The invention relates to the technical field of
mine safety engineering and hydrogeological
information processing, and discloses a multi-scale
coal seam water spatial variation analysis and main
control factor identification system for group-mine combined mining, which comprises a multi-source spatio-temporal data benchmark unification and preprocessing module for generating a standardized
data set; the
macro-micro orthogonal collaborative dynamic
permeability tensor field construction module is used for establishing a
permeability tensor field; a non-Euclidean hydraulic distance calculation module based on Riemannian manifold calculates a manifold hydraulic distance; a variable-
lag embedded convergence cross-mapping causal identification module calculates causal
coupling strength; a causal topology feedback driving
model correction module iteratively corrects the model to convergence according to the difference; and the spatial variation characteristic analysis and main
control factor output module outputs a main
control factor list. According to the method, a causal topology feedback-driven model self-correction
closed loop is constructed, a causal intensity reverse constraint
physical model obtained through convergence cross mapping is used, and the problems that a traditional
physical model cannot respond to mining-induced fracture evolution in real time, and pure
data mining lacks physical
interpretability are solved.