Information-theoretic Analysis of Unsteady Data

摘要: The temporal evolution of scientific data is high relevance in many fields application. Understanding the dynamics over time a crucial step understanding underlying system. availability large scale parallel computers has led to finer and resolution simulation data, which makes it dicult detect all relevant changes system by watching video or set snapshots. In recent years, algorithms for automatic detection coherent structures have been developed that allow an identification interesting areas steps unsteady data. With such techniques, user can be guided subsets automatically created does not occlude aspects simulation. this paper, we give overview dierent show how their combination helps gain deeper insight look at directions further improvement. Two CFD simulations are used illustrate techniques. 1998 ACM Subject Classification G.3 Time series analysis, E.4 Coding information theory