Gaining the edge in automation

Edge computing advances give users more options for architecting automation systems as well as flexible communication and programming choices.

By Benson Hougland March 11, 2019

Traditional automation architectures are built around centralized programmable controllers connected to remote field devices and instruments. However, this concept is shifting as computing power is progressively embedded near the edge of automation systems using new types of intelligent components.

Here are some of the advantages offered by edge computing designs as compared to traditional and more centralized strategies.

Combine and conquer

Edge computing benefits designers and manufacturers through consolidating components and configurations. An edge programmable industrial controller combines input/output (I/O), control, data processing, communications, and human-machine interface (HMI) functionality that can be located near or on machines, process trains or smart equipment (Figure 1).

An on-board touchscreen display, with options for a larger local monitor, is useful for machine manufacturers since control and visualization are effectively merged. This makes commissioning, operating and troubleshooting easier since a separate PC is not required. Local operator functions such as pushbuttons and indicators can be handled on the HMI more effectively than traditional pilot devices.

Edge computing makes it easier for users to incrementally roll out system improvements. Instead of affecting large in-service installations, an edge computing system can be installed and tested locally and merged into the larger system via a quick hot cutover.

Shifting communications and data processing to the edge makes sense for many reasons. Control decisions are made in real time where they are needed. Data is obtained, pre-processed, and analyzed near the source. This reduces the required network bandwidth, data storage, and processing power upstream of the edge component.

Programming options

Programming for these edge components can take many forms. Traditional programmable logic controller (PLC) users usually will look for ladder logic or other IEC 61131-3 programming languages. However, a flowchart-based programming language is often better suited for the application.

Python or C/C++ might be preferred for more advanced calculations and data processing. Some edge computers can accommodate these programming languages and others.

Universal translator

Communications flexibility is another hallmark of edge computing. An edge programmable industrial controller is equipped with various communication ports and supports a wide range of protocols so it can connect with numerous local intelligent systems such as PLCs.

Edge computing components support operations technology (OT) protocols such EtherNet/IP, Modbus, BACnet, those from OPC Foundation, and others. It also supports information technology (IT) protocols and development tools such as TCP/IP, simple network management protocol (SNMP), message queuing telemetry transport (MQTT), and Node-RED. This suite of interfaces effectively “flattens” and simplifies the system architecture (Figure 2).

Figure 2: Edge computing solutions, like the Opto 22 groov EPIC, combine I/O, control, communications and HMI functionality to flatten the control system architecture. Courtesy: Opto 22

With these integration abilities, edge computing is ideally suited for gathering the right data in the field, processing it into useful information, and securely communicating it to higher-level data architectures. Edge computing can interact with other supervisory on-site automation platforms, enterprise databases, or even exchange data with cloud services.

Familiar features

Classic automation systems relied on centralized and dedicated HMI hardware and/or software, sometimes proprietary and usually very expensive. Today’s users have been trained by their home computers, consumer devices and smartphones to expect rich HMI options almost everywhere. Because edge computing can deliver HMI content via a mobile-ready web server, it delivers the options users expect for interfacing with the automation system anytime and anywhere. This gives users a more personal and familiar connection with the automation system than previously available.

Picking the right control system

When designing a control system, users select the best combination of features from a variety of options. Historically, that has meant relying on more centralized platforms, which are functional, but also impose constraints.

However, the expanding abilities of edge computing give designers and engineers many new options. Edge computing can combine as much or as little I/O connections, control, communication, data processing and HMI visualization functionality as needed.

As a result, edge computing is often an ideal solution for users desiring granular, modular, and scalable architectures and ones that fulfill many requirements.

When specifying edge computers ask if it can it:

Serve as a standalone controller and HMI for one machine?
Replicate one machine to many and enable them to communicate with each other and with local plant systems?
Extend control to a larger supervisory system and publish data to the cloud? If so, can it publish some data now and the rest later?

While it might not be perfect, edge computing’s flexibility makes it a very viable option for many different applications.

Benson Hougland is vice president of marketing, Opto 22. Edited by Chris Vavra, production editor, Control Engineering, CFE Media, cvavra@cfemedia.com.