpolation are studied to continuous the discrete curvature.The interpolation effect of each method is compared with the plane sine function and spatial spiral function as the experimental functions,and the best cubic B-spline interpolation method is preferred.Meanwhile,the planar curve reconstruction algorithm and the spatial curve reconstruction algorithm based on different local activity frames are derived in detail.Then the differences and connections between the Frenet frame and GTF frame are analyzed.(3)The basic principle of GRU recurrent neural network is explained,and the predictive interpolation model of scraper conveyor running trajectory curve based on the GRU model is constructed.The model purposes to solve the problem that the trajectory curve can be reconstructed with high accuracy even if a small amount of FBG points are damaged.In addition,based on the coordinate data from the numerical simulation of sensor under "S" deformation,the GRU prediction model is trained and learned.The effects on the prediction curves of the model under two conditions of different amounts of missing data and different locations of missing data are summarized,thus the prediction accuracy of the model is verified.(4)The structural model of the curvature sensor is established by Creo,and the load is applied to it by ANSYS,so that the sensor forms unidirectional bending and like arctangent function deformation in the horizontal and vertical planes respectively.Then its strain and displacement nephograms are obtained by simulation,and the strains of corresponding measurement points are extracted and substituted into the algorithm for curve reconstruction.The feasibility and accuracy of the algorithm for plane and spatial curve reconstruction are verified according to the evaluation indexes such as end absolute error,relative error and RMSE.(5)The experimental system of curve reconstruction algorithm is designed.By adjusting the rotation and displacement platforms respectively,the sensor is bent successively into arcs with curvature radii of 6m,7m and 8m.Based on the wavelength change data of the demodulator,the curve reconstruction algorithm is used to reconstruct the bending curve on the PC.The absolute error of the end is about 2mm.Moreover,taking No.12307 intelligent working face of Wangjialing mine as engineering background,the real-time sensing system of scraper conveyor running trajectory based on FBG curvature sensor is introduced.The reconstruction algorithm is practiced in the field,and the reconstructed curve is basically consistent with the actual shape of the underground.Finally,the reconstruction errors in the experiment and field application are analyzed in terms of encapsulation angle,reference coordinate system,strain transfer efficiency and other factors,and the corresponding solutions are proposed to provide theoretical support for further improving the curve reconstruction accuracy.There are 77 figures,5 tables,and 85 references in this thesis.