LEAD FEATURE LOCATION TRACKING DELIVERS CRITICAL DATA
The bigger data picture
The ‘departure board’ is one aspect of the data mining that Larder has overseen,
there is much more that can be gleaned. Currently, intranet access to some 21
KPIs generated from ERP/SFDC data is possible, with the underlying data also
available to be viewed – Microsoft Business Intelligence is a tool also employed to
mine and present data. These KPIs help drive the company’s activities, shining a
spotlight on areas that require attention.
For example, production ef ciencies are one of the measures. Explains Larder:
“A focus that we gave the guys was production ef ciencies. However, you can’t look
at ef ciencies until you are ensuring all time is being accounted for. You clock on
to direct activities – operations on job cards – and indirect activities such as
housekeeping, preventive maintenance or training, whatever. The sum of the direct
and indirect clockings is then compared to paid hours, we call this measure
‘utilisation’ which runs at around 97%. Paid time has to be accounted for before
you can look at ef ciencies, so a worker clocks onto a job with a planned time of
two hours, completes in 1 hr, is 200% ef cient but his other 7 hrs are not
accounted for.
“We wanted to give the production teams the data to see what was going on in
the area and where efforts to drive improvements could be made. Each
department has an average ef ciency that allows the team leaders to work with
those performing under the average. Generally, improvements come from knowing
how best to process a given operation, which one worker may have mastered but
this knowledge has not been cascaded through. Then we look at the department in
relation to the division, and then we look at the division in relation to the group,
so you basically get this golden thread of performance running throughout the
business, with focused effort in the areas that need to perform.
“Because it’s demonstrable, you get away from subjective opinions around
performance, because you can show what the rest of the team is achieving. It’s
objective, it’s just data, so we don’t have to fall out about it. But an unintended
consequence of this measure was that guys were just working on whichever job
was in front of them.” So, some jobs could sit around for many tens or even
hundreds of days, which was making the departure board system ineffective as
people were not able to keep on top of data input, so things were spiraling out of
control. Location tracking and focus on priority parts and not just production
ef ciencies has brought order to the system, making it as effective as was
intended.
Dyer Engineering’s head of digital innovation is a fan of raw date, hence the
departure boards and KPIs, but the inclusion of the zone information into the
departure boards came subsequent to the idea of jobs being located visually
using the Think Inside platform. It was another piece of data, ultimately, so was
incorporated it into the existing system. It is not a silver bullet on its own, however,
but part of the whole data enterprise, he stresses.
The company is working on using existing data even more effectively. “One
thing we are looking at is understanding what needs to happen in any department.
We have 10-15 departments across 10 buildings within which there are 10,000
open operations. There’ll be a priority. There’ll be the most important job in the
factory that must be moved from place to place, the second most important, third,
etc. Currently, we just move things as best we can, but what we want to get to is an
intelligent priority list that’s dynamic – real-time tracking. As soon as a job
completes, the guy in the shop will clock off and it disappears from one departure
board and is now ready to go to the next one, but it’s in the wrong location. It’s the
hottest job in the factory, right, so go and get that now. And here’s a noti cation.
So, you’ve got someone who’s surgically moving bits
around the factory.” The system supporting noti cations is likely to be Microsoft
Flow through the Of ce 365, which would re either an email or an MS Team
noti cation – MS Teams is a mobile app.
That’s better use of existing data and systems, but the collection of more is
targeted. A currently held investment is that of machine monitoring. This will focus
attention on issues causing machine downtime, such as machine reliability, tool
life/reliability, material availability, engineering queries and so on. It is reckoned to
be able to drive up running hours, delivering an extra 10% of capacity.
KPI information is gradually being made visible to staff via the intranet, so that
they can see the business issues and react more quickly to business priorities
more autonomously. (The intranet is also home to other areas of information,
including audio books, digital training management, FAQs, document library, forms,
in-built apps, instant individual performance data, wellbeing resources and there’s
an aspiration to bring in a real-time pro t share scheme – not all yet available, this
is part of a ‘Digital Future’ landscape).
Applying arti cial intelligence (AI) to collected data is next. Dyer Engineering is
working with Durham University on a European programme that will see a PhD
student join the company for a three-year funded project looking at how AI can be
leveraged in the manufacturing sector. Larder again: “AI needs big datasets, that is
where it really comes into its own. Our quoting process is probably a primary
target. We have thousands of engineering drawings and we have the corresponding
data from the shop oor. So, I’ve got two things, the predicted or estimated bill of
operations and a bill of materials and the actual gures. And there’s going to be
some sort of sweet spot in there, work that we get the highest business bene t
from. We’ve got some feel, but we are going to turn that data into a tool that
highlights what is the most pro table out of all that we currently do.
“I might have a part that doesn’t have a great gross margin, but you know what,
it ies through the factory. It doesn’t hit the ground and is out within a week, no
hold ups. But here’s another part that does make quite a lot of money, but it sits
around blocking up all the other jobs. Everything has an impact on something
else. You might say ‘that’s great’, but it stopped all this other work getting through
it. So, you might be better to drop the one that you think is great. It’s not what you
think you should do. Data is key to this and how you manage it. We’re hoping that
the PhD student will join us in October this year.”
What of the future, then? Well, the work that the company does is not going to
fundamentally change, Larder advises. And the operations currently undertaken
likewise will remain similar. However, it’s additive manufacturing operation for its
end-of-line jigs will likely become a separate operation, with this also supporting
an external service. Within its main operations, he suggests cobots will take on
repetitive tasks, other automation could be used to move materials over short
distances between operations and there will be a larger IT department, with digital
engineers concerned with tting sensors, capturing data and driving business
improvement from its analysis. And if the company gets really good at this, there is
nothing to stop Dyer selling a service to other rms wanting to tread a similar path.
There is no doubt that Dyer Engineering, although ostensibly a traditional
engineering company at the operational level, is very de nitely moving into the
Industry 4.0 world of data-driven manufacturing. The visible hardware of SFDC,
location sensing hardware and display screens at the shop oor only hint at
something more; the resulting data and the use to which it is put are not
immediately discernible, but that invisible resource is now of key importance.
14 April 2020 | www.machinery.co.uk | MachineryMagazine | @MachineryTweets
/www.machinery.co.uk