Slurry density measurement, specially in cyclone feed streams, is
fundamental to grinding circuit control. With the purpose of increasing
the availability of such a measurement, the design of a soft-sensor in an
actual grinding plant is addressed in this paper. The system is designed to
substitute the momentary unavailability of the real density sensor, either
because it has failed or because it has been removed for any other reason. The ARMAX model structure is determined using the stepwise
regression method. Starting with a list of the candidate variables, the
method selects, one by one, and in a systematic manner the variables
which best model the measurement to be replaced, choosing them from
the set of presumably correlated candidate variables. In this paper not
only the single measurements are included in the list of candidates, but
also combinations of such measurements having physical significance. A
phenomenological model has been built with the purpose of determining
these combinations. It turned out that in most of the cases the components
of combined measurements having physical significance were selected by
the stepwise regression method, instead of the single measurements.
Actual grinding plant data is used to determine the model structure, to
estimate the model parameters and to test the predictive capability of the
developed soft-sensors. Good results are shown to be obtained during
periods of up to 30 hours of the density soft-sensor operation.