The next BriefingsDirect Voice of the Customer discussion revisits the drive to define the “refinery of the future” at Texmark Chemicals.
Texmark has been combining the best of operational technology (OT) with IT and now Internet of Things (IoT) to deliver data-driven insights that promote safety, efficiency, and unparalleled sustained operations.
Stay with us now as we hear
how a team approach -- including the plant operators, consulting experts and
latest in hybrid IT systems -- joins forces for rapid process and productivity optimization results.
Listen to the podcast. Find it on iTunes. Read a full transcript or download a copy.
To learn how, are are joined by our panel, Linda
Salinas, Vice President of Operations at Texmark Chemicals, Inc. in Galena
Park, Texas; Stan
Galanski, Senior Vice President of Customer Success at CB Technologies (CBT) in Houston, and Peter
Moser, IoT and Artificial Intelligence (AI) Strategist at Hewlett Packard Enterprise
(HPE). The discussion is moderated by Dana Gardner, Principal Analyst at Interarbor Solutions.
Here are some excerpts:
Gardner: Stan, what are the trends, technologies, and operational methods that have now come together to make implementing a refinery of the future approach possible? What’s driving you to be able to do things in ways that you hadn’t been able to do before?
 |
| Galanski |
Galanski: I’m
going to take that in parts, starting with the technologies. We have been
exposed to an availability of affordable sensing devices. These are
proliferating in the market these days. In addition, the ability to collect
large amounts of data cheaply -- especially in the cloud -- having ubiquitous Wi-Fi,
Bluetooth, and other communications have presented themselves as an opportunity
to take advantage of.
On top of this, the
advancement of AI
and machine learning (ML)
software -- often referred to as analytics -- has accelerated this opportunity.
Gardner: Linda,
has this combination of events dramatically changed your perspective as VP of operations?
How has this coalescing set of trends changed your life?
Salinas: They
have really come at a good time for us. Our business, and specifically with Texmark,
has morphed over the years to where our operators are more broadly skilled. We
ask them to do more with less. They have to have a bigger picture as far as
operating the plant.
Today’s operator is not just sitting
at a control board running one unit. Neither is an operator just out in a unit,
keeping an eye on one tower or one reactor. Our operators are now all over the
plant operating the entire utilities and wastewater systems, for example, and
they are doing their own lab analysis.
This technology has come at a
time that provides information that’s plant-wide so that they can make more
informed decisions on the board, in the lab, whenever they need.
Gardner: Peter,
as somebody who is supplying some of these technologies, how do you see things
changing? We used to have OT and IT as separate, not necessarily related. How
have we been able to make those into a whole greater than the sum of their parts?
OT plus IT equals success
Moser: That’s
a great question, Dana, because one of the things that has been a challenge
with automation of chemical plants is these two separate towers. You had OT
very much separate from IT.
The key contributor to the
success of this digitization project is the capability to reboot those two domains
together successfully.
Gardner: Stan,
as part of that partnership, tell us about CBT and how you fit.
Galanski: CBT
is a 17-year-old, privately owned company. We cut our teeth early on by fulfilling
high-tech procurement orders for the aerospace industry. During that period, we
developed a strength for designing, testing, and installing compute and storage
systems for those industries and vendors.
It evolved into developing an
expertise in high-performance
computing (HPC), software design platforms, and so forth.
About three years ago, we
recognized the onset of faster computational platforms and massive amounts of
data -- and the capability for software to control that dataflow -- was
changing the landscape. Now, somebody needed to analyze that data faster over
multiple mediums. Hence, we developed a practice around comprehensive data
management and combined that with our field experience. That led us to become a
systems integrator (SI), which is what we’ve been assigned to for this refinery
of the future.
Gardner: Linda,
before we explore more on what you’ve done and how it improves things, let’s
learn about Texmark. With a large refinery operation, any downtime can be a big
problem. Tell us about the company and what you are doing to improve your
operations and physical infrastructure.
 |
| Salinas |
Salinas: Texmark
is a family-owned company, founded in 1970 by David Smith.
And we do have a unique set of challenges. We sit on eight acres in Galena Park, and we are surrounded
by a bulk liquid terminal facility.
So, as you can imagine, a
plant that was built in the 1940s has older infrastructure. The layout is
probably not as efficient as it could be. In the 1940s, we didn’t have a need
for wastewater treatment. Things may not have been laid out in the most
efficient ways, and so we have added these things over the years. So, one, we
are landlocked, and, two, things may not be sited in the most optimal way.
For example, we have several control
rooms sprinkled throughout the facility. But we have learned that siting is an
important issue. So we’ve had to move our control room to the outskirt of the
process areas.
As a result, we’ve had to reroute
our control systems. We have to work with what we have, and that presents some infrastructure
challenges.
Also, like other chemical
plants and refineries, the things we handle are hazardous. They are flammable, toxic,
and they are not things people want to have in the air that they breath in
neighborhoods just a quarter-mile downwind of us.
So we have to be mindful of
safe handling of those chemicals. We also have to be mindful that we don’t
disrupt our processes. Finding the time to shut down to install and deploy new
technology, is a challenge. Chemical plants and refineries need to find the
right time to shut down and perform maintenance with a very defined scope, and on
a budget.
And so that capability to come
up and down effectively is a strength for Texmark because we are a smaller facility
and so are able to come up and down and deploy and test and prove out some of
these technologies.
Gardner: Stan,
in working with Linda, you are not just trying to gain incremental improvement.
You are trying to define the next definition, if you will, of a safe, efficient,
and operationally intelligent refinery.
How are you able to leapfrog
to that next state, rather than take baby steps, to attain an optimized refinery?
Challenges of change
Galanski: First
we sat down with the customer and asked what the key functions and challenges
they had in their operations. Once they gave us that list, we then looked at
the landscape of technologies and the available categories of information that
we had at our disposal and said, “How can we combine this to have a significant
improvement and impact on your business?”
We came up with five solutions
that we targeted and started working on in parallel. They have proven to be a
handful of challenges -- especially working in a plant that’s continuously operational.
The
connected worker solution is garnering a lot of attention in the
marketplace. With it, we are able to bring real-time data from the core
repositories of the company to the hands of the workers in the field.
Based on the feedback we’ve received
from their personnel; we feel we are on the right track. As part of that, we
are attacking predictive maintenance and analytics by sensoring some of their
assets, their pumps. We are putting video analytics in place by capturing video
scenes of various portions of the plant that are very restrictive but still need
to have careful monitoring. We are looking at worker safety and security by
capturing biometrics and geo-referencing the location of workers so we know they
are safe or if they might be in danger.
The connected
worker solution is garnering a lot of attention in the marketplace. With it,
we are able to bring real-time data from the core repositories of the company
to the hands of the workers in the field. Oftentimes it comes to them in a
hands-free condition where the worker has wearables on his body that project
and display the information without them having to hold a device.
Lastly, we are tying this all
together with an asset management system that tracks every asset and ties them to
every unstructured data file that has been recorded or captured. In doing so, we
are able to put the plant together and combine it with a 3D model to keep track
of every asset and make that useful for workers at any level of responsibility.
Gardner: It’s
impressive, how this touches just about every aspect of what you’re doing.
Peter, tell us about the
foundational technologies that accommodate what Stan has just described and also
help overcome the challenges Linda described.
Foundation of the future refinery
 |
| Moser |
Moser: Before
I describe what the foundation consists of, it’s important to explain what led
to the foundation in the first place. At Texmark, we wanted to sit down and
think big. You go through the art of the possible, because most of us don’t
know what we don’t know, right?
You bring in a cross-section
of people from the plant and ask, “If you could do anything what would you do?
And why would you do it?” You have that conversation first and it gives you a
spectrum of possibilities, and then you prioritize that. Those prioritizations
help you shape what the foundation should look like to satisfy all those needs.
That’s what led to the foundational
technology platform that we have at Texmark. We look at the spectrum of use
cases that Stan described and say, “Okay, now what’s necessary to support that
spectrum of use cases?”
But we didn’t start by looking
at use cases. We started first by looking at what we wanted to achieve as an
overall business outcome. That led us to say, “First thing we do is build out pervasive
connectivity.” That has to come first because if things can’t give you data,
and you can’t capture that data, then you’re already at a deficit.
Then, once you can capture
that data using pervasive Wi-Fi with HPE Aruba, you need
a data center-class compute platform that’s able to deliver satisfactory
computational capabilities and support, accelerators, and other things
necessary to deliver the outcomes you are looking for.
The third thing you have to ask is, “Okay, where am I going to put all of this computing storage into?” So you need a localized storage environment that’s controlled and secure. That’s where we came up with the edge data center. It was those drivers that led to the foundation from which we are building out support for all of those use cases.
Gardner:
Linda, what are you seeing from this marriage of modernized OT and IT and
taking advantage of edge computing? Do you have an ability yet to measure and
describe the business outcome benefits?
Hands-free data at your fingertips
Salinas: This has
been the perfect project for us to embark on our IT-OT journey with HPE and CBT,
and all of our ecosystem partners. Number one, we’ve been having fun.
Two, we have been learning
about what is possible and what this technology can do for us. When we visited
the HPE
Innovation Lab, we saw very quickly the application of IT and OT across other
industries. But when we saw the sensored pump, that was our “aha moment.”
That’s when we learned what IoT and its impact meant to Texmark.
As for key performance
indicators (KPIs), we gather data and we learn more about how we can employ IoT
across our business. What does that mean? That means moving away from the
clipboard and spreadsheet toward having the data wherever we need it -- having it
available at our fingertips, having the data do analytics for us, and telling
us, “Okay, this is where you need to focus during your next precious turnaround
time.”
The other thing is, this IoT
project is helping us attract and retain talent. Right now it's a very
competitive market. We just hired a couple of new operators, and I truly
believe that the tipping point for them was that they had seen and heard about
our IoT project and the “refinery of the future” goal. They found out about it
when they Googled us prior to their interview.
We just hired a new
maintenance manager who has a lot of IoT experience from other plants, and that
new hire was intrigued by our “refinery of the future” project.
Finally, our modernization
work is bringing in new business for Texmark. It's putting us on the map with
other pioneers in the industry who are dipping their toe into the water of IoT.
We are getting national and international recognition from other chemical
plants and refineries that are looking to also do toll processing.
They are now seeking us out
because of the competitive edge we can offer them, and for the additional data and
automated processes that that brings to us. They want the capability to see
real-time data, and have it do analytics for them. They want to be able to
experiment in the IoT arena, too, but without having to do it necessarily
inside their own perimeter.
Gardner:
Linda, please explain what toll processing is and why it's a key opportunity
for improvement?
Collaboration creates confidence
Salinas:
Texmark produces dicyclopentadiene,
butyl alcohol, propyl
alcohol, and some aromatic solvents. But alongside the usual products we
produce and sell, we also provide “toll processing services.” The analogy I
like to tell my friends is, “We have the blender, and our customers bring the
lime and tequila. The we make their margaritas for them.”
So our customers will bring to
us their raw materials. They bring the process conditions, such as the
temperatures, pressures, flows, and throughput. Then they say, “This is my
material, this is my process. Will you run it in your equipment on our behalf?”
When
we are able to add the IoT component to toll processing, when we are
able to provide them data that they didn't have whenever they ran their
own processes, that provides us a competitive edge over other toll
processors.
When we are able to add the
IoT component to toll processing, when we are able to provide them data that
they didn't have whenever they ran their own processes, that provides us a competitive
edge over other toll processors.
Gardner: And,
of course, your optimization benefits can go right to the bottom line, so a
very big business benefit when you learn quickly as you go.
Stan, tell us about the
cultural collaboration element, both from the ecosystem provider team support
side as well as getting people inside of a customer like Texmark to perhaps
think differently and behave differently than they had in the past.
Galanski: It’s
all about human behavior. If you are going to make progress in anything of this
nature, you are going to have to understand the guy sitting across the table
from you, or the person out in the plant who is working in some fairly
challenging environments. Also, the folks sitting at the control room table with
a lot of responsibility for managing the processes with lots of chemicals for
many hours at a time.
So we sat down with them. We
got introduced to them. We explained to them our credentials. We asked them to
tell us about their job. We got to know them as people; they got to know us as
people.
We established trust, and then
we started saying, “We are here to help.” They started telling us their problems,
asking, “Can you help me do this?” And we took some time, came up with some
ideas, and came back and socialized those ideas with them. Then we started
attacking the problem in little chunks of accomplishments.
We would say, “Well, what if
we do this in the next two weeks and show you how this can be an asset for you?”
And they said, “Great.” They liked the fact that there was quick turnaround
time, that they could see responsivity. We got some feedback from them. We
developed a little more confidence and trust between each other, and then more
things started out-pouring a little at a time. We went from one department to
another and pretty soon we began understanding and learning about all aspects
of this chemical plant.
It didn’t happen overnight. It
meant we had to be patient, because it’s an ongoing operation. We couldn't
inject ourselves unnaturally. We had to be patient and take it in increments so
we could actually demonstrate success.
And over time you sometimes
can't tell the difference between us and some of their workers because we all
come to meetings together. We talk, we collaborate, and we are one team -- and
that’s how it worked.
Gardner: On
the level of digital transformation -- when you look at the bigger picture, the
strategic picture -- how far along are they at Texmark? What would be some of
the next steps?
All systems go digital
Galanski: They
are now very far along in digital transformation. As I outlined earlier, they
are utilizing quite a few technologies that are available -- and not leaving
too many on the table.
So we have edge computing. We
have very strong ubiquitous communication networks. We have software analytics
able to analyze the data. They are using very advanced asset integrity
applications to be able to determine where every piece, part, and element of
the plant is located and how it’s functioning.
I have seen other companies
where they have tried to take this only one chapter at a time, and they
sometimes have multiple departments working on these independently. They are
not necessarily ready to integrate or to scale it across the company.
But Texmark has taken a corporate approach, looking at holistic operations. All of their departments understand what’s going on in a systematic way. I believe they are ready to scale more easily than other companies once we get past this first phase.
Gardner:
Linda, any thoughts about where you are and what that has set you up to go to
next in terms of that holistic approach?
Salinas: I
agree with Stan. From an operational standpoint, now that we have some sensored
pumps for predictive analytics, we might sensor all of the pumps associated
with any process, rather than just a single pump within that process.
That would mean in our next
phase that we sensor another six or seven pumps, either for toll processing or
our production processes. We won’t just do analytics on the single pump and its
health, lifecycle, and when it needs to be repaired. Instead we look at the
entire process and think, “Okay, not only will I need to take this one pump
down for repair, but instead there are two or three that might need some
service or maintenance in the next nine months. But the fuller analytics can tell
me that if I can wait 12 months, then I can do them all at the same time and
bring down the process and have a more efficient use of our downtime.”
I could see something like
that happening.
Galanski: We
have already seen growth in this area where the workers have seen us provide
real-time data to them on hands-free mobile and wearable devices. They say, “Well,
could you give me historical data over the past hour, week, or month? That
would help me determine whether I have an immediate problem, not just one spike
piece of information?”
So they have given us
immediate feedback on that and that's progressing.
Gardner:
Peter, we are hearing about a more granular approach to sensors at Texmark, with
the IoT edge getting richer. That means more data being created, and more
historical analysis of that data.
Are you therefore setting
yourself up to be able to take advantage of things such as AI, ML, and the
advanced automation and analytics that go hand in hand? Where can it go next in
terms of applying intelligence in new ways?
Deep learning from abundant data
Moser:
That’s a great question because the data growth is exponential. As more sensors
are added, videos incorporated into their workflows, and they connect more of
the workers and employees at Texmark their data and data traffic needs are
going to grow exponentially.
But with that comes an
opportunity. One is to better manage the data so they get value from it,
because the data is not all the same or it’s not all created equal. So the
opportunity there is around better data management, to get value from the data
at its peak, and then manage that data cost effectively.
That massive amount of data is
also going to allow us to better train the current models and create new ones.
The more data you have, the better you can do ML and potentially deep learning.
Lastly, we need to think about
new insights that we can’t create today. That's going to give us the greatest
opportunity, when we take the data we have today and use it in new and creative
ways to give us better insights, to make better decisions, and to increase
health and safety. Now we can take all of the data from the sensors and videos
and cross-correlate that with weather data, for example, and other types of
data, such as supply chain data, and incorporate that into enabling and empowering
the salespeople, to negotiate better contracts, et cetera.
So, again, the art of the
possible starts to manifest itself as we get more and more data from more and
more sources. I’m very excited about it.
Gardner: What advice
do you have for those just beginning similar IoT projects?
Galanski: I
recommend that they have somebody lead the group. You can try and flip through
the catalogs and find the best vendors who have the best widgets and start
talking to them and bring them on board. But that's not necessarily going to
get you to an end game. You are going to have to step back, understand your
customer, and come up with a holistic approach of how to assign
responsibilities and specific tasks, and get that organized and scheduled.
There are a lot of parties and a lot of pieces on this chess table. Keeping them all moving in the right direction and at a cadence that people can handle is important. And I think having one contractor, or a department head in charge, is quite valuable.
Salinas: You
should rent a party bus. And what I mean by that is when we first began our
journey, actually our first lecture, our first step onto the learning curve
about IoT, was when Texmark rented a party bus and put about 13 employees on it
and we took a field trip to the HPE Innovation Lab.
When Doug
Smith, our CEO, and I were invited to visit that lab we decided to bring a
handful of employees to go see what this IoT thing was all about. That was the
best thing we ever could have done, because the excitement was built from the
beginning.
They
saw, as we saw, the art of the possible at the HPE IoT lab, and the
ride home on that bus was exciting. They had ideas. They didn't even
know where to begin. The buy-in was there from the beginning.
They saw, as we say, the art
of the possible at the HPE IoT lab, and the ride home on that bus was exciting.
They had ideas. They didn’t even know where to begin, but they had ideas just
from what they had seen and learned in a two-hour tour about what we could do
at Texmark right away. So the engagement, the buy-in was there from the
beginning, and I have to say that was probably one of the best moves we have
made to ensure the success of this project.
Listen to the podcast. Find it on iTunes. Read a full transcript or download a copy. Sponsor: Hewlett Packard Enterprise.
You may also be
interested in:
- How the composable approach to IT aligns automation and intelligence to overcome mounting complexity
- How HPC supports 'continuous integration of new ideas' for optimizing Formula 1 car design
- Advanced IoT systems provide analysis catalyst for the petrochemical refinery of the future
- Want to manage your total cloud costs better? Emphasize the ‘Ops’ in DevOps, says Futurum analyst Daniel Newman
- A new Mastercard global payments model creates a template for an agile, secure, and compliant hybrid cloud
- Where the rubber meets the road: How users see the IT4IT standard building competitive business advantage
- IT kit sustainability: A business advantage and balm for the planet
- Industrial-strength wearables combine with collaboration cloud to bring anywhere expertise to intelligent-edge work
- How the data science profession is growing in value and impact across the business world
