By Rich Merritt
What’s a distributed control system? We asked end users, systems integrators and vendors and got answers ranging from “a control system with two PLCs connected by a network,” to “DCS doesn’t exist any more. Modern systems are (insert DCS vendor’s euphemism here).”
One reason we can’t precisely define a DCS is because they have changed so much since 1975 when Honeywell and Yokogawa introduced them. In those early days, DCSs used multiloop digital controllers to perform regulatory control.
But much has changed in the past 33 years. What were once simple control systems are now decision-support and decision-automation systems. Users want systems that will put more people and functions—engineers, managers, maintenance, planning, environmental, and the supply chain—in touch with the process. We want to gather more data, hold it longer, and turn it into useful information. We also want to report problems on an exception basis in a timely manner to the right people—not just the operator—and to automate complex decisions, such as real-time process optimization.
Going Open with COTS
A modern sub-$1,000 PLC can perform regulatory control, probably as well as any DCS—but it can’t perform all the high-level functions. For that, it needs a PC supervisor and a lot of software. Enter the HMI/SCADA vendors, who have added all the necessary high-level software and emerged as a formidable competitor to DCSs.
Figure 1. A “typical” DCS system, depending on whom you ask.
The days of proprietary control systems are over—almost. According to Peter Martin, vice president for strategic ventures at Invensys Process Systems (IPS), the trend toward commerical off-the-shelf (COTS) systems began in 1987 when Foxboro introduced the I/A Series control system and with it the concept of using standard hardware, software and communications in a control system. (For a discussion of “Should Control Engineers Do Windows?” scroll to the bottom of page 3 or use the following link www.controlglobal.com/0812DCS.html.)
Todd Stauffer, PCS 7 Marketing Manager at Siemens Energy & Automation, says, “In 1998, the DCS was coming out of the proprietary age. The move toward open systems and commercial off-the-shelf technology was driven by users and suppliers alike. This transition has allowed the pace of technology innovation to increase. We would have never made it to this point if we had stayed proprietary.”
Proprietary networks have been replaced with various versions of Ethernet and COTS routers, switches and wireless devices. Computer hardware is all standard stuff, all based around the same basic hardware and Windows-based software. In fact, some DCSs use conventional controllers and HMI/SCADA software, says Jim Ford, director at system integrator Maverick Technologies. “DCSs are thrown together all the time with PLCs for basic control and software, such as Wonderware used for the HMI.”
Thanks to their conformity with fieldbus, HART and other industry standards, most process sensors and transmitters have nothing proprietary about them. You can use anybody’s hardware in your control system.
The major DCS vendors also use fieldbus because they have the necessary hardware and software. While fieldbus is not proprietary, only the big DCS vendors have made the necessary investment in things like fieldbus protocol stacks and H1 cards. Few, if any, of the competing HMI/SCADA control systems offer fieldbus, but this will change as soon as fieldbus hardware becomes available for PC-based systems.
So what’s still proprietary? The hardware controllers and I/O from most major DCS. vendors.
And software, of course. Or is it? Some DCSs use historians, asset management programs, loop-tuning programs and ERP software from the same suppliers—as do several of the HMI/SCADA-based control systems, such as Wonderware, Citect, Indusoft and Iconics. Both the big DCS and the HMI/SCADA-based system vendors claim to have completely integrated software from the loop controllers to ERP.
From an architecture standpoint, the major difference between the two is that DCS systems use proprietary hardware, while HMI/SCADA systems use conventional hardware, such as PLCs, panel PCs and hybrid controllers.
HMI/SCADA: A Viable Competitor of DCS
From a functionality viewpoint, an HMI/SCADA system can offer exactly the same capabilities as a DCS (except for fieldbus). An HMI/SCADA system uses the same field devices, the same Ethernet networks and the same COTS-based HMIs as a DCS. An HMI/SCADA system has the same type of software available as a DCS.
To counter claims that HMI/SCADA can’t handle “big jobs,” Scott Kortier, marketing communications manager at InduSoft, points to Brookhaven National Labs, in Brookhaven, N.Y., with its Relativistic Heavy Ion Collider (RHIC). Indusoft recently upgraded the RHIC’s control system. “This system includes over 10,000 I/O points, a dozen operator stations plus 15 security-enabled, browser-based displays,” says Kortier.
Tim Van Wyk, a development engineer with Citect, agrees. “SCADA can provide the offerings that have been perceived to only come from a DCS solution,” he says. “With an open SCADA system, customers can use partners that bring the best industry experience and advanced technologies to their SCADA solution without relying solely on a single source, as with a DCS. When customers need to quickly adapt to changes in their business environment, an open SCADA provides the flexibility, which can be difficult and slower with a closed DCS system.”
Naturally, DCS vendors disagree. Randy Reiss, a DeltaV developer at Emerson Process Management, says, “There are inherent limits to the throughput of the communication of SCADA level products. The vendors of SCADA-level products are subject to the operation of technology below them to send and receive data, resulting in a less-coordinated communication subsystem. A major benefit of a DCS is that the system is all developed together. What is usually the truth is that there are a bunch of little systems that have minimal communication between them. That means that an operator cannot call up a display from outside the immediate area or trend data across the plant to look for correlations. Try and get end-product quality side by side with raw material data…good luck!”
Citect’s Van Wyk counters, “There is no separation of data across little systems. Users can leverage data from components onwards to the open MES product expanding the total solution to the business enterprise.”
DCS vs SCADA
The “classical controls vendors” who produce DCSs bring extensive industry experience, an army of engineers and a completely integrated hardware and software solution to the table. They are formidable competitors to HMI/SCADA systems, and they are at the cutting edge of control.
“It’s the DSCs that are at the forefront of technology and are the future of control,” says Randy Reiss. “It’s the DSCs that are blazing the trail and designing the systems that enable advanced technologies like multivariate statistical process monitoring, MPC and wireless. It’s all in a DCS before it’s in a SCADA product. So if you want to see the future of control, look at the big DCS venders. They are the ones throwing things on the wall to see what sticks.”
That knife cuts both ways, says Van Wyk. “With open SCADA systems running on Windows, users get to leverage not only their industry-specific software, but also access worldwide software developments. When running a DCS, users have to wait for software to be redeveloped within the closed DCS software environment.”
DCSs also can do top-to-bottom integration for both hardware and software. Harsh Chitale, vice president, strategy and global marketing at Honeywell Process Solutions, says a DCS, “…marries information management with knowledge-based technologies, including optimization, model-based control, simulation, abnormal situation management, scheduling, wireless and advanced sensing, and advanced applications, and delivers that them in a form suited to industrial manufacturers.”
Roy Tanner, Systems Marketing Manager at ABB, says, “DCS advantages include pre-integrating/testing various components and having single stores of data and information. This has expanded beyond traditional HMI and control to other functional areas such as safety-instrumented systems, electrical system integration and asset optimization. Also, DCS systems are released as a pre-tested, pre-integrated ‘system’ which reduces complexity by NOT having multiple database synchronization, backup and disaster recovery issues, various software revision levels, and multi-vendor support issues.”
“There will always be a need for the more tightly integrated system that a traditional DCS provides,” says Dave Emerson, Director of Yokogawa’s U.S. Development Center, adding “This scalable level of integration provides the benefits of an integrated DCS and the connectivity of traditional SCADA designs.”
On the other hand, HMI/SCADA suppliers typically rely on a network of systems integrators to install and configure their control systems. This is not a bad thing, because many systems integrators are just as qualified as the DCS vendors’ engineers. However, SIs are often constrained by their customers’ preferences toward certain vendors.
Ed Diehl, executive director of Concept Systems, says, “We work with our customers to identify the best architecture for their application and circumstance. In virtually every case, there are numerous control technologies that will do the job. The key is to find the best one for that customer and to integrate it properly.”
But in the end, “A key ingredient to the concept of DCS is homogeneity,” says Dennis Kilgore, president of DLL Solutions. ”This comes somewhat from the need to have a ‘single neck to choke,’ instead of vendors pointing fingers in order to escape blame.”
And therein may lie the primary reason for the continued success of DCSs, in spite of their high cost—three times that of a comparable HMI/SCADA system, by one estimate. A DCS vendor can supply a completely integrated system—all from a single vendor who takes responsibility for the delivery, configuration, startup and maintenance.
So what’s in store for future distributed control technology? Clearly, the march to open systems continues. There are no significant hardware differences between the proprietary controllers of DCS vendors and the PLCs/PACs of the HMI/SCADA systems. So it is safe to predict that a universal open controller platform will emerge.
Mark Nixon, chief architect of DeltaV at Emerson Process, says one of the catch phrases these days is smart plants. “These smart plants operate to tighter specifications, involve a much greater understanding of the processes, greater automation and decision support, more expansive use of automation and data interpretation. ”
ABB’s Roy Tanner says control is moving to the field. “Due to technology and the adoption of standards, one of the trends we’ve seen is the distribution of control elements further into field devices,” he says. “If technology continues, unified fieldbus architectures with even more intelligent devices will change the role of the today’s process controller to that of a process ‘coordinator.’”
Peter Martin of Invensys calls this new breed of control system an enterprise control system (ECS) instead of a DCS. He says an ECS could include a conventional DCS, SCADA or other control technology, but it must also involve an architecture which enables full plant floor interoperation; open communication access across the enterprise; support for asset performance management tools that enable unfied maintenance and operations management; and a unified engineering environment across all plant floor domains.
“But there are forces in play that could affect the evolution of the DCS,” Stauffer warns. “Cybersecurity issues are responsible for ongoing changes in control system architecture and are making people think twice about implementing wireless. In addition, experienced process control engineers, operators and maintenance people are leaving the workforce. The next generation of DCS will have to compensate for this loss of knowledge, so it will be much more of a knowledge-based system.”
In the future, we will continue to see DCSes and HMI/SCADA solutions coming closer together until you can’t see the difference in technology. At that point, the purchase decision will come down to this: Who do you trust?
Rich Merritt is a contributing editor at Control.
Why Control Users Diss Windows
One fly in the ointment for both DCS and SCADA-type systems is that almost all of that high-level software is Windows-based.
The first DCSs were developed to solve the problem of computer failures in minicomputer-based controllers. When the central system failed, it usually shut down the process. A DCS “distributed” the control to multiple devices, so no single failure could take down the system. Today, we have, in a sense, come full circle—instead of unreliable computer hardware we have—in some users’ minds, at least—unreliable software.
Many of today’s DCSs use the Windows operating system in their PC-based HMI workstations, and some even use versions of embedded Windows in their hardware controllers. The reason, of course, is because Windows makes it easy to connect to networks and software at all levels, from field controllers to high-level ERP and asset management programs. This has some users concerned about reliability.
“In the power industry, Windows is hated,” says a control engineer in the power industry. “It is being shoved down our collective windpipes, and we fight back. This is not an ideological fight, as the open-source zealots make it appear. It is a quality issue. Windows is the Wal-Mart of operating systems and built only for profit and market share, in my opinion, and in the opinion of many, many others.”
“Our Brand new Emerson Ovation DCS does not have Windows in it in any way,” he continues. “I, as many others, will not allow Windows in our DCS system. It is not for my mission-critical applications. Our system is based on Sun Solaris.”
An anonymous process control engineer in Alabama echoes these concerns. “I just replaced a 1995-vintage Bailey OIS42 with a modern DCS,” he reports. “The system is up and running and our operators are using it. But none of them have thanked me for replacing the OIS42s.”
He says the primary reasons are that graphics and faceplates take one to two seconds to appear after selection, which is much slower than the old system, and the Windows-based software is not robust enough to avoid the need for occasional restarts. “The vendor’s engineer told me, ‘Remember, it is a Windows system. It may need rebooting occasionally.’ Hmph! DEC Alphas don't need periodic rebooting!” he says.
“In their desire to be able to interface with or replace any system, and to say that they're using off-the-shelf hardware, they've created monsters which are more difficult to configure, slower to operate and possibly less reliable,” he adds.
Even a systems integrator agrees. Dennis Kilgore, president of DLL Solutions does data historian integration and interfaces with both HMI/SCADA systems and DCSs. “We work closely with customers in the power generation, wind generation and T & D sectors,” says Kilgore. “We get an earful every now and again from customers sharing their frustrations as they relate to DCS and controls. I would say that Windows is ‘not ready for prime time’ in the role of a process controller. The concept of real-time and 24x7x365 is more than a slogan―it is an absolute requirement.”
It is probably fair to point out that Windows is not running the regulatory control part of a DCS. That’ s done in the fieldbus devices and the hardware controllers. Windows just runs the HMI and high-level software. If it fails, as is its wont, all it takes down are the operator displays.
As you might expect, DCS vendors are quick to respond.
Todd Stauffer, PCS 7 Marketing Manager at Siemens Energy & Automation, says that Windows has gotten much better over the years. “The stability of the Windows platform in process control applications has improved greatly with the transition from NT to 2000 to XP,” says Stauffer. “The majority of the user feedback we have received is around the extra effort that is required to maintain Windows-based systems, particularly software patch management. This is one of the tradeoffs that goes with the adoption of COTS [commercial off-the-shelf] supplier PC hardware and open systems. The PC hardware is much cheaper, but the maintenance and upkeep of the PC is more significant, especially with growing concerns around cybersecurity and the need to keep up to date with the latest security patches and virus profiles. Much of our effort has been around making this software patch management process easier for our customers.”
Harsh Chitale, Vice President of Marketing and Strategy for Honeywell Process Solutions, agrees. “Open systems need more care and feeding and also discipline in setup, backup, network management, security, etc. In older proprietary systems, this was less cumbersome and often taken for granted.” He says that while the user industry quickly adopted open systems for their obvious benefits of cost, choice and larger resource pool, it is only now that it is adjusting to the additional support requirements that come with it.
The author of the Citect white paper, “From Telemetry to the Enterprise” (available on ControlGlobal.com) say, “While SCADA systems are subsuming more and more of the original functionality of the DCS, including in-plant, closed- loop control, alarm management, process optimization and data analysis, the DCS vendors are producing systems that look arguably just like SCADA systems, but continuing to call the systems DCS. For everything except the most critical control functions in the petrochemical industry, there is now no longer a difference between the capabilities of a DCS provided by a classical controls vendor and a SCADA system put together by a control system integrator.”