Baelz Automatic 6490B Device Manual page 12

Device manual
3. Target setpoint is not reached during the optimisation
Immediately after the optimisation is finished ( "tunE" does not appear anymore) the actual value is not close to the
target setpoint. It is recommendable to reach the target setpoint as exactly as possible at the end of the optimisation to
get really good results.
Reasons:
a) The process gain P.G, defined before starting the optimisation, did not correspond to the actual process gain P.G of
the process. Frequently this happens during the first optimisation when the process gain P.G is still set to the standard
value = 100%.
Solution: Restart optimisation. The setpoint value will be reached more exactly this time, because the process gain P.G,
which was also calculated during the previous optimisation process, is used as a base of the following
optimisation process. If the process gain is known or measured, it can be adjusted manually already before
starting the first optimisation run.
Measuring the process gain P.G in manual mode:
Change the actuator about a fixed rate ∆Y (%) and determine the given change of the actual value ∆PV.
Then the process gain can be calculated by P.G = (∆PV / ∆Y) * 100%. If the controlled system has got a
linear behaviour, the process gain is constant all over the entire control range.
I.e. the actuator is changed from 30% to 70% → ∆Y = 40%. Thereby the actual value rises from
50°C to 110°C → ∆T = 60°C. At a measuring range of 0°C...300°C this corresponds to the change of the
actual value ∆PV = 20%. The process gain can be calculated then by P.G = (20% / 40%) * 100% = 50%.
Depending on the process gain, the controller calculates the necessary change of the actuator at start of the
optimisation for reaching the target setpoint at the end of the optimisation.
A small process gains causes a bigger change of the actuator instead of a bigger process gain.
If the temperature rises up to a not permitted high value, it could be necessary to cancel the optimisation
(see also 5.).
b) In non-linear controlled systems, even by proceeding a following optimisation, the target setpoint can not be reached
exactly enough.
Solution: Let the optimisation run a couple times until the target setpoint is reached exactly enough. The process gain will
be defined then by an iterative method, what means, with every run the process gain comes closer to the actual
process gain.
In non-linear controlled systems for different sub-ranges within there will be optimised, different optimisation
results will be created. Therefore it is necessary to determine the most important range for the control which
should be optimised. If all ranges do have the same importance, we recommend you to optimise the sub-range
with the slowest time behaviour (see also 6.a) and 6.b)).
c) The prime energy is not sufficient to reach the target setpoint.
Solution: Increase prime energy or chose a target setpoint that can definitely be reached.
d) The actuator does not move to the new position given by the controller.
Solution: Check function of the actuator and its control.
4. The optimisation "does not" finish or just after 42 minutes respectively
The maximum time of optimisation is limited up to 42 minutes. In case that the conditions to finish the optimisation are not
given even after 42 minutes, the optimisation process will be cancelled automatically.
Reasons:
a) The limited time of 42 minutes for optimisation might be too short for several, very slow processes.
Solution: Switch over to the configuration level just before the 42 minutes are elapsed and change the setting "OPt = 1"
to "OPt = 0". Therefore the optimisation is cancelled manually and the control parameters will be recalculated.
b) At processes with no stable state (drift, post heating, ...) cancelling the optimisation after 42 minutes is possible just as a
later ending.
Solution: The movement of the actual value has to be observed to recognize the approximate end of the settling.
Afterwards in the configuration level the setting "OPt = 1" has to be changed to "OPt = 0" to cancel the
optimisation with recalculation of the control parameters.
If there is a drift, the optimisation has to be started from the manual mode before the drift starts.
c) Because of the change of the manipulated variable at start of the optimisation, the change of the actual value ∆PV is
too small, so the balance of the controlled system is not recognized.
The change of the actual value ∆PV has to be at least 1/4 of the difference between target setpoint and actual value at
beginning of the optimisation.
W. Bälz & Sohn GmbH & Co. Koepffstraße 5 D-74076 Heilbronn
6490B / 6490B-y / 6590B
Telefon (07131) 1500-0
Page 12
Telefax (07131) 1500-21