The choices for a controller are as diverse now as they have been at any time in the history of industrial control systems. Many types of controllers are available for discrete, process, and hybrid control applications, including programmable logic controllers (PLCs), programmable automation controllers (PACs), distributed control systems (DCS), industrial PCs (IPCs) or PC-based control, and embedded (board-level) controllers. And the lines that differentiate them continue to blur. What criteria should be used to choose among them and why?
As control system engineers who have designed many utility-based control systems, we see a large variety of controllers even within the same plant. This diversity seems to indicate a lack of standards, which leads to a lack of efficiency and reliability with regard to installation and ongoing maintenance of control systems. Because controllers sometimes last for more than 20 years, they often are forgotten behind gray panel doors where no one sees them and they tend to run until failure. Out of sight, out of mind, as the saying goes. Many times, there are plant expansions or retrofits where a vendor or contractor may install a later version of a controller-or even change to a different controller-guided more by personal preference than overall control system maintenance. Such was the status of operational technology (OT) for many years. Reliability was the key. The emergence of Windows-based human-machine interfaces (HMIs) required controllers to add Ethernet connectivity, but even then, control systems were largely isolated from the rest of the business. [subhead]
Controls for the 21st century
However, Internet connectivity now drives many technology decisions. More people are relying on smartphones and tablets to access data while on the go. Information technology (IT) changes have exploded and OT and IT are converging. The Industrial Internet of Things (IIoT) also has moved to center stage and businesses need to combine OT and IT to make it work. Control systems engineers are well-poised to help this merger because they understand plant floor operations as well as real-time data that operates in millisecond timeframes.
The convergence of OT and IT now drives the decision as to which controller best fits a particular application. Originally, control system engineers and technicians decided which controller best fit their application and others in the organization didn't interfere. Hardware selection was based on factors, such as processing capability, controller memory, and modularity where input/output (I/O) could easily be added as needed. Consideration was given to availability of spare parts or support. It was up to the programmer or system integrator to use the device with which they were familiar. Sometimes as long as a device was "Ethernet capable" or "Modbus capable," that was good enough and that is why companies ended up with so many different controllers.
That decision model is no longer suitable for industrial control system (ICS) technology because it now evolves so quickly. And the data affects and requires input from more people. As a result, control system selection priorities have changed.
Historical controller selection criteria
In the past, criteria for control system selection decisions focused primarily on hardware and how it could be supported. They included:
Modern controller selection criteria
Today, control system selection decisions focus less on hardware-centric criteria and more on networking, security, and standardization. Modern ICS selection priorities include:
Control systems with inherent networking capabilities that include embedded cybersecurity features: It's hard to imagine not having data at our fingertips. That would not be efficient. Companies need staff at all levels to have access to real-time data about operations-from the plant floor to the C-suite. Air gaps are no longer viable. In fact, as long as humans are involved, no true air gap exists. Controllers have been introduced within the past year that have deep encryption methods, yet can continue to provide local control functionality during denial of service attacks on their communications capability. They also may include component supply chain manufacturing security keys to ensure the highest levels of hardware and software authentication.
These features alone have changed the game in terms of control systems. A manager of a utility made this eye-opening statement, "Now, the control system decisions are no longer made just by operations and technicians staff, but it is up to the management to protect our utility's infrastructure by selecting secure control systems platforms." He was wise to arrive at that conclusion, move to new control system standards, and be willing to present that challenge to the utility's board so it could take proper action.
Robust programming languages for rapid application development: By having a complete tool set for controller programming, the integrator already has many advantages for efficiency and reliability. These are best expressed in IEC 61131 with function block (FB) diagram, structured text (ST), sequential function chart (SFC), and ladder diagram (LD). Having open (nonproprietary) programming standards that can now be incorporated with deeply-embedded security keys will aid control system engineers and system integrators to confidently deliver the solutions that businesses need. Be sure to check the licensing cost of your programming software. Some manufacturers are breaking free from the licensing models and offering integrated development environments that are free, allow immediate downloading, and can be used immediately.
Scalability, standardization across the enterprise, and reduced part count: If this appears to be competing interests, check again. At least one manufacturer whose systems we install has broken the mold when it comes to leveraging what was once competing standards. Scalable architecture with one backplane (in five-slot, 10-slot, and 20-slot options), one power supply, one blazing fast CPU module, and a few I/O modules (including a universal module that is software-selectable as input or output and discrete or analog). The platform is scalable to any size while reducing part counts to a dozen or so components, rather than having to select from thousands of part numbers like some vendors expect of you.
Standardizing on one platform throughout the enterprise can have a significant benefit. Moving data between controllers for proper automation functions can be a hassle. Having a single platform that can reach from field remote terminal unit (RTU) all the way up to plant controllers with DCS functionality of programming and HMI integration could yield benefits for years to come.
Even the protocols have changed: Even as fast as software applications change, historically, the base-level protocols were fairly constant-not so anymore. Publish/subscribe protocols have changed the way we implement networks. No more hammering the networks by polling for data continuously. With the pub/sub model, we have the ability to push the data-when changing data necessitates-to anywhere in the enterprise: to the ICS, HMIs, enterprise resource planning (ERP), inventory control, and so on. Protocols that do this are available today with message queuing telemetry transport (MQTT) and are in process with OPC UA (OPC-UA stands for OLE for process control, unified architecture).
The right expertise: Having the right expertise on hand to implement modern control technology is still key. Whether that expertise is in-house staffing, or outsourced engineering and integration, users will benefit from having a team that is responsive and agile at quickly meeting business needs. Training in industry standards such as IEC 61131-3 also will be a plus.
There are many disruptive technologies that affect the ICS. We are asking more of the industrial control system ICS than ever before. Staying with the status quo is no longer good enough. For digital natives who grew up in this space, the selection criteria will be second nature to them. This requires us to think differently about control system platforms, the vendors, and partners to determine which systems can now meet these challenges as we look ahead to increased growth, reliability, and cybersecurity of the ICS and the infrastructure it controls. It is time to stop thinking about hardware and start demanding the open and secure automation platform and tools required for the digital enterprise.
Read the case study to learn how a utility company upgraded their SCADA and RTU.