Beyond Process Control - DCS as Integration Platform

下載 (2.9 MB)

Think for a moment about what a DCS looked like 25 years ago. In 1990, Yokogawa's CENTUM system had already been available for 15 years. Like others at the time, it was a simple system by today's standards. It operated largely with proprietary hardware, software and networking protocols, and it interface with lots of analog field devices.

Functionally, it read information from field devices, executed loops, and adjusted valve actuators. It was isolated from other networks, and its ability to collect historical data was limited but it had moved past pen-and-paper chart recorders. If company managers wanted to know what was going on in the plant, they viewed it via reports printed on green-bar paper or visited the control room.

Life didn't stay that way for long in the DCS world. New technologies were emerging, things like fieldbus networks and HART. More sophisticated historians were gathering additional data, and management expected more kinds of information. HMIs were growing in sophistication and there was less need for operators to go out and perform manual operations. Computers [M.T1] and other off-the-shelf hardware were replacing proprietary equipment along with the invasion of Microsoft Windows and Ethernet.

Those kinds of changes have continued and DCS suppliers have had to keep pace. Some suppliers and users have taken an entirely defensive attitude, resisting developments whenever possible. Others have led the evolution, seeing and seizing opportunities to bring new functionality and improve manufacturing capabilities.

A DCS today has to help make a plant competitive by allowing it to operate safely and reliably at a high performance level, all without increasing headcount. The best systems also provide tools to capture institutional knowledge and build lessons learned into the automation systems.

Listening to Customers

Any automation supplier that wants to remain in business has to listen to its customers, especially when so many technologies are emerging simultaneously and there are so many pressures on end-user companies. There have been dozens of articles published on the pressures driving refineries, chemical processors, and the like. Feedstock sources are changing constantly, energy prices can fluctuate hourly, and highly-networked plants seem wide open to cyber criminals.

One element that comes into the discussions with customers again and again is the need for one system that can serve as the single integration platform that can pull information together from a wide variety of data sources as listed in Table 1.

When considering all the items listed in the table, this is a huge task that's only getting bigger. And, more often than not, that integration task is falling on the DCS. It still has to control the process—that basic functionality remains but there is so much more that's expected.

For at least some of these external sources, for example process skids, the DCS may be expected to perform real-time control in addition to data gathering.

While the list of functions grows longer, there is also a demand for simplicity as experienced engineers and operators retire, only to be replaced with a new generation that has much different expectations of life in a process plant. Anything that isn't immediately accessible intuitively can be a difficult learning process for these newer employees.

Integrating the Data

To many managers and operators trying to survive in a process manufacturing environment, the mention of "big data" probably draws a painful look brought on by the thought of terabytes of manufacturing data. Is it necessary to have second-by-second readings recorded from every process sensor? Must our company management in Singapore really be able to see real-time data on what's happening in a distillation column in Ahmedabad?

Some companies are drowning in data, while others are using it to improve manufacturing and profitability. One way to improve plant performance is to use diagnostic data provided by smart instrumentation. That capability is now available through many wired and wireless communication protocols: HART, foundation fieldbus, Profibus PA, ISA100.11a and others.

The DCS can support analysis tools to monitor the performance of field instrumentation, particularly valves via smart controllers. The result can be a condition-based evaluation program integrated with the maintenance and reliability areas. Such programs have led to substantial cost savings, increased uptime and greater productivity in many process plants.

All those communication protocols have their specific applications and more than one will be found in most every plants. The DCS has to be able to communicate with all of these protocols, either directly or through converters, and handle information in a way that makes the transmission method transparent to the operator.

The person in the control room doesn't care how a critical reading gets to the screen so long as it gets there accurately and reliably. The I/O capabilities of a DCS have to be all encompassing with the least amount of hardware and wiring. Software configurable marshaling is a major step in this direction, as are fieldbus networks with their built-in multiplexing capabilities.

More Subsystem Integration

Oil is flowing to refineries all around the world from offshore and onshore wells that were impossible to economically exploit just a few years ago. The complexity of these production sites has created new types of subsystems that must all be brought together as part of the DCS.

Control systems on the sea floor connect to systems on the platform which connect with onshore facilities. For fracking, numerous subsystems must be employed, including full-scale water treatment plants at some sites. These are all specialized systems, but seamless integration with the DCS is critical to profitable production.

More commonplace subsystems in conventional plants can include many areas that come under control of the DCS.

All of these can present their own challenges for integration, but a capable DCS should be able to make them part of a unified whole. As with field devices, operators simply want the data, they don't care how they get it as long as it works. In some cases, they also need to make adjustments to real-time control schemes, and consistency of data presentation and operator actions is even more important in these instances.

Integrating Humans

Most of the discussion so far has been how control systems and other computing systems work with the DCS, but humans have to figure into the process, and that means HMIs. There has been a longstanding sense that while technologies may change, people remain basically the same. However over the last few years, that has been disproved in many ways.

Plant managers have had to deal with large-scale personnel turnover, and it's only getting started. It's common to hear customers say things like, "Almost half our employees are eligible to retire and could walk out of here tomorrow. And those people will take an enormous amount of knowledge with them." The DCS has to provide a means to capture that knowledge and integrate it into the automation strategy.

While it's important to design HMIs for the people that are working now, it's even more important to design them for the next group. As mentioned before, millennials live with interactive technology, and they expect it to be natively intuitive. A typical 23-year-old probably won't be able pull a manual off the shelf, start reading and comprehend the content well enough to perform his or her job in a timely fashion.

One of the biggest integration tasks for a DCS is to pull all the different parts and subsystems together in a way that creates a fully consistent look and feel from one end to the other in the control room. The thought that some subsystem only works from its own HMI with its own method of interaction will fail with operators in daily use. Instead, all data presented to the operators has to be uniform, as well as expected operation interactions.

To make matters worse, HMIs have to include many more types of capabilities. Why is the valve at the storage tank not opening? Touch a spot on the screen and bring up a video camera in that part of the plant. Zoom in and get a better look at the valve before sending out a technician, and make that happen in the control room in a matter of seconds.

Where are the technicians in the plant now? Call up the employee locater using active RFID and see where they are. Touch another spot on the screen and talk to them via a Wi-Fi radio. All of these technologies and capabilities and more must be integrated into the operator screens and the DCS.

Consistent look and feel is also critical to operator effectiveness, not just to fulfill some aesthetic sense. If an operator has to stop a pump in an emergency, he or she has to know how to do that without thinking. If the operation to stop is shown differently on one screen than another, the operator now has to stop and think, "What am I looking at now?" That delay can be the difference between a safe shutdown and an incident.

Graphic guidelines following ISA101 and other similar standards helps develop AOG (advanced operating graphics) that can ensure clarity and consistency from screen to screen for operator effectiveness. Providing this advanced decision support helps overcome the knowledge drain when experienced operators retire.

These critical aspects of system design should follow applicable standards to take advantage of the knowledge gathered, and for the sake of consistency from plant to plant. Applicable standards include Management of Alarm Systems for Process Industries (ISA18.2), Human-Machine Interface (ISA101), Procedure Automation for Continuous Process Operations (ISA106) and others.

A simulation environment can help train operators and engineers [M.T2] if the models are reliable. When an individual closes a valve, turns on a pump or changes a setpoint, the simulator needs to respond just as the plant would in the same situation. This kind of training will grow in importance given the way that millennials tend to learn. Traditional classroom methods are being replaced with more interactive techniques that allow younger trainees to follow their impulse to try things. Simulators avoid having to deal with questions like, "What happens if I push this?" in a working control room.[u3]

The capabilities of training simulators are advancing all the time. The idea of creating a second virtual plant that mirrors the actual installation in every meaningful way is now a real, practical and even cost-effective way to take operators into a training environment that is indistinguishable from the real thing. Engagement on any level is possible.

Growing Role for Simulation

Just as simulation capabilities for operator training are advancing, the same technologies are helping optimize a plant's operation, or even create a new plant virtually before the first piece of pipe is cut. That mirror plant used as a training tool can also determine how a change in feedstock might affect production, and how operational parameters can be adjusted to accommodate a new energy source. New setpoints can be determined in advance making the changes quick and smooth.

An accurate process model can fill in knowledge gaps in the plant. What's the temperature at a specific level in the distillation column? That answer could come from adding a new sensor, with all it's expense and potential production disruption. Or it could come from a simulation that can provide a full temperature profile at all levels of the column based on known measurements combined with values calculated from the process model.

A new plant or major upgrade project can be built in a virtual environment, operated and characterized before any concrete is poured or steel welded. Large oil and chemical companies are asking for ways to shorten and simplify new projects, and this is a major element of such efforts. The ability to create a virtual plant and see it working allows designers to troubleshoot problems before they're built and eliminate surprises during construction. A new plant or upgrade can be completed on time and on budget, delivering a plant and process that behaves just as expected.

Implementing a Project

The call for faster and less expensive capital projects extends to specific elements within the plant automation infrastructure. Some of those elements are technical and can streamline implementation. For example, more flexible I/O systems are less dependent on hardware with more configurability via software.

I/O system[M.T4] can be landed in a marshaling cabinet at any point,[u5] and then the system uses software to assign the device type and location. That can reduce hardware and installation costs while allowing greater flexibility when changes surface in the final stages of a project.

Similarly, smart I/O and fieldbus installations allow for an instrument or a valve to be assigned to any controller, instead of having to make all controller assignments very early in the design process. Adding a field device to a fieldbus system is pretty simple if there's spare capacity, much easier than hard wiring from the device to the DCS I/O.

Wireless instrumentation can extend this strategy even farther. When there is no marshaling cabinet at all because wireless devices communicate with a gateway, the location of a device and its connection to the control system can be anywhere. Wireless protocols such as ISA100.11a allow a broad range of device types to be incorporated. Even those with large data streams or specific protocols are no problem with capabilities like data tunneling.

Both wired and wireless fieldbus networks can incorporate mechanisms to diagnose and configure an individual transmitter or other sensor from an engineering station rather than going out in the plant and doing it at the device. At the same time, the instrument database can be automatically updated so that it stays current with the way the plant really is. All those capabilities result in improved performance and lower costs.

It Still Runs the Process

After all that, the DCS does have to run the process, and all these new technologies work in support of that most basic objective.

Effective instrumentation and field actuators support advanced process control (APC) capabilities to maintain maximum production. High-fidelity simulation, in addition to training operators, can help characterize the process and determine where improvements can be made, all without interfering with actual production. If feedstock or energy sources change, the simulator can also help determine how the process might need to be changed.

All the operating elements can interface with enterprise control systems, reporting production and other operational data in any form necessary to calculate KPIs.

These elements are part of the CENTUM VP platform, Yokogawa's ninth generation DCS over nearly 40 years. To some degree or another, they are also characteristic of automation systems from other suppliers in the same space (Table 2).

Today, users have a wide variety of capabilities and choices available from a number of suppliers. There is no technological reason why any facility cannot exploit these new capabilities and enjoy new levels of productivity, profitability and sustainability.


  • 分散式控制系統(DCS)

    我們的分散式控制系統 (DCS) 可自動化和控制製程,提高績效。 已經超過 10,000 家工廠利用Yokogawa DCS 實現生產目標。