New Innovation for Process Automation Architecture

Overview

ExxonMobil’s program to develop and prototype a highly interoperable multi-vendor process automation system has attracted attention in both the automation supplier and end user communities.  At ARC Advisory Group’s 2015 Orlando Forum, ExxonMobil presented the broad outlines of the idea and its underlying motivation and business case for its new innovation for process automation architecture   At this year’s ARC Forum, just after awarding a contract to Lockheed Martin for the prototype development, ExxonMobil explained both the motivation and the architecture in much greater detail. Since then, the program has been a major talking point within the process automation industry. For more background, please refer to the recent ARC Insight ExxonMobil’s Quest for the Future of Process Automation.

One result of all this discussion and publicity is that other similar initiatives have come to light.  Indeed, ExxonMobil is not alone either in its vision for a future process automation system or in undertaking research and development to bring about such a system. Another company doing very similar work (but quietly and privately) is state-owned, Saudi Aramco. Aramco is the world’s largest oil producer and like ExxonMobil easily has the size to undertake such an effort.

This report outlines Saudi Aramco’s motivations, architecture, research, and development efforts and compares these to the corresponding work done by ExxonMobil. From a purely technical perspective, the two company’s initiatives are startlingly similar.

Saudi Aramco has a long history of automation and DCS mega-projects, both in-Kingdom and internationally.  In addition to its huge upstream production capacity, Aramco has a refining capacity of roughly 4,000,000 barrels/day. The company also provides feedstocks to huge new petrochemical facilities, some of which it partially owns via joint ventures. Many of the company’s older mega-projects are now in their second generation of automation systems.  As a result, the company has experience with both greenfield and brownfield installations, most of them very large.  Saudi Aramco also has experience with multiple automation vendors in its supply chain and, in this respect, has broader and more typical experience than ExxonMobil. Aramco’s experience with large process automation systems is decades long, broad, and deep.

Aramco Program Motivation

The main motivation for Aramco’s experimentation was frustration at the inherent limits of a multi-loop DCS controller.  Such controllers naturally have fixed limits on the available compute, storage, and sensor I/O resources. While these resources are typically designed in a balanced manner, the designed-in resources are far from optimal for applications such as compute-intensive advanced process control (APC) and optimization strategies. 

Aramco wants to easily and effectively develop applications that extend beyond the boundaries of single DCS controller or beyond the bounds of a single supplier’s DCS installation, while at the same time retaining the virtues of existing DCS controllers – a stable platform with very high availability and rigidly deterministic performance. In the company’s view, the way to achieve this was to move to a “flatter” system architecture and away from the rigidly hierarchical layered model of DCS (I/O, controller, server).  Doing so would avoid concentrating (and isolating) critical real-time process data within DCS controllers.

Aramco Architecture

Figures 1 and 2 show a typical present day Saudi Aramco DCS system and the envisioned future Collaborative Process Automation system.  The obvious difference is in the field equipment. The existing DCS I/O and controllers have been replaced by sets of “smart” junction boxes.  These handle field device I/O and serve as a conduit for device management.  Otherwise, all the functions performed by DCS controllers now take place in servers located in a unit-level or plant-level server facility. In the new architecture, all the functions handled in the server room relate to mission-critical real-time operational technology (OT)-related activities, just as DCS controller functions are today.  This includes I/O processing and computation of inferential measurements. 

Saudi Aramco has made no specific references to virtualization technology, but given these critical functions, a high-availability server capability (with six or seven “9s” of system reliability) would be required.  In ARC’s view, Aramco has not yet defined the requirements for this server environment because it instead chose to demonstrate the feasibility of existing network communication tools to serve such an architecture.

With regard to networks, it is also worth noting that the “control network” of the new architecture is far more mission-critical than even the “plant control network” of the existing systems.  Unlike the existing design, the control network is required to be in service for any process automation to operate.  It is absolutely mission-critical.

Comparing the Programs - Similarities

The technical similarities between the Aramco and ExxonMobil programs are striking and worth noting, especially since to the best of ARC’s knowledge the two efforts have been conducted independently:

• Motivation - In both cases the rigid hierarchy of field device>controller>server was judged to be an unnecessary legacy constraint given today’s technology, and one that persists in present-day DCS products.
• Server Role - Both programs assign major roles to high-availability servers.  Aramco certainly counts more on this, but both programs envision what today are “base level” process automation functions moving to servers operating in a high availability on-premise cloud service.
• DDS - Such a system requires a scalable and highly reliable real-time data distribution service for networked communication between physical devices.  Both programs have chosen the publicly-specified DDS technology for this role.
• Safety Systems - Both programs have limited their scope to automation (DCS) functions rather than safety shutdowns or other safety applications.

Comparing the Programs - Differences

While the automation systems envisioned by these two programs are remarkably similar, ARC notes some key technical differences:

• Automation Function Hosting - By far the most striking difference between the two architectures is that Aramco envisions all process automation functions moving to servers, while ExxonMobil sees process automation tasks migrating to either servers or the new “Distributed Control Node” (DCN) module, which is essentially a networked single-loop digital controller.  This difference does not make the architectures incompatible.  On the contrary, the system extensibility envisioned by both would enable any such modules to be added.  The ExxonMobil architecture does not specify a module type for “monitoring only” process measurements.  The Aramco architecture does not specify any local control functionality. 
• Architectural Difference - ARC believes that this architectural difference may be caused by the layout of the existing systems that each company plans to replace.  In ExxonMobil’s case, existing field device wiring has been brought back to DCS rack rooms.  ARC is not privy to the configurations of existing automation systems in Aramco’s installed base, but the trend toward using remote DCS I/O would likely be present in newer parts of Aramco’s installed base, but not in the older DCS installations that ExxonMobil has targeted for replacement.
• Program Stage - Based on first impressions, it appears to ARC that Aramco has a slightly larger experience base with this technology. Much of Aramco’s work has been to test the performance level of such a system at larger scale and to prove to its satisfaction that existing commercial DDS software products are suitable for DCS-sized process automation applications.  Aramco also has tested a pilot application in a non-critical system, but this size of pilot system is significantly smaller than a typical unit DCS.

Recommendations

In ARC’s view, the amazing similarity of these automation visions and programs, plus the fact that they were developed independently adds credibility to the concept of a process automation system consisting of such networked devices (of various sizes and form factors) and servers.

Note also that despite the architectural differences, the type of systems envisioned by both the Aramco and ExxonMobil programs are highly compatible. 

Finally, it is clear that these companies (and others that have done similar work) should find a means to collaborate that will enable them (and their suppliers) to further develop this vision and begin to work out best practices for employing it in industry.  What is unclear at this stage is how Saudi Aramco plans to partner with other end users and/or technology suppliers for this initiative.  ARC looks forward to learning more about the Saudi Aramco Collaborative Automation program in the future and sharing this information with our clients.

If you would like to buy this report or obtain information about how to become a client, please Contact Us

Keywords: Saudi Aramco, DCS, DDS, ExxonMobil, High Availability, Operational Technology, Process Automation, ARC Advisory Group.