基于RBF神经网络的土石坝渗透系数反演及演化规律研究OA

The dynamic variation of the permeability coefficient in earth-rock dams affects the seepage characteristics and stability of the dam body.Taking an earth-rock dam of a reservoir as the research subject,a seepage model is constructed based on finite element numerical simulation.Combined with Radial Basis Function(RBF)neural network training and inversion,the evolution patterns of the horizontal and vertical permeability coefficients of the dam body are analyzed.The results show that:(1)Among the inverted anisotropic permeability coefficients,the horizontal permeability coefficient gradually decreases with the increase of years,while the vertical permeability coefficient gradually increases.(2)The inverted isotropic permeability coefficient gradually increases over the years.(3)The seepage pressure curve obtained from the isotropic permeability coefficient is close to the measured seepage pressure curve,indicating that the isotropic permeability coefficient obtained through the RBF neural network model training has higher accuracy.