A new architecture for control systems being developed
Engineering and IT Insight: ExxonMobil is working to define a new architecture for distributed control systems (DCSs) by creating a system made up of programmable devices and distributed control nodes (DCNs) that can collective execute multiple layers of control strategies.
A chance to change the entire control industry is on the horizon with the ExxonMobil initiative to define a new architecture for the next generation of control systems. The current architecture for distributed control systems (DCSs) and programmable logic controllers (PLCs) were defined in the 1970's and 1980's when computers were much bigger, slower, and far more expensive than today's standard commercial systems.
The software industry has also changed in the past 30 years as it has moved from simple text editing in assembly language, Fortran and C, to fully integrated development, test, configuration control, automated build, and release tools supporting multiple scientifically developed computer languages. Today we also have deep libraries and tools for databases, HMI, and networks. Today, the average programmer is 10 to 20 times more productive than 30 years ago. The classical DCS/PLC system (CD/P), a box with one or more processors and a lot of inputs and outputs (I/O), has become smaller, cheaper, and faster but is still based on a 30-year-old architecture when computers were slow and expensive.
These advances that are shaking up the world through connected devices have not penetrated the barrier of obsolete industrial system architectures, however. The ExxonMobil initiative is working to break down the barrier and fix the major limitations of the CD/P architecture. One of the major limitations addressed is the inability to incrementally update and expand the control capabilities of the CD/P systems. Over a period of 30 years, most major and minor pieces of equipment in a facility will be replaced and upgraded. The part that essentially remains the same is the CD/P system. Replacing a DCS or PLC usually involves a complete shutdown of production, for an extended period of time, and a rip and replace of the current CPU boxes with new CPU boxes. It may even require a rewrite or redesign of current control strategies. These upgrade efforts are so difficult, expensive, and time consuming that many companies just live with the limitations of the memory and performance constraints of their 20- to 30-year-old systems.
The ExxonMobil initiative defines reference architecture for a DCS. A system is made up of intelligent, programmable devices (pumps, valves, motors, etc.) and distributed control nodes (DCNs) that collectively execute multiple layers of control strategies. Each control strategy, defined using any IEC or information technology (IT) language, is split into independent and sequenced control elements. Each element is then executed in one of the intelligent devices or DCNs.
Each intelligent device and DCN contains less computing capability and memory than a current low-end smartphone, but this is enough for a DCN to execute part of a control strategy and hold an entire facility's control strategy. Even a small amount of available memory, less than 1 GB, is enough space to hold all of the control strategies for a large refinery. Just as a biological cell holds all of the "code" for a plant or animal, but expresses only the part of the code for its "role," a DCN holds all of control strategies for a facility, but executes only the parts assigned to its "role."
Over time, as DCNs and intelligent devices are replaced with newer and faster versions, the overall system's control capability and capacity grows. Each new DCN or intelligent device imports the entire control strategy when it joins the network, and then assumes its unique individual role. Additional layers of control strategies can be incrementally created, distributed to DCNs and executed in the sea of processors that make up the truly distributed control system. There is no need for "big bang" upgrades and corresponding production shutdowns, and there is no need to restrict control strategies to those provided by a single vendor.
As with any "world changing" initiative, there are a lot of naysayers talking about the ExxonMobil initiative. There are naysayers in both end-user companies and vendor companies. End users, almost exclusively older engineers, may not understand the impact that the 4, 5, and 6 order of magnitude changes in capabilities have brought. Some have even said that what they have is perfectly fine and there is no reason to change. Vendors, especially those in market leadership roles, also see little reason for change. These vendors are afraid of the change because it will change customer expectations and place their market position at risk.
It is time for end users and vendors to dare to dream about new control system architectures. The naysayers said you could not sail around the world without falling off the edge, that humans would never fly, and man would never walk on the moon. Only because some dared to dream of better solutions did we achieve what was considered impossible. Today we are at the precipice of a new control system landscape. It will take the bold and adventurous to take advantage of the incredible and once-in-a-lifetime opportunity offered by the new information technologies. Be bold, and dare to dream about what can be done using today's new IT capabilities.
Dennis Brandl is president of BR&L Consulting in Cary, N.C., www.brlconsulting.com. His firm focuses on manufacturing IT. Contact him at firstname.lastname@example.org. Edited by Chris Vavra, production editor, Control Engineering, CFE Media, email@example.com.
The ExxonMobil initiative has been created to define a new architecture for the next generation of control systems.
Each control strategy is split into independent and sequenced control elements and then executed in one of the intelligent devices or DCNs.
What other developments or tweaks could be done to make this initiative even more useful for engineers?
This posted version contains more information than the print/digital edition issue of Control Engineering.
At www.controleng.com, search Brandl for more on related topics.
See other articles for 2016 at www.controleng.com/archive. See other Manufacturing IT articles.