Comparison and a possible source of disagreement between experimental and numerical results in a czochralski model

Experimental and numerical observations of oscillatory instability of melt flow in a Czochralski model are compared, and a disagreement observed at small crystal dummy rotation rates is addressed. To exclude uncertainties connected with flow along the free surface, the latter is covered by a no-slip thermally insulating ring. Experiments reveal an appearance of oscillations at temperature differences smaller than the numerically predicted critical ones. At the same time, a steep increase of the oscillations amplitude is observed just beyond the computed threshold values. By increasing thedummy rotation gradually, we are able to qualitatively confirm the numerically predicted flow destabilization. A good quantitative comparison is reached only with a rather strong rotation of the crystal dummy. Focusing on thedisagreement in the non-rotating case, we argue that the experimentally observed instability is triggered by an external excitation that results from low-amplitude temperature oscillations in thermostatic baths. This argumentis supported by a numerical simulation of the parametrically excited model.