Study On The Fast Computation Technique Of Seismic Traveltimes With Finite-difference Method

Seismic wave traveltime is an important attribute parameter of seismic waves. It has important applications in the forward modeling and inversion research of seismic exploration and practical production. It is used in many of the approaches that image the earth's internal structure and detect the earth's internal material composition. In many practical applications, traveltime calculation usually costs a lot of time because of the large quantity of computation, therefore, it is very necessary to improve the efficiency of computation when the stability of algorithms is guaranteed.In this paper, the main object of our research is the finite-difference calculation of seismic wave traveltime along the expending wavefronts. By considering both the algorithm and the programming of computation, we further improved the speed of calculation and computation of the seismic wave traveltimes. The major tasks of this paper include the following five parts:①Firstly, summarize various of traveltime algorithms, then, explain various of finite-difference traveltime algorithms. According to the expending mode of finite-difference calculation, classify the finite-difference traveltime algorithms as expending box method and expending wavefronts method. The expending box method includes the eikonal equation finite-difference traveltime algorithms and the upwind finite-difference traveltime algorithms. The expending wavefronts method includes the expending wavefronts finite-difference traveltime algorithms and the fast marching traveltime algorithms.②Improve the difference format of traveltime calculation. The new difference format avoids the problem of negative square root and therefore improves the stability of the algorithm.③Presents a grid densification method near the source which effectively improves the calculation accuracy near the source without destroying the causality of the algorithm.④Improve the computation speed by considering the data structure which is used to store and sort the wavefront points. Presents five data structure including the stack data structure, queue data structure, minheap data structure, untidy priority queue data structure and segmentation minheap data structure.⑤Present an efficient, neat and reasonable programming to realize the method, the programming also plays an important part in improving the calculation speed.