Tools for the optimisation of mixing protocols and geometries
| Autor: | Kruijt, P. G. M., Patrick Anderson, Galaktionov, O. S., Gerrit Peters, Meijer, H. E. H. |
|---|---|
| Přispěvatelé: | Polymer Technology |
| Jazyk: | angličtina |
| Rok vydání: | 1999 |
| Zdroj: | Proceedings of the conference, 15th annual meeting / Polymer Processing Society / Ed. P.D. Anderson, P.G.M. Kruijt Pure TUe |
| Popis: | Computational tools available for the analysis of mixing behaviour are usually memory and time consuming and generally not flexible enough to allow for optimisation of mixing processes. These tools include the Poincare method, periodic point and manifold analysis, and interface tracking techniques. The Poincare method is generally used to locate periodic islands in a chaotic flow. It usually gives a good indication of how large unmixed regions can be. Manifold analysis reveals the stable and unstable regions in a chaotic flow. As a result it can be used to determine regions in the flow where high stretching occurs or where islands exist. Tracking reveals the distribution and local stretching of subdomains after a number of periods. After tracking, mixing properties can be studied by analysing these distributions. If the description of the boundary! of a tracked sub-domain is accurate, mixing measures as e.g. the intensity of segregation and scale of segregation can be applied on a discretised flow domain. These mixing measures can subsequently be used to compare different mixing protocols or changes of geometry. Although the tracking method is accurate, it is limited in its use, since the number of points required to accurately describe the boundary of a tracked subdomain, may increase exponentially every period. In optimisation of mixing flows, parameters of the flow have to be adapted and the analysis needs to be repeated. Since optimisation generally requires numerous analyses, the methods listed above are not efficient and a cheaper, more flexible method to evaluate mixing behaviour is needed. The proposed method subdivides an arbitratry domain into a number of subdomains, each with a boundary. Every boundary is tracked for a certain amount of time. For every subdomain it is then computed what ! fraction of the area is advected to the other subdomains. These fractions are stored in a concentration distribution or mapping matrix. By left-multiplying this matrix by an initial concentration stored in a concentration column, the concentration distribution after that certain amount of time will result. In this presentation the lid driven, rectangular cavity flow serves as an example. The coarse grain, or locally averaged, density is mapped. Intensity and Scale of segregation are chosen as mixing measures. The 2D algorithm for this time periodic 2D flow can, however, be expanded to 3D spatial periodic flows. Also residence time distribution can be mapped, but this is only of interest in spatially periodic flows. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|