An Adaptive Node-Distribution Method for 2D Radial Basis Function Finite Difference Modelling
P. Duan, B. Gu, Z. Ren and Z. Li
Event name: 81st EAGE Conference and Exhibition 2019
Session: Seismic Modelling II
Publication date: 03 June 2019
Info: Extended abstract, PDF ( 1.67Mb )
Price: € 20
Radial-basis-function-generated finite difference methods (RBF-FDMs) has been proven successfully in modeling seismic wave propagation. Node-distribution is typically the first step in RBF-FDMs, but it’s still a subject of ongoing research in seismic because the actual geological structure cannot be easily modeled, and the continuity properties cannot be accurately reflected. We design an adaptive node-distribution method to overcome above issues for seismic forward modeling. The proposed method consists of two novel points. First, a robust and effective node-generation algorithm has been built using node distribution method and assignment of velocity values to scatter nodes, which improve the modeling stability greatly with small computational cost because no node repel algorithm for interfaces is involved. Second, we propose an adaptive grain-radius generation method, which can automatically provide wider scope of grain-radius in seismic forward modeling under the premise of ensuring dispersion relation and stability condition. Numerical results suggest that our method can effectively improve the stability of RBF-FDMs and accurately describe complex geological structures.