Maintworld 1/2020
ROTATING EQUIPMENT SERVICES: A COMPREHENSIVE, WORRY-FREE PACKAGE // SELF-INFLICTED RELIABILITY PROBLEMS OF ROTATING MACHINERY // VIEWING MAINTENANCE AS A SYSTEM TO OPTIMIZE PERFORMANCE
ROTATING EQUIPMENT SERVICES: A COMPREHENSIVE, WORRY-FREE PACKAGE // SELF-INFLICTED RELIABILITY PROBLEMS OF ROTATING MACHINERY // VIEWING MAINTENANCE AS A SYSTEM TO OPTIMIZE PERFORMANCE
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
1/<strong>2020</strong> www.maintworld.com<br />
maintenance & asset management<br />
Rotating Equipment Services:<br />
A comprehensive,<br />
worry-free package p 8<br />
SELF-INFLICTED RELIABILITY PROBLEMS OF ROTATING MACHINERY PG 12 VIEWING MAINTENANCE AS A SYSTEM TO OPTIMIZE PERFORMANCE PG 40
Your best weapon against poor<br />
reliability is knowledge.<br />
You need knowledge. Your colleagues need knowledge. Techniques,<br />
solutions, strategy and the business case - it is all critical knowledge.<br />
MOBIUS CONNECT is your gateway to knowledge. Whether your<br />
focus is condition monitoring or the bigger picture of reliability<br />
improvement our websites, live events and worldwide communities<br />
provide easy access to the information you need.<br />
We invite you to be part of MOBIUS CONNECT. With MOBIUS<br />
CONNECT you can...<br />
• Learn and share<br />
• Customize your learning pathway<br />
• Continue your educational journey, live or online<br />
• Engage with industry leaders from around the world<br />
The #1 Worldwide Training Event<br />
Get hands-on learning experience, participate<br />
in interactive workshops, network with peers,<br />
talk with industry experts and see the latest<br />
technologies.<br />
Learn more at www.thecbmconference.com
Find<br />
Your<br />
Learning<br />
Path<br />
15,000+<br />
Live Session Attendees<br />
1500+<br />
Educational Videos<br />
and Articles<br />
600+<br />
Contributors<br />
www.mobiusconnect.com
Break the Break/Fix Cycle<br />
Stop being reactive. Shift to a truly predictive maintenance model. Anticipate<br />
problems before they cause costly downtime, and keep your equipment<br />
and facilities running more efficiently than ever before.<br />
ICONICS’ Fault Detection and Diagnostics (FDD) technology significantly reduces costs and improves<br />
operational efficiency. It incorporates a standard library of fault rules that can be customized to predict<br />
equipment failures and advise personnel of preventive actions. It also digitizes the knowledge of your<br />
expert workers so even new technicians are able to leverage tried-and-true techniques for addressing the<br />
toughest issues. Evolve to a connected field service operation, where scheduling the right person to be in<br />
the right place at the right time happens without you having to lift a finger.<br />
Check out the<br />
solution video at:<br />
iconics.com/cfsvideo
Automatically Detect Faults<br />
ROI Typically Within 12 to 18 Months<br />
Library of Preconfigured Fault Rules<br />
Rich Visualization and Reporting<br />
Predict, Reduce, and Eliminate<br />
Equipment Downtime<br />
Reduce Maintenance and Determine<br />
Probable Causes
EDITORIAL<br />
The COVID-19 Crisis,<br />
a Maintainer’s View<br />
‘NOUS SOMMES EN GUERRE!’ I can’t state any<br />
better than the French President Emmanuel<br />
Macron, that humanity is currently at war with<br />
COVID-19. This type of lung infection is caused<br />
by the Severe Acute Respiratory Syndrome<br />
CoronaVirus 2 (SARS-CoV-2), which is spreading<br />
around the world after an initial outbreak in<br />
the Chinese Wuhan region in December 2019.<br />
Many governments are taking severe measures,<br />
including travel bans, school closures, lockdowns<br />
requiring people to stay at home, factory shutdowns,<br />
etc. The world economy has been impacted<br />
severely; stock markets colour deep red. I don’t<br />
need to explain to you that this situation is unprecedented. Let’s look at the current<br />
situation through the eyes of a maintenance, reliability and asset manager.<br />
Chronicle of a Problem Foretold<br />
Coronavirus infections have already been lurking for several years. Usually these<br />
viruses cause relatively mild symptoms, such as colds in winter and early spring. As<br />
a matter of fact, 5 to 10 percent of colds are caused by coronaviruses. But we knew<br />
from the 2002 SARS and 2012 MERS outbreaks, that coronaviruses could potentially<br />
become very dangerous. However, even after two major wake-up calls, policy<br />
makers and companies were still not willing to make the investment in preventative<br />
measures. We had plenty of time to develop medicines and vaccines to prevent<br />
today's outbreak, but it just did not happen. Sounds familiar, doesn’t it? How many<br />
times do maintainers need to remind upper management that it is better to prevent<br />
than to repent (= feel regret and remorse)? Also, the asset manager knows all<br />
too well that the risk of doing nothing needs to be incorporated into investment<br />
decisions. Unfortunately, this is often not the case. A small short-term gain usually<br />
wins over over long-term benefit. As a result, we now live in times of repent.<br />
System Overload<br />
The risk of developing a severe COVID-19 lung infection through the Coronavirus<br />
increases with age – for the geeks: it is a failure pattern B. Taking into account that<br />
in Europe 18 percent of the population is aged (compared to only three percent in<br />
Africa), hospitals are hit by an exponentially increasing number of patients.<br />
Just as technicians are firefighting an ever-increasing number of failures<br />
caused by overaged assets, doctors and nurses are overwhelmed by the huge demand<br />
on intensive care. Add to this the fact there is no proper treatment and it<br />
becomes clear we are headed for a complete health system overload.<br />
Valuable Assets Need to be Protected<br />
Just as it is the maintenance technicians that keep things running in a factory, it is<br />
the doctors and nurses (and hospital technicians) who are our most valuable assets<br />
at the front of the war against COVID-19. It goes without saying that, if we want to<br />
win the battle, we need to protect those doctors and nurses from being infected.<br />
Enter the facemasks. The global shortage of facemasks has revealed another major<br />
risk in today’s society. The relentless offshoring of manufacturing has led to very<br />
vulnerable global supply chains. If there is a single lesson to be learned from the current<br />
Corona Crises, then it is the fact that decision takers need to do more and better<br />
long-term thinking. All of us can help by making the right choices ourselves.<br />
Stay healthy!<br />
Wim Vancauwenberghe<br />
Maintenance Evangelist<br />
34<br />
A<br />
good root cause<br />
elimination program<br />
needs a process, practical<br />
training in critical<br />
thinking, and coordination<br />
and follow-up of actions.<br />
6 maintworld 1/<strong>2020</strong>
IN THIS ISSUE 1/<strong>2020</strong><br />
20<br />
There<br />
are many ways Asset<br />
Performance Management<br />
(APM) 4.0 helps you gain<br />
insights from data and<br />
optimize asset performance.<br />
=<br />
46<br />
Chemicals plants often<br />
have plenty of good<br />
data on equipment<br />
performance and reliability.<br />
A predictivemaintenance<br />
program might be the worst<br />
way to use it.<br />
8<br />
Rotating Equipment Services: A<br />
comprehensive, worry-free package<br />
12<br />
16<br />
20<br />
24<br />
Self-Inflicted Reliability Problems of<br />
Rotating Machinery<br />
Cloud-Enabled, On-Premises, or<br />
Both?<br />
DATA IS ABUNDANT Insights and<br />
Actionable Information are Hard to<br />
Find<br />
Experience Feedback – Rotating<br />
Machinery<br />
28<br />
32<br />
34<br />
36<br />
38<br />
RESONANCE - The Hidden Threat<br />
How predictive maintenance<br />
enhances plant safety<br />
Root Cause of an Electrical Problem,<br />
Did You Find the Systematic<br />
Problem to Solve?<br />
The Day After Tomorrow in Asset<br />
Performance<br />
A standardised methodology with<br />
factory specific outcome Multi-site<br />
approach with VDMXL<br />
40<br />
43<br />
46<br />
48<br />
Viewing Maintenance as a System<br />
to Optimize Performance<br />
Use of High-speed Thermography<br />
in Laser High-temperature Capillary<br />
Gap Brazing<br />
Predictive Maintenance: The Wrong<br />
Solution to the Right Problem in<br />
Chemicals<br />
Industrial AI in Maintenance: False<br />
Hopes or Real ACHIEVEMENTS?<br />
Issued by Promaint (Finnish Maintenance Society), Messuaukio 1, 00520 Helsinki, Finland tel. +358 29 007 4570<br />
Publisher Omnipress Oy, Mäkelänkatu 56, 00510 Helsinki, tel. +358 20 6100, toimitus@omnipress.fi, www.omnipress.fi<br />
Editor-in-chief Nina Garlo-Melkas tel. +358 50 36 46 491, nina.garlo@omnipress.fi, Advertisements Kai Portman, Sales<br />
Director, tel. +358 358 44 763 2573, ads@maintworld.com Layout Menu Meedia, www.menuk.ee Subscriptions and<br />
Change of Address members toimisto@kunnossapito.fi, non-members tilaajapalvelu@media.fi Printed by Painotalo Plus<br />
Digital Oy, www.ppd.fi Frequency 4 issues per year, ISSN L 1798-7024, ISSN 1798-7024 (print), ISSN 1799-8670 (online).<br />
1/<strong>2020</strong> maintworld 7
PARTNER ARTICLE<br />
Rotating Equipment Services:<br />
A comprehensive, worry-free package<br />
for pumps, turbines, compressors, etc.<br />
Many processing plants include rotating equipment that is in continual use and<br />
contributes significantly to the enterprise’s productivity. This equipment must be<br />
operated in a cost-effective manner and must not break down. To ensure this,<br />
Bilfinger offers rotating equipment services covering the entire lifecycle of the<br />
systems involved. A project currently underway at the fertilizer producer Yara<br />
Porsgrunn in Norway shows how this works in practice.<br />
BERNARDO SEQUEIRA, Business Development Rotating Equipment, Bilfinger SE;<br />
ØYSTEIN HALDORSEN, Department Manager, Rotating Machinery, Bilfinger Industrial Services Norway AS.<br />
BILFINGER’S NETWORK for the provision<br />
of rotating equipment services covers all<br />
of Europe and consists of several centers<br />
of competence; the intention is to<br />
further expand this network in future.<br />
It comprises a large number of technical<br />
specialists like engineers and technicians<br />
and allows Bilfinger’s national<br />
companies to share their expertise along<br />
with the experiences they have gained<br />
with regional particularities. Customers<br />
benefit from having a service provider on<br />
call who can offer all the required services<br />
from under one roof, across national<br />
boundaries and regardless of the type of<br />
machinery involved. Bilfinger aims to<br />
position itself as a complementary service<br />
provider alongside OEMs (original<br />
equipment manufacturers).<br />
The target industries in this context<br />
are chemical & petrochemical, oil & gas,<br />
as well as energy & utilities. Obviously,<br />
every plant features different types of<br />
equipment. Bilfinger has classified rotating<br />
equipment into two categories and<br />
has set up a center of competence for each<br />
one. The first is “heavy rotating equipment,”<br />
which includes turbomachinery,<br />
piston compressors as well as multistage<br />
high-pressure pumps and gear units. The<br />
second is “small rotating equipment,”<br />
which comprises standard pumps and<br />
drive motors. In all cases, Bilfinger provides<br />
end-to-end support specifically<br />
tailored to its customers’ needs, including<br />
manufacturer-independent consulting,<br />
installation and commissioning of new<br />
equipment; maintenance, repair and ongoing<br />
optimization during operation; as<br />
well as the final decommissioning.<br />
Investment phase, installation<br />
and commissioning<br />
The lifecycle of a piece of machinery<br />
essentially consists of four phases: Its<br />
initial acquisition in the investment<br />
phase, its installation and commissioning,<br />
the operating phase, and the final<br />
decommissioning. Each of these lifecycle<br />
phases is fully covered by Bilfinger’s rotating<br />
equipment services.<br />
8 maintworld 1/<strong>2020</strong>
PARTNER ARTICLE<br />
During the investment phase, Bilfinger<br />
advises its customers as to which<br />
equipment will best suit their needs,<br />
while offering feasibility and engineering<br />
studies along with the planning to be<br />
submitted along with the building permit<br />
application. If so requested, Bilfinger<br />
can also provide the full range of asset<br />
management services for small rotating<br />
equipment. In this case, Bilfinger offers<br />
its customer an equipment-rental pool<br />
from which roughly 20,000 devices such<br />
as pumps, electric motors or frequency<br />
converters, etc. can be leased at reasonable<br />
cost. This ensures the highest possible<br />
level of availability, as do comprehensive<br />
support services for the repair<br />
or replacement of faulty devices – which<br />
can be provided in a matter of hours, depending<br />
on location.<br />
The arrangement offered by Bilfinger<br />
known as the “Value Performance<br />
Contract” is unique on the market. It<br />
offers customers a guarantee for a respective<br />
pump’s availability in return for<br />
Operating phase and<br />
decommissioning<br />
During the operating phase, Bilfinger’s<br />
task is to provide detailed maintenance<br />
services and regular repairs for the customer’s<br />
rotating equipment. Since these<br />
services are performed on a continual<br />
basis, any long or unexpected downtimes<br />
are prevented that potentially could result<br />
in huge losses of production. Upkeep<br />
and repair measures of a more extensive<br />
nature are planned far in advance and<br />
in constant coordination with the customer.<br />
This ensures that any protracted<br />
downtimes can be kept to a minimum.<br />
Likewise, the developments in the<br />
field of industrial automation (“Industry<br />
4.0”) are highly relevant to rotating<br />
equipment services, specifically when it<br />
comes to prescriptive maintenance. For<br />
example, vibration sensors can be used<br />
to monitor the status of rotating machines<br />
and to alert the competent maintenance<br />
team should the monitored parameters<br />
deviate from their target range. This<br />
allows the maintenance team to intervene<br />
promptly and head off any impending<br />
damage. One service Bilfinger Industrial<br />
Services Norway offers its customers is<br />
the evaluation of the vibration signals<br />
emitted by compressors and turbines. On<br />
that basis, irregular running noises are analyzed<br />
and attributed to a specific cause,<br />
so that appropriate action can be taken<br />
immediately if required. This type of status<br />
monitoring protects the machinery<br />
involved from unexpected breakdowns.<br />
In Germany, Bilfinger distributes a vibration<br />
sensor specifically developed for the<br />
monitoring of pumps, and also provides<br />
supervision and evaluation services.<br />
The optimization and modification of<br />
rotating equipment is another field that<br />
Bilfinger looks after. Here, the main objective<br />
is to realize potential energy-savings –<br />
a significant aspect given that energy consumption<br />
accounts for about 80 percent<br />
MANY PROCESSING PLANTS INCLUDE ROTATING<br />
EQUIPMENT THAT IS IN CONTINUAL USE AND CONTRIBUTES<br />
SIGNIFICANTLY TO THE ENTERPRISE’S PRODUCTIVITY.<br />
a fixed price, along with contractually<br />
assured, progressive reductions of the<br />
maintenance costs. Bilfinger delivers the<br />
desired rotating equipment and provides<br />
support services to ensure it continues<br />
to be functional over a longer term.<br />
of the lifecycle costs of many pumps. Up to<br />
60 percent of these costs potentially could<br />
be saved by adjusting the respective pump<br />
to run at its optimal operating point.<br />
Technicians from the Bilfinger network<br />
analyze the relevant systems and draw<br />
on their decades of experience to prepare<br />
and implement proposals for technical<br />
improvement.<br />
A case in practice: Yara<br />
Porsgrunn in Norway<br />
Depending on a customer’s requirements,<br />
each service will be provided by a specific<br />
team drawn from the network. In Norway,<br />
Bilfinger is in a unique position since the<br />
center of competence for heavy rotating<br />
equipment located there constitutes the<br />
largest trove of experience in both onshore<br />
and offshore services. One example<br />
is Yara Porsgrunn, a Norwegian fertilizer<br />
producer running four nitric acid plants, a<br />
key input material for fertilizer. These facilities<br />
are among the biggest of their kind<br />
in the world, with a total output capacity<br />
of 1.8 million tons per annum. Bilfinger’s<br />
task was to recondition a compressor<br />
train during a scheduled downtime, while<br />
concomitantly performing a chemical<br />
10 maintworld 1/<strong>2020</strong>
PARTNER ARTICLE<br />
cleaning and oil-flush of the lubricating/<br />
hydraulic system. The compressor train<br />
consists of a steam turbine, a centrifugal<br />
compressor, a gear unit, an axial compressor<br />
and an expander, all of which<br />
were coupled to a drive shaft.<br />
The biggest challenge in this case was<br />
to carry all out the activities simultaneously<br />
in order to keep the stop time as<br />
brief as possible. The work was performed<br />
by roughly 3 engineers and 50<br />
technicians supplied by Bilfinger, who<br />
worked as a team with Yara Porsgrunn<br />
with added assistance from MAN Energy<br />
Solutions as the OEM. A local Bilfinger<br />
workshop manufactured a number of<br />
replacement parts and was able to perform<br />
repairs on site. This was a significant<br />
advantage, since the delivery times<br />
of the OEM were too long. Thanks to<br />
Bilfinger’s seasoned experts, the project<br />
ran smoothly. Despite the unexpected<br />
discovery of worn-out parts, the scheduled<br />
downtime of three weeks was only<br />
slightly exceeded as a consequence of the<br />
flexible approach taken by Bilfinger and<br />
its technical expertise.<br />
Rotating equipment services in<br />
the future<br />
The end-to-end package offered by<br />
Bilfinger bundles all the required services<br />
under one roof, while leaving them customizable<br />
to each customer’s needs. The<br />
advantage this service delivers over OEMs<br />
is that it allows customers to obtain independent<br />
advice and to tap into expertise<br />
transcending that of any one manufacturer.<br />
Compared to its competitors in the<br />
field of rotating equipment, Bilfinger also<br />
is able to deliver a greater amount of qualified<br />
manpower. Roughly 200 technical<br />
specialists for heavy rotating equipment<br />
currently belong to the network.<br />
In other words, Bilfinger is in an<br />
excellent position to handle upcoming<br />
developments on the market, particularly<br />
in view of the fact that cooperation<br />
models will become increasingly important<br />
in future: Global customers are<br />
looking for providers who are able to<br />
offer services of consistently high quality<br />
across the globe, while ensuring that<br />
these services are steered centrally. The<br />
demand for the rotating equipment services<br />
offered by Bilfinger as a leading international<br />
industrial services provider<br />
is certain to grow steadily in the years<br />
to come – also because customers need<br />
to fill the prevailing shortfall in skilled<br />
personnel with the required expertise<br />
in rotating equipment.<br />
1/<strong>2020</strong> maintworld 11
ASSET MANAGEMENT<br />
Self-Inflicted Reliability Problems<br />
of Rotating Machinery<br />
The root cause of poor reliability can come from many sources. You may<br />
experience reliability issues due to the age of your plant. Or perhaps poor design<br />
decisions were made. Or the original construction crew cared nothing for reliability.<br />
And there may be other reasons, outside of your control, that resulted in the<br />
reliability problems you experience today.<br />
IN ANY RELIABILITY IMPROVEMENT<br />
initiative, you will need to address<br />
these issues, but first, you need to address<br />
the self-inflicted reliability issues.<br />
“But we don't have self-inflicted reliability<br />
problems."<br />
It is a bitter pill to swallow, but yes,<br />
you do. But that is good news because<br />
it is much easier to deal with the selfinflicted<br />
root causes than the inherent<br />
reliability problems you adopted.<br />
What are self-inflicted<br />
reliability problems?<br />
In order to determine why equipment<br />
fails prematurely (or why you experience<br />
slowdowns, safety incidences, or quality<br />
problems), you could go through a<br />
detailed failure modes, effects, and criticality<br />
(FMECA) analysis process, or you<br />
could perform root cause failure analysis<br />
12 maintworld 1/<strong>2020</strong><br />
JASON TRANTER,<br />
ARP-E/L, CMRP<br />
Mobius Institute<br />
(RCFA) after each failure occurs. Or you<br />
could learn from the experience gained<br />
at thousands of plants around the world<br />
and consider some of the most common<br />
root causes of equipment failure - we will<br />
focus on rotating machinery.<br />
The number three cause of<br />
reliability problems<br />
Let’s start with the most obvious problems<br />
and then we will work backward to<br />
their root causes.<br />
Most equipment, like motors, pumps,<br />
fans, compressors, and turbines, are<br />
designed to run for many, many years<br />
without unplanned downtime. Yes, they<br />
may have some components that wear<br />
out, but many of the components, such<br />
as the bearings and gears, are designed to<br />
give years of trouble-free operation. But<br />
that assumes that all of the parts were<br />
installed correctly, the components are<br />
precision-aligned, the bearings and gears<br />
are correctly lubricated, all fasteners are<br />
tightened to the correct tension, there<br />
is no resonance, belts are tightened to<br />
the correct tension, and the rotors are<br />
precision-balanced.<br />
And it assumes that the equipment<br />
is operated as per design. Pumps, for<br />
example, should be operated at their<br />
“best efficiency point.”<br />
What happens at your plant? Do these
The<br />
The<br />
The Uptimization<br />
Uptimization Experts.<br />
Experts.<br />
What does<br />
DOWNTIME<br />
mean to you?<br />
marshallinstitute.com<br />
marshallinstitute.com
ASSET MANAGEMENT<br />
root causes exist? If you are not sure,<br />
then they almost certainly do.<br />
We will now take a quick look at just<br />
a few of those areas so that you can see<br />
why seemingly minor issues cause such<br />
serious problems.<br />
Shaft alignment<br />
When two shafts are “collinear” (no<br />
angle or offset between their centerlines),<br />
it reduces the stress on the bearings,<br />
couplings, shafts, and the rest of<br />
the machine components. Research was<br />
performed that revealed that just 5/60 th<br />
of a degree of angular misalignment can<br />
halve the life of your bearings.<br />
If you use laser alignment with<br />
appropriate tolerances, and you remove<br />
soft foot in all its forms (base issues, pipestrain,<br />
etc.), then you will have eliminated<br />
a common root cause of failure.<br />
Balancing<br />
When you balance to ISO 21940-11 grade<br />
G 1.0, the cyclical forces on the bearings,<br />
shaft, and structure are minimized, and<br />
thus you gain greater reliability. If you<br />
do not have a balancing standard, then<br />
unbalance will be a root cause of failure.<br />
And if you wait until the unbalance<br />
generates “high” vibration, “forcing”<br />
you to perform corrective maintenance,<br />
then you will have reduced the life of the<br />
equipment and supporting structure.<br />
Why is that? The life of a bearing is<br />
inversely proportional to the cube of the<br />
load. That sounds very complicated, but<br />
an easier way to say it is that if you double<br />
the load, the life will be reduced to<br />
one-eighth (23).<br />
Therefore, while the rotor is out-ofbalance,<br />
the bearings are being stressed,<br />
and their life expectancy will be reduced.<br />
Misalignment also generates these forces,<br />
and that is why it must be minimized.<br />
The unbalance is also generating<br />
forces that stress the structure, potentially<br />
resulting in fatigue failure of the<br />
structure itself or its foundations.<br />
The unbalance forces are also amplified<br />
by resonance. The structure will<br />
“naturally” vibrate back-and-forth, or<br />
side-to-side, or in other ways at certain<br />
frequencies. If the vibration generated<br />
by unbalance (or misalignment, or<br />
pump-vane vibration, or other avoidable<br />
“forcing frequencies”) is close to one of<br />
these natural resonant frequencies, the<br />
motion will be amplified. That is not<br />
good for the machine or structure.<br />
Lubrication<br />
When you correctly lubricate bearings<br />
and gears, whether you use grease or<br />
oil, and that lubricant is free of contaminants,<br />
you will achieve maximum<br />
life. But if bearings are not adequately<br />
greased, their life will be reduced. If the<br />
oil is contaminated, or the viscosity is<br />
incorrect, or the additives are depleted,<br />
then the life of gears and bearings will be<br />
greatly reduced.<br />
Research was performed to determine<br />
which particles caused the greatest<br />
damage. It was not the 40 µm particles,<br />
or the 10 µm particles - it was the tiny<br />
“3-5 µm” particles.<br />
And you may think that if you can’t<br />
see the water in oil, then the oil must<br />
be fine. Sadly, that is not correct. By the<br />
time you can see the water, the life of the<br />
bearing has been reduced by 70 percent.<br />
We could continue the discussion,<br />
but suffice to say that there is a great deal<br />
we can do to avoid problems that arise<br />
due to imperfect maintenance and operating<br />
practices.<br />
The number two cause of<br />
reliability problems<br />
It is one thing to understand all of the<br />
root causes we have just discussed – and<br />
there are many others – but it is another<br />
thing to be able to get approval to establish<br />
standards and purchase all of the<br />
tools, such as laser alignment systems,<br />
that enable the technicians and operators<br />
to do the job correctly. But owning<br />
the tools and having standard operating<br />
procedures will not solve the problem.<br />
The problem will only be solved when<br />
the maintenance technicians and operators<br />
want to use them properly, and they<br />
are given the time and encouragement to<br />
use them.<br />
So we will need to address the desire,<br />
i.e., the culture. Culture is the key to<br />
success.<br />
The number one cause of<br />
reliability problems<br />
A strong case could be made that the root<br />
cause of all failures ultimately derives<br />
from the lack of senior management support<br />
for a culture that values reliability.<br />
Without their support, it will be impossible<br />
to change the culture and thus<br />
change behaviour.<br />
Just think of the initiative to improve<br />
safety at your plant. If senior management<br />
did not support it, do you think<br />
your plant would have made the gains<br />
14 maintworld 1/<strong>2020</strong>
ASSET MANAGEMENT<br />
that it has made? Senior management<br />
enabled people to be employed in safety<br />
roles, it invested in the training and tools,<br />
it agreed to signage that provided warning<br />
and feedback on progress, it stood<br />
strong when there were opportunities to<br />
cut corners that would risk safety, and it<br />
made it quite clear how important safety<br />
is to the future of the organization. (Well,<br />
I hope that is the case at your plant.)<br />
You need the same thing to happen<br />
with reliability improvement.<br />
Everyone within the organization<br />
needs to understand that reliability is<br />
critically important to the organization<br />
and that senior management will stand<br />
strong when shortcuts that compromise<br />
reliability are proposed.<br />
Therefore, you need to gain senior<br />
management support so you can change<br />
the culture and thus successfully implement<br />
a reliability improvement initiative<br />
that eliminates the self-inflicted root<br />
causes.<br />
But wait, is there more?<br />
Since we are discussing root causes, let’s<br />
consider the root cause of the lack of senior<br />
management support. Is it their fault<br />
for not appreciating the opportunity<br />
to improve reliability, or is it your fault<br />
because you have not presented the business<br />
case for reliability improvement?<br />
It is common for people to talk about<br />
the “commonsense” benefits of reliability<br />
improvement. It is also common<br />
for reliability and condition monitoring<br />
teams to assume that senior management<br />
appreciates the benefits of what<br />
they have achieved, without ever<br />
communicating the financial benefits<br />
of their actions.<br />
Therefore, perhaps the true root<br />
cause of poor reliability is the inability<br />
(or unwillingness) of reliability and<br />
condition monitoring team leaders to<br />
establish a business case, sell the business<br />
case, and continually communicate<br />
the value of the reliability improvement<br />
initiative.<br />
Where does condition<br />
monitoring fit into this?<br />
Many people will believe that if they<br />
have a condition monitoring program,<br />
the reliability will be optimized. Sadly,<br />
that is not true. Most faults detected are<br />
avoidable, and while it is important to<br />
get an early warning, it is much more important<br />
to avoid the problem in the first<br />
place. Condition monitoring can help by<br />
detecting the root causes of failure: misalignment,<br />
unbalance, lubrication issues,<br />
looseness, and so on. If those problems are<br />
cost-effectively nipped in the bud, then we<br />
will avoid future failures.<br />
Another way that condition monitoring<br />
can play an important role is by performing<br />
acceptance testing. As part of the<br />
purchase agreement, the condition monitoring<br />
specialists can perform tests to ensure<br />
the new or overhauled equipment is<br />
“defect-free.” You may be surprised at how<br />
many problems you bring into the plant.<br />
Learning more<br />
As you can imagine, there is a great<br />
deal more that could be said about all<br />
of these topics. In an attempt to clarify<br />
the process we are discussing in this<br />
article, we developed a process called<br />
Asset Reliability Transformation, or ART.<br />
You can learn more, without charge, at<br />
www.reliabilityconnect.com.<br />
Conclusion<br />
The condition monitoring group has an<br />
important role to play. Providing an<br />
early warning minimizes the impact of<br />
premature failure, and detecting and<br />
eliminating the root causes ensures that<br />
we achieve the greatest life and value<br />
from our precious assets.<br />
The reliability improvement team<br />
has an even more important role to play.<br />
Proactively eliminating the root causes<br />
of failure ensures there will be fewer<br />
failures.<br />
But trying to improve reliability<br />
without aligning every activity to the<br />
goals of the organization, and thus<br />
gaining support from senior management,<br />
which then drives the necessary cultural<br />
changes, will never achieve the true<br />
potential of the initiative.<br />
Asset Reliabiity<br />
Transformation.<br />
The Key to a<br />
Happy Life.<br />
1/<strong>2020</strong> maintworld 15
PARTNER ARTICLE<br />
Cloud-Enabled,<br />
On-Premises,<br />
or Both?<br />
Which Data Structure<br />
Works Best for<br />
Your Maintenance<br />
Operations?<br />
ICONICS develops automation software that visualizes, archives, analyzes, mobilizes,<br />
and cloud-enables organizations’ data. With the expansion of the Industrial Internet<br />
of Things (IIoT), the data processes involved in many automation-based capabilities<br />
can be performed either on-premises (as has traditionally been done by the<br />
majority of organizations worldwide) or via the cloud (which has increasingly and<br />
rapidly become a more attractive option).<br />
THERE ARE MULTIPLE reasons why organizations’<br />
maintenance operations<br />
would select IIoT-integrated monitoring<br />
and control solutions. Among these are:<br />
• Reduced on-site hardware obsolescence<br />
• •Secure access across multiple<br />
locations<br />
• Expanded connectivity<br />
Reduced On-site Hardware<br />
Obsolescence<br />
A business starts or an organization<br />
forms. It begins to amass the equipment<br />
needed to perform its functions, as<br />
16 maintworld 1/<strong>2020</strong><br />
TOM BUCKLEY,<br />
IoT Business<br />
Development Manager,<br />
ICONICS<br />
well as the automation solutions it will<br />
deploy. Oftentimes, these same entities<br />
also begin their own maintenance operations<br />
in order to keep their hardware in<br />
good working condition. Just as a company/organization<br />
accumulates equipment<br />
to run the business itself, it also<br />
accumulates IT hardware (servers, PCs,<br />
networking equipment, etc.) that is used<br />
to perform related tasks (such as those<br />
critical to automation and maintenance<br />
operations).<br />
However, just like business-focused<br />
machinery (e.g., an aging piece of manufacturing<br />
equipment), an organization’s<br />
IT equipment can also start showing<br />
signs of aging. What may have been sufficient<br />
even just a few years ago may no<br />
longer be comparable to newer machines<br />
that can perform more advanced tasks.<br />
Organizations are now able to consider<br />
the costs associated with upgrading/
PARTNER ARTICLE<br />
retrofitting/replacing their onsite IT<br />
machinery versus the cost of moving applications<br />
to the cloud, where the cloud<br />
service operators are responsible for<br />
agreed-upon performance of their equipment.<br />
Many customers of cloud service<br />
providers are able to take advantage of<br />
an increase in processor performance to<br />
be able to take care of the heavy lifting of<br />
such tasks as advanced data analytics or,<br />
specifically for maintenance operations,<br />
predictive maintenance / fault detection<br />
and diagnostics (FDD) applications.<br />
Cloud service customers also appreciate<br />
the increase in capacity for rapid big data<br />
storage and retrieval.<br />
There are even ways for existing production<br />
equipment to take advantage of<br />
the cloud. Even though newer modern<br />
equipment may come with the ability<br />
to tie in directly to the IIoT, some older<br />
hardware can be used with emerging<br />
edge devices (IoT gateways) to make it<br />
easier and more cost-effective to become<br />
“cloud-enabled” instead of undergoing<br />
a full replacement. These gateways<br />
provide multiple other benefits, as well,<br />
including the integration of multiple<br />
devices, sensors, and other equipment to<br />
publish messages to the cloud independently<br />
from subscribers. Software modules<br />
built into such gateways decrease latency,<br />
provide edge data processing, and<br />
THERE ARE WAYS FOR<br />
EXISTING PRODUCTION<br />
EQUIPMENT TO TAKE<br />
ADVANTAGE OF THE CLOUD.<br />
empower edge analytics with onboard<br />
FDD and workflow technologies with<br />
real-time visualization of KPI data.<br />
Secure Access Across<br />
Multiple Locations<br />
Cloud security is accomplished by<br />
methods between both the applications<br />
themselves and the cloud service<br />
being used. For instance, this could be<br />
between the applications being accessed<br />
through an on-site edge device/IoT<br />
gateway and Microsoft Azure, to name<br />
one of many cloud services. Software,<br />
such as ICONICS IoTWorX, running<br />
on the IoT gateway can be provisioned<br />
and can communicate data securely via<br />
the Microsoft Azure IoT Hub, taking advantage<br />
of the inherent security features<br />
that come with an Azure subscription.<br />
Additionally, cloud-based automation<br />
1/<strong>2020</strong> maintworld 17
PARTNER ARTICLE<br />
(HMI/SCADA, historian, analytics, etc.)<br />
systems are backed up securely off site<br />
from where the data originates, keeping<br />
it safe from equipment damage or natural<br />
disasters.<br />
Such decentralized security measures<br />
are compelling for businesses/organizations<br />
that outgrow one central geographical<br />
location. Global enterprises,<br />
especially, see the benefits in trusting the<br />
security options provided through their<br />
cloud service providers and the integrated<br />
software that utilizes them. ICONICS<br />
IoTWorX, for instance, can connect multiple<br />
buildings, factories, and equipment<br />
through secure TLS encryption and popular<br />
cloud platforms, such as Microsoft<br />
Azure and Amazon Web Services. Data<br />
can be accessed from anywhere through<br />
a pub/sub architecture for real-time<br />
visualization of KPI data at the edge.<br />
IoTWorX delivers an efficient, secure<br />
connection to the cloud through bidirectional<br />
AMQP for Microsoft Azure,<br />
as well as MQTT, REST, and WebSockets<br />
for third-party cloud providers.<br />
Expanded Connectivity<br />
Another benefit of utilizing cloud-based<br />
automation solutions is that there is often<br />
an increase in the number of available<br />
communication protocols that can be<br />
used. This is in addition to the advanced<br />
security measures (bidirectional AMQP<br />
transport protocol [for Microsoft Azure]<br />
and MQTT, REST, and WebSockets [for<br />
third-party providers]) that cloud-based<br />
solutions provide.<br />
For maintenance operations, it’s definitely<br />
a benefit to be able to “talk to” as<br />
many of the machines within the organization<br />
as possible. As an example, ICON-<br />
ICS IoTWorX software is compatible<br />
with multiple standard communication<br />
protocols. These include protocols specific<br />
to plant floor applications; such as<br />
OPC Classic, OPC Unified Architecture<br />
(OPC UA), and Modbus; as well as those<br />
specific to building automation (BACnet)<br />
and IT hardware (SNMP). This provides<br />
users with the ability to communicate<br />
with a wider array of connected<br />
equipment, ultimately enabling users to<br />
better detect potential issues and utilize<br />
an organization’s data, wherever it might<br />
be created, transmitted, or stored.<br />
Cloud Contingencies<br />
For those concerned about the viability<br />
of cloud-based solutions during interruptions<br />
to internet service, there are<br />
measures that can be put into place to<br />
help ensure data doesn’t go missing or<br />
get corrupted. ICONICS has solutions<br />
that provide rapid data archiving and retrieval,<br />
including a "store-and-forward"<br />
feature that is useful when a network<br />
connection is unavailable; one specifically<br />
edge-based (IoT Hyper Collector)<br />
and the other traditionally on-premises<br />
or cloud-enabled (Hyper Historian).<br />
IoT Hyper Collector is part of IoT-<br />
WorX, the previously mentioned micro-<br />
SCADA software suite installed on a<br />
third-party IoT edge device. The collector<br />
has the ability to replay buffered<br />
data back locally, as well as to store and<br />
forward to the cloud when connectivity<br />
is present. For a more traditional<br />
on-premises approach, ICONICS Hyper<br />
Historian Collector also utilizes a similar<br />
store-and-forward feature. If a collector<br />
has lost connectivity to the logger, it will<br />
continue to buffer the data until connectivity<br />
is reestablished.<br />
Best of Both<br />
While the IoT Hyper Collector and<br />
Hyper Historian Collector are examples<br />
of how to retain data integrity both via<br />
the cloud and on-premises, respectively,<br />
there is nothing to prevent an organization<br />
from taking a hybrid approach.<br />
This bridges the gap between OT and<br />
IT and alleviates any "silo effect" of the<br />
organization’s data collection, storage,<br />
and retrieval. The same can be said for an<br />
organization’s entire automation solution<br />
and related data, as a whole. Some<br />
may benefit from a strictly cloud-based<br />
solution. Others may still have reason<br />
to remain with an entirely on-premises<br />
one. However, neither has any restriction<br />
towards using elements of the other<br />
in such a hybrid scenario.<br />
Each organization will make its own<br />
determination regarding what works<br />
best for their business processes and operations<br />
(cloud, on-premises, or hybrid),<br />
as well as to the automation software<br />
vendors that can best support it.<br />
18 maintworld 1/<strong>2020</strong>
YOUR PARTNER IN<br />
ULTRASOUND<br />
INSTRUMENTS<br />
Leak Detection<br />
Bearing Condition Monitoring<br />
Bearing Lubrication<br />
Steam Traps & Valves<br />
Electrical Inspection<br />
TRAINING<br />
CAT & CAT II Ultrasound Training<br />
Onsite Implementation Training<br />
Application Specific Training<br />
CONTINUOUS SUPPORT<br />
UE SYSTEMS<br />
Windmolen 20<br />
7609 NN Almelo<br />
The Netherlands<br />
T: +31 546 725 125<br />
E: info@uesystems.eu<br />
W: www.uesystems.eu<br />
Free support & license-free software<br />
Online Courses<br />
Free access to our Learning Center<br />
(webinars, articles, tutorials)<br />
CONTACT US FOR AN<br />
ONSITE DEMONSTRATION
DIGITALISATION<br />
DATA IS ABUNDANT<br />
Insights and Actionable<br />
Information are Hard to Find<br />
Plant Managers Can<br />
Improve Decision Making<br />
with Asset Performance<br />
Management 4.0 and<br />
Digital Twins<br />
SAVVY PLANT EXECUTIVES have noticeably<br />
changed how they approach topline<br />
revenue growth opportunities and<br />
bottom-line cost reductions. In assetintensive<br />
industries, digital advancement<br />
strategies like Asset Performance<br />
Management 4.0 and digital twins are<br />
becoming critical strategies for operational<br />
excellence.<br />
Within a plant, digital twins empower<br />
engineering, operations, and maintenance<br />
to collaborate and capitalize<br />
on the opportunity of Industry 4.0. If<br />
common pitfalls can be overcome, such<br />
as data overload, the lack of systems<br />
integration and interoperability, inconsistent<br />
business processes, and siloed<br />
data, then this open and connected data<br />
environment can become a sustainable<br />
competitive advantage.<br />
Maximizing the<br />
Value of Assets<br />
The reason plant staff collect, store, and<br />
analyze plant data is to improve decision-making<br />
and operations. Ultimately,<br />
they want to maximize the return on<br />
assets, minimize production, labor, and<br />
material costs, reduce the total cost of<br />
asset ownership and investment, and deliver<br />
predictable performance, on time.<br />
We can improve the performance of<br />
the physical assets used for production<br />
through the way people operate equipment,<br />
through improved maintenance<br />
20 maintworld 1/<strong>2020</strong><br />
SANDRA DIMATTEO<br />
Sandra DiMatteo is<br />
the Global Director<br />
of Marketing<br />
for Digital Twin<br />
Solutions, Asset and Network<br />
Performance at Bentley Systems.<br />
She has more than 20 years<br />
of experience in digitalization<br />
solutions in asset performance<br />
management and reliability<br />
software, asset lifecycle information<br />
management and EAM enterprise<br />
asset management operating in<br />
a connected data environment in<br />
energy and process industries,<br />
utilities and public infrastructure. She<br />
is an advocate and speaker on digital<br />
twins, IIoT analytics, AI and machine<br />
learning, BIM and asset management<br />
technology solutions. Sandra is on<br />
the Reliability Leadership Institute<br />
Board of Advisors and founded<br />
the Ontario Chapter of the Society<br />
of Maintenance and Reliability<br />
Professionals.<br />
practices, or through design and engineering.<br />
This is not a new concept, as an<br />
average plant might have 30 to 50 design<br />
optimizations, modifications, or additions<br />
per year.<br />
However, the siloed approach separating<br />
engineering, operations, and<br />
maintenance is no longer good enough.<br />
Success now requires a new way forward<br />
with a collaborative focus on asset health<br />
to attain performance targets. A reliability<br />
strategy actively resolves asset health<br />
shortcomings, leading to eliminating<br />
performance shortfalls.<br />
Asset Performance<br />
Management<br />
Asset Performance Management (APM)<br />
focuses on meeting performance requirements<br />
through reliability and takes<br />
a cross-discipline approach to fulfil plant<br />
goals and customer needs. Asset performance<br />
management is the plant manager’s<br />
opportunity to extend asset life<br />
safely and reliably, which avoids capital<br />
expenditures.<br />
Some of the key characteristics of an<br />
Asset Performance Management system<br />
include:<br />
• Asset strategy and risk analysis<br />
• Condition-based or reliability<br />
centered maintenance processes<br />
and practices<br />
• A mobile inspection platform with<br />
augmented reality and virtual reality<br />
capabilities<br />
• Predictive analyses including<br />
statistical modeling, neural networks,<br />
artificial intelligence, and<br />
machine learning<br />
• Spare parts optimization<br />
• Asset lifecycle information management<br />
•<br />
The Tipping Point for IIoT<br />
APM 4.0 is effectively deployed within<br />
a digital twin because the Industrial Internet<br />
of Things (IIoT) movement has<br />
arrived at the tipping point. The cost of<br />
sensors, data connection, and data storage<br />
is now a fraction of what it used to be.<br />
As a result, the amount of raw data being<br />
generated in plants from IIoT sources is<br />
growing exponentially, and many organizations<br />
cannot keep up. Every sensor you<br />
add produces thousands of additional<br />
data points. As a result, making sense<br />
of the data to gain meaningful insights
DIGITALISATION<br />
and get to the right decisions can be<br />
time-consuming and difficult if you are<br />
not sure of its relevance or accuracy.<br />
For most plant managers, the vision of<br />
a completely autonomous plant is still a<br />
pipe dream.<br />
Industry 4.0 can connect physical assets<br />
in the plant to their digital counterparts<br />
to improve the automation of plant<br />
operations and maintenance. Using edge<br />
computing to implement artificial intelligence<br />
and automated rules is a fast and<br />
easy way to alert personnel of problems<br />
that must be addressed. However, edge<br />
computing might not monitor all aspects<br />
of every asset over the long-term. To fully<br />
oversee a facility, you need a systematic,<br />
sustainable approach for tracking<br />
asset health over time with visible, accessible,<br />
and trusted engineering data.<br />
Plant managers have embraced<br />
IIoT out of a desire to eliminate human<br />
senses from inspections, such as seeing a<br />
leak or hearing a malfunctioning motor.<br />
Even with the explosion of sensors that<br />
can detect changes in operating conditions,<br />
the ugly truth is that plants remain<br />
highly people dependent. To improve<br />
automation, plants need an efficient,<br />
effective, and comprehensive program<br />
that goes beyond what people can notice<br />
themselves. Such a program should include<br />
a playbook of fully defined organizational<br />
and business processes, proactive<br />
and predictive asset management<br />
practices, and the right technology that<br />
enables the implementation and execution<br />
of real-time asset performance.<br />
Gain Insights to Make<br />
Effective Decisions<br />
There are many ways APM 4.0 helps you<br />
gain insights from data and optimize<br />
asset performance. Automated rules,<br />
calculations, artificial intelligence, and<br />
machine learning are all valuable methods<br />
to enable faster and more effective<br />
decisions. But, engineering information<br />
for each must be complete, accurate, and<br />
available to ensure you are making the<br />
right decision at the right time. Otherwise,<br />
it becomes harder to mitigate costs<br />
and downtime when the asset fails.<br />
In short, effective decision-making<br />
depends on always knowing the current<br />
state of the asset and becoming informed<br />
immediately when that state changes.<br />
This knowledge should include essential<br />
engineering information, as well as how<br />
to bring the asset back to the as-built,<br />
as-commissioned, or as-designed states.<br />
Asset lifecycle information management<br />
is the backbone of APM 4.0. Components,<br />
structures, systems, and operating<br />
states all change over time due to<br />
wear and tear, operator decisions, and<br />
overall plant conditions. Changes in any<br />
single asset can negatively impact wider<br />
systems and processes. Trustworthy<br />
engineering data enables plant engineers<br />
to determine why a change occurred and<br />
who caused the change.<br />
However, raw sensor data alone might<br />
not be useful as the complexity and<br />
interconnections of piping and process<br />
equipment, systems, instrumentation,<br />
and control devices has increased. Operations<br />
technology relies on analytics<br />
visibility as well as subject matter experts<br />
that can act based on the massive<br />
amount of data being generated. Digital<br />
twins, together with asset performance<br />
management 4.0, can harness that raw<br />
data and create a trusted system of<br />
systems. They can connect data with<br />
processes and identify, consolidate, and<br />
analyze all relevant sources of data to<br />
make asset health more visible and drive<br />
informed decisions and measurable<br />
business results.<br />
Digital Twins Enable<br />
Collaboration between<br />
Engineering, Operations and<br />
Maintenance<br />
A digital twin is a digital representation<br />
of a physical asset, process or system, as<br />
well as the engineering information to<br />
understand and model its performance.<br />
Typically, a digital twin can be continuously<br />
synchronized from multiple federated<br />
sources, including sensors and<br />
continuous surveying, to represent its<br />
near real-time status, working condition<br />
or position. Digital twins enable users to<br />
visualize the asset, check status, perform<br />
analysis and generate insights in order to<br />
predict and optimize asset performance.<br />
1/<strong>2020</strong> maintworld 21
DIGITALISATION<br />
As a result, digital twins eliminate silos<br />
of data to deliver situational awareness<br />
and intelligence.<br />
Additionally, a continually updated<br />
digital twin provides proof of accurate<br />
information needed for regulatory compliance.<br />
Engineering, operations, and<br />
maintenance greatly benefit from the<br />
combination of APM and digital twins.<br />
APM provides the strategy and analytics,<br />
while digital twins unify the data, provide<br />
situational awareness and insights,<br />
and deliver actionable information in the<br />
hands of those who need it when they<br />
need it.<br />
Don’t Forget the ET in Your<br />
IT - OT Strategy<br />
Engineering data is always evolving.<br />
As assets are designed, commissioned,<br />
and operated, new information becomes<br />
generated. Individual assets<br />
also evolve through maintenance and<br />
modifications during the operational<br />
phase. The convergence of information<br />
technology, operations technology, and<br />
engineering technology (or IT-OT-ET)<br />
feeds the digital engineering model and<br />
creates a comprehensive digital twin of<br />
the working assets in the plant, facility,<br />
or network. In addition to communicating<br />
the current state of the asset, the<br />
digital twin can perform operational<br />
and engineering simulations to model<br />
the performance of an asset over time<br />
and evaluate options to improve performance.<br />
Essentially, the digital twin connects<br />
the data from IT-OT-ET in a single portal<br />
view, allowing the team to validate,<br />
visualize, and analyze all plant data in<br />
any format and any data storage location.<br />
A digital twin environment that is<br />
open, interoperable, connected, and contextualized<br />
enables true collaboration<br />
between engineering, operations and<br />
maintenance.<br />
A Day in the Life… Who Needs<br />
a Digital Twin and Why?<br />
From the plant floor to the boardroom,<br />
digital twins quickly give plant staff the<br />
information they need to make important<br />
decisions.<br />
GLOBAL AND REGIONAL EXECUTIVES –<br />
Executives overseeing a division, region,<br />
or the entire global operation need to<br />
track and compare plants and fleets. This<br />
includes identifying high-performing<br />
and under-performing plants, then determining<br />
what makes them succeed or<br />
fail. Executives must also provide stakeholders<br />
with proof that the plants are in<br />
control and that assets are safe and reliable.<br />
Using digital twins enables them to<br />
generate informative visualizations of<br />
large-scale assets and grant a regional or<br />
companywide perspective.<br />
PLANT MANAGERS – Plant managers<br />
must ensure plant production is predictable,<br />
safe, and efficient. They need<br />
continual access to key performance<br />
indicators to identify and evaluate units<br />
that are underperforming. When issues<br />
are reported, managers rely on accurate<br />
information for audits and course-corrections.<br />
By providing a unified digital<br />
twin that provides a complete, consistent<br />
view of plant data, managers allow engi-<br />
22 maintworld 1/<strong>2020</strong>
DIGITALISATION<br />
neering, maintenance, and operations<br />
to collaborate and solve problems more<br />
effectively and efficiently.<br />
ENGINEERS – Engineers must quickly<br />
identify potential operational problems<br />
and consider solutions. They need to determine<br />
what has changed in engineering<br />
models, piping and instrumentation<br />
documents, drawings, or the maintenance<br />
and reliability program. Most<br />
importantly, they need to know that<br />
the data is trustworthy so that they can<br />
investigate, troubleshoot, and make fast<br />
and informed decisions. Digital twins<br />
can provide this peace of mind, especially<br />
when engineers can update digital<br />
twins as needed and view relevant IIoT<br />
and engineering information in an open,<br />
connected data environment.<br />
OPERATIONS – Operators need access<br />
to performance and maintenance data<br />
without having to waste time determining<br />
which data is relevant and which is<br />
not. They must review as-operated data<br />
and historical data alike to understand<br />
what field changes and engineering decisions<br />
were made and why. Digital twins<br />
help operators see the big picture and<br />
optimize production. By supplementing<br />
control systems with performance dashboards<br />
containing real-time performance<br />
data, operators receive a holistic overview<br />
of the complete facility including areas or<br />
units not available in the local DCS.<br />
MAINTENANCE AND RELIABILITY – Maintenance<br />
managers must review upcoming<br />
work before attending production<br />
scheduling meetings. Maintenance, reliability,<br />
and integrity engineers need to<br />
monitor and manage equipment health<br />
as well as piping and vessel integrity to<br />
easily spot trends and bad actors. They<br />
need to know what engineering changes<br />
were made in the past and use rules,<br />
calculations, AI, and machine learning<br />
to analyze data. Digital twins combine<br />
analysis methods and provide valuable<br />
insights to ensure safety, reliability, and<br />
asset integrity.<br />
Digital Transformation<br />
For any successful team, operational<br />
excellence depends on a clear strategy,<br />
skilled players trained on the best techniques,<br />
efficient and effective equipment,<br />
and leadership that can coach the<br />
team to cohesively execute the game<br />
plan. When done correctly, this leads to<br />
consistent wins.<br />
APM 4.0 and digital twins are digitally<br />
transforming plants to help them stay<br />
ahead of the competition. Everyone from<br />
the plant floor to the boardroom needs<br />
insights to make more informed decisions.<br />
Digital twins provide a federated<br />
portal view of all necessary systems and<br />
data, which gives workers at all levels the<br />
insight needed for overall success.<br />
SEVERAL CASE STUDIES<br />
OMAN GAS COMPANY implemented<br />
a digitalized, automated framework<br />
for its reliability and integrity program.<br />
Establishing a connected data<br />
environment and digital workflows<br />
reduced failures and improved reliability<br />
performance by 9%. The technology<br />
transformed the team and<br />
how they manage assets, resulting in<br />
significant economic gains.<br />
ARCELORMITTAL USA successfully<br />
implemented an equipment reliability<br />
initiative, saving USD 2.1 million during<br />
a year-long pilot program at one<br />
of its hot strip mills. The success of<br />
the pilot led to a wider implementation<br />
across 10 key focus areas and a<br />
savings of more than USD 14 million<br />
nationwide over two years. Arcelor-<br />
Mittal adopted asset performance<br />
management best practices, processes,<br />
methodologies, and digitalization<br />
technology to efficiently share<br />
trusted information and change the<br />
company’s maintenance culture from<br />
reactive to proactive.<br />
EPCOR UTILITIES INC. implemented<br />
an ISO 55000-aligned; risk-based<br />
asset management process supported<br />
by Bentley System’s APM technology.<br />
Using asset health indexing, they<br />
gained an understanding of the consequence<br />
of asset failures, including<br />
replacement costs, damage to adjacent<br />
assets, impacts to safety, and<br />
environmental cleanup costs. Coupling<br />
this information with outage<br />
times and electrical load data, they<br />
could better predict the annual risk<br />
cost. The result was lowering their<br />
SAIDI Interruptions Duration Index<br />
score to 0.833, well below the regulated<br />
threshold of 1.15 hours/customer.<br />
BP created a central information<br />
store (CIS) to manage information<br />
needed for operations, including all<br />
documents, tags, metadata, and 3D<br />
model visualization. Using a Microsoft<br />
Azure-based cloud deployment<br />
of Bentley Systems’ AssetWise asset<br />
performance software, the project<br />
team seamlessly integrated engineering<br />
information into operations.<br />
Doing so supports safe, reliable, and<br />
efficient operations throughout the<br />
life of their assets. With safety at the<br />
heart of all operations, BP ensures it<br />
continuously maintains the integrity<br />
of operational information.<br />
1/<strong>2020</strong> maintworld 23
ASSET MANAGEMENT<br />
Experience Feedback –<br />
Rotating Machinery<br />
PATRICE DANNEPOND, SDT Ultrasound Solutions<br />
Display of the database in the UAS software<br />
Experience feedback on<br />
the implementation of an<br />
ultrasound-based preventive<br />
maintenance program.<br />
Introduction<br />
In this article, we would like to share an<br />
experience feedback from a paper mill that<br />
has implemented a preventive maintenance<br />
program to monitor rotating machines using<br />
ultrasound technology. SDT International<br />
helped implement this monitoring program<br />
by training and coaching the teams in charge<br />
of mechanical maintenance reliability. The<br />
purpose of this program is to monitor a fleet<br />
of about 70 rotating machines during the year<br />
after its implementation and to extend it to<br />
100 machines over the second year.<br />
Issue<br />
This paper mill has relied on preventive<br />
maintenance for many years, using known<br />
and proven technologies for the monitoring<br />
of rotating machines. In 2018, they decided to<br />
extend this monitoring to equipment with rotation<br />
speeds up to 30 RPM, as well as to their<br />
speed-reducing gears.<br />
They purchased an SDT270 type ultrasound<br />
detector, in DU version, along with its<br />
UltrAnalysis (UAS) software, and SDT International<br />
and the Reliability department of<br />
the paper mill developed a training program<br />
suited to the rotating machinery monitoring<br />
program. The first step consisted in creating<br />
the database including these 70 machines,<br />
and then in recording an initial measurement<br />
of the mechanical status of each bearing<br />
and each gear. After a simple onsite analysis<br />
(ultrasonic listening) and a more detailed<br />
analysis (overall or static measurements and<br />
spectral or dynamic measurements) using<br />
UAS, pre-alarm, alarm and danger thresholds<br />
were assigned to each measurement point.<br />
This background work, which is required,<br />
allows technicians of the reliability department<br />
in charge of the measurements routes<br />
to get a quick overview of the asset hierarchy<br />
and immediately see the machines that have<br />
an alarm status.<br />
Display of the asset hierarchy including all rotating machines under monitoring<br />
and alarm statuses for each piece of equipment. In the present case, 19 rotating<br />
machines are monitored, 2 of which have exceeded the danger threshold for the<br />
bearings of the rear motor.<br />
ISSUE<br />
This preventive maintenance program has 3 objectives:<br />
• Highlight the efficiency of ultrasound measurements on rotating machines.<br />
• Issue a relevant diagnosis.<br />
• Offer preventive maintenance with reliable indicators.<br />
Experience feedback after onsite measurement sessions from<br />
October 2018 to November 2019<br />
Monitoring of a parallel reduction gear<br />
• Machine: Decanter – High-speed input bearing of the reduction gear<br />
Measurement carried out on 16/10/2018 Measurement carried out on 20/09/2019<br />
24 maintworld 1/<strong>2020</strong>
ASSET MANAGEMENT<br />
On the time spectra (same scale), we can observe the occurrence<br />
of shocks compared to the first measurement carried out<br />
in 2018. The trend curves show the evolution of the RMS static<br />
value: from -4.3 dBµV in 2018 to +16.8 dBµV in 2019.<br />
Based on SDT criteria, this increase corresponds to the early<br />
failure of a mechanical part of the reduction gear (bearings and/<br />
or gears).<br />
Zooming in on the FFT frequency spectrum allows highlighting<br />
dissymmetry of the modulation around the meshing frequency,<br />
which is characteristic of a damaged gear mesh.<br />
Listening to the bearing and analysing the frequency spectra<br />
(see chart above) has allowed confirming this diagnosis by observing<br />
the emergence of significant peaks associated with this<br />
gear damage.<br />
We can observe repeated shocks associated with the frequency<br />
of the high-speed input drive pinion of the reduction<br />
gear (24.93 Hz and its harmonics) with demodulation at each<br />
peak. Broken tooth and teeth clearance. Replacement of the reduction<br />
gear during a scheduled production shutdown, which<br />
avoided an untimely breakdown which could have generated<br />
significant expenses due to production losses.<br />
The endoscopic video inspection of the worm screw of the reduction<br />
gear confirmed the ultrasound diagnosis.<br />
The customer took the reduction gear down during a production<br />
shutdown.<br />
Wear detected on the tooth of the worm screw of a reduction gear,<br />
wheel and screw:<br />
• Machine: Lime mud filter agitator – High-speed input<br />
bearing of the reduction gear<br />
We can observe that between each revolution of the worm<br />
screw, there is a phenomenon occurring, which can be heard<br />
through the ultrasound detector as a sliding of the gear (worm<br />
screw/bronze wheel).<br />
Measurement carried out on<br />
10/07/2018<br />
Measurement carried out on<br />
29/10/2019. After replacement of<br />
the reduction gear<br />
Monitoring of the degradation of the bearing of a low-speed reduction<br />
gear (opposite transmission):<br />
• Machine: Vertex separator reduction gear – 4-train<br />
parallel reduction gear<br />
Measurement carried out on 11/12/2016<br />
From the beginning of the monitoring of this reduction gear<br />
(August 2018) using ultrasound technology, the occurrence of<br />
shocks can be observed on the time spectrum.<br />
1/<strong>2020</strong> maintworld 25
ASSET MANAGEMENT<br />
By zooming in on the time spectra, one can observe repeated<br />
shocks at 9.756 Hz (see the table of characteristic frequencies<br />
below) related to the frequency of the inner ring of the bearing<br />
of the low-speed reduction gear (opposite side of transmission).<br />
Diagnosis confirmed on the frequency spectrum (see below).<br />
Monitoring of the degradation of the bearing of a low-speed reduction<br />
gear (opposite transmission):<br />
• Machine: Lime mud filter – Bearing opposite transmission<br />
23140 CCK – 14.28 RPM<br />
Measurement carried out on<br />
16/10/2018<br />
Measurement carried out on<br />
07/01/2019<br />
The amplitude of the scaling-type defect is modulated by the<br />
rotation speed.<br />
On the spectrum, this is evidenced by a peak at the frequency<br />
of the defect of the inner ring of the output bearing of the reduction<br />
gear and by sidebands at the rotation frequency of the<br />
shaft, i.e., 0.5 Hz – 30 RPM (low-speed reduction gear).<br />
On time spectra (same scales), one can observe the occurrence of<br />
shocks right from the beginning of the monitoring of this bearing.<br />
After replacement of the bearing, all shocks disappeared. The<br />
rolling elements were no longer held in their housings.<br />
Trend curves show the decrease of the RMS value after servicing<br />
(from +10.7 dBµV in 10/2018 to -4.4 dBµV in 01/2019).<br />
Trend curves show the decrease of the RMS value after servicing,<br />
from +12 dBµV in 08/2019 to +1.2 dBµV in 11/2019.<br />
Measurement carried out on<br />
26/08/2019<br />
Measurement carried out on<br />
25/11/2019. After replacement<br />
of the reduction gear<br />
Replacement of the reduction gear during a scheduled production<br />
shutdown, which avoided an untimely breakdown which<br />
could have generated significant expenses due to production<br />
losses.<br />
Conclusion<br />
The challenge was taken up by the Reliability team of this paper<br />
mill. The implementation of a preventive maintenance program<br />
for 70 rotating machines has had a beneficial and decisive outcome.<br />
It will be extended to 100 other machines over the course<br />
of <strong>2020</strong>. SDT International offered a simple solution and suitable<br />
measurement tools, along with a LEVEL1 ASNT certified training<br />
program. Users have acquired a comprehensive mastery of<br />
this technology, which was new to them.<br />
1. Using ultrasound technology to monitor low-speed rotating<br />
machines, this paper mill was able to avoid a number of unscheduled<br />
shutdowns (see examples below) and highlight the complementarity<br />
of ultrasound and vibration technologies.<br />
2. Based on this experience, the Reliability department has<br />
decided to initiate ultrasound-aided greasing campaigns. Using<br />
suitable equipment (software and hardware), this acoustic lubrication<br />
program will ensure perfect greasing by indicating:<br />
• the right grease,<br />
• the right greasing location,<br />
• the right greasing interval,<br />
• the right quantity of grease to add,<br />
• the right indicators for the lubrication condition.<br />
Thus, full traceability of the lubrication program will be ensured.<br />
3. The versatility of ultrasound detector SDT270DU also allowed<br />
for the implementation of:<br />
• An energy-saving policy (detection of compressed air leaks,<br />
control of steam traps).<br />
• Control of tubeblowers (detection of leaks on steam valves).<br />
• Preventive maintenance of high voltage electric systems (corona,<br />
tracking, arcing).<br />
26 maintworld 1/<strong>2020</strong>
Be a LUBExpert<br />
®<br />
A COMPLETE ULTRASOUND SOLUTION TO MANAGE YOUR ACOUSTIC LUBRICATION PROGRAM<br />
Poor greasing practices are<br />
a leading cause of bearing failure.<br />
Many lube departments re-grease on a wasteful<br />
calendar-based schedule. This leads to over and<br />
under greased bearings that fail to deliver their<br />
engineered value.<br />
LUBExpert tells us when to grease...<br />
and when to stop.<br />
Grease reduces friction in bearings. Less friction<br />
means longer life. LUBExpert alerts you when<br />
friction levels increase, guides you during<br />
re-lubrication, and prevents over and under<br />
lubrication.<br />
Grease Bearings Right<br />
Right Lubricant<br />
Right Location<br />
Right Interval<br />
Right Quantity<br />
Right Indicators<br />
Ultrasound Soluons<br />
sdtultrasound.com/lubexpert
PARTNER ARTICLE<br />
RESONANCE -<br />
The Hidden Threat<br />
Text: Adash Ltd.<br />
You have done proper balancing (twice for sure), you have done alignment, you<br />
have checked the mountings and your machine is still vibrating like an old washing<br />
machine for no obvious reason… The reason is probably resonance.<br />
MECHANICAL RESONANCE is the tendency of a mechanical<br />
system to respond at a greater amplitude when the frequency<br />
of its oscillations matches the system's natural vibration frequency<br />
than it does at other frequencies.<br />
For every mechanical object, there are natural frequencies<br />
at which that object vibrates more easily and harder than at<br />
other frequencies.<br />
You can see this in everyday life. For example, there might<br />
be something in your car that does not vibrate very much, but<br />
if you make a small change in engine RPM, you can hear much<br />
stronger vibrations.<br />
This is an example of where the engine RPM is near the<br />
natural frequency of the thing that is vibrating.<br />
It is very dangerous to operate machines near their natural<br />
frequencies because even a small unbalance can generate extremely<br />
high vibrations. This can destroy the machine very easily.<br />
We usually encounter the resonance problem on the machine<br />
foundation. When the machine speed is near the natural<br />
frequency, then we measure high vibrations without a visible<br />
reason. The maintenance team usually carries out all the<br />
standard procedures like balancing, alignment and bearing<br />
mounting checks but vibrations still remain high. The resonance<br />
problem is the reason.<br />
Now I would like to describe to you how to use the bump<br />
test measurement. It is the perfect measurement for situations<br />
where you suspect a resonance problem. We used a steel beam<br />
for this simple demonstration. Imagine that it is a machine<br />
foundation.<br />
We used a standard 100 mV/g accelerometer and a hammer.<br />
The Adash VA4Pro and VA5Pro vibration analyzers contain<br />
a super easy bump test mode. There is no need for any settings.<br />
Just place the sensor on the object to be measured and hit it<br />
with a hammer.<br />
Let's look at the final graph. [1]<br />
High peaks on the spectrum represent natural frequencies,<br />
in our case 100 Hz and 280 Hz. There will be a resonance problem<br />
if the machine speed is near these two frequencies.<br />
We can move the natural frequency to solve the problem by<br />
reinforcing the construction in order to change the natural frequency.<br />
For example, we could add another pillar in the middle<br />
and the natural frequencies will decrease. [2]<br />
A fixed pillar would be welded to the machine foundation in<br />
28 maintworld 1/<strong>2020</strong>
PARTNER ARTICLE<br />
Graph 1: initial spectrum<br />
Graph 2: changed spectrum<br />
1/<strong>2020</strong> maintworld 29
PARTNER ARTICLE<br />
Figure 1. Every mechanical object has its mode shapes.<br />
Figure 2. The length of each arrow is<br />
proportional to the g value at that point.<br />
Figure 3. The first natural frequency is 10 Hz.<br />
Figure 4. Second natural frequency.<br />
the real world and natural frequencies would not only be<br />
reduced but also re-tuned.<br />
Every mechanical object has its mode shapes. I will<br />
explain it on this simple free beam (Figure 1).<br />
The first mode shape is this: we can see two nodes and<br />
three anti nodes. The next mode shapes have more nodes<br />
and more anti nodes. But for now, we will only consider<br />
the first one, whose natural frequency is the lowest<br />
natural frequency from the measured graph. We marked<br />
several points on our steel beam and measured the amplitude<br />
on each mark.<br />
The length of each arrow is proportional to the g value<br />
at that point. Now you can see the first mode shape,<br />
which we have got from real measurements (Figure 2).<br />
For simple objects you can calculate the mode shapes,<br />
but for a complicated construction it is not possible, you<br />
have to measure it.<br />
We analyzed the frame by hitting it with hammer.<br />
When the machine is operating, you can also measure<br />
the vibration levels at every point. In this case the frame<br />
is not excited by the hammer but by operational speed.<br />
The vibration levels will not be the same on all points of<br />
30 maintworld 3/2019
the machine. Measure all of them and draw the arrows<br />
again, you will get the first operational deflection shape<br />
of the machine. For finding the next operational deflection<br />
shapes, you have to know the spectrum of vibration<br />
on each point. Knowing the machine shapes is important<br />
for machine understanding.<br />
The ADASH analyzers contain the ADS mode. It enables<br />
you to measure these shapes very simply and then<br />
illustratively animate the results.<br />
In the next example I show you why it is important to<br />
know the mode shapes which are the cause of the vibration<br />
problem. We used the shaker and the rubber cord.<br />
The first natural frequency is 10 Hz and you can see the<br />
first mode shape (Figure 3).<br />
If I need to decrease the vibration level, then I can add<br />
the pillar to many places and it will work.<br />
But sometimes the second mode shape could be the<br />
problem rather than the first. Now you can see the second<br />
natural frequency (Figure 4).<br />
The location of the pillar is now much more important.<br />
If I add it in the middle, then the vibration remains<br />
unchanged.<br />
Back to our first example with the steel beam. Initial<br />
overall vibration was 6.34 g. We added a pillar in the middle<br />
and the main natural frequency on 100 Hz decreased<br />
approximately 4 times. But the second natural frequency<br />
on 280 Hz decreased only about 2 times and now is influencing<br />
our frame almost as much as the first frequency.<br />
The overall vibration is now 3.19 g.<br />
Then we moved the pillar from the middle to 1/3 of<br />
the way between the end pillars. The new natural frequencies<br />
look like this. [3]<br />
You can see that the first frequency remained the<br />
same, but the second was more reduced.<br />
Overall vibration decreased to 2.56 g.<br />
Graph 3: final spectrum<br />
BUMP TEST<br />
MEASUREMENT<br />
IS THE PERFECT<br />
MEASUREMENT FOR<br />
SITUATIONS WHERE<br />
YOU SUSPECT A<br />
RESONANCE PROBLEM.
ASSET MANAGEMENT<br />
How Predictive<br />
Maintenance<br />
Enhances<br />
Plant Safety<br />
without sacrificing productivity<br />
ADRIAN MESSER,<br />
CMRP<br />
adrianm@uesystems.com<br />
Ultrasonic sensors<br />
improve safety by<br />
allowing assets to be<br />
inspected at a distance<br />
An effective predictive maintenance strategy<br />
leverages technology and analytic data to allow for<br />
optimized scheduling of corrective maintenance tasks.<br />
The goals of these programs are to reduce asset<br />
downtime, prevent unexpected failures, promote<br />
productivity and increase the safety of personnel.<br />
IN THE PAST, safety investments often<br />
meant a reduction in overall productivity<br />
or an increase in cost structure. However,<br />
modern technology has created<br />
new opportunities for facility managers<br />
to maintain or even increase productivity<br />
levels as the plant implements new<br />
safety protocols.<br />
How maintenance impacts<br />
facility safety<br />
Maintenance work can be dangerous,<br />
especially when assets fail in unexpected<br />
or catastrophic ways. In fact, a<br />
report from BLR (Business & Legal Resources)<br />
revealed that between 25 and<br />
30 percent of workplace deaths in the<br />
manufacturing industry are related to<br />
maintenance activities. Shocks, burns<br />
and injury from moving parts are all<br />
common causes of fatal accidents.<br />
A predictive maintenance strategy<br />
seeks to reduce or completely<br />
eliminate unexpected failures. Using<br />
technology such as ultrasound sensors,<br />
stakeholders can detect minute<br />
changes in sonic output to determine<br />
when an asset is likely to malfunction.<br />
Given enough data, stakeholders can<br />
determine expected life spans for all<br />
32 maintworld 1/<strong>2020</strong>
ASSET MANAGEMENT<br />
facility assets and build a maintenance<br />
schedule around this information. This<br />
way, maintenance personnel can power<br />
down assets during a period that is not<br />
likely to interfere with the facility’s<br />
productivity goals. They can complete<br />
maintenance work well before an asset<br />
begins to behave erratically, reducing<br />
the risk of severe injury.<br />
Combined with safety training and<br />
other fail-safes, a predictive maintenance<br />
program gives facility managers<br />
the ability to mitigate risk with precision.<br />
Predictive maintenance can improve<br />
asset uptime without sacrificing safety.<br />
Why predictive solutions<br />
optimize risk management<br />
Reactive maintenance programs carry<br />
a lot of financial risk. According to the<br />
International Society of Automation,<br />
PREDICTIVE MAINTENANCE<br />
REQUIRES FORETHOUGHT,<br />
STRATEGY AND INTELLIGENT<br />
TECHNOLOGY INVESTMENTS.<br />
manufacturers across all industries<br />
lose approximately $647 billion annually<br />
to asset downtime, thereby losing<br />
out on a corresponding $13 trillion in<br />
production value. To put those numbers<br />
in context, ISA explained that the<br />
average cement mill loses roughly $7<br />
million per year in lost production.<br />
Traditionally, matters of financial<br />
risk and safety risks have been a balancing<br />
act. More safety measures often<br />
meant fewer resources available for<br />
production. Predictive maintenance<br />
changes this paradigm significantly. In<br />
a predictive scheme, both productivity<br />
and safety increase with the strategic<br />
scheduling of maintenance tasks. The<br />
program not only allows maintenance<br />
staff to know when an asset is likely to<br />
fail, but also specifically how the asset<br />
is likely to deteriorate. Combined, this<br />
knowledge gives stakeholders an unprecedented<br />
strategic advantage.<br />
At the end of the day, safety issues<br />
and maintenance issues stem from the<br />
same problem: a lack of planning and<br />
analysis. Both types of risk are much<br />
easier to manage when stakeholders<br />
have the resources to plan for contingencies<br />
and eliminate the potential for<br />
risk before it manifests in a workplace<br />
accident.<br />
How technology improves<br />
safety and productivity<br />
An effective predictive maintenance<br />
strategy begins with good data. According<br />
to Control Engineering, this<br />
data can come from multiple sources,<br />
such as thermal readings and remote<br />
ultrasound sensors. The more performance<br />
indicators a facility manager<br />
has to work with, the more optimized<br />
the maintenance strategy.<br />
For example, say a facility manager<br />
knows one of his or her assets enters<br />
a fail state approximately every six<br />
months. The manager could schedule<br />
maintenance every five months to prevent<br />
failure. A smarter approach might<br />
be to place sensors along the moving<br />
parts of the asset and continually<br />
monitor decibel readings for changes.<br />
These sensors can show maintenance<br />
staff exactly which part of the asset<br />
is beginning to fail, so they can take<br />
targeted action in a timely manner.<br />
Over the years, the system would<br />
improve continually as stakeholders<br />
understand exactly when and how an<br />
asset needs to be serviced for optimal<br />
performance.<br />
Predictive maintenance requires<br />
forethought, strategy and intelligent<br />
technology investments.<br />
Sensors can be<br />
permanently<br />
mounted into<br />
an asset for an<br />
easy remote<br />
inspection<br />
The 4Cast monitors<br />
assets 24/7, taking<br />
decibel readings and<br />
sound recording<br />
1/<strong>2020</strong> maintworld 33
ASSET MANAGEMENT<br />
Root Cause of an Electrical<br />
Problem, Did You Find the<br />
Systematic Problem to Solve?<br />
Root cause failure analysis<br />
is a common term used<br />
within reliability and<br />
maintenance. Working<br />
with reliability and<br />
maintenance, we see<br />
organizations do root<br />
cause analyses but very<br />
few leads to corrective<br />
action and improvements.<br />
So why don’t we solve<br />
problems?<br />
A GOOD ROOT CAUSE elimination program<br />
needs a process, practical training<br />
in critical thinking, and coordination<br />
and follow-up of actions. IDCON has<br />
developed a process called Root Cause<br />
Problem Elimination (RCPE) where we<br />
emphasize solving the problem, documenting<br />
and executing tangible actions<br />
and visible results.<br />
I’ll give you a real-world example of a<br />
facilitated Root Cause Problem Elimination<br />
event.<br />
Wednesday at 08:42 AM Paper Machine<br />
#8 shutdown due to power loss in<br />
the press section. The paper machine<br />
was down for more than 9 hours before<br />
the circuit breaker was replaced and<br />
started up.<br />
The suggested approach is to use the<br />
same tools as in a murder investigation,<br />
because the circuit breaker was murdered!<br />
We started by collecting the data<br />
for the investigation. The data included:<br />
• Understanding how this equipment<br />
works<br />
• What happened, where, when, and<br />
similar objects<br />
• Changes in time – before, during<br />
and after the paper machine shut<br />
down?<br />
• Physical evidence - gathered all<br />
OWE FORSBERG,<br />
Senior Management<br />
Consultant with<br />
IDCON INC.<br />
Figure 1: Failed Power Circuit break for<br />
Paper machine press section.<br />
Figure 2: The How Can diagram using<br />
Post-it® notes. All you need is a wall, no<br />
software needed.<br />
the parts of the breaker, took pictures,<br />
and did a forensic analysis<br />
• Conducted interviews with personnel<br />
involved<br />
As mentioned earlier we train people<br />
to eliminate problems using critical<br />
thinking tools and a structured method.<br />
Much like RCA, the process starts with<br />
the trigger. In this case, the trigger was<br />
defined as<br />
“machine downtime event >30-minutes<br />
then initiate formal root cause analysis”<br />
The next step is to clearly state the<br />
problem. The problem statement should<br />
follow the rule “one object and one problem”<br />
in this case “circuit breaker to press<br />
section shorted”<br />
The problem was now identified. The<br />
next step was to determine “How Can”<br />
the circuit breaker short. To find alternatives<br />
that may have caused the problem,<br />
we use a tool called the How-Can Diagram<br />
(Figure 3).<br />
Start from the right, spell out the trigger,<br />
problem statement and then you<br />
ask the questions “how-can the circuit<br />
breaker short? Answers “phase to phase”<br />
or “phase to ground”<br />
Next question How-Can this short<br />
happen? “Dust on the terminal connects<br />
(stabs)” or “Corrosion on the terminal<br />
connections (stabs)” and keep going until<br />
you have exhausted the alternatives.<br />
Most important, provide all alternatives<br />
that make sense but don’t jump to conclusion.<br />
The best way to start the creative<br />
process is not to use a spreadsheet or software<br />
but post-it notes on the wall. The<br />
post-it notes engages the problem-solving<br />
team and keeps them actively involved –<br />
trust us on this one!<br />
Next step is to check the facts and find<br />
the most likely cause, in this way you can<br />
eliminate or confirm parts of the How-<br />
Can diagram. State each fact for the each<br />
of the boxes in the How-Can Diagram and<br />
ask the question “This is a fact because?”<br />
34 maintworld 1/<strong>2020</strong>
ASSET MANAGEMENT<br />
Figure 3: The How-Can diagram for<br />
this investigation<br />
Example the dust and dirt on the terminal<br />
connects (stabs) is a fact because we could<br />
see it and confirmed that it was present<br />
on other circuit breakers too. We could<br />
not confirm that there was any mechanical<br />
damaged or alignment problem of the<br />
terminal connects (stabs) of the circuit<br />
breaker. The Electrician and Electric Engineer<br />
confirmed that the arcing occurred<br />
between the breaker phase terminal<br />
connects (stabs) and not to ground. The<br />
investigation confirmed that the Technical<br />
Cause for the breaker to short is the<br />
missing dust and chemical filters.<br />
There were other causes we needed<br />
to figure out – How Can a filter not be in<br />
place? What we found was that the filters<br />
weren’t in stock because they weren’t approved<br />
for purchase! The cost for the filters<br />
were $50K each. It was deemed too<br />
expensive to the maintenance budget.<br />
The Human Cause is that the maintenance<br />
team should have known the<br />
impact of not replacing the filters to save<br />
$100k. The decision to save $100K on the<br />
filters, ended up costing them a lot more<br />
to repair the “murdered” circuit breaker.<br />
The Systematic Cause is the lack of<br />
training for how the filters impact the<br />
long-term reliability of the MCC room<br />
electrical equipment. The other systematic<br />
issue, changes to the equipment<br />
should be reviewed and approved by the<br />
qualified technical expert through the<br />
Management of Change process before<br />
being implemented.<br />
Eliminating the problem<br />
The RCPE team presented the investigation<br />
to the mill leadership and put a<br />
plan in place that would eliminate the<br />
systematic cause, the human cause and<br />
the technical cause. A training plan was<br />
put in place so that everyone understood<br />
how important the filters were to the reliability<br />
of the plant and a MOC process<br />
was put in place ensuring that no changes<br />
were made without proper approval.<br />
Reveal Your Potential<br />
Get a Reliability and Maintenance Assessment<br />
Call us +1 919-847-8764
EVENT<br />
The Day After Tomorrow<br />
in Asset Performance<br />
Peter Hinssen is an<br />
entrepreneur, who has<br />
focused on startups for<br />
almost 20 years. He is<br />
a technologist at heart.<br />
Peter will be talking at the<br />
Asset Performance 4.0<br />
Conference on September<br />
16th in Ghent about how<br />
companies and technical<br />
services can prepare for<br />
the future.<br />
PETER, YOU WILL BE PRESENT AS A<br />
KEYNOTE SPEAKER AT THE ASSET PER-<br />
FORMANCE AWARDS. CAN YOU EXPLAIN<br />
WHERE YOUR INTEREST IN THE INDUS-<br />
TRY COMES FROM?<br />
It is a little bit personal because my father<br />
worked in the oil and gas industry<br />
his entire life, specifically Maintenance<br />
and everything that deals with process<br />
control. So, when I was a kid, it was all I<br />
heard from my dad coming home. And I<br />
think the evolution that you see in this<br />
industry is fascinating. New technologies<br />
are changing, in my opinion, tremendously:<br />
dealing with assets, managing<br />
performance and thinking about prediction<br />
is going to change tremendously.<br />
So, I am very excited to be part of this.<br />
Can you give some examples of technologies<br />
that will impact our world?<br />
Big Data. I mean, this is an industry<br />
that has always been interested in information.<br />
But now Big Data is becoming<br />
abundant. We have technologies to deal<br />
with that. We have machine learning,<br />
artificial intelligence, all these mechanisms<br />
of connectivity. I think if you put<br />
it all together, it is piling up technology<br />
upon technology that is fundamentally<br />
changing how we think about how to<br />
deal with data. And I think it will have a<br />
tremendous impact on this industry.<br />
PETER HINSSEN,<br />
Author of The Day<br />
After Tomorrow and<br />
keynote speaker at the<br />
Asset Performance 4.0<br />
Conference in Belgium.<br />
WHAT DO YOU THINK THE MAIN CHAL-<br />
LENGE OF THE INDUSTRY TODAY IS?<br />
We are currently in a disruptive era. I use<br />
this word carefully, but it indicates a constant<br />
acceleration and the need to follow<br />
that speed. Lots of companies see a huge<br />
conflict between possibilities and reality.<br />
So this gap and tension between what is<br />
possible and what you do day-to-day is<br />
a big challenge. We need to take a huge<br />
leap in skills and technology. This is also<br />
an opportunity to become more critical<br />
of your company. Performance plays an<br />
important role. And the reason why your<br />
company exists is absolutely core. But be<br />
careful what you wish for, because once<br />
you enter the spotlight, you have got to<br />
deliver. Take up your role and realise it.<br />
WHAT MAKES IT SO DIFFICULT TO TAKE<br />
UP THIS ROLE?<br />
Being able to tell the story and carry it<br />
out. Storytelling is key. IT people should<br />
be rockstars, but most of the time they<br />
are not so communicative. That is because<br />
they don't have the skills to tell<br />
their story. If you go from predictive<br />
maintenance to Asset Performance in a<br />
connected world, then you have so many<br />
touch-points, that you have to broaden<br />
your gaze. You need more skills and<br />
competences. Your suppliers change.<br />
Your partners change. And everything<br />
becomes more fluid.<br />
IN YOUR BOOK ‘THE DAY AFTER TOMOR-<br />
ROW’, YOU'RE TALKING ABOUT WHAT IS<br />
GOING WRONG IN COMPANIES TODAY.<br />
WHAT IS YOUR VISION?<br />
Well, I have a very simple idea of how<br />
much time companies spend on today,<br />
tomorrow, and the day after tomorrow.<br />
Most companies are very busy with today.<br />
And when they look at the future,<br />
they often extrapolate today, they think<br />
that tomorrow is approximately the<br />
same. But we are now facing so many<br />
different changes that there might be<br />
changes in business models or in technologies<br />
or new players coming onto the<br />
market. We have to think about this disruption,<br />
'this is the day after tomorrow',<br />
and how you deal with that. When I talk<br />
about today, tomorrow, and the day after<br />
tomorrow, many people say they dedicate<br />
70-20-10 percent of their time on it.<br />
The reality is we spend 93 percent of our<br />
time on today, maybe 7 percent thinking<br />
about tomorrow and virtually none<br />
on the day after tomorrow. And I think<br />
in many industries, this was okay and in<br />
the 20th century. But we are now fully in<br />
the 21st century. That doesn't work anymore.<br />
We have to be much more flexible<br />
and agile. And that is why the day after<br />
tomorrow is more important than ever<br />
before.<br />
YOU WILL ALSO TALK ABOUT TWO IN-<br />
TERESTING CONCEPTS: STAYING ESSEN-<br />
TIAL AND STAYING RELEVANT. HOW CAN<br />
WE ACHIEVE THAT?<br />
Of course you want to be essential. You<br />
want to do something that makes sense.<br />
If you do predictive maintenance, that<br />
is essential. If you work for customers,<br />
you are hoping that you are vital for that<br />
customer. But the other question is, how<br />
relevant are you? And I think there is a<br />
very clear difference between essential<br />
and relevant. Think about the telecom<br />
industry. If you look at telecom 10 years<br />
ago, a telecom operator was essential.<br />
You needed a SIM card, and they were<br />
relevant, they gave you added value. Today,<br />
in this world, they are still essential,<br />
because you still need that SIM card. But<br />
the relevance has dropped. And therefore,<br />
if whatever capacity you have in an<br />
36 maintworld 1/<strong>2020</strong>
EVENT<br />
Asset Performance 4.0<br />
Conference & Exhibition<br />
• Asset Performance 4.0 Conference,<br />
15-17 September <strong>2020</strong>,<br />
ICC Ghent, Belgium<br />
• More info and registrations :<br />
www.assetperformance.eu<br />
organization, whatever position you have<br />
in dealing with the outside world, that<br />
is the core question, are you essential? I<br />
hope you are. But how can you make sure<br />
that your relevance doesn't go down?<br />
AND THAT BRINGS US TO THE MAINTE-<br />
NANCE AND QUALITY DEPARTMENT.<br />
THEY DON'T ADD VALUE DIRECTLY TO<br />
THE PRODUCT. THE CUSTOMER DOESN'T<br />
PAY FOR MAINTENANCE THAT WAS NEC-<br />
ESSARY TO PRODUCE HIS PRODUCT. BUT<br />
IT IS ESSENTIAL. SO WHAT WOULD YOUR<br />
TIP BE TO STAYING RELEVANT?<br />
Well, I think it is that we are in an age<br />
where everything is interconnected.<br />
So, if you look at an organization, they<br />
are not silos anymore. We are in this<br />
network age, everything is connected<br />
to everything else. So when you say that<br />
your customer doesn't pay for the maintenance<br />
directly, that is true, but your<br />
customer will feel, see and understand<br />
whether this is something which is integral<br />
in terms of quality thinking or performance<br />
management. And in the end,<br />
if your company wants to be flexible and<br />
fast and agile, you have to incorporate<br />
that into every part of the organization.<br />
Every fibre, every node, every element<br />
has to understand that you are part of<br />
a bigger picture, and you have to keep<br />
reinventing yourself to be both essential<br />
and relevant for the outside world.<br />
THE DAY AFTER TOMORROW<br />
IS MORE IMPORTANT THAN<br />
EVER BEFORE.<br />
THIS CONNECTIVITY IS ENTERING A LOT<br />
OF FACTORIES. THEY ARE EXPERIMENT-<br />
ING, BUT WE SEE SOME RELUCTANCE<br />
AND DIFFICULTIES TO SCALE UP. HOW<br />
CAN YOU SOLVE THAT?<br />
Well, I think we are in a phase where a<br />
lot of the technologies emerging take<br />
something like AI or machine learning<br />
that is relatively new. Most people<br />
don't understand it very well, there<br />
is a huge skill gap that we need to fill,<br />
because we need to train and prepare<br />
people for that. But a lot of things are<br />
just trying out, companies are experimenting<br />
and figuring out how to apply<br />
this, but it is very early in the game.<br />
If you compare that to the PC industry,<br />
this was the time where we had<br />
Commodores and Ataris, and not the<br />
established industry like we have it<br />
today. We are going through that phase.<br />
And if you are too early, you are going<br />
to burn a lot of money and not get a<br />
lot of results. But if you wait too long,<br />
you probably run the risk of becoming<br />
completely obsolete. I think it is making<br />
sure that you are constantly in tune that<br />
you are constantly alert that you follow<br />
this as closely as possible and make the<br />
right move at the right time. But you<br />
can only do that if you are prepared.<br />
HOW CAN PEOPLE PREPARE FOR THIS<br />
SKILL CHALLENGE?<br />
I think they have to make time for it.<br />
Time is the biggest issue. To put your<br />
day-to-day work at the side is difficult,<br />
but you really have to invest in these<br />
skills. Experiment at first, like tinkering<br />
with fuel until somethings blows<br />
up. There are so many possibilities,<br />
also online, to test and try out different<br />
stuff. Everybody talks about life long<br />
learning, but most managers don't do it<br />
themselves.<br />
WHY SHOULD PEOPLE ATTEND YOUR<br />
KEYNOTE AT THE ASSET PERFORMANCE<br />
4.0 CONFERENCE?<br />
I think this is one of the most fascinating<br />
industries that for a long time, has<br />
already worked with data. But it's now<br />
making a quantum leap. I'd love to talk<br />
about how I see that evolving, I hope to<br />
inspire you to maybe even do more than<br />
what you're doing today. But above all,<br />
to prepare ourselves for, I think, a very<br />
disruptive wave that is going to affect<br />
everyone. And I think if we understand<br />
this, we can all actually come out even<br />
better as a result.<br />
1/<strong>2020</strong> maintworld 37
PARTNER ARTICLE<br />
A standardised methodology<br />
with factory specific outcome<br />
Multi-site approach with VDM XL<br />
Every factory is unique. Think of differences in<br />
product and manufacturing process, the technical<br />
condition of the assets or the way we do maintenance.<br />
Then it appears to be impossible to implement one<br />
standardised improvement method still enabling each<br />
Technical Services Department to add value to the<br />
operating result. It is possible though, with VDM XL .<br />
VDM XL STANDS for Value Driven Maintenance<br />
and Asset Management. VDM XL<br />
explains how to extract maximum economic<br />
value from an existing plant, fleet<br />
or infrastructure using a professional<br />
management approach. This worldwide<br />
recognised method was developed by<br />
Mark Haarman and Guy Delahay from<br />
consultancy firm Mainnovation. With<br />
this methodology capital-intensive companies<br />
can professionalise their Technical<br />
Services Department and transform<br />
it from a cost centre into a business<br />
function that continuously improves<br />
business performance.<br />
Customisation<br />
What do you need, to improve Maintenance<br />
& Reliability in your organisation?<br />
How to create value with asset<br />
management? How do you manage your<br />
maintenance organisation and make<br />
sure your decisions benefit the company?<br />
“The answers to these questions<br />
vary per factory”, says Mark Haarman,<br />
managing partner from Mainnovation.<br />
“We have applied the VDM XL method in<br />
factories in various industries like Food<br />
& Beverage, Life Science and (Petro)<br />
Chemicals. But also in Energy & Utilities<br />
and Fleet & Transportation you need<br />
to manage your assets, preferably in a<br />
way that creates value for the company.”<br />
Haarman explains how every factory,<br />
every client has specific needs because of<br />
the uniqueness of the assets. “Working<br />
with a standardised method and making<br />
sure that every specific factory executes<br />
maintenance in the same way, seems impossible.<br />
But we provide a solution with<br />
VDM XL . Our standard approach can be<br />
customised when it comes down to implementing<br />
improvements. We present<br />
a plant-specific solution, with a focus on<br />
the most dominant value driver per factory.<br />
VDM XL is a standardised methodology,<br />
but the outcome is always factory<br />
specific.”<br />
Value drivers<br />
VDM XL distinguishes four axes on which<br />
the Maintenance & Asset Management<br />
organisation can add value with an existing<br />
installation. They are called the four<br />
value drivers. Haarman: “With an audit<br />
we can measure the current performance<br />
and maturity levels of the Technical<br />
Services Department and determine<br />
MULTI-SITE APPROACH :<br />
THE MOST DOMINANT VALUE driver<br />
varies per factory. “But every factory<br />
can start with the same method”,<br />
Haarman explains. That's why VDM XL<br />
is the right decision for a multi-site<br />
approach. The method has already<br />
proven itself at large companies with<br />
multiple site locations. “By using one<br />
method, it's possible to compare sites,<br />
while each factory derives its own<br />
action plan. A strategy that creates<br />
value per factory and creates opportunities<br />
for continuous improvement.<br />
And I can tell you: that makes both<br />
senior management and the Technical<br />
Services Department happy.”<br />
what would be the winning Maintenance<br />
and Asset Management strategy for the<br />
future. This strategy describes the improvements<br />
in processes, organisation,<br />
IT systems, data and performance management<br />
needed to create maximum<br />
economic value for the business. Now<br />
we know on which value driver we need<br />
to focus: should the Technical Service<br />
Department be managed based on cost<br />
reduction, increased uptime, safety improvement<br />
or lifetime extension?”<br />
MAINNOVATION<br />
☏ +31 (0)78 614 67 24<br />
info@mainnovation.com<br />
www.mainnovation.com<br />
38 maintworld 1/<strong>2020</strong>
ASSET MANAGEMENT<br />
Viewing Maintenance<br />
as a System to<br />
Optimize Performance<br />
Tracy T. Strawn,<br />
Strategic Advisor,<br />
Marshall Institute<br />
In 1958, Mao Zedong of China ordered all sparrows killed because they were eating<br />
grain necessary for people, and he thought killing the sparrows would result in<br />
surplus food for 60,000 people. This campaign seemed successful, as the sparrow<br />
was nearly made extinct in China. Unfortunately, Mao did not realize that sparrows<br />
were natural predators to locusts and other insects. With the sparrows gone, the<br />
locusts multiplied, devastating Chinese agriculture. The ecological imbalance helped<br />
spur on massive food shortages and the death of an estimated 30 million people.<br />
WHY DID THIS HAPPEN? Mao didn’t have an appreciation of a<br />
system which maintained ecological balance. Once upset, the<br />
unintended consequences resulted in millions of deaths.<br />
Ecological systems and their mismanagement are only one<br />
facet of the study of systems. Systems or systems theory is an<br />
interdisciplinary field that studies the nature of systems—from<br />
simple to complex—in nature, society, engineering, technology<br />
and science. Some areas of study include systems engineering,<br />
systems analysis and systems thinking. This article will focus<br />
specifically on how systems concepts apply to the maintenance<br />
organization and why we should be concerned about it.<br />
What is a system?<br />
• A system is a group of interacting or interrelated entities<br />
that form a unified whole. A system is delineated by its<br />
spatial and temporal boundaries, surrounded and influenced<br />
by its environment, described by its structure and<br />
purpose and expressed in its functioning. Systems are<br />
the subjects of study of systems theory. ¹<br />
From this we can assume a system is comprised of smaller subsystems<br />
with a purpose or goal. Systems science can apply to a<br />
business or organization.<br />
• Blanchard defined system as “…a set of interrelated components<br />
working together with the common objective of<br />
fulfilling some designated need”. 2<br />
When applied to businesses, we can conclude that these “components”<br />
are working together to produce something of value<br />
to the user, and that their efficiency and effectiveness are<br />
dependent on how well they work together. If a system is inadequately<br />
designed or poorly connected and integrated, it most<br />
likely will be unable to achieve its goals.<br />
Symptoms of broken systems:<br />
• Silo mentality: an inward mindset that resists sharing<br />
information/resources with others<br />
• Processes plagued by waste and inefficiency due to poor<br />
workflow design, broken customer/supplier relationships<br />
and poor decision making<br />
• Variability in meeting customer requirements due to<br />
poor system design and standardization<br />
• Components or subsystems of an organization are not<br />
aligned with a clear purpose, characterized by inability<br />
to execute strategy and lack of understanding of what<br />
matters to performance.<br />
•<br />
These symptoms can lead to financial ruin.<br />
Business systems are frequently identified by department<br />
names: Procurement, Engineering, Finance or Human Resources.<br />
Their workflows, connectivity and relationships result<br />
in producing something of value. If their workflows, connectivity<br />
and relationships are optimum, the value they produce<br />
should meet or exceed the goals of the business.<br />
Systems and processes are the essential building blocks of<br />
a company. Every facet of business - storeroom, workshop,<br />
production - is part of a system that can be managed to produce<br />
something of value. The details of each business system vary<br />
by company, but the fundamentals remain the same.<br />
¹ Wikipedia, Systems<br />
² Systems Engineering Management, 2nd Edition, B. Blanchard, 199<br />
40 maintworld 1/<strong>2020</strong>
ASSET MANAGEMENT<br />
The Key Concepts of Systems:<br />
1. Systems are composed of interconnected parts<br />
2. Inter-connected parts are interdependent<br />
3. Every system has an aim<br />
4. The structure of a system determines its behavior<br />
5. How well the parts cooperate to support the aim determines<br />
how efficient and effective the system functions<br />
6. Ultimately we seek synergy in which “the whole is greater<br />
than the sum of the parts<br />
CORPORATE<br />
OBJECTIVE<br />
MAINTENANCE<br />
OBJECTIVE<br />
PLANT<br />
OBJECTIVE<br />
Is Maintenance a System?<br />
Based on definitions, maintenance is a system: usually a subsystem<br />
of the corporation, made up of a collection of elements<br />
organized to achieve a purpose. How well the elements are integrated<br />
and interact determines the efficiency and effectiveness<br />
of the maintenance system.<br />
Figure 1 is a simple depiction of a system view of maintenance.<br />
The maintenance objective will reflect and align with<br />
corporate and plant objectives, considering the plant structure,<br />
while defining equipment strategies for ensuring the level of<br />
reliability required. Equipment strategies will define spares<br />
policy and outline resource structure needed to plan, schedule<br />
and execute the workload. The administrative structure will<br />
define and support personnel policies and determine budgets/<br />
controls to manage costs. Subsystems will be reviewed/improved<br />
to optimize their contribution to the systems aim and<br />
goals: achieving the desired level of reliability while meeting<br />
cost and safety targets.<br />
MAINTENANCE<br />
CONTROL<br />
ADMIN<br />
STRUCTURE<br />
REVIEW AND<br />
IMPROVE<br />
RESOURCE<br />
STRUCTURE<br />
PLANT<br />
STRUCTURE<br />
EQUIPMENT<br />
STRATEGIES<br />
1/<strong>2020</strong> maintworld 41
ASSET MANAGEMENT<br />
VIEWING AND UNDERSTANDING MAINTENANCE<br />
AS A SYSTEM WITH AN AIM AND PURPOSE<br />
IS THE FIRST STEP IN DESIGNING AND<br />
DEVELOPING A MAINTENANCE SYSTEM.<br />
Figure 2 depicts maintenance as a system. This example<br />
shows how three sub-systems working as a system contribute<br />
to achieving optimum inherent availability. Three subsystems<br />
in this example are materials management, resource management,<br />
and reliability management. Materials management<br />
and resource management aid in optimizing the repair process.<br />
Both subsystems are critical in identifying and procuring<br />
spares required for the repair, planning/scheduling the repair,<br />
and mobilizing the resources to complete the repair. To ensure<br />
repair is made in the shortest time and at the lowest cost,<br />
information and communication must flow seamlessly and<br />
continuously between the two subsystems. The metric used to<br />
measure the repair process is ‘mean time to repair’ or MTTR.<br />
The third subsystem in the maintenance system example<br />
is reliability management, responsible for establishing the<br />
equipment strategies for achieving optimum reliability while<br />
considering safety and cost. Equipment strategies are defined<br />
by manufacturers’ recommendations or a risk based approach<br />
like reliability centered maintenance. The metric used to<br />
measure reliability performance is ‘mean time between failure’<br />
or MTBF. MTTR and MTBF will be used as illustrated in Figure<br />
2 to calculate inherent availability.<br />
In order to realize the goal, achieving the desired inherent<br />
availability, all three subsystems must work together.<br />
Achieving a systems aim must be managed with attention to<br />
the entire system. When we optimize subcomponents of the<br />
system, we don’t necessarily optimize the overall system. Suboptimization<br />
is the practice of focusing on one part of a system<br />
and making changes intended to improve that subsystem while<br />
ignoring the effects on other subsystems. This will lead to suboptimization<br />
of the whole system leading to waste, delays, and<br />
inefficiencies resulting in lost profits and lower plant throughput.<br />
Optimizing system performance should begin in the system<br />
design phase.<br />
8 Characteristics of a Good System Design:<br />
1. Designed with the customer (internal and external) in mind<br />
• Work products and services are handed off to internal<br />
customers who must meet their requirements.<br />
The objective is to prevent the internal customer<br />
from reworking or worse, passing through product<br />
that is fails to meet specifications.<br />
2. Represents your best known way of doing something<br />
• A well defined system should be documented with<br />
workflows of each subsystem and where the work is<br />
performed and handoffs are made, clearly describing<br />
roles and responsibilities.<br />
3. Has one primary aim, goal or purpose<br />
• Primary/secondary goals should be clearly defined<br />
and communicated to stakeholders.<br />
4. Has an owner, accountable for/reporting on system results<br />
5. Is as simple as possible, documented, understood by<br />
workers, and repeatable<br />
• System users should be trained/coached on subsystems<br />
and processes.<br />
6. Has performance standards and results are measured.<br />
• System should be thoroughly implemented and capable<br />
of meeting performance standards and goals.<br />
System users understand performance standards<br />
and can identify out of compliance conditions. Measures<br />
are monitored by all team members.<br />
7. Workers get ongoing feedback about system performance<br />
and are recognized for good results.<br />
8. Has sufficient focus on system details to eliminate most<br />
bottlenecks, inefficiencies, waste, and rework.<br />
Optimizing Maintenance System Performance<br />
These steps will aid in optimization of existing maintenance<br />
organizations and system with every process having an input<br />
from a supplier and an output to a customer. Process stakeholders<br />
understand how they add value to the goal of the system.<br />
A robust system will have clear specifications for each<br />
product handed off to internal customers along with feedback<br />
to suppliers.<br />
1. Process map and document your system<br />
• Identify your processes<br />
• Show how the processes work together to produce<br />
value and their interconnectivity<br />
• Ensure roles and process steps are clear<br />
2. Identify your products and services for each process<br />
3. Understand customer/supplier specifications/requirements<br />
4. Identify the suppliers/customers for each product/service<br />
5. Communicate process requirements to suppliers<br />
6. Identify customer specifications<br />
7. Demonstrate the process is capable/meets specifications<br />
8. Achieve a thorough implementation<br />
9. Continuously improve<br />
Summary<br />
The complexity of maintenance systems increases as new technologies<br />
are introduced. In today’s environment, there is an<br />
increasing need to develop/produce systems that are robust, reliable,<br />
high quality, supportable and cost effective. Viewing and<br />
understanding maintenance as a system with an aim and purpose,<br />
rather than a collection of disparate parts, is the first step<br />
in designing and developing a maintenance system that can be<br />
managed and optimized for sustained long term performance.<br />
42 maintworld 1/<strong>2020</strong>
PARTNER ARTICLE<br />
Use of High-speed<br />
Thermography<br />
in Laser High-temperature<br />
Capillary Gap Brazing<br />
Lasers are extremely versatile tools in industry and manufacturing technology.<br />
Due to their flexibility, they serve as a key technology for implementing<br />
the goals of industry 4.0. Although laser cutting and welding are nowadays<br />
regarded as turnkey technologies, most laser applications, for example<br />
joining of hybrid materials, 3D printing or ultra-short pulse processing,<br />
still require considerable research and development.<br />
Text: M. HOFELE, D. KOLB, S. RUCK, H. RIEGEL, LaserApplikationsZentrum of the Hochschule Aalen; InfraTec GmbH Infrarotsensorik und Messtechnik<br />
Photos: LASERAPPLIKATIONSZENTRUM OF THE HOCHSCHULE AALEN<br />
THE LASERAPPLICATIONCENTER (LAC)<br />
of Aalen University intensively researches<br />
and develops new methods of laser<br />
material processing. Thus, innovative<br />
materials for Additive Manufacturing<br />
are developed and investigated within<br />
public R&D projects, including magnetic<br />
materials or electrical energy storage materials<br />
for electromobility. Another focus<br />
is lightweight construction. Here, among<br />
other things, mixed metallic compounds<br />
and hybrid lightweight structures<br />
made of aluminium and CFRP for CO₂efficient<br />
mobility concepts are investigated.<br />
The newly developed processes<br />
aluminium laser polishing and high-temperature<br />
capillary gap brazing are already<br />
being used in industrial projects.<br />
Making Heat Flow Visible<br />
Laser processes are highly dynamic<br />
thermally induced processes that cannot<br />
be detected with the naked eye.<br />
High-speed visual cameras that are<br />
often used are not capable of visualis-<br />
ing the heat flow in the component. A<br />
contacting temperature measurement<br />
of the moving very small liquid metal is<br />
not possible. In addition, the processing<br />
zone should remain free of influences<br />
from the measurement system. This is<br />
exactly what thermographic cameras<br />
do, which provide high frame rates with<br />
high spatial resolutions at the same<br />
time.<br />
Specially Configured<br />
Thermographic Camera within<br />
Laser Material Processing<br />
Laser-based production involves special<br />
requirements on the use of a thermographic<br />
camera. One reason for this are<br />
the processing temperatures of typically<br />
500 °C to 2,000 °C. In addition, the components<br />
of the camera must be protected<br />
against sputters, caused by the machining<br />
process. If the laser manufacturing<br />
processes take place in process chambers<br />
under a specific atmosphere, the<br />
measurement section is enriched with a<br />
gas. Especially the optics of the camera<br />
must be protected from the reflected<br />
laser radiation. It is therefore equipped<br />
with a laser protection lens for solidstate<br />
lasers and a filter for through glass<br />
and high-temperature measurement.<br />
Due to these precautions, the camera can<br />
be used near the laser beams without any<br />
problems.<br />
Configured appropriately, the thermographic<br />
camera ImageIR® 8300 hp<br />
from InfraTec supports the LAC with<br />
its high spatial and temporal resolution.<br />
Using the camera's MicroScan function,<br />
images can be taken with a spatial resolution<br />
of more than one megapixel. The<br />
10 GigE interface allows fast data transmission<br />
up to 355 Hz in full frame mode.<br />
Due to the optical package consisting of<br />
a telephoto lens with 50 mm focal length<br />
and the close-up lens for reducing the<br />
minimum focusing distance down to 170<br />
mm, the researchers can easily adapt the<br />
camera to changing working distances<br />
and sizes of the measured objects.<br />
1/<strong>2020</strong> maintworld 43
PARTNER ARTICLE<br />
Laser camera: ImageIR® 8300 hp<br />
Fig. 1 (a) Test setup in laser cell<br />
TLC 1005, (b) 3D model process<br />
chamber, (c) Dimensions of<br />
solder sample<br />
LASERS ARE EXTREMELY<br />
VERSATILE TOOLS<br />
IN INDUSTRY AND<br />
MANUFACTURING<br />
TECHNOLOGY.<br />
Analysing the Temperature<br />
Control Behaviour During<br />
Laser Beam High-temperature<br />
Soldering<br />
The LAC of Aalen University, together<br />
with its industrial partner conntronic<br />
Prozess- und Automatisierungstechnik<br />
GmbH from Augsburg, is investigating<br />
laser high-temperature capillary gap<br />
brazing of corrosion-resistant steels<br />
for tube assemblies in automotive and<br />
mechanical engineering within the<br />
framework of the publicly funded research<br />
project “enAbLe”. In contrast to<br />
induction and furnace brazing, the laser<br />
beam serves as a flexible and highly efficient<br />
tool. The challenge lies on the one<br />
hand in the required highly pure reduction<br />
process environment to remove<br />
the oxide layers for good wetting of the<br />
copper solder and on the other hand<br />
in the homogeneous tempering of the<br />
joining zone. For homogeneous heating,<br />
the laser beam is controlled to the<br />
desired process temperature of 1,300 °C<br />
by means of a coaxially integrated highspeed<br />
pyrometer with sampling rates of<br />
several kilohertz. In addition to temperature<br />
control, the exposure strategy has<br />
a decisive influence on the formation of<br />
the temperature zones. Besides the FEM<br />
simulation, the thermographic camera<br />
is used for process development in the<br />
empirical experiments.<br />
The experiments take place in a sixaxis<br />
TLC 1005 laser cell with an infrared<br />
4 kW disc laser TruDisk 4002. The test<br />
geometry consists of a tube plug connection,<br />
austenitic chrome-nickel steel<br />
1.4301, with outer tube diameters of 10<br />
mm and 7.9 mm. Three weld spots offset<br />
by 120° fix the pipe-plug connection. The<br />
solder is pure copper from Voestalpine<br />
in the form of a solder ring (Fig. 1c). The<br />
oxygen-reduced process chamber used<br />
for the tests has a laser-permeable beam<br />
entry window in the lid. The soldering<br />
tests are carried out in a forming gas atmosphere<br />
with a residual oxygen content<br />
of less than 150 ppm. The atmosphere is<br />
monitored by means of a residual oxygen<br />
measuring device. During the process,<br />
the soldering assembly, clamped in a<br />
three-jaw chuck, rotates around the pipe<br />
axis by an external rotary machine axis.<br />
The laser beam is defocused to a diameter<br />
of 9 mm and radially tempering the<br />
outer surfaces of the joining zone (Fig.<br />
1b). The beam centre is oriented centrically<br />
to the solder ring.<br />
The laser soldering process is divided<br />
into three phases: heating, soldering with<br />
component rotation and cooling. During<br />
the heating phase of 10 s (Fig. 2 upper<br />
row), the laser heats the facing component<br />
surface to the control temperature<br />
of 1,300 °C in a stationary manner. During<br />
the soldering phase (Fig. 2 middle<br />
row) the assembly performs a complete<br />
rotation with an angular speed of 540 °/<br />
min. After the copper solder depot has<br />
been molten, the solder gap filling starts<br />
at 31.9 seconds. Due to the lower emission<br />
of copper, the forming fillet appears<br />
cooler than the surrounding steel surface<br />
of the joining partners. Because of the very<br />
good thermal conductivity of the solder,<br />
the through heating of the inner tube also<br />
starts at this point (difference between<br />
picture at 16.3 s and picture at 31.9 s in<br />
Fig. 2). At the end of complete rotation<br />
and shutdown of the laser, the component<br />
cools to below 600 °C within 18 s (Fig. 2<br />
lower row).<br />
The measurement data of the thermographic<br />
camera allows a versatile process<br />
analysis afterwards. The diagram in figure<br />
3 shows the temperature-time sequence<br />
of two measurement points, P1 in the laser<br />
spot centre and P2 on the inner tube. The<br />
dynamic temperature changes in the gap<br />
filling process stand out in this context.<br />
Fig. 2 Process sequence for laser beam high-temperature capillary gap soldering of a<br />
tube plug-in connector<br />
44 maintworld 1/<strong>2020</strong>
PARTNER ARTICLE<br />
Fig. 3 Analysis of the temperature-time sequence on the basis of the thermal image data<br />
Fig. 4 Cutted joining zone of a lasersoldered<br />
tube-plug-in connection in<br />
longitudinal direction.<br />
Figure 4 shows the longitudinal<br />
cross section through the faultless laser<br />
soldered tube plug connection. The presented<br />
sample shows a complete gap filling<br />
without porosity. The evenly formed<br />
grooves on both sides ensure a good<br />
distribution of force and low-turbulence<br />
flow inside the pipe.<br />
Due to the evaluation of further thermographic<br />
data, the temperature control<br />
strategy could be further optimised<br />
and soldering times of less than 10 s<br />
could be realised.<br />
Collecting Important Process<br />
Knowledge<br />
Regarding laser soldering, thermography<br />
provides important insights into<br />
process development. Thus, optimised<br />
path planning for optimal heating can<br />
be identified and at the same time<br />
the thermal load of the surrounding<br />
zones can be reduced. The LAC aims<br />
to achieve comparable results with<br />
similar tasks. With the help of the thermographic<br />
camera, for example, the<br />
temperature distribution gets recorded<br />
in each layer at the 3D metal printing<br />
process of selective laser melting. The<br />
focus lies primarily on the heating and<br />
cooling behaviour of the used materials,<br />
which has a direct effect on the<br />
structure quality that is formed.
ASSET MANAGEMENT<br />
Predictive Maintenance:<br />
The Wrong Solution to the<br />
Right Problem in Chemicals<br />
WIM GYSEGOM, SVEN HOUTHUYS, and JOEL THIBERT, McKinsey & Company<br />
Chemicals plants often<br />
have plenty of good data<br />
on equipment performance<br />
and reliability. A predictivemaintenance<br />
program<br />
might be the worst way to<br />
use it.<br />
IN THE CHEMICALS INDUSTRY, like many<br />
others, there is considerable excitement<br />
about the potential of advanced predictive-maintenance<br />
(PdM) approaches.<br />
The promise of these new techniques<br />
is tantalizing. Using machine-learning<br />
technologies to comb through historical<br />
performance and failure data, they<br />
aim to tell operators when and how a<br />
component is likely to go wrong in the<br />
future with a high level of predictability.<br />
This should reduce the impact of equipment<br />
failures—and the cost of efforts to<br />
prevent such failures—by turning inefficient,<br />
unplanned maintenance activities<br />
into efficient, planned ones.<br />
At first sight, chemicals plants seem<br />
like the ideal environment for PdM.<br />
High levels of automation and instrumentation,<br />
combined with rigorous<br />
maintenance record-keeping, create the<br />
rich data that machine-learning systems<br />
require. Moreover, most plants strive for<br />
stable operating conditions, potentially<br />
making it easier to spot patterns and<br />
trends. There’s also a compelling business<br />
case for improved reliability. Overall<br />
equipment effectiveness (OEE) losses<br />
due to unplanned maintenance range<br />
from 3 to 5 percent across the industry.<br />
Predicting Poor Results<br />
Take a closer look, however, and the<br />
potential of PdM in chemicals begins<br />
to evaporate, for four main reasons.<br />
AUTHORS:<br />
WIM GYSEGOM is a partner<br />
in McKinsey’s London office, where<br />
SVEN HOUTHUYS is an associate<br />
partner, and JOEL THIBERT is an<br />
associate partner in the Santiago office.<br />
• TOO LITTLE DATA. In a chemicals<br />
plant, predicting failures is harder<br />
than it first appears. Unplanned<br />
downtime is typically concentrated<br />
in a small number of large events.<br />
That means there are typically too<br />
few datapoints for PdM systems to<br />
learn from.<br />
• TOO LITTLE TIME. Even when it’s possible<br />
to create models with predictive<br />
power, they often work over time<br />
horizons that are too short to be useful<br />
in chemicals manufacturing. Predicting<br />
that a part will fail in two days<br />
or two weeks is useful in a truck or<br />
machine tool, but it may not help in a<br />
plant where shutdowns take several<br />
days and maintenance teams require<br />
months to plan interventions and<br />
source spare parts.<br />
• TOO LITTLE IMPACT. The impact from<br />
PdM is often low because plants<br />
operate critical assets with a high<br />
degree of redundancy and few single<br />
points of failure. If a pump stops<br />
unexpectedly, operators can often<br />
switch to a backup unit with little impact<br />
on production.<br />
• Too little savings. Finally, a focus<br />
on reducing unplanned downtime<br />
ignores the largest source of throughput<br />
losses in most plants. Shutdowns<br />
for planned maintenance events<br />
cause OEE losses of 5 to 10 percent on<br />
average, twice as much as unplanned<br />
stoppages.<br />
Towards Digital Reliability<br />
Do these challenges mean analytics<br />
provides little or no value in efforts to<br />
improve asset productivity in the chemicals<br />
sector? No. The industry is achieving<br />
considerable success with a range of<br />
digital reliability techniques, many of<br />
which are far cheaper and less complex<br />
to implement than advanced PdM. Take<br />
three prominent examples:<br />
46 maintworld 1/<strong>2020</strong>
ASSET MANAGEMENT<br />
Condition Monitoring<br />
Improving condition monitoring<br />
through better remote sensing can cut<br />
mean time-to-repair, significantly reducing<br />
the impact of equipment failures. At<br />
one chemical plant, a few critical pumps<br />
suffered repeated failures. No backups<br />
were available for these units, and the<br />
issue was a significant source of production<br />
losses at the plant.<br />
“We decided we couldn’t wait for the<br />
plant and reliability engineers to identify<br />
the root cause, redefine the pump’s technical<br />
specifications, and then procure<br />
replacements,” the plant’s maintenance<br />
manager told us. “So, we focused on mitigating<br />
the impact of the failures, rather<br />
than avoiding them.”<br />
The plant’s reliability team installed<br />
a handful of new sensors on the pumps<br />
and started to monitor their condition<br />
online in real time, allowing them to detect<br />
imminent failures a few hours before<br />
they occurred. By enabling maintenance<br />
personnel to be ready to intervene, this<br />
intervention reduced the mean time-torepair<br />
on these pumps from 6.5 2 Predictive<br />
maintenance: the wrong solution to<br />
the right problem in chemicals to around<br />
3 hours, cutting OEE losses by almost<br />
half and saving approximately 120,000<br />
US dollars for each failure.<br />
Smarter Capex Decision–<br />
Making<br />
Better data means better investment decisions,<br />
especially when it comes to the<br />
allocation of sustaining capex costs—or<br />
avoiding equipment failure by making<br />
the right, risk-informed capex decisions.<br />
Most chemical companies struggle to<br />
set the right level of sustaining capex,<br />
as they find it difficult to allocate funds<br />
across multiple plants and disparate asset<br />
types.<br />
This problem is ultimately about ensuring<br />
that resources are used to their<br />
maximum potential—which is exactly<br />
the question that zero-based budgeting<br />
(ZBB) has successfully addressed,<br />
through internal disciplines that assess<br />
all spending in terms of return on investment.<br />
The same techniques apply<br />
to capex investments in assets, or “asset<br />
ZBB,” which combines available historical<br />
data with local expertise to assess the<br />
potential impact of either replacing or<br />
not replacing particular equipment. The<br />
new approach allows all equipment renewal<br />
projects to be compared using the<br />
same yardstick: spend efficiency.<br />
In the words of a manager responsible<br />
for running one chemical manufacturer’s<br />
capital-planning process, “We<br />
used to have a formal process to capture<br />
and assess sustaining capex projects, but<br />
we had no clear way to rank-order projects<br />
and always ended up prioritizing those<br />
with immediate, visible impact. Now we<br />
are able to have a fact-based, data-driven<br />
discussion about risk and trade-offs, which<br />
has led us to spend less overall—and to<br />
manage what we do spend more wisely.”<br />
Root-Cause Problem Solving<br />
Better data also means better root-cause<br />
problem solving. That helps companies<br />
to prevent the recurrence of failures, to<br />
improve their failure-modes and-effects<br />
analysis (FMEA) processes, and to optimize<br />
preventative-maintenance plans.<br />
Together, those actions address the critical<br />
aspects of reliability performance, reducing<br />
both the impact of failures and the cost<br />
of preventing them.<br />
At one chemicals plant, for example,<br />
failures in a critical piece of equipment<br />
caused operators to activate an emergency<br />
shutdown three times in as many months.<br />
These shutdowns were inconvenient<br />
enough—but when the site team attempted<br />
to restart the plant, they found that the<br />
abrupt shutdown of the unit led to an accumulation<br />
of solids in key vessels and pipes.<br />
Fixing that problem led to lengthy delays<br />
in start-up and significant losses in output.<br />
To address the issue, the company applied<br />
a combination of traditional rootcause<br />
problem solving and smart analytical<br />
techniques. Analysis of process data helped<br />
them understand how and why solids were<br />
accumulating under emergency shutdown<br />
conditions. The issue was fixed with a<br />
combination of enhanced monitoring and<br />
changes to preventative maintenance plans.<br />
But the data driven insights also allowed<br />
the plant to revise its emergency shutdown<br />
procedures to stop the plant safely without<br />
causing the solids problem. That change reduced<br />
start-up time after any kind of emergency<br />
shutdown by 90 percent.<br />
The potential for digital reliability extends<br />
far beyond predictive maintenance.<br />
And for chemicals companies, we believe<br />
that these other digital approaches are<br />
both easier to implement and offer greater<br />
value. The highly instrumented nature of<br />
most chemical production facilities means<br />
many companies already have a rich, and<br />
largely untapped, source of data to support<br />
digital reliability efforts. For those plants<br />
that still don’t, then it’s time to “sensor<br />
up”: better data is the vital first step on the<br />
digital-reliability journey.<br />
1/<strong>2020</strong> maintworld 47
TECHNOLOGY<br />
Industrial AI in Maintenance:<br />
False Hopes or Real<br />
ACHIEVEMENTS?<br />
Artificial intelligence (AI) is an umbrella term for a set of technologies in which<br />
computer systems are programmed to exhibit complex behaviour in challenging<br />
environments. AI is regarded as the major force driving innovation today.<br />
Authors: UDAY KUMAR, DIEGO GALAR and RAMIN KARIM, Luleå University of Technology<br />
FROM AN INDUSTRIAL point of view, AI technologies should be<br />
understood as methods and procedures that enable technical<br />
systems to perceive their environments through context and<br />
situation awareness. They are able to process what they have<br />
monitored and modelled, solve certain problems, find novel<br />
solutions never found by humans, make decisions, and learn<br />
from experience to be better able to manage the processes and<br />
tasks put under AI supervision, Figure 1.<br />
Machine learning (ML) is one area of artificial intelligence<br />
used by industry. Machines need data to learn, either large<br />
quantities of data for one-time analytical purposes, or streams<br />
of data from which learning is continuously taking place. Based<br />
on acquired data either on line or off line, machine learning<br />
can reduce complexity and detect events or patterns, make<br />
predictions, or enable actions to be taken without explicit programming<br />
in the form of the usual ‘if-then’ routines or without<br />
classic automation and control engineering, Figure 2.<br />
Figure 2:<br />
Roadmap<br />
from<br />
traditional<br />
automated<br />
process to<br />
Industrial AI<br />
Figure 1: Solution and knowledge extraction form asset data<br />
48 maintworld 1/<strong>2020</strong>
TECHNOLOGY<br />
AI technologies are expected to increase the efficiency<br />
and effectiveness of industrial processes. The primary goals<br />
are to reduce costs, save time, improve quality, and enhance<br />
the robustness of industrial processes. However, AI is not as<br />
well-used in industry as we might expect, given its potential.<br />
Enormous changes and high costs are needed to integrate AI<br />
applications into corporate structures and along the entire<br />
value-added chain. At this point, AI applications tend to be<br />
found in the areas of robotics, knowledge management, quality<br />
control, and maintenance analytics shifting from traditional<br />
approaches to predictive ones.<br />
A good field for AI in maintenance in industrial environments<br />
is the analysis and interpretation of sensor data, distributed<br />
throughout equipment and facilities. The Internet<br />
of Things (IoT), i.e. distributed data-suppliers and data-users<br />
capable of communicating with each other, is the basis for this<br />
use of AI. IoT acquires the data after pre-processing, records<br />
the status of all different aspects of the machines, and performs<br />
actions in process workflows on the basis of its analysis. Its<br />
central purpose is to identify correlations that are not obvious<br />
to humans to enable predictive maintenance (Figure 3), for example,<br />
when complex interrelated mechanical setting parameters<br />
have to be adjusted in response to fluctuating conditions<br />
in the environment to avoid compromising the asset’s health.<br />
Figure 3:<br />
Predictive<br />
maintenance<br />
supported<br />
by AI<br />
ASSET RELIABILITY<br />
PRACTITIONER<br />
[ARP] Training & Certification<br />
Improving the reliability and performance of an industrial facility is important. A top performing facility will be<br />
safer, have fewer environmental incidents, provide for job satisfaction and support financial reward for owners<br />
and shareholders.<br />
The Mobius Institute Asset Reliability Practitioner [ARP] training and certification program provides the<br />
knowledge, qualifications, and growth path to enable a program to be run successfully.<br />
[ARP-A]<br />
ARP-Advocate:<br />
Teaches the big picture<br />
including terminology.<br />
[ARP-E]<br />
ARP-Engineer:<br />
Educates on technical<br />
aspects of a reliability<br />
maintenance program.<br />
[ARP-L]<br />
ARP-Leader:<br />
Trains leaders on how to<br />
successfully implement<br />
reliability and improvement<br />
initiatives involving an<br />
entire organization.<br />
Choose the ARP course that is right for you!<br />
Visit www.mobiusinstitute.com/arp for more information
TECHNOLOGY<br />
Industrial AI’s capacity to analyze very large amounts of<br />
high-dimensional data can change the current maintenance<br />
paradigm and shift from preventive maintenance systems to<br />
new levels. The key challenge, however, is operationalizing predictive<br />
maintenance, and this is much more than connecting<br />
assets to an AI platform, streaming data, and analyzing those<br />
data. By integrating conventional data such as vibration, current<br />
or temperature with unconventional additional data, such<br />
as audio and image data, including relatively cheap transducers<br />
such as microphones and cameras, Industrial AI can enhance<br />
or even replace more traditional methods. AI’s ability to predict<br />
failures and allow planned interventions can be used to<br />
reduce downtime and operating costs while improving production<br />
yield. For example, AI can extend the life of an asset beyond<br />
what is possible using traditional analytics techniques by<br />
combining data information from designer and manufacturer,<br />
maintenance history, and Internet of Things (IoT) sensor data<br />
from end users, such as anomaly detection in engine-vibration<br />
data, images and video of engine condition. This information<br />
fusion during the lifecycle of the asset is called product lifecycle<br />
management (PLM).<br />
Explainable AI in Maintenance<br />
Advances in AI for maintenance analytics are often tied to<br />
advances in statistical techniques. These tend to be extremely<br />
complex, leveraging vast amounts of data and complex algorithms<br />
to identify patterns and make predictions. This<br />
complexity, coupled with the statistical nature of the relationships<br />
between input data that the asset provides, makes them<br />
difficult to understand, even for expert users, including the<br />
system developers, Figure 4. This makes explainability a major<br />
concern.<br />
Figure 4: Engineers and data scientist must co-create the AI<br />
solution for maintenance together<br />
While increasing the explainability of AI systems can be<br />
beneficial for many reasons, there are challenges in implementing<br />
explainable AI. Different users require different<br />
forms of explanation in different contexts, and different contexts<br />
give rise to different needs. To understand how an AI<br />
system works in the maintenance domain, users might wish to<br />
know which data the system is using, the provenance of those<br />
data, and why they were selected; how the model and prediction<br />
work, and which factors influence a maintenance decision;<br />
and why a particular output is obtained. To understand what<br />
type of explanation is necessary, careful stakeholder engage-<br />
ment and well-thought-out system design are both necessary<br />
as can be seen in figure 5.<br />
Figure 5: Architecture of explainable AI systems for Maintenance<br />
Decisions<br />
There are various approaches to creating interpretable systems.<br />
Some AI is interpretable by design; these systems tend to<br />
be kept relatively simple. An issue with them is that they cannot<br />
get as much customization from vast amounts of data as<br />
more complex techniques, such as deep learning. This creates a<br />
performance-accuracy trade-off in some settings, and the systems<br />
might not be desirable for those applications where high<br />
accuracy is prized. In other words, maintainers must accept<br />
more black boxes.<br />
In some AI systems – especially those using personal data or<br />
those where proprietary information is at stake – the demand<br />
for explainability may interact with concerns about privacy. In<br />
areas such as healthcare and finance, for example, an AI system<br />
might be analyzing sensitive personal data to make a decision<br />
or recommendation. In determining the type of explainability<br />
that is desirable in these cases, organizations using AI will<br />
need to take into account the extent to which different forms<br />
of transparency might result in the release of sensitive insights<br />
about individuals or expose vulnerable groups to harm.<br />
In the area of maintenance, when the AI recommends a<br />
maintenance decision, decision makers need to understand<br />
the underlying reason. Maintenance analytics developers need<br />
to understand what fault features in the input data are guiding<br />
the algorithm before accepting auto-generated diagnosis<br />
reports, and the maintenance engineer needs to understand<br />
which abnormal phenomena are captured by the inference algorithm<br />
before following the repair recommendations.<br />
One of the proposed benefits of increasing the explainability<br />
of AI systems is increased trust in the system. If maintainers<br />
understand what led to an AI-generated decision or recommendation,<br />
they will be more confident in its outputs. But the<br />
link between explanations and trust is complex. If a system<br />
produces convincing but misleading explanations, users might<br />
develop a false sense of confidence or understanding. They<br />
might have too much confidence in the effectiveness or safety<br />
of systems, without such confidence being justified. Explanations<br />
might help increase trust in the short term, but they do<br />
not necessarily help create systems that generate trustworthy<br />
outputs or ensure that those deploying the system make trustworthy<br />
claims about its capabilities.<br />
REFERENCES:<br />
Galar, Diego, Pasquale Daponte, and Uday Kumar. Handbook of Industry 4.0 and SMART Systems. CRC Press, 2019.<br />
Galar, Diego. Artificial intelligence tools: decision support systems in condition monitoring and diagnosis. Crc Press, 2015.<br />
Galar, Diego, Uday Kumar, and Dammika Seneviratne. Robots, Drones, UAVs and UGVs for Operation and Maintenance. CRC Press, <strong>2020</strong>.<br />
50 maintworld 1/<strong>2020</strong>
VIBRATION ANALYSIS<br />
THERMAL IMAGING<br />
ULTRASOUND<br />
MEASUREMENT<br />
eyesight – hearing – sensitivity<br />
we have in common<br />
MASTER THE LANGUAGE OF YOUR MACHINERY<br />
WWW.ADASH.COM
VIBGUARD® IIoT<br />
The world‘s first and only state-of-the-art solution for<br />
predicitive maintenance in Industry 4.0 environments<br />
Online Condition Monitoring System:<br />
• Available with 6 up to 20 analogue channels<br />
• Applicable in any Industry 4.0 environment<br />
• Featuring MQTT protocol for global<br />
remote status overview<br />
• Process quality enhancement due to<br />
seamless integration into control systems<br />
www.VIBGUARD-IIoT.com