Fig. 1.2. Computational domain for flow past a cylinder Thus, we are interested in computing accurate error estimates for solutions that are relatively inaccurate. This is an important consideration, given that much of classical error analysis is derived under conditions that amount to assuming that the numerical solution is in the "asymptotic range of convergence", meaning that the solution is sufficiently accurate that the rate of convergence can be observed (1.10) With a little thought, we

more »

... cognize that this algorithm has the potential to be a multiscale solution procedure since we can now resolve the solution of each component on independent scales. That is one benefit of using operator decomposition. Unfortunately, this decomposition has unforseen effects on both accuracy and stability. The reason is that we have discretized the instantaneous interaction between the reaction and diffusion components. ∆t