Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore an advanced conference talk that revisits and extends adaptive mesh refinement (AMR) algorithms beyond traditional P4est approaches, presenting a refined strategy where finer elements are superimposed on their parents rather than replacing them. Discover how this innovative approach eliminates the need for explicit continuity constraints like hanging nodes, significantly simplifying refined mesh handling in finite element computations. Learn about the algorithm's caching mechanism for local element contributions that reduces assembly overhead, making it particularly beneficial for local time-stepping schemes in discontinuous Galerkin (DG) methods where block-diagonal mass matrices enable per-element timestep execution without global assembly requirements. Gain insights into practical alternatives to traditional AMR frameworks specifically designed for Ferrite.jl workflows, with detailed explanations of implementation strategies and performance advantages. Understand the theoretical foundations and computational benefits of this superposition-based refinement strategy, including its impact on mesh adaptation efficiency and finite element method implementations within Julia's scientific computing ecosystem.
