Transform From the Edge Up
EP Editorial Staff | May 1, 2022
Start your enterprise-level digitization project by empowering end users to develop valuable IIoT insights at the edge.
By Derek Thomas
Manufacturers in all industry segments desire quicker access to the wealth of data available in their facilities so they can make better-informed production and business decisions. Accomplishing this is frequently a significant challenge.
Most of these sites have a wide range of production equipment, utilities, and facility systems, installed over many years from various suppliers. The only consistency is inconsistency. These systems produce a lot of information as they operate—status indicators, inputs and outputs, and alarm indications—that, with the right know how, could be transformed into key performance indicators (KPIs). Combined with other analyses, the operational data can also generate insight into how operations can be improved, but only if the information can be extracted, collected, and presented in a useable way.
Legacy infrastructures strand data
Most companies don’t have available teams of data scientists or extra IT experts, but they do have experienced OT personnel who know the systems and where improvements are needed. However, they must be able to easily deploy solutions, access data, and analyze the relevant information. They also need actionable information they can use immediately, not time-delayed feedback from a complex system and network architecture.
Source data may be stranded for a variety of reasons, such as:
• isolated, with no network connection
• ignored, since no system accesses it
• under sampled, providing an insufficient data rate
• inaccessible, in a format unusable by traditional systems
• non-digitized, such as information captured on paper.
There are solutions to obtain stranded data, but no end user wants to implement dozens of proprietary connectivity methods, software products, and cloud packages to access their numerous assets, controllers, sensors, and protocols. In fact, most end users likely don’t have the budget, time, or personnel to tackle massive enterprise-wide IIoT projects.
Some companies try to start with a pilot project using a “home grown” solution, but end up disappointed when they find it won’t scale up because it introduces significant support and development issues. Needs and requirements change over time, and it can be costly or difficult to adapt custom and enterprise-grade solutions. Today’s users need flexible solutions that can start as simple OT data projects that meet unique needs, without morphing into a massive IT project, breaking down at scale, or limiting future needs.
Building from the edge up
For years, manufacturing IIoT initiatives started as massive, top-down, enterprise or cloud-centric projects that required immense amounts of money, time, and infrastructure. These projects took years to define and implement, with results often realized much too late, if ever, because 75% of these IIoT projects failed to achieve success.
A better approach, made possible with today’s modern edge solutions, is for end users to begin their digital-transformation journey by harvesting the most pertinent data locally at the edge, and then using it to create immediate results. This method is incremental, manageable, and effective—but only if the solution can be scaled up, and even connected to the cloud when it makes sense to do so.
The cloud can and should be considered in any long-term vision, but these types of higher-level resources must have a clear purpose. The cloud is the best option when massive amounts of unstructured data need to be analyzed over a long period of time. Using the cloud in conjunction with the edge delivers data analysis in full fidelity at the edge, with the most pertinent information forwarded to the cloud. This enables the best of both worlds—rapid improvements on the plant floor, along with long-term continuous improvement empowered by the cloud.
Manufacturing-centric edge connectivity
Turnkey hardware/software edge solutions are the key to taking the risk and complexity out of selecting and deploying IIoT platforms, and for helping end users quickly develop beneficial digital transformation projects starting at the edge.
The most advanced version of these solutions are edge controllers. These are a modern evolution of programmable logic controllers (PLCs), able to perform traditional deterministic control, but also able to support extensive data connectivity and processing at the edge. With the right suite of apps, an edge controller can collect, store, process, share, and visualize data, and even provide closed loop, autonomous control functions.
When more extensive processing capabilities are needed, or when dealing with legacy equipment with an existing controller, an edge computer is the logical choice. These applications include advanced analytics, artificial intelligence/machine learning (AI/ML) applications, human-machine interface (HMI) functionality, or even locally hosted supervisory control and data acquisition (SCADA) systems.
An edge solution should provide easy-to-use tools for creating insights, and built-in dashboard elements to display results. While many open-source tools are available today, they are often met with concerns by IT personnel about long-term support and security.
However, a purpose-built IIoT enablement platform, incorporating tools such as Grafana for visualization and Node-RED for visual data-flow configuration, offers the same benefits, along with the assurance of interoperability, regular upgrades, and recognized security that users have come to expect. These solutions can also be augmented with more-advanced connectivity, HMI, and SCADA applications hosted on an industrial PC (IPC). This flexibility allows users to direct their efforts completely toward the IIoT projects they need, while avoiding the burden of experimenting with and proving-out basic functionality.
Building from the edge up using proven, fit-for-purpose hardware/software solutions is the best way for operating plants to use available resources to quickly achieve localized results. As confidence is gained and successes achieved, a proper solution can be seamlessly scaled up within the plant, across the enterprise, or to the cloud. EP
Derek Thomas was formerly Vice President of marketing and discrete sales for the Emerson, St. Louis (emerson.com), machine automation solutions business. Prior to Emerson, Thomas also worked as an engineering leader at Procter & Gamble. He holds an MBA from Washington Univ., St. Louis, and a Bachelor of Science in mechanical engineering from Purdue Univ., W. Lafayette, IN.