During the 20th Annual ARC Industry Forum in Orlando, representatives from ExxonMobil put forth a vision of a more open process control architecture—depicting it as a “system of systems” with an open method of communicating between existing control systems, new control systems and supervisory systems. The presenters provided little detail beyond that, and that was on purpose.
Open system goals
Apparently, ExxonMobil’s primary motivation is cost. Though the company recognizes that newer systems indeed offer value, it believes the cost of moving its intellectual property (IP) from existing systems to these newer systems overshadows the potential benefits gained. Therefore, part of the vision is a system that allows IP to be retained, regardless of the hardware.
The company also wants to address the cost of incorporating third-party solutions. By convincing the industry to adopt a more widely accepted open standard, ExxonMobil argues that integration costs will drop. The idea is to decouple those applications with high IP content from the specifics of the control systems installed.
ExxonMobil also wants to capitalize on mixing and matching best-in-class components from various suppliers, instead of compromising with the features of one supplier's all-inclusive package.
What about best-in-class hardware?
The announcement says little about hardware beyond its ability to host various functions. It focuses heavily on integrating control applications and hardware independence. The thinking here is that the IP is difficult to move from one system to another, but so are wires.
The distributed control node (DCN) in ExxonMobil's vision is a single-loop controller with universal I/O capability and the ability to communicate on a high-speed deterministic network with peers and supervisory nodes.
The building blocks have existed for decades, with a number of single-loop controllers still on the market. Any of these could be enhanced with the same I/O technology available in the leading distributed control systems (DCSs) to support universality and smart field communications. Most offer some sort of protocol to allow a remote user to configure and/or view the device information. This protocol could easily be changed.
Many suppliers of hybrid DCSs and programmable automation controllers (PACs) promoted their advantages over single-loop controllers by highlighting the ease of adding supervising logic, redundancy, HMI integration, and ease of engineering and maintenance. But are these differences insurmountable with today's existing technology? The ExxonMobil proposal harkens back to the days of racks of panel-mounted analog controllers or a 1970s-vintage DCS connected to a central minicomputer with lots of advanced computations coordinating across the unit.
In most current systems, most I/Os have similar specifications: environmental, D/A conversion, electrical, etc. Most major DCS manufacturers offer some form of late-binding I/O. A handful of manufacturers also sell generic I/O that could be used by a multitude of different controllers or even a software-based controller hosted elsewhere.
There is not much use of third-party I/O in process control systems, though. Why? The integrated I/O offers better functionality. Or does it? Given ExxonMobil's desired architecture, I/O for indication alone or even multivariate control could also be added to the real-time service bus. With a properly specified standard, I/O from multiple vendors could easily be integrated to meet a given need.
Technology already available
Hardware vendors seeking to meet the needs expressed through the ExxonMobil initiative have plenty of options. Much of the technology is sitting on the shelves now. The biggest challenge is finding the “secret sauce,” the unique value proposition, in the open systems environment. In the proposed architecture, there appears to be little value attributed to the frontline controls. The configurations are envisioned to be easily portable and the communications universal; the epitome of a commodity item.
When we consider today’s DCS architectures, the trend is toward consolidation. Controllers today can accommodate more I/Os than ever. The DCN concept would actually reverse this trend. The realities of today’s facilities would likely continue the centralized approach where control hardware is located in a room, because the wiring costs are too high and even hardened electronics survive longer under a controlled climate. Greenfield installations would permit greater distribution. Distributed I/O can be supported now, yet makes up only a small percentage of the installations.
Several available technologies could open up the architecture further. Current wireless technologies would allow the service bus to be extended to the field devices themselves, with the DCN functionality hosted in the field device electronics (similar to Foundation fieldbus control in the field). There is also an initiative that could extend to the service bus via a hardwired network solution to field devices, but it would need to support the current wiring infrastructure to be useful for existing facilities.
Though nearly all the technologies required exist now, vendors could take several paths to meet the vision. The question that end users really need to answer are: “What problems am I trying to solve?” and “What will I really buy?” The market has proven that if the upgrade to a new system doesn’t present appreciable short-term benefits and payoff, the architecture and long-term lifecycle costs won’t make a difference in the buying decision.
>>Mark Sen Gupta, [email protected], is senior consultant at ARC Advisory Group. He has more than 25 years of expertise in process control, SCADA and IT applications with companies such as Mobay, Honeywell, Plant Automation Services (PAS), CygNet Software and Invensys. He holds bachelor’s and master’s degrees in electrical engineering from the Georgia Institute of Technology.
