Abitibi boosts newsprint output and quality at lower production
cost with PC-based control system, fieldbus technology.
Richard Hatten is the Electrical and Instrumentation Superintendent
for Abitibi Consolidated at Steilacoom, just outside of Seattle,
WA.
 The
Steilacoom mill, with around 210 employees, produces up to
127,000 tons of newsprint per year. Major customers are large
West Coast metropolitan daily newspapers, including the Seattle
Times, the Orange County Register, the Sacramento Bee, the
Everett Herald, and the Tacoma News Tribune.
The facility, which dates back to the turn of the century,
began producing newsprint in the '40s. The plant's currently
operating machine dates back to the mid-'60s. So maintenance,
repair, and upgrading equipment are a continual process, and
a key part of Hatten's job. In 1999 Hatten oversaw a three-area
changeout of the plant's automated control system, to a PC-based
digital system, including field-based architecture and FOUNDATION
Fieldbus technology. Here, Hatten recounts the system's reduced
installation and maintenance costs, and the resultant increased
control, performance, and product quality at the mill.
The Control challenge
Up until September of 1999, the Abitibi Steilacoom mill had
issues with reliability and troubleshooting of control hardware.
The distributed control system we'd been using in most of
the mill dated back to the late eighties and was the industry
standard. The racks of stand-alone, digital analog controllers
on our paper machine were micro-processor-based, but difficult
to configure and maintain, and the controls were poorly documented.
It was very difficult for us to manage, and we were largely
dependent on contractor help for that management.
Intermittent hardware failures in the wet end of the paper
machine had caused 10 to12 hours of shut down time in the
month of June alone. The other major issue was correcting
improperly implemented control strategies. When a controller
was programmed, it did not generate a configuration document,
so there was no accurate record of how it was last configured.
The only way to tell was to manually scroll through all of
the setup menus.
Our first challenge was in the reliability of several stand-alone
controllers. The wet end cabinet is installed ten feet from
the headbox -- obviously a warm and humid environment. The
system that pressurized the cabinet had failed and the air
conditioning had failed, so the controllers were subject to
this adverse operating environment, which contributed to failures.
A second challenge lay in some improperly set up control
strategies and processes -- specifically on our dry end pulper.
These caused weight swings on the wet end of the machine,
and often low-consistency stock back through the broke system.
If we added water to correct consistency, somewhere down the
line, we'd have to take that water back out again. This manual
correction invariably threw off the white water balances in
the mill. This introduced variability on all the consistency
loops that were dependent on a common white water source,
compromising quality coming out of the pulper.
A third challenge was with the five sets of low volume condensate
receivers, which reside in a very high vibration environment.
These receivers drain the condensate out of the dryer section
of the paper machine. The old pneumatic controls in this area
were very susceptible to failure from vibration. As a consequence,
we seemed to have a technician in every other day to repair
those controllers. Adjustment was difficult and imprecise,
and often the job was improperly done. The receiver levels
would go unstable, and condensate would either back up into
the dryer cans, which reduces drying capacity and slows the
machine down -- or they'd run the receivers dry and we'd cavitate
and damage the pumps that remove the condensate.
The switch to Fieldbus
When we began looking at alternative systems, we'd been paying
a six-digit annual service contract to maintain the existing
system, and I didn't feel we were getting a good return for
that fee. I'd had bad experiences with incompatibilities in
gateways between control systems and PLCs, so I wanted one
clean, integrated system.
Around February 1999, Representatives from control engineering
company PCE Pacific asked me to review Austin, TX-based Fisher-Rosemount's
DeltaV system, and its field instrument integration path.
We had a considerable investment in intelligent field instruments
and an asset management system. I reasoned that if any system
was capable of integrating these components, I wanted it.
I'd first heard about the DeltaV system among the emerging
PC-based control systems a couple of years earlier. But at
the time, I didn't believe it offered all the features I wanted
in a control system. Now, two years after its inception, I
saw a system designed to work in conjunction with the instrumentation
that I was putting on the machine today.
Implementation
In May 1999 we began installing the new system to control
the 55 I/O in the machine's wet end. We had several days outage
to pull out the old controllers, modify the cabinet to accept
the new hardware, and reterminate the existing wiring. We
contracted our local representative to implement the existing
control strategy in the new platform. To minimize impact on
the operators, we emulated the existing faceplates in the
new graphics. Everyone was nervous about the new system's
startup, but it went very smoothly. When we did have a process
problem that took the machine off line, we were able to use
the new historian to systematically eliminate potential causes
for the problem. Since it presents time-stamped events in
the same view as the graphical history, determining cause
and effect was easy.
Then in September of 1999, we implemented the first phase
of the dry end. This was a perfect opportunity to get our
feet wet with fieldbus, because it was a stand-alone application
with pneumatic controllers and some Fisher valves. We purchased
a cabinet and the configuration service. We developed all
the graphics and did all the loop drawings and the training
documentation for our operations staff. To resolve the ongoing
control headache in the area with its high vibration level,
we put new fieldbus differential pressure transmitters and
valve positioners on the existing valves, all connected via
one fieldbus segment going out to all five receivers. In each
receiver, the PID control runs between the transmitter and
the valve, so we now have five loops running on this one field
bus segment. They all execute at about 400 milliseconds, for
extremely fast control. The whole system for five level control
loops and a serial interface to the PLC for motor control
was installed in an eight-hour shift. By then our people had
had some training on the new system and we took over the project's
configuration.
The next area was the steam and condensate system. Here,
we put in the cabinet we'd purchased for the condensate receiver.
We did all the fieldbus and hardware engineering internally.
We did all the graphics, and after some initial fieldbus configuration
help, we did most of the control configuration internally
as well. The commissioning was that easy.
Reduced installation cost
On the dry end, there are 23 FOUNDATION fieldbus transmitters,
and five fieldbus positioners, five that are analog, and five
motors hardwired to the PLC. We cut the number of twisted
pairs required from thirty-three to eight, and because of
the serial interface to the PLC, we did not have to re-wire
any of the motors. In addition, we were able to eliminate
our traditional hard-piped conduit to each transmitter by
using pre-fabricated Fieldbus quick-connect cables with stainless
steel connectors. These cables have excellent temperature
and environmental ratings, and are standing up well to the
dryer section's hot, humid environment. The serial interface
was written in the native communication language of our PLC,
so it was very easy to implement.
Reduced training time
Due to an oversight on my part, the
operations staff was out of the mill on a two-week curtailment
when they put this phase of the system in. I had written up
some narratives and made some graphics, but hadn't scheduled
any formal training. So when these operators returned to the
plant, they found a totally new system to operate, with absolutely
no training. To their credit, everyone on the staff picked
up on the system right away. In fact by the next day, they
were offering suggestions for improvements. The system is
so simple and intuitive that operators, based on a little
understanding of the process, were able to navigate their
way around the whole system, and get the information they
needed to control the process, with just a few simple key
strokes.
Similarly in the wet end, when we implemented the system
changeover, we programmed the new system's faceplates to look
similar to the old single-loop controllers. So when we made
the transition, the operators came in and, with only a couple
of hours training, were able to navigate and control the machinery.
Reduced downtime
We immediately eliminated down time losses. And we calculate
payback on the system, based on previous downtime, at 1.8
years -- a rather quick return on our capital expenditure.
Reduced Maintenance
Perhaps the best maintenance result of the changeout is in
the steam and condensate area. Here, I can now watch the positioners
hold their condensate levels within a fraction of an inch.
Since the changeover, this area has gone from being the number
one maintenance problem on the entire machine, to being a
non-issue. It's just basic control, but the control is being
executed right there on the valve positioner in that high
vibration environment, and they stay dead on.
Reduced Operating Costs
I've also been working on a program to eliminate all of the
old control system hardware over a period of time equal to
the amount we would pay on an annual basis for the system's
support.
The second thing we've seen -- though not a high cost impact
for us -- is big savings in steam consumption, which we have
documented at 200,000 pounds per month. The original control
strategies were improperly implemented on the thermo-compressors.
One of the first things the dry end operators pointed out
to me was that we were no longer using any make-up steam.
And the reason is simply that the old single loop controllers
could not maintain levels and differential pressures adequately
by flashing condensate from high-pressure receivers into the
next, lower pressure, receivers, thus saving energy. So the
mill had been using makeup steam at every set of dryers. It
was very simple in the new control system to get the thermo-compressors
working properly to provide the flash steam.
Improved product quality
The side benefit from increased level and pressure control,
which will ultimately be far more important than the steam
savings, is the quality of the paper itself. Though difficult
to quantify, the quality was evident right away in the form
of more uniform paper. With better level and differential
pressure control comes better temperature control. During
the first month's operation, this has resulted in dramatic
moisture variability reduction, which in turn greatly impacted
the quality of sheet caliper, basis weight and moisture content,
and bottom-line profitability of the machine.
Improved Control
With the installation of the new system, I got a huge surprise
when I saw the tuning numbers in the dryer section. In the
past, we had to rely on experience and gut feeling to tune
the loops in the dryer section. With the tuning information
provided by the new system, we can now see the hard data and
make truly informed decisions. This aspect of the system has
empowered our operators. We've given them information that
they'd always lacked before. They're able to use that information
to make good decisions. And the reliability of making grade
changes is an immediate win.
Improved production
Possibly the most important benefit we can attribute to the
new system is an output improvement. A sheet break on the
machine puts it into an unstable situation. When you thread
back up you've got to cycle out the variabilities before regaining
normal inspector grade. With the new system, we've been able
to reduce the number of sheet breaks. And when we do get a
break, we're now able to thread up and get back on grade faster.
Conclusion I'm convinced that we've met every criterion we
started out to meet in terms of reduced down time, and improved
quality and reliability. None of the controllers on the fieldbus
side has given us the slightest problem, and we feel that
the maintenance cost of the system will be greatly reduced
as a result of all the off-the-shelf equipment that comprises
the new system. And we fully expect to see a big payback in
reduced long-term maintenance of this system. Combine that
with the supportability from a local sales force from the
vendor, and I think we've got a perfect solution.
In summation, the new system enabled our maintenance and
operations staff with some tools that allow them to take our
paper machine to the next level of performance and quality.
With the improvements we've implemented in this old plant,
the quality of our product is on a par with any of our competitors.
We now have a direction in which to continue cost-effectively
increasing our quality and output. Our future steps include
continuing to phase out the older generation system, opening
further opportunities for variability reduction and production
increases we can capitalize on.
back to top
|
