Focus on data tasks to define IIoT

There can be a lot of confusion about how the Industrial Internet of Things (IIoT) should be defined, especially because it’s evolving so quickly. Because of this uncertainty, many experts emphasize that IIoT can be understood by focusing on the data processing it’s trying to perform.

AspenTech

“IIoT and artificial intelligence (AI) unlock value from industrial data. This is where insights, predictive analytics and AI come from,” says Yaacoub Awad, senior product manager for AIoT at AspenTech. “IIoT often means industrial machines generating data and being accessible in the cloud, which can lead to users having more data than they know what to do with. AspenTech has always helped customers gather, organize and analyze their industrial data. The cloud then becomes an enabler for the industrial data foundation because Aspen Enterprise IP.21, our data historian, can be deployed in the cloud. Aspen Connect makes it very easy to transfer data to the cloud, so that data analysts have information and insights they need in one place.”

Hugo Redon Rivera, principal software developer at AspenTech, reports that companies are making use of cloud-based versions of historians that run in conjunction with services like Microsoft Azure and Amazon Web Services (AWS), and can send back analytics results on dashboards.

“With industrial data from multiple plants in one place, users can examine their industrial data and draw insights from it,” explains Rivera. “They might not just need IIoT. They may need to digitalize other operations or centralize their workflows.”

Once users decide what results they need to see and how often, Rivera adds they can be displayed on AspenTech’s Aspen Enterprise Insights real-time dashboards, which likewise use Aspen Connect to migrate IIoT data to the cloud. In addition, Aspen AI Model Builder software combines production data with first-principles simulation functions to develop useful calculations. It also employs neural networks to reduce the computational complexity of these calculations, which can shorten the time needed to complete the calculations for a refinery reactor or other processes.

“In addition to reducing the computational complexity of the simulations, it makes predictions without customized programming, and lets users test different raw materials and scenarios without customized modeling,” adds Rivera. “They can also build a quasi-infinite amount of data by running simulation models in the cloud with Aspen Multi-Case to generate analytics and results.”

Emerson

“To me, IIoT is about connectivity and integration, and what comes with them. Much of IIoT has been getting field devices on the TCP/IP stack, but it also involves connecting and integrating all automation assets in a way that’s more IT friendly and enables OT and IT convergence,” says Claudio Fayad, technology VP of Emerson’s process systems and solutions business. “Many of today’s industrial technologies started in IT, but now they can be used on the OT side. These include cloud-computing services, cybersecurity, virtualization and software containerization, and they’re enabled by OPC UA, MQTT, RESTful API and others.”

Fayad reports that examples of IIoT in action include PLCs that have dual operating systems in edge applications, with one side handling real-time interlocks and control, and the other side managing connectivity workloads and analytics. “This increased technical integration helped our DeltaV DCS platform work with Moderna to accelerate creating vaccines using a more digitalized process,” adds Fayad. “It gives us a more digitalized enterprise. Employing DeltaV with IIoT connectivity brings systems together more effectively and improves production.”

For instance, Moderna produced its COVID-19 vaccine in less than a year, even though previous vaccines were never previously developed, tested and produced in less than four years. One reason for this success was the flexibility Moderna gained by implementing a more digital framework, and increasing its agility by working closer with Emerson and other suppliers.

“We tried to build a native digital manufacturing site with mature toolkits like Syncade MES and DeltaV,” said Roland Smith, business and GXP systems VP at Moderna, who presented at the Emerson Exchange Virtual Series in 2021. Moderna also benefitted from previously starting to build a digitalized manufacturing facility in 2016, and included advanced analytics tools into its processes from the beginning, ensuring its team would have comprehensive, contextualized data to support its processes. “Today we have almost a billion rows of data that we’ve learned from.”

For example, Smith added that flexibility and agility due to digitalization let Moderna carry out a “minimum viable product” approach to drive innovation in its process designs. Its teams designed part of a batch record to solve an operation, and immediately put theory into action by testing those processes in a live environment. The feedback gained from those tests enabled Moderna to rapidly improve and design future batches.

Consequently, Moderna’s approach to life sciences manufacturing and digital technology methodology paid off. Even in pre-COVID manufacturing, the team:

• Reduced cycle times from more than 10 days to five or six days;
• Eliminated stops and starts in operations with parallel-path workflows;
• Reduced batch reviews from days to hours; and
• Increased error proofing, reducing manual error rates by about 80%

Fayad adds that DeltaV and Syncade MES enable more sitewide analytics and innovation to support digital transformation by combining the cloud and edge and the field, which means connecting all of a user’s sensors and hardware all the way to the cloud. Likewise, Emerson’s other digital technologies such as cloud-hosted digital twins, cloud engineering and remote virtual factory acceptance tests (FAT) enable greater collaboration to reduce project hours, costs and risks.

“This is one of the most exciting times because IIoT and digitalization allow so much innovation and create so many technical possibilities,” says Fayad. “However, it can be overwhelming to try and bring all these technologies together, especially when you’re just starting to use them for digital transformation to drive optimization. The key is to start small, decide what questions you want to answer, make improvements in operations and maintenance in one area, and scale up later.”

Fayad reports that if users start with an IIoT project that’s too big, it can generate too much information and become unwieldy. “If you end up with terabytes of data, you may ask ‘what do I do now?’ ” adds Fayad. “This means the proof of concept (PoC) may not work, and the project’s participants may give up on it.”

To avoid information overload, Fayad adds it’s important to start small and address the underlining data on analytics, too. “For instance, if you want to do advanced control, you’re not going to do it with loops in manual, so you need to make sure they’re on automatic first,” explains Fayad. “The same concept applies when trying to use IIoT for digital transformation, where the value of analytics will highly depend on the quality of the data coming in. Distributed control systems (DCS) are a great platform to execute your POCs. For 25 years, they’ve been collecting data through industrial protocols like HART, Foundation Fieldbus, OPC and other protocols and making the data available to controllers and higher-level applications.

Fayad notes that IIoT and digitalization are ingrained into distributed control systems as an edge platform, seeking to put computing power as close as possible to where production data is produced. “The edge has always been OT’s domain, but now the edge is expanding its infrastructure to include more connections and secure the right data for analytics,” says Fayad. “How much edge computing and where to put it also depends on whatever problem users are trying to solve. The edge can also be the first layer of analytics, as well as where process control occurs. OT is the core of the edge.”

Fayad concludes that IIoT and digitalization are causing cultural shifts that are reorganizing and collapsing departments, but suppliers can help IT and OT develop joint partnerships. “Many participants are recognizing these shifts and convergences, and are getting all their voices together and allowing them to be heard,” adds Fayad. “They’re learning to get IT the data it needs, enabling OT to connect via secure protocols and interfaces, and cooperating to avoid doubling up on infrastructure and adding unnecessary complexity.”

Honeywell

Even though many IIoT technologies are relatively new, the needs they can fulfill are not. “Users need improved ways to run their operations by virtualizing their controls, bringing data to the cloud, getting real-time input from their machines and processes, and securing other types of information, including financials, work orders and personnel data,” says David Trice, chief product officer and general manager, Honeywell Connected Enterprise. “IIoT is a way to input data and get strategic feedback on performance for better decisions.”    

Trice reports that HCS recently launched several IIoT solutions based on its Honeywell Forge enterprise performance management (EPM) software. These include “control tower” tools for optimizing performance among connected buildings and other facilities, migrating automated asset maintenance tasks, and balancing energy use with comfort in HVAC with cloud-based monitoring and controls.

“Honeywell Forge gets data from sensors, controls, historians and other devices, and delivers it to cloud-computing services like Microsoft Azure,” explains Trice. “They also be assisted by Honeywell Forge Asset Performance Management and Honeywell Forge Connect software. These solutions let users pull data without a lot of work because they’re usually already tied to an installed base of Honeywell devices, and can get connected, add assets, and link to the cloud in weeks instead of months. In the past, if a production issue occurred, users might not be able to access data about it until the end of the day. With IIoT, if an issue happened at 9 a.m., they can learn what they need to know by 9:30 a.m. IIoT also enables modeling, data integration, consuming the results, and providing insights for users at more useful points during their production processes.”

Novum Automation and Phoenix Contact

Reese Gallagher, lead systems engineer and owner, Novum Automation in Traverse City, Mich., reports it specializes in industrial automation, panel building and machine control, and released its Promea process measurement and analysis product in April, which is designed to improve on the traditional, repetitive and cumbersome process of acquiring process/machine data, and analyzing it—typically via manual means, such as pen and paper, manual audit checks, Excel spreadsheets, manual entry, etc. Promea uses commercial, off-the-shelf (COTS) software such as Grafana, Amazon LightSail, mySQL and Ignition to provide cost-effective, turnkey, automated process measurement and analysis.

“We put COTS pieces together to create an inexpensive, rapid-deployment process measurement solution that can bring Industry 4.0 tools to small-to-large SCADA and measurement applications,” says Gallagher. “Users can spend $2 million on other process measurement applications, and still have to write SQL scripts for pulling data (or submit requests to a data analyst to create them), which doesn’t help engineers and operators on the plant floor diagnose and solve problems in real-time. Promea can also connect to a user’s pre-existing data historians/databases, such as OSI PI, FactoryTalk, AWS RDS, MS SQL Server, and mySQL.”

Gallagher defines one aspect of Industry 4.0 as an efficient way to put data into engineers’ hands or visualize it to solve problems, so they can get away from slower, manual methods, and instead automatically collect and distribute information to a local database or the cloud, shrink the traditional process measurement loop, see dashboards and trendlines in real-time from anywhere, and make better decisions faster.

“We connect the dots with the latest reliable technology that’s available, from sensors and electrical panel hardware to efficient cloud database storage, and real-time dashboards to provide a cost-effective turn-key solution to our customers.,” explains Gallagher. “We install and wireback sensors, set up Grafana/Ignition dashboards, connect to the cloud, provide complete, turnkey, point-of-process systems with the data that users value most, and do it immediately with real-time dashboards and trends.”

Gallagher adds that Novum uses push-in terminal blocks, PTFix distribution blocks, TC routers and power supplies from Phoenix Contact to build enclosures, and integrate its Promea Edge hardware and software faster and more reliably. Promea Edge takes data from sensors, performs preliminary analysis, pushes it to a SQL database or cloud-computing service without intruding on the IT side, and can display it on a web-based dashboard.

For instance, Graham Packaging Co. in Lancaster, Pa., needed inline monitoring and 100% inspection of plastic bottle wall thicknesses. Instead of pulling 10-20 bottles per hour as it did in the past. Novum installed a Keyence light-based sensor for non-opaque bottles, set upper and lower thickness limits, and pushed its data via Promea to a local client and dashboard using Grafana and Tableau software.

“This gave the client information about every bottle, which they needed to find out-of-spec bottles soon enough to adjust their machines,” says Gallagher. “For example, they could now identify out-of-spec bottles sooner, and make machine adjustments faster to reduce scrap rates.”

Gallagher adds that users running other applications can make similar gains with IIoT tools if they’re determined to understand how it works. “This is a big paradigm shift, so users must go through the process of understanding how much easier it can be,” explains Gallagher. “Five or 10 years ago, they’d likely need several specialists to do an IIoT project. Now, they may just need one controls engineer to review and assemble COTS products to attain a process measurement and analysis solution. Today’s users may not need formal schooling in database management or networking. If they just pick reliable products, they can do far more without a technical background.”

Rockwell Automation

“Users want to understand what their devices and processes doing and why to optimize them, and the IIoT is making more of these connections and performance data available. Users just need to decide how to scale up to whatever level is optimal for their process or organization,” says Chirayu Shah, head of product for cloud and edge analytics at Rockwell Automation. “For instance, users baking cookies at five sites previously gathered and distributed data via an Excel spreadsheet, but in the past few years, more are implementing unified approaches that enable real-time optimization, while others are trying closed-loop data gathering and analytics. This allows better recipe changeovers, and tweaks to time and temperature that are also closer to real-time for better performance and outcomes. We’re also seeing more citizen data scientists, who are learning from their controllers, and writing updated instructions back to their processes. This is why Rockwell Automation is investing more in self-learning and self-healing processes.”

Shah reports that higher-speed data automation is also emerging, which is enabling analytics that can make plant-floor life easier. For example, a company experiencing quality issues with the coffee pods it was manufacturing needed sub-second batch data to do a root-cause analysis across its dataset, so Rockwell Automation built an information model that gave it better access to information and more time to store it. This allowed more accurate production sequencing, and let it find the root-cause of failures.

“They used our FactoryTalk (FT) Edge Gateway to inherit datasets from the information model, including recipes and time-stamped batch data from the controller,” explained Shah. “This gave the data layer the common context it needed. Usually, customers build IIoT and MQTT networks to reach cloud-computing services, but find it also takes a lot of time and effort to clean and add context to the data they want to analyze. In fact, 80% of an analytics project’s time can go to preparing data, while only 20% is left the develop the application. FT Edge Gateway can reduce time needed for data cleaning, and get application development up into the 40-60% range.”

As usual, Shah adds that IIoT and analytics projects should begin with identifying a scalable production problem, and using it to point out the right technology. However, equally adaptive and scalable change management and people management are also crucial for enabling IIoT technologies and projects to succeed by making production issues more visible. “A few seconds faster or few degrees of temperature in the right direction is what more accessible data and faster analytics are trying to accomplish,” says Shah. “This is why we also try to replicate tribal knowledge, and build bridges to the awesome know-how that’s out there. There’s no magic to it. It’s just preserving domain expertise, and predicting issues and failures—so future users won’t have to put their hands on a pipe to try and determine what’s going on.”