Rockwell Automation 1746-BTM Barrel Temperature Control Module User Manual - Manual (Page 56)

Publication 1746-UM010B-EN-P - April 2001
5-10 Control and Autotune a Loop
BTM Auto Tune The 1746-BTM Auto Tune procedure was designed to be performed
as a one-time event from which all characteristics of the system being
controlled could be identified and incorporated into the control
scheme.
The identification procedure has two critical points:
Before exercising the system with the identification procedure the
system in question must be as stable as possible. Essential to a good
system identification routine is the assumption that there is a high
correlation between excitation and system response. The identification
procedure tries to identify the system by assuming that any excitation
used to disturb the system is solely responsible for the observed
reaction of that same system. If heat or any other form of input is
exciting the system other than what the identification routine is aware
and in control of, the routine will draw erroneous conclusions about
its observations and misidentify the system in question. Thus at the
onset of auto tune it is desired to let all of the influences other than
direct auto-tune excitation dissipate before excitation is applied.
In the BTM there is a wait period at the beginning of the auto tune
procedure. Once it is initiated for all of the aforementioned effects to
dissipate. Then and only then will the routine qualify the system as
stable and begin excitation. This, however, does not mean that the
system must have zero output. It does mean that the system must
remain at a consistent temperature with very little fluctuation. There
are two ways in which this condition can be achieved. The first and
most typical, is that of a cold-start condition. In this case there is no
energy being inputted to the system and the system is stable with
respect to ambient conditions. With regards to the BTM this is
achieved easiest by setting all active zones to Manual Mode with 0%
(zero percent) output. The second most common way to achieve
stability is with a non-zero output being applied to the system, either
in Manual or Automatic Modes. It is perfectly acceptable to have a
non-zero output applied to the system as long as it is stable: in
Automatic Mode this is stability at a setpoint, in Manual Mode this is a
single manual output value and a settled response (no temperature
variation). A common mistake is to have heat applied to the system
prior to auto-tune and turn it off just prior to auto-tune initiation. In a
high lag situation where it takes some finite amount of time for that
energy to manifest itself as temperature, the result is as follows. The
auto-tune routine waits for stability and qualifies it, it then attempts to
exercise the system by giving a step output, the heat that was
previously introduced into the system finally produces a temperature
change, the auto-tune routine marks the rise in temperature as being a
direct result of its excitation of the system and records the amount of
time from the beginning of its step output to the start of the
temperature rise as the deadtime. In this case the system is now
misidentified with an artificially short deadtime.