CI²MA - Publications | Preprints

Preprint 2010-01

Stefan Berres, Raimund Bürger, Rodrigo Garces:

Centrifugal settling of flocculated suspensions: A sensitivity analysis of parametric model functions

Abstract:

The centrifugal settling of a flocculated suspension in a rotating tube can be modelled by a strongly degenerate parabolic partial differential equation whose coefficients depend on two material-specific model functions, namely the hindered settling function and the effective solid stress function. These model functions are usually given by certain nonlinear algebraic expressions that involve a small number of parameters. The present work is related to the problem of determining these parameters for a given material. This problem of parameter identification consists in minimizing the distance between observed and simulated concentration profiles by successively varying the parameters employed for the simulation, starting from an initial guess. The feasibility and robustness of this procedure, which does not necessarily lead to a unique solution, decisively depends on the sensitivity of the solution of the direct problem to the different scalar parameters. These sensitivities are evaluated by a series of numerical experiments. It turns out that the model is extremely sensitive to the choice of the so-called critical concentration marking the transition between hindered settling and compression. Moreover, the robustness of the parameter identification method depends significantly on whether intermediate (i.e., transient) or stationary concentration profiles are used for identification.

Download in PDF format PDF

This preprint gave rise to the following definitive publication(s):

Stefan BERRES, Raimund BüRGER, Rodrigo GARCES: Centrifugal settling of flocculated suspensions: A sensitivity analysis of parametric model functions. Drying Technology, vol. 28, no. 7, pp. 858-870, (2010)

 

 

  CI²MA, CENTER FOR RESEARCH IN MATHEMATICAL ENGINEERING, UNIVERSIDAD DE CONCEPCIÓN - MAILBOX 160-C, CONCEPCIÓN, CHILE, PHONE: +56-41-2661324