It was observed during early experiments with original
electronics control that in instants shortly after the grate
sweeping all measured quantities vary quickly and thus
disturbing steady trend (Fig. 2, on the left). Such quick
variations made automatic evaluation of signals trends very
difficult. Minor success was achieved using various low-pass
filters with longer time constants. However, such filters also
added unwanted dynamics to signals and thus made quick
control impossible. Also identification of the process with
such disrupted trends in signals would be very difficult. For
these reasons it was important to eliminate the peaks before
continuing with the research.
It was supposed that the peaks are caused by strong
mixing of fuel pellets in various state of burning during the
grate sweeping. The length of the grate sweeper movement
was therefore shortened from original 30 mm to 6 mm. Due
to necessity of maintaining the original burning rate of the
fuel on the grate, the time period between the grate
sweeping was shortened five times – from original 10
minutes to 2 minutes.