Prediction Of The Water Yield Property In Sandstone Roof Using Multi-factors In Deep Buried Mining Area With Complex Structure

The mine roof water disaster is a kind of mine water disasters, it is a serious threat to the safety mining of coal. If the roof water gushes out, working face production environment will be worsen, and it will bring unsafe factors to the mining production. The No.6 mining area is located in the southern wing of syncline axis in Yanzhou coalfield, there are many faults in the No.6 mining area, there are many secondary folds, the structure is complex; mining depth ranges from 600 m to 1000 m, constant stress is high. Water abundance distribution of sandstone aquifer in No.3 coal roof is uneven, mining hydrogeological condition is complex, there is a possibility of disastrous water burst in local district. Based on material of hydrography in mining area and neighboring hydrogeologic boundary conditions and the relationship of Coal seam roof aquifer and water-resisting layer are analyzed, cranny water in sandstone is mainlystatic reserves in the mining area. The height of water fractured zone in the mining area was analyzed and simulated according to the data of drilling and coal seam roof lithology, and the thickness of sandstone statistics is 100 m.Based on the study of study, the influence of stress level on sandstone fissurewas researched, it comes to the conclusion that cracks tend to be closed and water abundance is relatively lower with thee increase of buried depth of roof sandstone. Mining area was qualitatively analyzed according to the stability of faults, conclusion was got and verified combined with different depth of pumping test using volumetric strain formula.Using the software of Arcgis and matlab and the factors of lithology, fractures structures and stress level, the analytic hierarchy process(AHP) water model and neural network(ANN) model of the waterwere built up. Impact factor coefficients was acquired using expert experience rating method,relative water abundance partition map was acquired using weighted superposition approach; neural network was obtained adopting water gushing points in the No.14 and No.4 mining area andsafety points as the learning and training samples.the roof sandstone water was partitioned using neural network. The relationship between analytical hierarchy process(AHP) water abundance partition map andartificial neural network(ANN) water abundance partition map was compared and studed.The partition results of water abundance partition map was verified using mining drilling combined with situation of mining exploitation. The estimation of water inflow and prevention and control of water solutionin the mining area were output.