The Treibacher Industrie AG produces ferromolybdenum according to a metallothermic production procedure. The primary intention of this work was to introduce an extensive new calculation model for the burden, to receive an optimised composition of the raw material for every possible concentrate. Based on thermo-dynamic calculated data a neural network has been provided, which should deliver the burden composition in order to gain the best possible yield of molybdenum. Different optimisation campaigns were carried out and examined on their effectiveness. In this work an optimisation variation could be found which leads to low molybdenum values in the slag and to less loss of molybdenum in the slag.
|Translated title of the contribution||Neuronal networks for the production of ferro-molybdenum|
|Award date||29 Jun 2012|
|Publication status||Published - 2012|
Bibliographical noteembargoed until 24-05-2017
- neuronal network