Help get this topic noticed by sharing it on Twitter, Facebook, or email.

Control on Stochastic or Erratic Signals

Filtering Stochastic or erratic signals may be quite troublesome.
The spread in signal level may be over 10-50% of the reading in successive samples.
These wild varying signals come from field instrumentation with a natural spread by origin. Such as pH / O2 which are driven by scarse electrons.
By nature Wind Speed or Direction, (Blood) Pressure etc. or derived from resulting calculations or by sample instrumentation (GC / QMS) cq. optical imaging.
Others can be due to bad or ageing of input conditioners eg. capacitors or rubber seal/membrane. (Should be resolved by replacement if possible ).

Filtering techniques as eg. (weighed) averaging do not lead to proper results in control. Control on these raw and wild signals will require serious dampening or detuning otherwise leading to deterioration of the control element.
We have had successes by applying hardware first order filters and by simple low pass filters. Filtering parameters could be selected / adapted runtime.
Filters bypassed when bringing plant to operation condition manually.
The FoxboroPAC FBD library allowed to make optimal use of its contents for our instrumentation.
Fast trends and distribution histograms presented on MMI / HMI aid in getting grip on the situation.
1 person likes
this idea