Figure 8: Response of the model in steady state (left hand side) and an
increase of the spinning band rotational speed (right hand side).
Subsequently, after approx. 30 seconds, the speed of the spinning band
is increased from 300 rpm to 400 rpm (see Figure 8, right hand side).
This results in an increase of the pressure difference and the forecast
predicts a stronger rising trend for the next 20 seconds. Due to the
existing learned correlations, the classification of the model indicates
already at this early stage that the column is entering a flood point,
which enables the operator to take appropriate countermeasures.
It can thus be shown that the ML model is able to clearly predict the
flooding of the SBDC, therefore in the following it will be analyzed in
more detail what happens when the speed of the spinning band is reduced
(see Figure 9).