Detection for Liquidus Temperature by Embedded System with Data-density Algorithm

Abstract

Liquidus temperature (TL) of molten aluminum electrolyte accurately and reliably identified is very difficult on-site when under such the cases: 1: molten aluminum electrolyte is a complex impurities system and multi-component system; 2: low molecular ratio of NaF/AlF3 (CR); which implies there is small or no inflection point on the temperature curve and is very difficult and nearly impossible to identify the TL of molten aluminum electrolyte by both the step-cooling first derivatives and step-cooling second derivatives methods with conventional equipment. To solve this challenging issue, we developed a concrete and feasibility measurement equipment with embedded system and a novel data-density algorithm to direct detection TL of molten aluminum electrolyte on-site, which is total different from conventional off-line method with samples was measured indirect. A various of strictly experiments and tested in four big famous enterprises of China with six years over a wide range of furnaces with a varied of complex impurities system and multi-component system and derivation molecular ratios are carried out and the results some extent exhibit better performances than off-line setups in the respect of fast evaluation and determine the true TL.