19
9.4. AUTOMATIC SELECTION OF PID PARAMETERS
In order to use the PID parameter selection function for a given control output (1/2/3), first select the PID
parameters set (using the method described in chapter 9.3) to which the calculated data will be saved, and then
set the type of auto-tuning (with the parameter tvn1/2/3 description in chapter 8, Table 8, point IV). The auto-
tuning is activated at the start of the control (automatically after switching the supply on or manually using the
[F], [SET] function keys or the BIN binary input programmed as the start/stop of outputs operation, chapter 7.1).
Autotuning is performed independently for each of the outputs with the maximum available power (defined by
the parameter oPF1/2/3 Table 8, point II) and is signaled by cyclic messages tun1/2/3 (for the StEP method ) or
tuo1/2/3 (for oSct) or flashing of the upper right dot during the object analysis for Auto (chapter 7, point d).
The value of the parameter 36/40/44: tvn1/2/3 determines the choice of the PID parameter selection method:
a) tvn1/2/3 = Auto (continuous mode, smart logic) - the controller continuously checks whether there are
conditions for starting the tuning and tests the object in order to select the appropriate method. The algorithm
continuously forces operating in the PID mode. The necessary condition to initiate the PID parameter selection
procedure is the location of the PV current measured value outside the dead band defined as the sum of the
parameters value of the Pb proportional band and the associated hysteresis H in relation to the SP setpoint , as
in figures 9.4.
Fig.9.4.
Location of the deadband for the type of
heating type control (Fvn1/2/3 = indH)
and cooling type control (Fvn1/2/3 = dirC)
In order to avoid unnecessary activation of tuning, which may delay the course of the process, it is
recommended to set the hysteresis H to the highest possible value, not less than 10 ÷ 30% of the process value
variation range (e.g. measured temperature). Testing of the object with the short disactivation of the output and
the flashing of the upper right dot also takes place in the dead band in case of detecting sudden changes of the
measured value PV or the setpoint SP. The choice of the parameter selection method depends on the nature of
the initial conditions. For a stabilized controlled quantity, the step response method (fast) will be selected,
otherwise the oscillation (longer) method will be activated.
Automatic selection (continuous mode) enables the optimal selection of PID parameters for the current
conditions on the site, without user intervention. It is recommended for variable value control (disturbance of
set conditions during operation by modification of e.g. the setpoint SP or the furnace batch).
b) tvn1/2/3 = StEP (step response method, fast)- selection of parameters in the step stage (response to step-
forcing function). While determining the characteristics of the object, the algorithm does not cause an
additional delay in reaching the setpoint SP. This method is dedicated to facilities with stabilized initial
controlled value (e.g. temperature in a cold furnace). In order not to disturb the initial conditions, before starting
the auto-tuning, turn off the power supply of the actuator (eg heater) with an external switch or use the control
start/stop function (keys [F], [SET] or BIN input). The power supply should be switched on immediately after
starting the tuning, in the delay phase of switching on the output. Turning on the power at a later stage will
result in an incorrect analysis of the object and, as a result, incorrect selection of PID parameters.
c) tvn1/2/3 = oSct (oscillation method, longer)- selection of parameters using the oscillation method. The
algorithm involves the measurement of the oscillation amplitude and period at a slightly lower level (for heating
or a slightly higher level for cooling) than the setpoint SP in order to eliminate the danger that the target value
will be exceeded during the object test stage. While determining the characteristics of the object, the algorithm
causes additional delays in reaching the setpoint. This method is dedicated to objects with unstable initial
controlled value (e.g. temperature in a hot furnace).
The algorithms from subpoints b and c consist of the following stages:
- delay in activating the output (approx.15 seconds - time for switching on the power of the actuator, i.e. heating/
cooling power, fan, etc.) and determining the characteristics of the object,
To Next Page
Loading...