Model-Free Adaptive Control
of pH Loops
|• Manipulates reagent flow intelligently
||• Improves pH control by at least 50% reduction
|• Does not over-dose reagents.
||• Chemical (acid and caustic water) consumption
is sharply reduced.
|• Enables automatic control of pH value
in all ranges.
||• Environmental pollution and equipment
corrosion are tremendously reduced.
|• Reduces product pH variation.
||• Product quality and production efficiency
|• Improves efficiency & productivity.
||• Full investment is returned in months
if not sooner.
The trends compare MFA and PID control on the same
strong-acid-strong-base pH processes. MFA (top)
controls pH value tightly in all ranges at setpoint
7, 11, and 3. PID (bottom) is either sluggish or oscillating.
MFA pH Control at Rohm & Haas reported in Chemical
Rohm and Haas, a leading chemical company,
is successfully using an MFA control system to control
a problematic pH control loop in neutralization of an
organic process stream. The cost savings is estimated
at $170,000 per year. They also expect a reliability
improvement due to reduced formation of solids, according
to Teshome Hailu, control engineer at Rohm & Haas.
The stream to be neutralized was a two-phase stream
with varying concentrations of acidic species. In general,
pH control is difficult due to the nonlinearity of a
pH loop and this measurement had significant noise.
The original system was designed with only one valve
controlled by a Fisher Micro DCI controller. This introduced
time delay problems. Because the plant is already running
it would have been prohibitive to shut down operations.
Ordinarily, re-engineering the process would require
mathematical trans formation of process data to create
a linear solution. The recommended pH set point was
10.6; operators typically ran the process at 12 because
the pH loop became unstable close to the recommended
set point. Excess caustic from the higher pH resulted
in solids formation in the downstream separation equipment.
Setting the PID controller gain low enough to ensure
stability near the re commended pH point also resulted
in an extremely sluggish control response when a large
upset pushed the pH far away from the neutrality region.
Staff decided to retain the Fisher controller and feed
it a new input signal from a CyboCon
CE Model Free Adaptive controller to
compensate for the time delay. Once proper communication
was established there was no complicated tuning, step
testing, or data collection involved.
Improved pH control enabled them to lower the pH set
point from 12 to 11. Not only were cost benefits achieved,
operators also like the improved process upset handling
capabilities. In addition, reduction of excess caustic
and reduced solids fo rmation meant an unquantified
improvement in overall system reliability.
Other pH applications with CyboCon
CE controllers, such as Chiron in California
and Ultrafertil in Brazil, achieved similar results
within a short ROI period.