Study of mining induced overlying rock fissure numerical simulation based on intelligence parameters inversion

In order to master the law of mining induced overlying rock fissure and provide a reasonable scientific basis for safety mining under water. Orthogonal experimental design method and FLAC (FLAC, Fast Lagrangian Analysis for Continuum) program were used to obtain effective training samples for SVM (Support Vector Machine). Through reasonable selection of SVM parameters, SVM model was established between height of different position overlying rock fissure and simulation mechanical parameters. Mechanical parameters were calculated by measured overlying rock fissure height with SVM model. A new FLAC numerical simulation model were build with inversion parameters. Trill data by field measured in a coal mine in Shanxi was shown as example, Fissure heights calculated by numerical simulation were fit with the results of field measured, the maximum absolute error of overlying rock fissure height is 10m and maximum relative error is 13%, its show that the mechanical parameters which obtain by inversion is right and effective. Numerical simulation results which tally with the field measured value can be got with mechanical parameters acquired from intelligent inversion, and then laws of overlying rock movement and deformation can be holistic mastered from numerical simulation. The research provides a scientific basis for safety mining under water.