(Editor's Note: There are three figures that accompany this article that can be downloaded in PDF format via the "Download Now" button at the bottom of the page.)
Industrial field networking technologies as represented by the various flavors of fieldbus and industrial Ethernets represent a step change in the way automation and control systems are designed, installed and operated. They are still relatively new and unknown quantities. Even though the Fieldbus Foundation (FF) celebrated its 10th anniversary in 2004, a lot of confusion remains about how to effectively use the technology.
Nevertheless, there are compelling incentives for adopting the approach. A January 2003 Automation Research Corp. study, “Best Practices for Maximizing the Value of Fieldbus Implementations,” concluded that fieldbus enables superior return on assets and life-cycle cost savings. The report also stated that fieldbus requires intensive upfront planning.
The key point to “getting it done right” is managing the change. James Rhame of Shell Chemical Co. during his keynote address at the FF’s 2004 General Assembly emphasized that the biggest challenges in managing that change are:
- understanding the importance of people, and the fact that different people have different reactions to change;
- treating change as a mental, physical and emotional process, not as an event; and
communicating openly and honestly.
Fieldbus technology is more complex than traditional analog control systems with their point-to-point connections. It is also far more powerful. To capture that power requires that the decisions made when a project is being conceptualized and formulated are the right ones. In this respect, the project is just like any other one in which decisions made at the beginning can have a significant impact on final success or failure. It is for this reason that front-end engineering design (FEED) processes are used in most projects today. Of course, without sufficient knowledge or experience, it is almost impossible to be able to make informed decisions; consequently, one of the most critical components of a project’s success is another personnel issue: training.
Fortunately, a number of companies, technical societies, educational institutions and the fieldbus technology organizations themselves provide vendor-neutral training on fieldbus and industrial field-networking technologies.
Obviously, a one- or two-day course will not make people experts in fieldbus technology. It will, however, make them aware of some of the differences between a conventional analog system and a distributed digital control system as represented by fieldbus technology. It should convince them of the value of the approach for their project and also make them realize that they are not familiar enough with the technology to make the best decisions by themselves.
The team involved in the project probably is familiar, though, with the “decade rule,” which was proved by IBM in the 1970s and is still valid today. It states that $1 spent making a decision at any stage in the project is multiplied by an order of magnitude if the decision is delayed until the next stage. That means $1 spent during conceptual engineering by a small design team grows to $10 during FEED, $100 at procurement, $1,000 during construction and $10,000 at startup. So, spending a few extra dollars upfront will yield significant savings later. Yes, a consultant might be expensive, but that cost pales in comparison to the price of an idle plant, field rework or several idle field mechanics later in the project.
Now that we appreciate the importance of getting it right early, what are some of the FEED issues that need to be considered for a fieldbus project?
Power supply issues
Fieldbus network signals differ from analog ones because they alternate above and below a reference point to indicate a “1” or “0,” creating a ripple in the power supply. A power supply, however, is designed to maintain a steady signal on the wire pair, and so does its best to remove the ripple. It is for this reason that fieldbus systems have power conditioners, which, in some cases, are incorporated into the power supply as a single unit. These power conditioners can be either active or passive, depending upon their componentry and design. Active conditioners use solid-state components, such as diodes and transistors, to achieve the necessary signal conditioning, whereas passive conditioners employ induction coils and resistors. Many papers have been written on the relative merits of each type. With the reliability of silicon-based devices today, there is minimal difference between the two approaches.
In 2004, the FF issued a standard on power conditioners that provides a baseline for users. Units that comply with the standard will receive a FF “check mark.” Purchasers still will have to consider the level of local support and inventory, as well as the form factor and active versus passive design. What is more important, though, is the amount of power that remains on the output terminals of the conditioner — because more power means more devices can be placed on a network.
The size or output rating of a power supply largely depends upon the environment in which the devices will be placed. Just like in a traditional analog system, the area classification and the methods used to meet the electrical requirements of that classification drive a number of decisions. If the choice is to go with explosion-proof fittings for all devices, there really are no limitations on the power available to the network. However, if one of the other safety mechanisms — intrinsically safe (IS), fieldbus intrinsically safe concept (FISCO) or fieldbus non-incendive concept (FNICO) — is selected, it will have an impact on the design.
IS can be used in any area classification, but supplies the least amount of current. It can only provide about 85 mA to 90 mA to the network. Because each device takes anywhere from 15 mA to 30 mA, the network is limited to between four and six devices.
FISCO, which is based on work done in Germany, enables a usable output of up to about 130 mA, allowing several more devices to be added as long as they are certified as FISCO compliant. Fortunately, these compliant devices can be used in “normal” IS circuits as well.
FNICO is the most recent development. It is restricted to Zone 2 installations, which include the majority of chemical and petrochemical operations. Because Zone 2 has a lower probability of gases being present, FINCO can handle at least 180 mA while still maintaining the same overall level of risk as IS. This means that up to about 12 devices can be installed on a single network.
FF End User Council presentations with more information on these concepts are accessible via www.fieldbus.org/endusersupport/enduserpresentations/archiveeu03. Check, in particular, L. S. Towle’s “IEC 60079-27” paper.
With the current limitations defined, the next challenge is to design a network to make the best use of the technology at minimal cost.
Network layout options
Figure 1 shows common configurations that are possible for Foundation fieldbus — although it should be noted that any fieldbus can be installed “point-to-point” as well. The daisy-chain configuration is also included in the standards but is not used because if one device fails, all the downstream devices will also fail. It is possible to mix the various topologies on a single network if desired or required.
Network layout also includes factors based on the number of devices per network and overall cable length, so an approximate physical device location should be known when preparing the network design.
The table, which is based on the IEC 61158 standards, summarizes the impact the number of devices has on the overall cable length that can be used. The maximum spur length equals the sum of the lengths of all the individual spurs plus the home run cable. Values are only guidelines, but can be used to determine the density of devices. The table indicates that it is possible to have 32 devices on a network, but this is not really practical because the system will encounter a bandwith limitation before this physical limit can be reached.
For more information, download the FF’s “System Engineering Guuideline,” AG-181, from its Web site.
Fieldbus technology provides the ability for truly distributed control by supporting regulatory control (e.g., PID) in field devices. So, the plant owner or operator should decide at the beginning of the project if it wishes to design for and implement this option. These are two separate decisions. It is possible to design for but not implement field-based control. Considering future implementation during the initial design and network layout often makes a lot of sense because it can minimize field rewiring later. In fieldbus, all devices that are part of the control loop must be on the same network so that the host system does not need to be involved in transferring data from one device to another. Fieldbus uses the publish/subscribe model to communicate between devices. The field sensor publishes its data, every Compel Data command it receives during the macrocycle, and the control device (as well as any other devices, including the host, that need this information) subscribes to the data for its use.
The main reasons for implementing field-based control are that it provides single-loop integrity — that is, if a single part of the control loop fails, the entire loop does not — and is faster than host-based control. So, it enables control closer to the set point or control point. Better control translates into higher profits! For more details, see the paper “Benefits of Field Based Control” by Kurt Zech in the FF End User Council Web archive.
The final link in the chain that makes these benefits possible is proper integration. Figure 2, based on a slide from Jonas Berge, shows the various levels of a fully integrated system in which data can flow, with appropriate security precautions, among the various data sources, repositories and users. The flow of all these data needs to be determined at the start of the project to ensure that the infrastructure and work processes are in place when they are required during startup.
Fieldbus technology can deliver substantial maintenance savings if companies adapt their practices to take advantage of the opportunities.
The two-way communication between the devices and the host and configuration system means that the devices no longer are separated from the control system — they are part of it. Any change in the host or a device is immediately communicated to the other. So, configuration, calibration and range-changing of devices can be performed remotely from the control room if desired. (Devices can be configured to fail in different ways depending upon the type of event initiating the fault. For example, a device could fail last on loss of communication signal, high on loss of power, and low on loss of process signal or instrument air.)
The other significant change that results from fieldbus devices is that an abundance of diagnostic information becomes available to technicians. This information should be captured in a computer-based maintenance system and should lead to a true predictive maintenance program, with resultant savings in staffing, spare parts and lost production.
Note that fieldbus maintenance must be done with the device connected to and communicating with a network; otherwise there is no way for the device to receive power or to monitor the device’s response to its external environment.
Use of multivariable devices can provide significant project savings. Fewer devices will need to be specified, purchased and installed, leading to a corresponding reduction in field terminations and pipe penetrations, each of which is a potential leak point. Moreover, each variable will be available with an associated status bit confirming its validity. Figure 3 shows a vortex meter in which flow and temperature are obtained from a single device. Coriolis meters are another example of flow meters with multiple outputs; mass, temperature and density are common, and one manufacturer offers a viscosity output as well.
Incorporating multiple sensors in a single device makes mass-flow calculations possible for any vapor stream. For a liquid stream, a temperature sensor allows the output of a temperature-compensated flow without the need for a separate flow computer, as well as a “raw” value.
Because fieldbus devices can transmit data about a number of process variables on a single wire pair, there is no additional investment required or compromise in the accuracy of the measurement. Plus, the devices eliminate the need for analog/digital and digital/analog signal conversion common in a typical analog circuit.
This raises the question as to why a plant should remain with analog single-point connections and their limitations when it is possible to achieve more accurate measurements of a variety of parameters from a single device.
In addition, fieldbus technology, because of its plug-and-play capability, usually significantly reduces the commissioning and startup time for a facility’s control system, which is often the last piece of the plant to be installed.
Fieldbus represents both challenges and tremendous opportunities. With the right mind-set, skills and people, the many benefits of this technology can be realized through the entire life cycle of a facility.
Ian Verhappen, P.Eng., is director for ICE-Pros Inc., Ft. McMurray, Alberta, an independent instrument and control engineering consulting firm specializing in fieldbus, process-analyzer sample systems, and oil sands instrumentation and control. E-mail him at [email protected], or contact him through his Web site www.ICE-Pros.com.