Nanodac – Carbolite nanodac User Manual
Page 98
nanodac™
MC27 –EN– 1.04
96
EXAMPLE 4: WHEN TUNE R2G = R2GPD, AUTOTUNE FROM BELOW SETPOINT
Figure 6.4.6 Autotune from below setpoint
Periods A-F are largely unchanged from the 'Standard' algorithm, example 2 in
, with the following
exception:
- Changing the Target Setpoint during period A-B will not change the tuning setpoint.
Period F-H is replaced as follows:
F to G
Heat is applied for a period (F-G) of half the first heat cycle (D-E) to compensate for
the last cool cycle.
G to H
This is a period in which the controller is put into PD control.
The values of proportional term and derivative time for this period of PD control
are determined by the algorithm.
H
OPss is the output demand value at the end of this period and is used in the
determination of R2G.
B2.4.7 Manual tuning
If, for any reason, automatic tuning gives unsatisfactory results the controller can be tuned manually.
There are a number of standard methods for manual tuning, the Zeigler-Nichols method being described
here:
1.
Adjust the setpoint to its normal running conditions (assumed to be above the PV so that ‘heat only’
is applied.
2.
Set the integral and derivative times (Ti and Td) to ‘Off’
3.
Set High and Low cutback (CBH and CBL) to ‘Auto’.
4.
If the PV is stable (not necessarily at the setpoint), reduce the proportional band (PB) such that the
PV just starts to oscillate, leaving time between adjustments to allow the loop to stabilise. Make a
note of the PB at this point (PB
′
)
, and also note the oscillation period (‘T’).
If the PV is already oscillating measure the oscillation period (‘T’) and then gradually increase PB to
the point at which oscillation just ceases. Make a note of the PB (PB
′
)
at this point.
5.
If the controller is fitted with a cooling channel, enable this now.
6.
Observe the oscillation waveform and adjust ‘R2G’ until a symmetrical wave form is observed (Figure
6.4.7).
7.
Set PB, Ti and Td according to table B2.4.7