www.ipesearch.com INDUSTRY 4.0 23
FEATURE
Modern systems can reduce installation
costs, because data acquisition devices
– such as sensors – can be connected
directly to their existing wired networks.
Alternatively, they can easily go wireless
– as secure Wi-Fi networks are
increasingly common in factories.
Switching to wireless
One way of hooking into wireless is to
install a low-power wireless ‘mesh’
network. This makes it possible to install
sensors that run on battery power
alone. While the system is easy to install,
the energy cost may be difficult to
justify. However, modern wireless
condition-monitoring systems are
improving rotating equipment
performance programmes in a way that
was previously considered
uneconomical. This is done by taking
knowledge on machine-health
monitoring – collected over many
decades by manufacturers – and
combining it with network technology
from connectivity specialists. At the
same time, it minimises energy
consumption.
SKF has developed a wireless condition-monitoring
system that can automate
vibration data collection within a service
contract. In this case, a mesh network
protocol enables sensors to exchange
data by navigating around obstacles –
such as pipework and liquid storage
tanks – instead of trying to punch
through them.
Smart analytics
Condition-monitoring systems are also
becoming more economical to run,
thanks to new analytics approaches –
such as the use of machine-learning
technologies. These methods automate
the interpretation of machine condition
data to a higher degree than was
previously possible. This allows
companies to monitor more assets with
fewer skilled analysts.
Machine learning uses special
algorithms that give computer systems
the ability to learn without being
explicitly programmed. It relies on
gathering, categorising and interpreting
large amounts of data. Just as humans
cannot acquire new skills without
information – or examples to follow – so
machines are incapable of learning
without access to data. Machine
learning algorithms effectively learn
models of behaviour from the data sets
that are presented to them.
Advanced data analytics and machine
learning are already helping
manufacturers to raise productivity and
efficiency. In particular, they allow
companies to increase capacity without
significant capital investment.
New technology is also changing the
way that machine condition data is
used. While centralised – or remote –
data analysis is well established, the
internet and cloud computing have
made it far easier and cheaper to
implement. The ability to access real-time
data from a remote location has
caused a shift in productivity.
This can have huge benefits for
companies that operate multiple assets
around the world. These technologies
make the results of analyses far more
accessible. For instance, it gives a
factory manager instant access to the
status of a facility – from a glance at
their phone.
End of the walkaround?
The digital revolution delivers critical
data to even the most remote analyst or
factory manager. However, it is unlikely
to spell the end of the walkaround – but
it will change the nature of this age-old
manufacturing tradition.
In future, the raw data will already have
been logged and analysed – giving
maintenance specialists time for ‘value-added’
activities. Raw data does not
solve problems – such as a worn
bearing or malfunctioning machine –
but identifies the cause and points to a
solution.
For instance, the mass of data can help
a maintenance specialist prioritise
which machines need attention, and in
which order. At the end of the day, there
is no substitute for a hands-on
approach. It’s a little like when your car
breaks down. You may have all the fault
data at your fingertips, but only a
seasoned mechanic can make sense of
it all – and fix the car.
For more information:
www.skf.com/uk
Tel: 01582 490049
Machine-learning
algorithms
effectively
learn models
of behaviour
from the data
sets that are
presented to
them
/www.ipesearch.com
/uk
