Intelligent construction method and application of large-scale three-dimensional complex discrete fracture network model based on particle swarm optimization algorithm

Based on the Monte Carlo method, this study proposes a method for constructing a three-dimensional(3D) complex discrete fracture network (DFN) by introducing the particle swarm optimization (PSO) algorithm theory, which realizes the effective fusion of multisource survey data of on -site fractured rock masses and the highefficiency prediction of high -precision 3D DFN model in unknown intervals. Based on the Monte Carlo method, the DFN is randomly generated. Combined with the geological survey test results, the fitness function formula is constructed to determine the matching degree between the DFN and the actual fracture rock mass characteristics. The PSO algorithm is introduced to iteratively update the historical optimal fitness and parameters to achieve efficient prediction of the unknown interval fracture network. An underground water -sealed oil storage project was selected to verify the method, and the water -sealed safety was quantitatively assessed. The results show that the matching degree between the 3D complex DFN constructed by this method and the actual interval section reaches 85.37%, and the matching optimization rate is 21.29%. Therefore, this method can efficiently and accurately construct a 3D complex DFN, which provides an effective means for seepage analysis of underground rock mass structures.