At its most basic, IIoT is about forming connections that deliver data and allow beneficial decisions. However, as links, networks and data multiply, some kind of prioritization and triage is necessary, so users can focus on the most valuable tasks, and this likely means reevaluating and updating established connections, processes and infrastructures.

“For us, IIoT means more techniques for getting the right information to the right systems and people at the right time. What’s different now is we’re solving more problems in larger chunks over the past 10 years because we got better at getting more sensors and their data into storage,” says Chris Herrera, head of API and interoperability at Seeq. “Now, we have all kinds of data lakes that we’ve loaded everything into, but too often, we can’t find what we need. This is typically because IIoT and other networks are often so fragmented that a lot of data loses necessary context, especially as even more sensors, data and systems are brought online. These added silos can make it difficult to add, for example, something as simple as reason codes for downtime, and this can make it hard to test and qualify a system tag for a batch without identification, which is what makes it difficult to understand its context.”

Herrera reports that similar difficulties are increasing on IIoT’s physical network levels, such as operating without enough bandwidth, so the network can’t ingest all the new data sources coming in.

“There are a lot more plates to keep spinning now, and traditional, centralized infrastructures tied to specific network protocols or vendors are having a hard time surviving,” adds Herrera. “The primary problem is more data also means increasing complexity and more required maintenance. Because Seeq is at the decision-intelligence layer, we see these issues as a decision-velocity problem, and not just an infrastructure issue. Even before input from a data source arrives, it can still take three to five days for users to decide whether to bring all of the related players together, discuss, and decide what to do. If users don’t consider what they’re going to do with all their data, then they’ve just got a bunch of objects or things. This is why it’s so important to focus on what they’re trying to achieve.”

From calendar to condition

For instance, one of Seeq’s pharmaceutical clients recently migrated from calendar-based performance and maintenance to a condition-based procedure to reduce over- and under-utilization of equipment and systems. To gain this efficiency, however, the company needed a better understanding of its equipment’s capabilities within an overall architecture suitable for meeting its corporate goals. It needed data about performance, uptime, quantity, quality, sustainability, margins, energy utilization and other factors.

Get your subscription to Control's tri-weekly newsletter.

“These are the details we need to get to the right peak-operating devices and systems. Our pharmaceutical manufacturer’s goal was cutting energy use by 30%. It needed to understand the transient cycles of its compressors and pumps, then determine if it could run lines and batches longer, switch duty cycles on specific assets, or issue a work order to the operator. This would prompt the right actions, such as bringing one compressor down and another up, without needing to communicate the full corporate context all the way down to the plant floor,” explains Herrera. “All these analytical capabilities exist within IIoT. Users just need to make the necessary connections, and establish orchestration to realize beneficial outcomes.”

Adding an AI boost

While Seeq Analytics simplifies data access and evaluation, and Seeq Enterprise software streamlines process monitoring at scale, the company’s new Seeq Intelligence software integrates human-centered AI to amplify human expertise rather than replace it. Seeq Intelligence makes institutional knowledge persistent and scalable across the enterprise. This provides reassurance for decisions, enables more consistent reporting and actions, and delivers added context that can be used to measure goals and rationalize them more easily.

“Using a common analytics platform gives users a consistent way to seek root causes, identify problems, monitor processes, respond to production signals, and triage at scale,“ adds Herrera.

To solve the initial velocity problem, and make decisions quickly enough to make a difference, Herrera adds that users simply have to determine what they’re trying to achieve, and make sure they get the data needed to do it.

“If we just have an HMI and SCADA system, it’s not enough,” cautions Herrera. “Users must explore what IIoT can offer. This includes trying unified name space (UNS) to help solve data fragmentation issues; adopting data infrastructure solutions, such as Emerson’s AspenTech Inmation or Aveva’s Connect, which bring together data sources more quickly; or seeing how Databricks’ customers use systems like HighByte to let users build custom data flows. None of these are big bang solutions. They’re iterative processes that never really end, so it’s best to get started. Just use one of these tools to connect to the data you’ve got, and if you can’t do something you’d like to do, then that will lead to what’s needed next. As usual, it’s always the use cases that drive the necessity that determines what comes after.”