EMPIRICAL MODELING OF THE BAKING FURNACE TO PREDICT BAKED ANODE PROPERTIES

摘要

A soft-sensor model developed from historical carbon plant datarnand multivariate statistical methods was proposed in past work tornobtain quick predictions of individual anode properties right afterrnbaking for quality control purposes. It could only be used forrnanodes baked at the coldest and hottest positions within thernfurnace due to the core sampling method used at the partner plantrnand the way the data is collected. To complement the soft-sensor,rnthis work proposes a strategy to account for the thermal history ofrnanodes baked at any position. It is shown that combiningrncategorical variables for pit and baking positions and routinelyrnavailable firing equipment data is sufficient to predict therntemperature history of anodes baked in different positions, whichrnis typically not measured during normal operation. The model willrnbe integrated to the soft-sensor in the future to account for thernimpact of baking position on anode properties.