OCTOBER 2024

OFFICIAL PUBLICATION OF THE INTERNATIONAL SOCIETY OF AUTOMATION www.isa.org/intech

Total Automation: The Next

Frontier Is Sensor-to-Cloud

A Roadmap for Improved

Simulation Success

How to Harness Applied AI in

Industrial Manufacturing

Document Projects with the

Updated ISA5.1 Standard

Closed-Loop Control
Fundamentals
 OCTOBER 2024
DIGITAL MAGAZINE Advertisers Index
To obtain further information, please contact
the advertiser using the contact information
contained in their ads.
WELCOME
Wireless communications and remote asset management are the
themes of this issue of InTech digital magazine, the official publi-
Page..............................................................................15
cation of the International Society of Automation (ISA). Written
for engineers, managers, and other automation decision-makers,
InTech serves ISA members and the wider automation community
with practical, in-depth coverage of automation technologies, Page..............................................................................11
applications, and strategies that help automation professionals
succeed.

InTech is brought to you with the support of advertisers. Click

Page................................................................................ 3
the logos on the right to go to their ads, then click through to
their websites to discover information on automation hardware,
software and services.

Don’t Miss a Single Issue

Page................................................................................ 6
InTech is part of a family of ISA publications that keep you
informed and up-to-date on industrial automation, control
and security best practices, trends, new products and other
advances. Subscribe to InTech digital magazine, InTech Plus Page................................................................................ 4
newsletters and other resources through ISA’s automation news
and information subsidiary, Automation.com. To see bonus ISA
and InTech content, visit https://www.automation.com/en-us/
Page..............................................................................21
intech-international-society-of-automation.

Rick Zabel, Managing Director

InTech, Automation.com & Events Sponsorships
International Society of Automation (ISA)
Page................................................................................ 9

Page..............................................................................14
International Society of Automation
Setting the Standard for Automation™

company/internationalsocietyofautomation

InternationalSocietyOfAutomation
Page................................................................................ 8
ISA_Interchange

Trademarks used in this document are the property of their respective owners.
When we named our industrial application software “Ignition” fifteen years ago, we had
no idea just how fitting the name would become...

DIGITAL TRANSFORMATION
Ignition’s industry-leading technology,
unlimited licensing model, and army of
certified integration partners have ignited
a SCADA revolution that has many of
the world’s biggest industrial companies
transforming their enterprises from the
plant floor up.

With plant-floor-proven operational

technology, the ability to build a unified
namespace, and the power to run on-prem,
in the cloud, or both, Ignition is the platform
for unlimited digital transformation.

One Platform, Unlimited Possibilities

Visit inductiveautomation.com/ignition to learn more.

TRUSTED BY
 Future-Ready
Industrial Wi-Fi
Empower industrial applications with wireless connectivity, no matter
how challenging the environment or how rugged the requirements are.
Deploy Moxa industrial wireless solutions for tough situations that need
a fast, reliable, and secure network experience now and tomorrow.

Scan or Click to
Learn More

AWK-1165C AWK-1161A
802.11ax Wireless Client
Compact Industrial
Wi-Fi 6 AP/Client

Connect with Us
+1-888-MOXA-USA info.us@moxa.com
+1-714-528-6777 www.moxa.com
OCTOBER 2024 | VOL 71, ISSUE 4

FEATURES
SMART MANUFACTURING

16 Total Automation:
The Next Frontier Is
Sensor-to-Cloud
By Jack Smith
Building an all-inclusive approach to
the entire process automation strategy
of a manufacturing enterprise.

STANDARDS

22 Document Projects
Consistently with
the Updated ISA5.1
Standard
By Jim Federlein, PE
A new release of this widely used ISA
standard adds content and improves
readability.

DIGITAL TWINS

26 A Roadmap for
Improved Simulation ARTIFICIAL INTELLIGENCE

Success 32 How to Harness

By Dustin Beebe Applied AI in Industrial
The most effective simulation
solutions for process manufacturing
Manufacturing
By Michael J. Anthony, Jon A. Mills, and
are designed using integrated
David C. Mazur, Ph.D.
software to deliver value across the
Properly understand key concepts and
entire lifecycle.
use necessary contextual data to train
AI models.

INTECH OCTOBER 2024 5 WWW.ISA.ORG/INTECH

Blocking Out The Noise
So Your Safety and Control Systems Run Efficiently.
Avoid Downtime With The Right Signal Isolators, Converters and Splitters.
Call 800-999-2900 or visit www.miinet.com/signal-intech
 OFFICIAL PUBLICATION OF THE INTERNATIONAL SOCIETY OF AUTOMATION

EDITORIAL & PRODUCTION

CHIEF EDITOR: Renee Bassett, rbassett@isa.org
OCTOBER 2024 | VOL 71, ISSUE 4
SENIOR CONTRIBUTING EDITOR: Jack Smith, jsmith@isa.org
SENIOR CONTENT EDITOR, AUTOMATION.COM:

DEPARTMENTS Melissa Landon, mlandon@isa.org

EDITOR EMERITUS: Bill Lydon, blydon@isa.org
STANDARDS ADMINISTRATION DIRECTOR:
10 Talk to Me Charley Robinson, crobinson@isa.org
By Renee Bassett
ART DIRECTOR: Bonnie Walker
You Have Questions? ISA Has Answers
DIGITAL DESIGNER: Colleen Casper
12 IIoT Insights
By Guy Yehiav
ISA EXECUTIVE BOARD
How Remote Monitoring and IoT Devices Optimize ISA PRESIDENT: Prabhu Soundarrajan
Supply Chains ISA PAST PRESIDENT: Marty Bince
ISA PRESIDENT-ELECT & SECRETARY: Scott Reynolds
38 The Last Word ISA TREASURER: Ardis Bartle
By Jack Smith ISA EXECUTIVE DIRECTOR: Claire Fallon
Closed-Loop Control Fundamentals

ADVERTISING & SPONSORSHIP

Rick Zabel, PUBLISHER
rzabel@isa.org
Chris Nelson, ACCOUNT EXECUTIVE
chris@isa.org
Richard T. Simpson, ACCOUNT EXECUTIVE
richard@isa.org
Gina DiFrancesco, ACCOUNT EXECUTIVE
gina@isa.org
Cathi Merritt, ADVERTISING PRODUCT MANAGER
cmerritt@isa.org
Matt Davis, DIGITAL MEDIA PROJECT MANAGER
mdavis@isa.org

SM
2025
Media Planner

To order reprints of InTech print or digital articles, contact

reprints@mossbergco.com or 800-428-3340.

©2024 International Society of Automation (ISA) ISSN 0192-303X

Editorial and advertising offices are at 3252 S. Miami Boulevard, Suite 102, Durham, NC 27703; phone 919-549-8411;
email info@isa.org.

InTech digital magazine is published 4x per year: February, April, October, October. ISA Members receive InTech digital magazine as
part of their annual membership and get access to archived issues. Non-members can subscribe to InTech and InTech Plus newsletters
through ISA’s automation news and information subsidiary, Automation.com. InTech and the ISA logo are registered trademarks of ISA.

Trademarks used in this digital magazine are the property of their respective owners. Opinions expressed or implied are those of
the persons or organizations contributing the information and are not to be construed as those of ISA.

INTECH OCTOBER 2024 7 WWW.ISA.ORG/INTECH

We bring color into view!
Compact pressure sensors and switches with 360° custom-color
status display

256 colors
Individually selectable:

Measurement in progress
Sensor switching
Process malfunction

Compact design
15 cm

Hygienic
adapter system

Adjustment via
smartphone

$518
VEGABAR 39 Clamp 1"

www.vega.com/vegabar
 YOUR TICKET TO
COLLISION PROTECTION
Win a 3D safety sensor!

Register by 11/3/24
With a 3D time-of-flight
camera, you gain new perspectives.
Like 3D environment perception that
detects objects above the protective field
of a safety laser scanner. Enter to win the
safeVisionary2 from SICK, and your vehicles will be
one step closer to increased safety and availability.

SICK, Inc | Minneapolis, MN | USA | www.sick.com

 TALK TO ME | PERSPECTIVES FROM THE EDITOR

You Have Questions? ISA Has Answers

By Renee Bassett, InTech Chief Editor

Serving up member-generat- he’s met within ISA, McMillan answers ques-

ed technical content related tions like, “How can we compute startup
to standards, certifications, tuning for new control valves?” or “How to
training, control, safety, best migrate from obsolete to modern instru-
cybersecurity, systems inte- mentation and control systems?”
gration and other key topics is something the The newest way to access ISA expertise is
International Society of Automation has been MimoSM, an AI-powered large-language model
doing since the society began. Whether you educated on ISA content including back
need technical explanations or professional issues of InTech, standards, training materials,
development advice, ISA has answers. technical reports and more. Mimo has learned
ISA’s InTech magazine has a long, proud about industrial automation and operational
history of providing in-depth technical infor- technology cybersecurity from studying years
mation to instrumentation, automation and of ISA content available on Pub Hub,
control professionals. ISA’s Instrumentation the ISA member-only content
Technology magazine—launched as the ISA portal. This means
Journal in 1954 and renamed InTech in 1978— Mimo can access
was expanded when ISA bought Programmable files and informa-
Controls magazine in 1989. In subsequent tion not normally
years, ISA expanded its content offerings available widely or
further by launching its web presence at www. publicly.
isa.org (1994) and purchasing Automation.com Mimo also provides sources for its answers,
(2015) for industry news and new products, so users can go straight to the article, standard
newsletters, ebooks and other resources. or report from which Mimo has drawn its

The ISA Interchange blog has been around answer. This makes Mimo a richer and more

for more than 15 years. Upgraded in 2019, relevant resource for automation professionals

it has hosted a wide range of subject matter than ChatGPT or similar services. If you are

experts over the years. One of its longest-run- not an ISA member, or if you aren’t logged in,

ning components is the “Ask the Automation you’ll get a short and to-the-point answer. For

Pros” series. Monthly, ISA Fellow and 2010 more details and source files, log in to your

ISA Life Achievement Award recipient Greg member account or join ISA.

McMillan collects submitted questions and As an automation professional, you have

solicits responses from automation profession- questions. ISA has the people, the publica-
als. Using over 50 years of process automation tions and the resources to get your questions
experience and calling on the professionals answered.

INTECH OCTOBER 2024 10 WWW.ISA.ORG/INTECH

| AT20USA |

Your competition won’t

be able to handle this.
Get way ahead with adaptive automation

Levitating 6D
planar movers motion

Scalable Velocities
kg payload up to 2 m/s

Individual Free system

product transport layout

www.beckhoff.com/xplanar
Elevate your automation game with XPlanar. Developed in-house by Beckhoff, this mechatronic
system makes your material handling literally fly with magnetically levitating movers and base tiles
that you can arrange to fit your unique applications – automated warehouses, distribution centers,
microfulfillment, pharmaceutical fulfillment, material handling in manufacturing and beyond.
So, what will you do with XPlanar?
Movement and handling with up to 6 degrees of freedom
Transport and processing in one system
Wear-free, hygienic and easy to clean
Create nearly any layout and path you can imagine
Multi-mover control enables parallel and individual product handling
Fully integrated into the powerful and standardized Beckhoff control system
(TwinCAT, IEC 61131, Motion, Measurement, Machine Learning, Vision, Communication, HMI)
Flexibly change paths in software and easily modify shape, size, and number of movers
 IIOT INSIGHTS | WHERE THE INTERNET MEETS INDUSTRY

How Remote Monitoring and IoT Devices

Optimize Supply Chains
By Guy Yehiav

Several asset monitoring and inventory man- IoT-powered devices, sensors and cameras
agement challenges impact an enterprise’s throughout their supply chain which enables
supply chain efficiency. Collecting real-time remote and automated monitoring of various
data on asset location and condition on a asset conditions in real time. In previous
consistent basis is difficult without the proper generations of IoT, supply chain managers
technology infrastructure in place. Without would then analyze the data collected from
timely guidance and accurate intelligence, their IoT deployments to laboriously uncover
informed decision-making for successful actionable insights. However, now with the
supply chain operations is almost impossible. use of AI-driven analysis and leveraging
Optimizing asset and inventory management of ML models, organizations are unlocking
is crucial for reducing errors in inventory trends and patterns in inventory usage that
levels and avoiding out-of-stock and over- automagically enhance demand forecasting
stock situations. accuracy, reduce excess stock and minimize
Scalability issues pose another common out-of-stocks.
challenge. As inventories spanning multiple IoT sensing capabilities collect critical
locations, countries and continents become
data on the location and condition of an as-
larger and more globalized, the technology
set during transportation and storage while
platforms used to manage the value chain
reporting other variables such as light,
must scale accordingly. Unsurprisingly,
impact, temperature, humidity, vibration,
research from SAP found that 52% of busi-
etc. If an asset isn’t where it’s supposed to
ness leaders believe their supply chain needs
be or if any of these defined parameters
improvement. To overcome asset and inven-
exceed or dip below established thresholds,
tory challenges and bolster supply chains,
the IoT system will automatically notify the
companies must leverage advanced technolo-
pre-determined employee at the precise
gies and best security practices including the
time, empowering them to respond quickly
internet of things (IoT), integrated workflows
with corrective action before an issue
and real-time prescriptive analytics powered
results in inventory loss.
by artificial intelligence (AI) and machine
learning (ML). For example, an edge-to-cloud IoT solution
using fever tags for cattle monitoring tracks
Producing real-time intelligence the health of thousands of individual cattle
Enterprises optimize their asset and inven- which helps ranchers to identify, isolate and
tory management by strategically deploying treat sick cows before they infect the herd.

INTECH OCTOBER 2024 12 WWW.ISA.ORG/INTECH

 IIOT INSIGHTS | WHERE THE INTERNET MEETS INDUSTRY

Another example is having condition-sensitive

Best IoT security practices
food or medication products being shipped
To take advantage of the incredible value of IoT
across the country. IoT tags equipped with
technologies, protection against cyber-attacks is
multiple sensing capabilities are now able to
paramount as companies minimize vulnerabili-
empower retailers to combat theft and inter-
ties and safeguard overall network infrastructure
vene if refrigeration equipment fails along the
and data. IoT security best practices include:
route.
z Implementing secure access controls, such
The growing popularity of IoT adoption
as role-based access control and multi-
demonstrates that businesses recognize its
factor authentication.
critical role in optimizing asset and inventory z Using proper encryption to protect data
management. In fact, the global market worth
during transit and at rest.
of IoT in Supply Chain Management (SCM) will z Disabling unused services and changing
reach $41.8 billion by 2033. The real chal-
default settings, i.e., updating default
lenge is therefore not deciding to use IoT but
passwords, usernames and configurations.
figuring out how to effectively integrate these z Segmenting networks to separate edge IoT
IoT devices, sensors and cameras with existing
devices from critical internal networks; in
asset and inventory management systems at
the event of a breach, network segmenta-
scale.
tion will limit the blast radius.
For an enterprise to build a sophisticated z Leveraging cellular technology instead of
IoT solution on its own would require robust Wi-Fi for connectivity whenever possible.
network infrastructure and advanced console z Educating employees on best practices
servers that follow evolving standard commu- and suspicious links/phishing emails to
nication protocols. Integration with existing promote security across the supply chain.
IT infrastructure, enterprise resource planning z Performing regular firmware patches and
and inventory management systems is also a security updates that allow IoT devices to
heady task without the proper expertise and remain secure over their lifecycle.
commitment to ongoing support. Thankfully, Regular security practices like these are
there are flexible IoT solutions that remove only possible with integrated solutions that
the burden from the enterprise as they deploy support remote access and management to all
at scale. deployed IoT hardware.

ABOUT THE AUTHOR

Guy Yehiav is president of SmartSense by Digi. SmartSense was created to use
the power of the Internet of Things to help protect the assets most critical to their
customers. Over his 25-year career, Yehiav has built world-class technology companies
and prior to SmartSense by Digi, he served as general manager and vice president of
Zebra Technologies’ Zebra Analytics business unit.

INTECH OCTOBER 2024 13 WWW.ISA.ORG/INTECH

 No place to PROVEN

replace a battery. 40
YEAR
OPERATING

LIFE *

Highly remote locations call for Tadiran batteries.

Battery replacement is costly and often ANNUAL SELF-DISCHARGE
dangerous work. Reduce the risk with Tadiran
TADIRAN COMPETITORS
bobbin-type lithium thionyl chloride (LiSOCl2)
batteries. With an annual self-discharge rate of
just 0.7% per year, Tadiran LiSOCl2 batteries
enable low power consuming wireless devices to
operate for up to 40 years on a single battery, 0.7%
up to 4 times longer than the competition.
Our batteries also feature the highest capacity,
highest energy density, and widest temperature
range of any lithium cell, plus a glass-to-
metal hermetic seal for added ruggedness and
reliability in extreme environments.
Up to 3%
Tadiran Batteries
Take no chances. Take Tadiran batteries
2001 Marcus Ave.
that last a lifetime. Suite 125E
Lake Success,
NY 11042
1-800-537-1368
516-621-4980
* Tadiran LiSOCL2 batteries feature the lowest annual self-discharge rate of any competitive battery, less than 1% per year, enabling these
batteries to operate over 40 years depending on device operating usage. However, this is not an expressed or implied warranty, as each
application differs in terms of annual energy consumption and/or operating environment. www.tadiranbat.com
 It’s time to say “YES!” to LS PLCs
and get the advanced control you need for less!
FREE IEC
IEC-based
software

EtherCAT®
positioning

Advanced
motion
control

Pulse
positioning

PID PID Σ
Built-in & IN OUT
expansion
With the LS Electric XGB PLC by AutomationDirect, I/O
you can say “YES!” to...
High-
 Advanced motion control with up to 400 moves per axis and various single- and multi-axis speed HS I/O
operations including position and speed control, helical interpolation, and ellipse interpolation I/O
 FREE IEC programming with structured text, ladder logic, and more
 Multiple networking options including EtherNet/IP, Modbus TCP and RTU, and ASCII
USB
 EtherCAT support for precise, real-time motion control
 High-density I/O (up to 32 points embedded on CPU and select I/O expansion modules)
 High-speed I/O (integrated 2- or 6-axis pulse/direction motion inputs/outputs (up to 200kHz)
on the CPU module, and up to 4 channels of high-speed inputs at up to 500kHz on each Ethernet
high-speed counter module)
 FREE award-winning service and technical support
NEW! LS Electric 2-axis and Relay Output XGB PLC Models RS-485
Three new LS Electric XGB PLC models have been released RS-232
providing advanced control at an even lower cost.
• Two new 16-point 2-axis PLC versions support pulse/direction control
from the embedded I/O, featuring 4 high-speed inputs and 2 high-speed CPU with
outputs, as well as EtherCAT motion using EtherCAT expansion modules. 2-axis control, only:

$279.00
• New relay output PLC provides 8 inputs, 4 of which offer high-speed
capability, and 6 relay outputs supporting up to 2A per point. This PLC
version can also support EtherCAT motion using the optional EtherCAT
expansion modules.

Research, price, buy at: www.automationdirect.com/ls-electric-plc

Order Today, Ships Fast!

* See our Web site for details and restrictions. © Copyright 2022 AutomationDirect, Cumming, GA USA. All rights reserved. 1-800-633-0405 the #1 value in automation
SMART MANUFACTURING

Total Automation:
The Next Frontier Is Sensor to Cloud
Users need guidance to deploy sensors
appropriately. A new standard may be the answer.

The dictionary defines automation as “the By Jack Smith

technique of making an apparatus, a process,
or a system operate automatically.” ISA, the
International Society of Automation, defines to monitor and control the production and
automation as “the creation and application delivery of products and services.”
of technology to monitor and control the
During an ISA meeting where automa-
production and delivery of products and
tion concepts were being discussed,
services.”
Dennis Brandl, chief consultant at BR&L
Using ISA’s definition, the automation Consulting recommended that the term
profession includes “everyone involved in “total automation” be used to differentiate
the creation and application of technology it from Hyperautomation. Brandl formally
to monitor and control the production and presented total automation at COPERMAN
delivery of products and services.” An auto- 2023. The Conference on Performance and
mation professional is “any individual involved Management (COPERMAN) aims to bring
in the creation and application of technology together researchers and practitioners to

INTECH OCTOBER 2024 16 WWW.ISA.ORG/INTECH

SMART MANUFACTURING

present and discuss innovative contributions “The objective of total automation is to

concerning the measurement and manage- completely automate the processes in an
ment of organizational performance in a operational facility to increase efficiency and
modern business environment. productivity and reduce errors,” said Steve
Mustard, president and CEO at National
Total automation is an important aspect
Automation Inc. and former (2021) ISA presi-
of digital transformation because it serves
dent. “Even with advances in technology over
to use information technology/operational
the past few decades, many organizations
technology (IT/OT) to improve performance
continue to operate manual or semi-manual
for dangerous, dirty, demanding, delicate, and
processes. Examples include the manual col-
dull tasks (the five Ds of manufacturing).
lection of sensor readings, manual data entry,
and manual analysis of data.
The importance of total
automation “The concept is important now as organi-
According to Brandl, total automation is “a zations seek to squeeze out every last drop
disciplined and all-inclusive approach to the of efficiency from their operations so they
entire process automation strategy of a manu- can be competitive in the global marketplace,
facturing enterprise.” He said total automation responsive to changing customer demands,
is the next step beyond Hyperautomation, and be resilient to inevitable supply chain
which he explained as an IT initiative to disruptions.”
increase the automation of business processes
Mustard described the automation/
(production chains, workflows, marketing
manufacturing timeline. The first industrial
processes, etc.) by introducing artificial intel-
revolution of the 1800s transitioned processes
ligence (AI), machine learning (ML), and robotic
in labor-intensive industries such as mining
process automation (RPA). “Total automation
and textiles. The second industrial revolution
applies the concepts of automation to all ele-
in the 1900s introduced the internal combus-
ments of a company including OT and IT. The
tion engine and electrification, enabling mass
goal is to reduce the human errors that crop
production. The third industrial revolution saw
up in manual processes, and to use comput-
the rise of computers and telecommunications
ing resources to verify and validate business
enabling greater automation and digitalization.
operations,” he said.
“The fourth industrial revolution, or Industry
Brandl said that total automation allows for 4.0,” explained Mustard, “builds on these
performance management of all activities in a advances and seeks to reshape how industries
manufacturing enterprise. It is the combination operate through the use of disruptive technol-
of IT Hyperautomation, process automation, ogies such as AI, big data, and IIoT [Industrial
sensor automation, and OT task automation. Internet of Things]. Total automation leverages
Another way to look at it is by using new the disruptive technologies of Industry 4.0 to
technologies in a real-time environment. transform how organizations operate.”

INTECH OCTOBER 2024 17 WWW.ISA.ORG/INTECH

SMART MANUFACTURING

Mustard shared some examples of the use z Level 1: Use of IIoT to collect remote sen-
of these technologies to move toward total sor data; use of AI or ML to maintain cali-
automation: bration and report on sensor discrepancies.
z Level 2: Use of AI in expert systems sup-
z Using ML to automatically analyze images
porting operator decision making.
to detect corrosion or other defects to
z Level 3: Use of AI to optimize production
reduce the time and effort involved in
schedules and analyze machinery health.
manual analysis and improve accuracy and
z Level 4: Use of big data analytics, AI,
reliability by the removal of human bias.
and cloud to automate business decision
z Using IIoT to automatically collect sensor
making.
data to remove the need for manual data
“Through all layers, the objective of total
collection.
automation is to use disruptive technology
z Using AI to analyze sensor data to look for
to streamline and automate all processes,”
patterns that are not obvious to humans to
Mustard said.
reduce unplanned downtime.
Brandl agrees that a total automation
The need for a total standard would apply to all layers of the
automation standard ISA 95 model (Figure 1). “Layer 2 is covered
The industry already has ISA95. But how
by existing ISA and automation standards.
would a total automation standard fit the
Layer 1 is partially covered by the standards
various levels of the ISA95 model?
on maintenance and security (automated
According to Mustard, total automation calibration, automated cleaning, automated
applies to all ISA95 levels: alignment, automated error detection, etc.).

Figure 1. Automation in a manufacturing enterprise—ISA95 levels. Courtesy: Dennis Brandl

INTECH OCTOBER 2024 18 WWW.ISA.ORG/INTECH

SMART MANUFACTURING

Layer 3 is mostly covered by ISA 99, 95, and mechanical gauges should be replaced by
88. The concept of interoperable distributed sensors, where submetering is required, or
workflows helps fill in some of the missing what update period to set,” according to the
pieces, in my opinion,” he said. justification document. “A standard could help

Brandl and Mustard support the proposal plants—especially process plants. It would

of a new standard, or at least a revision to, or also make ISA more relevant in the digital

expansion of an existing standard. A justifica- transformation/Industry 4.0 megatrend.”

tion was submitted for the evaluation of such There seems to be a gap in the standards
a standard that lays out some considerations. with little to guide people in identifying,

According to the justification document, selecting, and validating sensing opportuni-

users need guidance in deploying sensors ties. Some examples may repurpose the data
appropriately throughout their plants. from existing sensors, while others require
“However, there are many equipment cat- new sensors. For the former, it is critical that
egories to cover and if we try add sensors to the additional dependencies be documented.
everything, we will never finish. A good start There are not any existing standards that
would be common asset types like pumps are relevant to the use of this technology,
and heat exchangers found in all plants. More or that must be followed in its application,
equipment types and other positions could according to the justification document. “It
be included in subsequent revisions or other could be somewhat related to ASME PTC,
sections. We could start with common asset which defines equations using data because
types like pumps and heat exchangers found the proposed standard will help users get the
in all plants. More equipment types and other right data. API670 is limited to vibration. The
positions can [be included In] subsequent proposed standard would be far broader in
revisions or other sections.” scope because it would automate all manual
Whether total automation, digital transfor- measurements (automate corrosion, acoustic
mation, Industry 4.0, or IIoT, getting real-time noise [leaks], mechanical gauges, and clip-
data begins with the sensors. “Users don’t boards). API682 is limited to pump seals. The
always know what to sense, what sensors [proposed] standard would be far broader in
are required on each equipment type, what scope.”

“Total automation leverages the disruptive technologies of

Industry 4.0 to transform how organizations operate.”
—Steve Mustard, president and CEO, National Automation Inc., and a former (2021) ISA president

INTECH OCTOBER 2024 19 WWW.ISA.ORG/INTECH

SMART MANUFACTURING

“There are not really any models or other The technology that supports a proposed
architecture-related information that helps standard does not define an architecture
to understand the technology and its ap- per se. It does, however, imply a definite
plication.” “The standard would recommend increased sensor count—more sensors in
sensors—not how these sensors are architec- existing architectures. Sensors are selected,
turally connected. These are sensors ‘beyond installed, configured, and supported by
the P&ID’—not for process monitoring or instrument and control personnel, many of
control. This would be related to the NAMUR whom are members of ISA.
NE175 standard; it is for equipment perfor-
In addition, this standard will make plants
mance and condition monitoring. In addition,
more sustainable. By using the appropriate
it would also be related to sustainability like
sensors, collected data would detect and
energy management, WAGES [water, air, gas,
pinpoint energy overconsumption, emis-
electric, and steam] submetering for EMIS,
sions, and equipment inefficiency. It could
and emissions monitoring like relief valves,
monitor cleaning optimization and help
flaring, and methane. It would also support
reduce flaring. Downtime would be reduced
equipment performance monitoring. It would
due to more predictive maintenance, failure
also fit nicely in the various layers in the
prediction, and reduced loss of containment.
ISA95 model. There are not really any other
Plants will be safer because of reduced
technologies related to a proposed total
human error, and fewer manual valves and
automation standard. The standard should
leaks. Finally, automating existing manual
recommend what sensors to deploy on each
data collection will enable plants to be more
type of equipment and in other places. It
productive.
would not define sensor or signal transmis-
sion. However, most sensors will be wireless
Looking ahead
using IEC62591 or other methods.”
Brandl said the concept of a “digital com-
The technology behind a proposed total au- panion” has started in the medical field. A
tomation standard drives functionality, which digital companion provides personalized
enables how it would be applied. Application assistance. “We need a digital assistant for
areas include (but are not limited to): everyone performing manufacturing opera-
z Reliability/maintenance of rotating equip- tions management tasks, either on the shop
ment, valves, etc. floor or in the production back office. A
z Integrity (corrosion/erosion) of piping and digital assistant that looks over your shoul-
vessels. der would manage your tasks, make remind-
z Safety (including health and environment): ers, bring up relevant information, record
safety showers, manual valves, etc. completions, walk you through manual
z Production/quality would require sensors steps in processes, and collect information
in place of mechanical gauges. from equipment; it is truly mobile. We

INTECH OCTOBER 2024 20 WWW.ISA.ORG/INTECH

 SMART MANUFACTURING

already have a name for it: Manufacturing productivity effectiveness (PPE) is the human
operations management [MOM]. But it’s equivalent of overall equipment effectiveness
your personal MOM, loaded with your tasks (OEE),” he said.
and schedules,” he said.
Standards require consensus. With so
Brandl advocates performance manage- many things to gain, and nothing to lose, total
ment—measuring and improving indi- automation stands to take automated manu-
vidual processes—for all activities. “Personal facturing to the next level.

ABOUT THE AUTHOR

Jack Smith is senior contributing editor for Automation.com and InTech digital
magazine, publications of ISA, the International Society of Automation. Jack is a
senior member of ISA, as well as a member of IEEE. He has an AAS in Electrical/
Electronic Engineering and experience in instrumentation, closed loop control, PLCs,
complex automated test systems, and test system design. Jack also has more than
20 years of experience as a journalist covering process, discrete, and hybrid technologies.

SEA-240013 Sealevel Flexio Print Ad InTech Digital 8.5 x 5.5.pdf 1 4/1/24 9:34 AM

INTECH OCTOBER 2024 21 WWW.ISA.ORG/INTECH

DIGITALIZATION

Document
Projects
Consistently
with the
Updated ISA5.1
Standard

A new release of this widely

used ISA standard adds content
and improves readability.
By Jim Federlein, PE

In project documents used to specify, devices. They are concise and function as
purchase, track, install and maintain instru- a specific means of communication for all
mentation and control system components, types and kinds of technical, engineering,
consistency is important. The International procurement, construction and maintenance
Society of Automation (ISA) has long known documents. This includes identification
that and promoted such consistency through schemes and graphic symbols for drawings
its standards. Seventy-five years after it was and documentation systems used in the
first introduced, ISA has published an update construction and maintenance of industrial
of its most widely used and internation- plants, including instrumentation and control
ally recognized standard: ANSI/ISA-5.1: diagrams, loop diagrams, electrical schematics
Instrumentation and Control -Symbols and and functional and binary logic diagrams.
Identification.
A common misconception is that this
The symbols and identification methods standard is a piping and instrument diagram
set forth in the standard are intended as con- (P&ID) standard. Although it does cover the
ceptualizing aids, design tools and teaching instrumentation and control portion of P&IDs

INTECH OCTOBER 2024 22 WWW.ISA.ORG/INTECH

DIGITALIZATION

and process flow diagrams (PFDs), it does resulted in the evolution from a hardware
not cover the piping, mechanical and other (instruments)-centric standard to a hardware/
aspects of these drawings. software (automation)-centric standard.
Key elements of ISA5.2-1976: “Binary Logic
A long and proud history Diagrams for Process Operations” were in-
The symbols and identification systems corporated. Binary logic symbols of Scientific
described in this standard accommodate Apparatus Makers Association (SAMA) PMC
advances in technology and reflect the col- 22.1-1981: “Functional Diagramming of
lective industrial experience gained since the Instrument and Control Systems” were also
original 1949 ISA recommended practice, RP- incorporated. Graphic symbol dimension
5.1, was revised, affirmed and subsequently tables were incorporated to establish mini-
published as ANSI/ISA5.1-1984. The 1949 mum mandatory dimensions for the symbols.
recommended practice and the 1984 stan-
Nonmandatory examples were moved to
dard were published as nonmandatory rather
a new Annex B: “Graphic symbol guidelines”
than as mandatory consensus documents.
(Informative) to provide some limited as-
The 1992 revision was published with sistance in the application or were removed
mandatory and nonmandatory requirements. for inclusion into future technical reports to
It incorporated key elements of ISA5.3-1983: provide special practices and requirements
“Graphic Symbols for Distributed Control/ of particular interest groups and/or specific
Shared Display Instrumentation, Logic, and industries.
Computer Systems.”
A significant change was to clarify the
The 2009 revision was published with meaning of the symbols circle-in-square
significant changes as technological advances and diamond-in-square. Previously, these

This example of the application of the standard is representative of the ISA standard and the two
companion technical reports, ISA TR5.1.02 and ISA TR5.1.03. Source: Figure 6 from ISA TR5.1.03.

INTECH OCTOBER 2024 23 WWW.ISA.ORG/INTECH

DIGITALIZATION

represented a distributed control system

The 2024 revision
(DCS) and a programmable logic control (PLC)
The 2024 revision of this standard changed
system, respectively. Given the evolution of
the title from “Instrumentation Symbols
control systems, a distinction was no longer
and Identification” to “Instrumentation
necessary, as both systems had evolved to
and Control Symbols and Identification” to
have similar capabilities.
emphasize that symbols for control are also
Thus, circle-in-square was redefined to included. This revision was published with
represent the basic process control system significant changes to improve the read-
(BPCS), regardless of the type of hardware ability of the document by organizing notes
used. In addition, safety systems (ISA84) were with corresponding tables; simplifying table
becoming prominent, and there were re- numbering; providing definition consistency
quests for a symbol to distinguish them. The and clarification; eliminating redundant text;
diamond-in-square symbol was redefined to adding new symbols; recognizing alternate
allow the user to choose it to represent either symbols, notation and identification for new
a safety instrumented system (SIS) or an automation technology; adding new sections
alternate (other than BPCS) control system. for normative reference, abbreviation and
ISATR5.1.01/ISA-TR77.40.01, “Functional bibliography; and adding the loop instrument
Diagram Usage,” issued in 2012 and reaf- diagrams symbol table. The changes made in
firmed in 2016, was published as the first the 2024 revision are listed in Annex A of that
joint technical report under ISA5.1 and ISA77. document.
The purpose of this technical report was to This revision moved nonmandatory Annex
provide guidance on documenting application A and Annex B from the 2022 revision into
software through functional diagrams by separate technical reports for easier mainte-
illustrating usage of function block symbols nance and to reduce the size of the standard.
and functions and to provide examples of The new technical reports are ISATR5.1.02,
complex function blocks. “Instrumentation and Control Identification
The 2022 revision of ISA5.1 was published System Guidelines,” and ISA-TR5.1.03,
as an interim revision to correct technical “Instrumentation and Control Graphic Symbols
and typographical errors and clarify known Guidelines.” These TRs provide examples and
usage questions. This provided a corrected information on the application of require-
standard for users and also served as a ments in the standard. Examples in these TRs
starting point for the ISA5.1 Working Group were updated with additions, revisions, and
to begin a new revision of this widely used deletions. The changes made in these TRs are
international standard. The corrections made listed in Annex A of the respective document.
in the 2022 revision are listed in Annex A of Users are encouraged to read and use these
that document. TRs together with the standard.

INTECH OCTOBER 2024 24 WWW.ISA.ORG/INTECH

DIGITALIZATION

and control they want to show on their PFDs

Looking ahead
and P&IDs and what information should be
The plan for the ISA5.1 Working Group of the
provided at a minimum.
ISA5 committee is to now develop a technical
report on instrumentation and control con- The updated ISA5.1 standard and its
tent of PFDs and P&IDs. While ISA5.1 deals accompanying ISA technical reports are avail-
with instrumentation and control content on able for purchase online. Anyone interested
various engineering drawings—including PFDs in participating in the ISA5.1 Working Group
and P&IDs—this new TR will provide guidance is asked to contact ISA Standards (standards@
and examples specific to PFDs and P&IDs. isa.org). For information on obtaining any of
This will include guidance to help users de- ISA standards documents and reports, visit
cide on the level of detail of instrumentation www.isa.org/findstandards.

ABOUT THE AUTHOR

Jim Federlein, PE, is Chair of ISA5.1 and a long-time leader in the ISA5 Standards
Committee for which he led the 2024 revision of ISA5.1. He is the winner of the
2024 ISA Enduring Society Service Award, one of ISA’s highest society-level awards,
in recognition of his decades of leadership and service in the ISA Pittsburgh section,
in ISA divisions, in several key ISA standards committees, and as a member of ISA’s
Standards and Practices Board.

ABOUT ISA
The International Society of Automation (ISA) is a nonprofit professional association founded in 1945
to create a better world through automation. ISA’s mission is to empower the global automation com-
munity through standards and knowledge sharing. ISA develops widely used global standards and con-
formity assessment programs; certifies professionals; provides education and training; publishes books
and technical articles; hosts conferences and exhibits; and provides networking and career develop-
ment programs for its members and customers around the world. Learn more at www.isa.org.

INTECH OCTOBER 2024 25 WWW.ISA.ORG/INTECH

DIGITAL TWINS

A Roadmap for Improved

Simulation Success
The most effective simulation solutions
for process manufacturing are designed
using integrated software to deliver value
across the entire lifecycle.
By Dustin Beebe

Today’s process manufacturers face a wide ar- them. New personnel are in short supply, and
ray of new complexities. An expanding global even when they are available, they take years
marketplace has made it more important than to upskill to a level where they can meet the
ever to increase efficiency and competitive- performance of their predecessors.
ness, while global events and trends continu- Further complicating the worker shortage
ally shift, making it harder to achieve those are new trends in the workforce, with mod-
goals. Many organizations around the globe ern plants seeing a trend of more transient
are facing a critical shortage of experienced workers. Gone are the days when an operator
workers. Plant engineers and operators are would sign on and stay for 30 years. Today’s
retiring at an unmatched pace, taking years or talent is typically ready to move to a new role,
even decades of institutional knowledge with or even a new location, in fewer than five

INTECH OCTOBER 2024 26 WWW.ISA.ORG/INTECH

DIGITAL TWINS

years. This is often less time than it takes to evaluate and validate process and automation
fully train them. Furthermore, today’s orga- designs, as well as enabling safer testing with
nizations are finding it hard to attract anyone improved results to help teams more easily
at all unless they offer modern working meet or even shorten project schedules.
environments. The new generation of workers
Yet even after project completion, simula-
was raised on digital technologies, and they
tion software continues to deliver value
expect to see those same capabilities in their
across its lifecycle. Dynamic simulation tools
workplace to help them learn more quickly,
provide the best possible training platform for
make better decisions and collaborate more
new operators, providing them the opportu-
effectively.
nity to work with systems that look, feel and
Global pressure to drive increased sustain- respond exactly like the controls they will use
ability while increasing performance is adding every day. These training simulations can be
additional complexity. Most teams must not built, deployed and used well before equip-
only focus on increasing production but also ment ever arrives onsite, ensuring operators
on doing so while reducing emissions and are ready to perform at their best on the very
curbing excessive energy use. Meeting those first day of operation. These simulations can
goals often means innovating on traditional then continue to be used to train new hires
operations—a big ask with fewer experienced throughout operations.
people and significantly reduced resources.
Dynamic simulation tools also provide a
To meet these challenges, simulation test bed where operations teams can test
software can be a game changer, but only if a new equipment, strategies, and configura-
project team approaches it thoughtfully. While tions to help them increase performance and
it is possible to build one-off simulations for drive more sustainable operations, without
each project and operational need, the result interrupting or risking operation of the plant.
will be costly and difficult to maintain. A better With the right simulation software in place,
solution is to evaluate simulation at every the dynamic simulation can be continually
stage of a project, building a cohesive simula- synchronized with the changing plant to
tion roadmap that will meet the organization’s ensure it is always available to empower
needs at every stage and continue to deliver operators and enhance the way they work to
value well after operations commence. meet their ever-changing goals.

The case for simulation Different stages have different

One of the primary benefits of simulation needs
software is that it provides proven results over One of the most important things to remem-
the lifecycle of a facility. In the earliest stages ber when developing a simulation software
of project design, simulation helps reduce roadmap is while there are different stages
capital expenditures by helping project teams of a project and operation, there are also

INTECH OCTOBER 2024 27 WWW.ISA.ORG/INTECH

DIGITAL TWINS

different types of simulation software, and Leveraging these high-fidelity models, the
they all must be paired strategically. The earli- team will develop their final visions for plant
est stages of project engineering will typically operations, define their operating philosophy,
require steady-state models. In the pre-front- set the project strategy and then use those
end engineering and design (PreFEED) stage, elements to continue to develop and validate
project teams will typically build a simplified their core process designs.
steady-state simulation they can use for the
As the team moves into project execution,
conceptualization of the plant.
their simulation needs will change. The team
These models have limited details, only will begin working with dynamic simulation
providing the general parameters of the software to execute detailed design, where
design of the plant the team wants to build. they develop the automation system and
Such a simulation might work in tandem begin testing procedures and controls, tune
with a capital cost estimator—effectively its control loops and eventually enter the com-
own style of simulation—to gather estimates missioning and training stage. Each of these
and calculations for what each element of elements can be performed on the simulation
construction might cost, allowing the team software to reduce risk and shorten time to
to scale certain elements up or down to stay full production.
within budget. Teams at this stage might
After project execution, the organization
even work with Monte Carlo simulation tools
will continue to use and update its dynamic
to see how different economic factors will
simulation, both to extend training new and
impact their design.
experienced operators as roles change, and
Later, in the FEED stage, the project as a test bed to define and test new operat-
team will further refine its steady-state ing strategies to unlock constant innovation
model using high-fidelity simulation software. (Figure 1).

Figure 1. Dynamic
simulation can be used
to train operators and
test control strategies
in the control room or
even in the field.

INTECH OCTOBER 2024 28 WWW.ISA.ORG/INTECH

DIGITAL TWINS

process but also provide objects that make

The right software at every stage
it easy to build out lower-fidelity objects as
The most impactful choice a project team can
users approach the edges of processes and
make is to select steady-state and dynamic
units. The advantage of such a solution is a
simulation tools that are designed for flexibil-
final product that is easier and more cost-
ity and seamless integration. Unfortunately,
efficient to manage.
many times these choices are made for the
immediate next steps without consideration
for needs later in the lifecycle of the facility,
or even later in the project.
Simulation software can be
One of the most important factors is re-
a game changer, but only if a
use because users need simulation to serve
as many purposes as possible to maximize project team approaches it
its value. For steady-state simulation, teams thoughtfully.
need ties to many tools, allowing them to
perform capital cost estimation—as well as
risk, economic and sustainability analysis.
Later in the lifecycle, a simulation that can be It is important to consider the required
fidelity to support individual use cases. For
compared to live plant conditions to look for
example, contemplate a dynamic operator
optimization opportunities can create value
training simulation where the operator
for many companies.
learns to monitor cooling water on an ex-
In dynamic simulation, one example of changer. The cooling water process could be
re-use is between steady-state simulation modeled in high fidelity, but doing so would
and dynamic simulation. When teams select be complex, and would require many vari-
integrated steady-state and dynamic simula- able changes in the simulation any time the
tion solutions, they can easily transfer their process changed. Those variables, however,
existing flow sheets, base configuration, offer little value in the training simulation.
equipment and instrumentation to their The operator does not need to know if the
dynamic simulation software. cooling water is 81 or 85 degrees. The user
Just as with the steady-state simulation, simply needs to know if there is cooling
teams want dynamic simulation to serve as water flow—a dynamic that can be modeled

many purposes as possible to maximize its and far more easily managed long term in

value. The best dynamic modeling tools also low fidelity.

empower project teams to work in multiple By contrast, if the simulation is training

fidelities. These solutions offer simulation an operator to know that a bioreactor needs
objects that allow teams to perform high- to run at 100 degrees versus 101, that
fidelity dynamic simulation at the core of the distinction might be critical, in which case

INTECH OCTOBER 2024 29 WWW.ISA.ORG/INTECH

DIGITAL TWINS

Figure 2. Simulation software allows users to quickly develop high-fidelity simulations using configurable
dynamic models of process unit operations.

the simulation might need to be high fidelity. Solutions built for success
However, that same high-fidelity bioreac- Whether an organization still has a deep
tor model might be in a training simulation bench of experienced operators or is trying
with many other low- and medium-fidelity to onboard a new generation of workers
elements as well. Every model that can be with limited experience, finding a safe
created in lower fidelity reduces the number way to test, train and tune new processes
of complex interconnections in the simulation is critical. New workers will need to gain
(Figure 2). experience as quickly as possible if the plant
hopes to meet the necessary performance
Dynamic simulation software with the
benchmarks dictated by competition in a
capability to easily incorporate high, medium
global economy.
and low fidelity empowers teams to custom-
ize their solution to the unique specifications Conversely, even experienced workers will
of their process. By eliminating unnecessary have to learn many new operating procedures
interconnections, teams reduce the likelihood (on very different, and often more complex
that the simulation will be too hard to main- equipment than they are used to) if they hope
tain as variables change due to equipment to help their plant meet new sustainability
swap-outs, degradation, fouling, or other benchmarks and comply with regulations.
changes. In either case, operators need a risk-free

INTECH OCTOBER 2024 30 WWW.ISA.ORG/INTECH

DIGITAL TWINS

Figure 3. As
operations become
more complex, plant
personnel will need
increasing access to
safe ways to learn
and test innovative
operations strategies.
Simulation will
be central to this
capability.

environment to learn, test and innovate. Such drive fast return on investment and continu-
an environment cannot be provided on live ous value over the lifecycle of the facility,
equipment (Figure 3). from the earliest stages of design, through
automation development and startup, and
Fortunately, today’s multipurpose steady-
even throughout operations as they change
state and multi-fidelity dynamic simulation
over the years. The key is selecting an inte-
tools are up to the task. Modern, best-in-class
grated solution upfront that is designed for
simulation tools offer the flexibility to meet
the unique needs of every stage.
the dynamic environment of today’s plants.
They also include the features necessary to All figures courtesy of Emerson

ABOUT THE AUTHOR

Dustin Beebe serves as vice president of Performance Software for Emerson. He is
responsible for the alignment of the Control Performance, Operator Performance
and Simulation businesses globally and the strategy synergy between Emerson
and AspenTech. Prior to joining Emerson, Beebe served as the President of ProSys
until it was acquired by Emerson in 2018. He has been in the industrial automation
business since 1996. Beebe has a bachelor’s degree in chemical engineering from the University of
Arkansas in Fayetteville, Ark.

INTECH OCTOBER 2024 31 WWW.ISA.ORG/INTECH

ARTIFICIAL INTELLIGENCE

How to Harness
Applied AI in Industrial
Manufacturing
By Michael J. Anthony, Jon A. Mills, and
David C. Mazur, Ph.D.

Properly understand key

concepts and use necessary
contextual data to train AI
models.

Industry 4.0 has continued to evolve and

grow its presence within the industrial
automation community. As a result, the com-
munity faces considerable challenges due to
the growth of Industrial Internet of Things
(IIoT) devices and technologies. One of these
technologies is artificial intelligence (AI).
Applied AI has been around as a concept for
many years in various fields, but industrial
automation has been cautious to adopt the
technology. This article will explore a brief
history of applied AI and its usage within
industrial automation.

INTECH OCTOBER 2024 32 WWW.ISA.ORG/INTECH

 ARTIFICIAL INTELLIGENCE

more than 30 hours of a typical work week

History of applied AI
The industrial automation and manufacturing would become automated by AI by the year

sectors are facing unprecedented levels of 2030. However, properly using AI in automa-

pressure. These pressures include increasing tion and manufacturing environments re-
demand for products, large backlogs due to quires first understanding some key concepts.
supply shortages from the COVID-19 pan-
demic and significant labor shortages due to How NLP solves the data
lack of skillset ready to enter these markets. problem
Natural language processing (NLP), a field at the
Due to the difficulty of hiring and retaining
intersection of computer science and linguistics,
talent and the skills gap shortage, businesses
has evolved significantly from its preliminary
need to turn to alternatives to help ease the
pressures they are facing. Many manufactur- concepts in 17th-century philosophy to its

ers turned to robotics to help solve the labor formal establishment with the dawn of comput-

challenges with various levels of success. ing in the 1940s. These early ideas laid the
groundwork for machine translation and the
first computational models of language.

There is much opportunity to use The field has seen steady progression, with
early rule-based (symbolic) methods being sup-
applied AI in manufacturing, but
plemented by statistical models and eventually
along with that opportunity come overtaken by today’s advanced neural network
many challenges. approaches. Among the most transformative
neural network architectures for NLP is the
“transformer,” introduced in the seminal paper
“Attention is All You Need” under the umbrella
A second problem that manufacturers are
of Google’s research initiatives in 2017.
attempting to solve is the data problem. With
computing and processing costs at the lowest Transformers have laid the foundation for
levels ever, data from devices and equipment the diverse array of NLP-powered AI now be-
is more prevalent. Manufacturers are strug- ing developed across the technology sector.
gling on their digital journeys with how to From compact models designed for budget
consume the growing volumes of data they’re devices to enormous architectures operating
producing. Extracting insights to drive useful on cutting-edge cloud computing resources,
outcomes is not easy. the scope and application of NLP models
The solution for many manufacturers is ap- have never been broader. Rather than only
plied AI. The impact of applied AI is promising being present in research fields, this is offered
in industrial automation and manufacturing. in forms that are relevant to industry adop-
A major consulting firm has estimated that tion or hyper-specific domain-bound tasks.

INTECH OCTOBER 2024 33 WWW.ISA.ORG/INTECH

ARTIFICIAL INTELLIGENCE

transformer-based models defy these con-

Nondeterminism neural
ventional practices due to their probabilistic
networks
outputs. The variability in responses means
Understanding the nondeterministic nature
that results from traditional testing can no lon-
of transformer-based neural networks is
ger guarantee consistent model performance.
fundamental to appreciating the challenges
highlighted in this article. These models, This inherent nondeterminism necessitates
which underpin contemporary advances in the development of novel testing and valida-
natural language AI, exhibit inherent stochas- tion frameworks attuned to the probabilistic
tic behavior often seen within manufacturing. nature of these AI models. Organizations
Not only does nondeterminism enable the must adapt by implementing strategies like
generation of diverse and fluent responses A/B testing, continuous monitoring and
across various domains, it also presents dynamic error analysis, which accommodate
unique challenges in standardization and the variability of responses.
quality assurance when deploying these
Moreover, product teams must be educat-
models into production environments.
ed in stochastic model behavior, establishing
Contrary to traditional software, where realistic expectations and a deeper under-
deterministic input-output relationships standing of the tradeoffs associated with
allow for standardized testing methods such leveraging these powerful but unpredictable
as unit, integration, and system testing, models in commercial applications.

An analytics module uses AI to detect production anomalies and alert workers so they can investigate or
intervene, as necessary.

INTECH OCTOBER 2024 34 WWW.ISA.ORG/INTECH

ARTIFICIAL INTELLIGENCE

Applied AI example in
industrial automation
Building on the discussion of prompt injection
techniques, such as retrieval-augmented gen-
eration (RAG), an experiment was conducted
to evaluate their efficacy in improving model
accuracy and reducing instances of generated
hallucinations. The authors used the GPT-
3.5-turbo model developed by OpenAI for
this investigation, structuring our prompts to
solicit specific information. The initial prompt
was constructed as follows:

“What is the name of alarm 10041 on the

PowerFlex 755T Variable Frequency Drive?
Provide only the alarm name, no other text.”

To guide the model’s responses, we incor-

porated an exemplar user/assistant exchange:

User: “What is the name of alarm 10012 on

the PowerFlex 755T Variable Frequency Drive?
Provide only the alarm name, no other text.”

Assistant: “Brake Slipped - Drive Stopped”

Upon presenting this structured prompt to

the model 10,000 times, we observed 6,934
unique responses, none of which were the
correct answer “Precharge Open Alarm.” This
result suggests an absence of the necessary
data within the model’s training corpus. The
most frequent responses are listed in Table 1.

DC Bus Overvoltage 173 Encoder Fault 61

Motor Overtemperature -
Drive Overtemperature 124 51
Drive Stopped
Motor Overtemperature 86 Analog Input Loss 48
Overvoltage Fault 79 Drive Overvoltage 45
Undervoltage Fault 68 Motor Phase Loss 44
Table 1. The initial GPT-3.5-turbo prompt yielded these responses.

INTECH OCTOBER 2024 35 WWW.ISA.ORG/INTECH

ARTIFICIAL INTELLIGENCE

To address this variation in response, the With this additional context, a subsequent
authors performed a second trial, introducing a set of 10,000 queries yielded only 19 unique
JSON object into the system context, derived responses. The majority appropriately identi-
directly from the relevant source material: fied the alarm as shown in Table 2.
{
Precharge Open Alarm 7,808
“Condition Type”: “Alarm 2”,
Precharge Open Alm 1,742
“Condition Code”: “10041\n11041”,
Precharge Open 344
“Display Text”:
“PrechargeOpenAlm”, PrechargeOpenAlm 44
Precharge Open Alrm 29
“Full Text”: “Precharge Open Alarm”,
Precharge Open Alar 12
“Fault”: “The internal precharge-
circuity-bypass relay (for drives) Precharge Open Alarm. 11
or main contactor (for CBIs) was
commanded to open while the drive
Precharge Open - Alarm 3
was stopped (PWM was not active) The name of alarm 10041 2
due to low DC bus voltage.”, on the PowerFlex 755T is
“Precharge Open Alarm”.
“Action”: “Investigate low DC bus
voltage or the reason the drive Precharge Open-Drive Stopped 2
entered precharge.”,
Table 2. Responses from the contextual prompt.
“Fault Action”: “—”,

“Configuration Parameter”: “0:37 The introduction of contextual data signifi-

[Prchrg Control]\n0:190 [DI cantly increased the frequency of the correct
Precharge]\n0:191 [DI Prchrg
response. However, despite the narrowed
Seal]\n”,
range of responses, the nondeterminism of
“Configurable Action”: “—” the model continued to produce slight varia-
} tions in the output.

INTECH OCTOBER 2024 36 WWW.ISA.ORG/INTECH

ARTIFICIAL INTELLIGENCE

model training. There is much opportunity to

Final thoughts
Harnessing applied AI in industrial manufactur- use applied AI in manufacturing, but along with

ing requires properly understanding key con- that opportunity come many challenges. We are
cepts and using necessary contextual data in AI in an exciting time to see how this evolves.

ABOUT THE AUTHORS

Michael J. Anthony graduated from Marquette University in Milwaukee with a
Bachelor of Science degree in computer and electrical engineering and started with
Rockwell Automation as a software development engineer in 2005. He worked on
a variety of information and human-machine interface (HMI)-focused products in
the FactoryTalk portfolio and has held roles as a product manager for a variety of
HMI, communication, and security software products in the Rockwell Automation
portfolio. Currently, Anthony is focused on applications communication technology in the Rockwell
Automation Strategic Development organization in the office of the CTO. He earned a Master’s degree
in 2019 and is pursuing a PhD in manufacturing systems focused on communication technologies
from Capitol Technology University in Laurel, Md.
David C. Mazur, PhD works as a senior manager for Rockwell Automation in
Milwaukee with a current focus on digital experiences for industrial automation
products. His experience includes application development in heavy industry
automation and infrastructure. Mazur received his BSEE from Virginia Polytechnic
Institute and State University, Blacksburg, VA in 2011. He graduated with his
MSEE degree in 2012 from Virginia Polytechnic Institute and State University. He
graduated with a PhD in mining engineering in September 2013 for his work with automation and
control of the IEC61850 standard. Mazur is an active member of the IEEE IAS and serves as working
group chair for the Communication-Based Protection of Industrial Applications Working Group.
He also serves as a member of the Mining Industry Committee (MIC) as well as the Industrial and
Commercial Power Systems Committee (I&CPS). Mazur is also an active voting member of the IEEE
Standards Association (SA).
Jon A. Mills graduated from Ohio University in Athens, Ohio with a Bachelor of
Science degree in computer science. He started at Rockwell Automation in 2013
with a focus on integrating intelligent devices into industrial control systems.
Working as a principal system architect, Mills continues to focus on device
integration within traditional operational technology (OT) networks in the OT/IT
(information technology) boundary.

INTECH OCTOBER 2024 37 WWW.ISA.ORG/INTECH

 THE LAST WORD | VIEWS FROM AUTOMATION LEADERS

Closed-Loop Control Fundamentals

By Jack Smith

A trip through past issues of InTech reveals a control means adjusting multiple single-loop
wealth of resources for understanding closed- controllers in unison to meet constraint
loop control fundamentals. control and optimization objectives of an
additional set of related process variables.
As I wrote in my August 2023 InTech article
Multivariable control is a central aspect of
on temperature measurement and control
nearly every industrial process operation.”
fundamentals, “Automatic control in continu-
ous processes uses industrial control systems While Kern’s article delves into more
to achieve a production level of consistency, complex subject matter that most of our read-
economy, and safety that could not be ers are familiar with, APC and multivariable
achieved by human manual control only. It control still depend on single-loop controllers.
is implemented widely in industries such as Understanding the fundamentals and/or re-
oil refining, pulp and paper manufacturing, viewing the basics can be beneficial to techni-
chemical processing, and power generating cians and operators who need a refresher.
plants, to name a few. The “big four” process
control parameters are temperature, pressure,
flow, and level.”
“APC continues to rely on the
Although that article was primarily about
lowly flow control loop, the most
controlling temperature, closed-loop control basic single-loop control, as the
concepts are fundamentally the same. Only
best rejector of unmeasured
the sensors and processes are changed.
disturbances and the most
Allan Kern, PE has 35 years of industrial
process automation experience and has
stable platform for the APC/
authored dozens of papers on more practical, optimization control hierarchy.”
reliable, and sustainable advanced process
control solutions. Kern helps companies
Jim Ford’s article in June 2019 InTech
improve process efficiency, quality, and
describes how single-loop control is still the
profits on-site or with online consulting
mainstay of advanced process control. He
complementing in-house resources, helping
says, “Today, even after 50 years, APC con-
bridge a skill shortage at many sites. He is the
tinues to rely on the lowly flow control loop,
founder of APC Performance LLC.
the most basic single-loop control, as the best
In his Feb. 2019 InTech article, Kern wrote, rejector of unmeasured disturbances and the
“Advanced process control (APC) refers pri- most stable platform for the APC/optimiza-
marily to multi-variable control. Multivariable tion control hierarchy.”

INTECH OCTOBER 2024 38 WWW.ISA.ORG/INTECH

 THE LAST WORD | VIEWS FROM AUTOMATION LEADERS

Again, although Ford was writing about into a strong standardized signal and transmit
flow, the concepts still apply to the other it to a control system. Sophisticated transmit-
three of the big four. ters can perform diagnostics on the sensor
to determine if there is degradation of the
It starts with the sensor actual element. The transmitter connects to
Process control parameter measurements the control system to provide the process
start with the sensor. The aforementioned variable (PV) measurements.
temperature control uses thermocouples,
Maintaining a digital signal to the control
resistance temperature detectors (RTDs),
system maximizes accuracy. Digital com-
and associated transducers and transmitters.
munications avoid the errors of converting
Pressure measurement requires pressure
the digital signal to analog 4-20 mA on both
transducers, flow requires flowmeters, and
the transmitter end and the control system
level requires a level measurement system.
end. Digital options include HART, Modbus,
Much can be—and has been—written on each
Profibus, and FOUNDATION Fieldbus.
of these technologies.
Accuracy and stability are fundamental
A transducer converts a physical phe-
traits of any process measurement. Although
nomenon into an electrical signal. In effect,
closed-loop control can be accomplished in
thermocouples and RTDs are types of trans-
many ways with many technologies, such
ducers. The use of the term is more common
as programmable logic controllers (PLCs)
in flow and pressure control.
or distributed control systems (DCSs), this
Transmitters convey a measured signal to article assumes a stand-alone single-loop
a control device. The signal coming directly controller (Figure 1). This theoretical con-
from the sensor is at a low level. The job of troller includes a signal processing front
a transmitter is to convert the sensor output end that converts low-level input from the

Figure 1. This single-loop controller can control

both heating and cooling simultaneously, and
can accept signals from a thermocouple or
RTD, or from a pressure/flow/level sensor, and
maintain a setpoint using a relay, voltage pulse,
current, or linear voltage output signal.

Courtesy: AutomationDirect

INTECH OCTOBER 2024 39 WWW.ISA.ORG/INTECH

 THE LAST WORD | VIEWS FROM AUTOMATION LEADERS

sensor to a usable signal, which is compared

Closing the loop
with a setpoint (SP). The resulting output
Regardless of the type of controller (PLC,
depends on the amount of error between
DCS, or single-loop controller), the measured
the measured temperature, or process vari-
signal from the sensor and/or transmitter is
able PV, and the setpoint.
compared to a setpoint. The resulting output
Single-loop controllers are used in small depends on the amount of error between
facilities, or for some isolated stand-alone the measured temperature, or PV, and the
processes. Large continuous process facilities, setpoint.
such as refineries and chemical plants, use
In addition to accurately measuring a
DCS to control pressure, temperature, flow,
process, there must be a way to control the
and level, and how they affect the opera-
amount of correction applied to that process.
tion of the plant. In some cases and some
The process itself “ties” the system together.
industries, PLCs instead of—or in addition
to—DCSs are used. Sometimes, PLCs control The output of the controller must have a
subprocesses via signals obtained from a means of actuating the controlled process.
main DCS. This can be heaters or burners, control valves,
or positioning devices. Then the closed-loop
PLCs have been used to control tempera-
control system begins again with the process
ture for decades. It should be noted that if
being sensed and the controller adjusting its
only temperature control is required, a DCS
output.
or a PLC is gross overkill. These systems are
designed to control the entire process plants Single-loop control gets more compli-
or parts of plants. Either of these devices is cated with the introduction of proportional-
capable of having hundreds of control loops— integral-derivative (PID) functionality. PID is a
temperature, flow, pressure, and level. topic for a future column.

ABOUT THE AUTHOR

Jack Smith is senior contributing editor for Automation.com and InTech digital
magazine, publications of ISA, the International Society of Automation. Jack is a senior
member of ISA, as well as a member of IEEE. He has an AAS in Electrical/Electronic
Engineering and experience in instrumentation, closed-loop control, PLCs, complex
automated test systems, and test system design. Jack also has more than 20 years of
experience as a journalist covering process, discrete, and hybrid technologies.

INTECH OCTOBER 2024 40 WWW.ISA.ORG/INTECH