background image
not fully understood. Therefore, a proven way to get a validated model is to compare the
model prediction to the measured process response of a given power modulation event. With
subsequent calibration or further improvement of the model, it is able to reproduce the
measured response.
3-hour total power shutdown event
In this paper, a comparison between the model results and measurements taken during a 3-
hour power shutdown event will be discussed. Case data from the past are used; this
corresponds to a model validation exercise conducted in 1996 [5].
The measurements were made by VAW in 1991 on a 240 kA prototype cell (see Figure 1),
which was about to be shutdown definitively when the hydro-electric power from the Töging
smelter's own dam was directed into the power grid. The 3-hour total power shutdown was
preceded by a 4-hour voltage treatment-preheating period and followed by an 8-hour voltage
treatment as reheating period.
Figure 1: CA240 reduction cell, Töging
During the whole period, the thermal response was recorded continuously from just before the
beginning of preheating to up to 4 hours after the end of the reheating period.
The drawback of using old data, however, is that the process data now required no longer
exist. The cell heat balance and the cell voltage breakdown were not measured at today's
standards, essential to the validation of the steady-state model on which the dynamic model is
based.