MODELOS MULTIVARIABLES PARA PREDECIR EL EXTRACTABLE DE NÍQUEL POR LA COMPOSICIÓN MINERALÓGICA DE LA MENA TECNOLÓGICA EN EL PROCESO CARON.

The investigation was made in a Loza (8 furnaces) of Punta Gorda Plant for turns of 8 hours during three months The purpose of the paper was to determine the multi-variable models to predict the Efficiency of Nickel reduction with greater effectiveness, when varying the chemistry-mineralogical composition of lateritic ore in the Herreshoff furnace, which constitutes the base for the advanced control of the technological circuit: Mine-Ore Preparation Plant- Reduction furnaces Plant, with the purpose of elevating the productivity, reducing energy consumption and diminishing the production costs. The data bases and their statistical processing were made in EXCEL. As a result, several multi-variable models were obtained , with effectiveness of prognosis superior than 91 %; the efficacy of COMPRAC Software was validated to obtain models based on the mineralogical composition; the utility of the diagnosis method was demonstrated to detect the causes of subjective errors and to select the reliable information in the evaluated technological circuit; An efficient research methodology for the empirical modeling on scale of Demonstrative Production Unit was validated , based in the securing of the technical-operative Similarity of great Plants with continuous regime of operating. [ABSTRACT FROM AUTHOR]