www.foodtechnology.co.nz 45
farms daily and transported to milk
powder production plants, where a
portion of the raw milk is separated
into skim milk and cream. The milk
is typically standardised by blending
raw milk, skim milk, lactose and
permeate to meet international
standards. The standardised milk
is then concentrated in multi-effect
evaporators and spray dried to form
powder. Optimal operation of these
processes is essential to reduce
costs, maximise capacity and meet
final product quality specifications.
When embarking on process
optimisation, Miraka called on
the specialist process control
capabilities of Rockwell Automation.
Leveraging the company’s Pavilion8
model predictive control software
platform, Miraka teamed closely
with engineers to specify and
deploy a system to optimise its milk
powder production plant. “There
was a high level of engagement
from all the key stakeholders
which facilitated a fast paced and
successful implementation,” Rockwell
Automation technical manager
Lynn Medich says. “Leveraging
our expertise in deploying model
predictive control solutions in dairy
standardisation, evaporation and
spray drying, we worked closely with
Miraka to engineer a solution that
generates operational improvements
and business value.”
All manufacturing processes have
variability that can be caused by
many factors. The primary objective
of the model predictive control
solution is to enhance stability in the
process unit through the reduction of
variability of key process parameters
and optimal control to desired
targets. The system includes modules
to control, analyse, monitor, visualise
and integrate information and
processes as well as encompassing
and managing dynamics and
changing disturbances that occur
minute by minute, delivering
enhanced process stability so that
defined control objectives can be
achieved. In the case of powdered
milk production, reducing the
variability of key process parameters
enables plants to drive to their
processing and specification limits
optimising yield and capacity.
The Miraka solution has been
implemented in phases to effectively
introduce model predictive control
technology to operators while
delivering value. Real time key
performance indicators were
deployed quantifying the value
being delivered as each unit was
commissioned. The evaporation
process was optimised in the first
phase, with variability reduced and
total solids targeting improved. The
evaporator total solids target was
lifted, resulting in better thermal
efficiency across the evaporation and
drying stages and increased capacity.
In the second phase spray dryer
model predictive control was
commissioned. “Deploying model
predictive control on Miraka’s spray
dryer allowed us to improve yield
through closed-loop moisture control
combined with targeting an increased
powder moisture,” Rockwell senior
application engineer Kerryn Sakko
says. “Capacity was additionally
maximised by pushing the process
safely to constraints.”
Implementation of multivariable
model predictive control on the
standardising process was delivered
as the final phase of the solution.
Use of multivariable models for
liquid composition control combined
with inferential models for tracking
composition through the entire
drying process resulted in reduction
of protein and fat variability in the
final powder. Additionally, the liquid
composition controller included
additional objective functions
to maximise use of lower cost
standardising ingredients.
Through consultation with Miraka,
a set of parameters were identified
that would provide information
about the effectiveness of the
process improvement. The identified
parameters were quality (powder
moisture yield); evaporator solids
(plant throughput improvement); dryer
thermal efficiency (dryer capacity
improvement); reduction in energy per
tonne of product (cost reduction); and
protein and fat optimisation (yield).
General manager of operations
at Miraka Paul Trewin says the
Pavilion8 solution has provided
the company with an exceptional
result. A capacity increase of more
than 4% was achieved during the
peak season, moisture targets were
raised by 0.04%, and improved
protein and fat control provided the
potential to reduce fat plus protein
giveaway by more than 100 tonnes
per year. In addition, optimisation
of the evaporator solids and drying
profiles resulted in an estimated
energy savings of greater than 1%
per tonne of milk powder during peak
production.
/www.foodtechnology.co.nz
