Chemical Engineering October 2010 - 54

Solids Processing
Figure 3. In-plant tests were conducted with a mixer
equipped with a high speed disperser and an anchor designed
to generate axial and lateral flow
(Figure 1). Since the UF-based adhesive
only reached 4,000-6,000 cP, feedstock
presented a huge technical shift
in this regard.
The team arranged a series of process
tests with Ross, the mixing equipment
partner. The tests were run on
laboratory- and pilot-scale mixing
equipment, using actual soy-based adhesive
ingredients to replicate conditions
on the process line. A successful
laboratory test identified a dual-shaft
mixer, equipped with a high-speed disperser
and a three-wing anchor agitator,
as the optimal solution (Figure 3;
see box, Equipment for Mixing HighViscosity
Materials). Columbia Forest
Products rented a 100-gal mixer for
the pilot phase of development.
The team established its development
center in the company's plant
in Klamath Falls, Oregon. With the
mixer operating alongside a dedicated
glue spreader, the team systematically
explored the influence of key process
variables, including pH, soy solids load,
cross-linker concentration and various
experimental additions designed to
modify tack and other properties.
Testing began in earnest in late 2004
using 24 × 24-in. samples of seven-ply
hardwood plywood made in the forest
products laboratories at Oregon State
University. A battery of tests was required
to fully assess the performance
of each sample, which made this a
laborious process. Industry-standard
tests for each trial panel included the
following: a dry shear test, a cyclic-boil
shear test, and a decisive three-cycle
boil test.
During this initial phase, viscosFigure
4. With systematic experimentation during in-plant
tests, the process is optimized - and later automated
ity immediately emerged as the most
persistent challenge the team would
face during testing and rollout. Coaxing
the 200,000+ cP material to flow
on the glue spreader was difficult, and
the adhesive was extremely difficult
to pump. On numerous occasions, with
pumps bogged down, lines locked up or
a hose blown, team members carried
glue to the spreader in 5-gal pails to
continue tests, while handling equipment
was being repaired.
Because of the higher viscosity of
the adhesive, breakdowns, repairs and
upgrades were routine. But the challenge
in handling higher-viscosity material
was really only half technical.
The team also had to overcome its own
expectations about the capabilities of
the existing equipment.
The team found that plant staff
accustomed to handling much lower
viscosities tend to underestimate the
challenge of pumping the thicker adhesive.
During the rollout, staff in
virtually all of Columbia Forest Products'
North American plants were
determined to move it with existing
equipment, but failed. Eventually, the
decision was made to upgrade to highcapacity
progressive cavity pumps and
similarly robust ancillary equipment
in every plant.
In-plant trials
In-plant trials are immensely beneficial,
because they generate data in
conditions that mimic actual production
(Figure 4). But in most companies,
where floor space is limited for nonproduction
activities, testing occurs
near ongoing production lines with
54 ChemiCal engineering www.Che.Com oCtober 2010
employees nearby who are not directly
involved in the tests. In such cases, the
development team should remain sensitive
to the image that testing presents
to others. Dramatic " failures " in
a test phase are usually not disturbing
to members of a development team.
After all, a " failure " is simply another
data point that helps to define process
limits. But when tests are conducted
in full view of others in the plant, the
sight of seemingly " unsuccessful " tests
can be demoralizing.
In the first full-scale mill tests, as
process variables were scaled up from
test sizes to full-size plywood sheets,
negative results were inevitable. The
moisture content in the first panel
was too high, for example, because
the solids content had been lowered
in order to lower viscosity - with offsetting
adjustments to other additives
to prevent a loss in performance. This
caused the panel to stick to the press.
In other tests, excessive steam pressure
in the hot press essentially blew
the panels apart - until the formulation
returned to a higher solids content
(and consequently, higher viscosity).
After this first round of tests, naysayers
predicted failure: " You'll never
be able to make plywood with glue
this thick. It just won't work! "
In fact, further changes to the adhesive
formulation soon produced
positive results. But along with the
optimal formula for mixing the adhesive,
another discovery was made: the
importance of anticipating the impact
on internal audiences when conducting
in-plant testing. The unflagging
confidence of plant employees, along
http://www.Che.Com
Chemical Engineering October 2010

Table of Contents for the Digital Edition of Chemical Engineering October 2010
