Linearize
Your Process - Get Optimal Response at All Production Rates
by John
Gerry, P.E.
Click
here to see the complete presentation in more detail as a web-cast.
A
non-linear process will have different response across the operating range. It
will be sluggish at one end of the range and comparatively faster or even
oscillatory at the other.
In the
graph taken from a flow loop, the setpoint is yellow, the process variable white
and controller output green. This loop is stable where the first setpoint change
is made. At the second setpoint change, the loop is getting oscillatory at the
higher value of the controller output.
This is
the same non-linear flow loop. The controller is in manual and colored green.
The process variable or flow is white. We changed the controller output by equal
intervals. At the lower part of the range the flow changed very little. When the
controller outputs gets higher, the flow changes a lot. This is another way to
see the non-linearity.
This
graph is created by taking the controller output from the previous graph and
plotting it against the flow or PV. This is called the process characteristic
and is another way to see the non-linearity in the loop.
For
example, a change in controller output represented by distance between the green
lines, causes a change in PV or flow represented by the distance between the
lavender lines.
You can
see the non-linearity by looking at the same change in controller output at
different parts of the range. At the low end of the controller output range, the
controller output change causes a small change in PV. This same change in
controller output at the higher end of the range causes a much larger change in
flow. The process has a small gain at lower controller outputs and a high gain
at higher controller outputs.
If
the maximum gain of the process divided by the minimum gain is larger than 2 or
3 then characterization will help the performance by a factor of 2 or 3.
Add a
characterizer block at the controller output as circled in green. Many control
systems already have characterizer blocks you can add to the controller output.
What
goes in the block is critical of course. The characterizer should re-map the
controller output so the net effect is that the controller sees a linear
process. In a sense, the characterizer cancels the non-linearity of the process.
It does this by using a shape that is opposite that of the process
characteristic.
Using
the code on the right in the characterizer block, linearizes our flow loop. The
last part of the code, underlined in red, handles where the controller output or
CO is large. A red arrow points to this same area on the process characteristic.
When the controller output or CO is large, the code multiplies the CO by 0.5627
to find the final output. 0.5627 is circled in green. This small multiplier
compensates for the large process gain at larger CO's.
Many
control systems let you enter the characterizer "code" as an X-Y pair list
as shown here on the right.
ExperTune's
Advanced
software includes a characterizer
wizard that will build the best characterizer for you based on plant data
you collect. It let's you choose the type of characterizer to match your control
system and you can graphically customize it by dragging, adding or deleting its
design boxes.
ExperTune's
software also includes a pH characterizer, but this is a subject for another
presentation.
Click
here to see the complete presentation in more detail as a web-cast.
This
article was written and provided by John Gerry P.E., president of Expertune.
Expertune designs pre-packaged industrial software which maximizes productivity
and efficiency and reduces waste in the process industries: chemical, pulp and
paper, utilities, refining, and food processing. For more information on
Expertune, please visit their website at: www.expertune.com.
(c)
copyright 1999-2003 ExperTune Inc.
