Fieldbus moves into the process control laboratory

Fieldbus moves into the process control laboratory

Rehg, James A

Abstract

The control hardware in manY process control laboratories uses single-loop digital controllers to control process parameters. While these older controllers are still used in some industries, they are often not the best choice for teaching students proportional, integral, and derivative control methods because of the cryptic programming technique required to set and view process variables. The newer field control system technology used in industry is a better choice for teaching process control.

The technology for industrial process control has evolved over the last forty years from direct digital control (DDC-1962) to programmable logic controllers (PLC1972) to distributed control systems (DCS-1976) and now to field control systems (FCS-1994). The latest field control system evolution is the implementation of Foundation fieldbus into the manufacturing environment. This paper describes field control systems and Foundation fieldbus, discusses their role in a plant network hierarchy, and compares Foundation fieldbus with the older DCS model. The instructional issues associated with teaching Foundation fieldbus in a system control course and laboratory are also discussed, and an overview of the hardware and software necessar to build a network of Foundation fieldbus devices on existing process trainers is provided.

Introduction

Foundation fieldbus is a digital control network that interlinks “smart” sensors and actuators in a manufacturing environment. It is the latest technology used to automate the capture of process data and the control of analog production systems. The evolution of control system architecture from direct digital control (DDC) to distributed control systems (DCS) and now to field control systems (FCS) is shown in figure 1. In every step of the evolution, the point of control for the process has moved closer to the sensors and actuators.1

Figure 2 illustrates the shift of the proportionalintegral-derivative (PID) function from the central system computer to the sensors and actuators at the point of measurement and control. This movement of the control process reduces wiring, helps troubleshoot process problems, and decreases maintenance costs for the industrial control system. It also allows fieldbus devices with the appropriate interface to be programmed by any fieldbus host computer on an existing local area network (LAN).1

Another feature of the FCS system is the capability of adding fieldbus devices to an existing fieldbus process while it is operational. This feature makes implementing new fieldbus networks into an existing system much less complex and does not require stopping the process while sensors are replaced or new field devices are added.

Fieldbus devices have built-in microprocessors and associated electronics that give them the ability to run process control loops internally without any need for processing power from a central computer or digital processor. The network that links FCS devices makes data from every device available to all the FCS devices and adds greater flexibility to control decisions. In addition to the data networking function, the FCS twisted-pair network cable can supply the power required to run all the sensors and actuators on the network. The FCS standard permits up to 32 devices on a single segment; however, with repeaters as many as 240 fieldbus devices can be attached to a single network.

The FCS standard of interoperability supports a “plug and play” architecture, in which field devices from different vendors can be mixed in a working fieldbus model and new field devices can be added without the need for major LAN reconfiguration.

Fieldbus Systems and Standards

There are numerous network control techniques that use the term fieldbus to describe their operation, which can make the selection and design of fieldbus-driven control systems rather confusing. An overview of the current network protocol choices available to the design engineer is shown in figure 3. Fieldbus LANs are divided into the two broad application categories of discrete (on/off) and process (analog). The automation level ranges from bit-level sensor (control of a single bit) to sophisticated proportional control and process information interfaces into the enterprise business unit for automation level control (supporting database management and inventory control.) In the discrete area, the Profibus Discrete Protocol (DP), Controller Area Network (CAN), Devicenet, and Smart Distributed System (SDS) protocols have good vendor support. In the process control area, two levels of LANs, called Hi and H2, are available and two protocols predominate: the Foundation fieldbus (FF) and Profibus PA.2

Development of the fieldbus standard started in the mid1980s when the Instrument Society of America (ISA) formed the SP50 fieldbus committee. In 1992, the number of variations in the standard narrowed when Fisher, Rosemount, Yokogawa, and Siemens created the Interoperable Systems Project (ISP) and major SP50 companies, including Honeywell and Allen Bradley, formed the World Factory Information Protocol (FIP) standards group.2 Further consolidation occurred in 1993 when the ISP and WorldFIP joined to form the Fieldbus Foundation. As a result, two protocols have evolved for LAN-based process control applications: the Foundation fieldbus standard supported in the United States and Asia and the Profibus PA standard popular in Europe.

Fieldbus Control Architecture

The architecture used with Foundation fieldbus configurations includes two LAN types known as HI and H2. The HI segment is a 31.25-kbit/s bus structure used to link fieldbus devices. As shown in figure 4, the HI bus can have four configurations: point-to-point, bus with spurs or multidrop, daisy chain, and tree. A fieldbus network can have a single topology or any combination of the four different options. Based on the International Electrotechnical Commission (IEC)/ISA physical layer standard, Type A shielded twisted-pair wire is preferred for H1 connections, with a maximum length of 1900 meters for H1 cabling.3 While new Foundation fieldbus installations would use Type A wire, current system implementations can convert to fieldbus technology with existing twisted-pair instrumentation wiring in most situations.

A terminator block, expansion block, or junction box is employed to create the tree or star/chicken-foot topology and to create multiple spurs of the H1 bus.4 A typical configuration for expansion blocks is shown in figure 5. Note that the initial block is a power conditioner plus field device terminator, while subsequent blocks are used only for termination of field devices. Since the twisted-pair network wiring carries both the data and device power, the power conditioner isolates the data from the power supply source. Also note that the initial junction block provides a connection point for the host computer. The H1 bus must have terminators (a series resistor and capacitor) placed at both ends of the bus to improve network data transmission by preventing data communications reflections from the ends of the network wire; as a result, some junction blocks have terminators built into the interfaces.

The H2 fieldbus LAN is a high-speed communications network that serves as a backbone for the HI segments.5 The H2 backbone can operate at 1, 2.5, or 100 Mbits/s. A typical LAN configuration with both H1 and H2 segments is shown in figure 6. Note that the HI network supports three plant processes and that data from the three H1 networks flows over an H2 network to a PLC, a data analyzer, and network server. The H2 network speeds are useful for transferring data between the smart field devices and other production hardware such as PLCs and process analyzers. The H2 LAN permits access to the fieldbus structure from any computer on the plant intranet, and gives process engineers and production planners direct access to process data and the ability to program the system from remote locations.

Comparison of FF and DCS

The DCS architecture shown in figure 2 is the current standard for large process control applications. In this type of system, a central processor is tied to numerous sensors and actuators and controls all of the system’s parameters. The DCS configuration is similar to the star topology with a computer at the center and terminals surrounding the central processor. In contrast, fieldbus architecture moves data filtering, conversion, tuning constants, and alarms into the field to be accomplished directly within the fieldbus sensor or actuator, which greatly reduces the need for central processing capability. Fieldbus technology also makes it possible to set all device configurations, including process parameters, process variables and set points, with corresponding software and hardware tools. Since parameter notation has been standardized across fieldbus device manufacturers, these devices are relatively easy to use and components can be mixed in a “plug and play” fashion.6

In comparison, DCSs have a single process control computer to which all of the sensors and actuators are connected. While the sensors and actuators can come from many different vendors, each device must be configured to work with the main process computer. The primary advantage of a Foundation fieldbus system over a DCS is the rich data set available to the network from the field device. In a DCS setup, a sensor has access only to data related to the process variable associated with that sensor. In contrast, every fieldbus device has access to the process parameters from every other device and actuator on the network. In other words, each fieldbus device has access to all of the data that would typically come from the DCS central process computer.

Foundation fieldbus systems can coexist with DCS systems, thereby protecting existing DCS investments. However, care must be taken in combining the two systems since no existing DCS system can match the functional characteristics of FF. The differences between FF and DCS characteristics can potentially cause confusion and undesirable system operation if devices for one system are used within the network design of the other.7

Introduction of Fieldbus into the Technology Lob

The first step in creating a Foundation fieldbus system is to select the process to control. Implementation of the fieldbus network in a control laboratory requires an interface to a number of process systems, both new and existing. Most process control laboratories in colleges and universities have process trainers for teaching control of temperature, pressure, flow, and level. Many of these training systems use the 4-20 mA loop current control system technology and stand-alone digital PID controllers to provide the closed loop process control. FF devices can be incorporated into these systems by just eliminating the stand-alone PID controller and using the existing 4-20 mA interface to the trainer actuators. In the simplest conversion, one or more of the standard process sensors are replaced with FF devices, and a fieldbus-tocurrent converter is used to link the new FF network into the trainer’s existing actuators using the 4-20 mA interface. Thus, implementing Foundation fieldbus with an existing trainer requires changing the process variable transmitters to fieldbus-type devices, adding wiring to provide the fieldbus network interfaces, and adding the appropriate jumpers between the fieldbus network and the trainer actuators.

The hardware and software needed to implement a FF system includes:8,9

1. computers with fieldbus-capable interface card and configuration software for configuring the fieldbus devices

2. a power supply with a dc voltage of 9-32 volts and a current capacity of about 20 mA per attached field device

3. the wiring for the 31.25 kHz Hi fieldbus network with a resistance of 100 SL/ft and an maximum attenuation of less than 3 dB/km. The cabling must be 16-26 gauge shielded twisted-pair. The recommended color code is orange or white (+) and blue or black (-)

4. a terminator at each end of the twisted-pair network composed of a 1-iF capacitor and a 100-SZ resistor to avoid reflection signals in the wire. Terminator blocks with the proper electronics can be purchased or constructed from available components

5. a power conditioner made from a 5-mH inductor and a 50-SL resistor must be placed between the power supply and the network

Field devices such as sensors and actuators can be purchased from a multitude of companies. The Fieldbus Foundation lists on their web page all of the companies with Foundation fieldbus standard equipment.9 Since fieldbus technology is relatively new, fewer components are currently available compared with the availability of the older 4-20 mA devices. However, the basic components used in college and university laboratories, such as temperature sensors, pressure sensors, flow sensors and valve positioners, can be easily obtained from fieldbus vendors. In addition, fieldbusto-current and current-to-fieldbus converters are available for interfacing to existing trainer components and actuators. Finally, electronic conversion modules known as “round cards” can be used to convert older 4-20 mA devices into fieldbus-compatible components.

Modifying an Existing Process Flow Trainer

The fieldbus network and control system designed for the process control laboratory at Penn State Altoona utilizes an existing flow process trainer, as shown in figure 7. The original unit had a differential pressure (DP) transmitter connected across an orifice plate with the 4-20 ma output of the DP unit connected to a variable speed motor controller, as shown in figures 8 and 9.

The motor controller changes the pumping capacity to control the fluid flow through the system shown in figure 10. A study of this flow schematic indicates that it is a straightforward control problem with a tank, single flow loop, pressure orifice, pump, pump motor speed control, inline visual flow indicator, and a series of manual values to manually adjust the flow rate through the system and to provide a system disturbance to a tuned system. In addition, a single-loop digital controller, visible in figure 7, provides the closed loop control. During operation, the process variable (4-20 mA pressure signal from the DP) is connected to the loop controller, and the control variable (4-20 mA output from the loop controller) is connected to the variable speed drive for the pump motor. Laboratory exercises required the student teams to identify P, I, and D control parameters using the process reaction and Ziegler-Nichols techniques, and then to determine how well the tuned system responded to changes in the set point and process load.

In the fieldbus solution, the variable speed pump is connected to a Smar FI-302 fieldbus-to-current converter that is networked to a Honeywell ST-3000 fieldbus DP sensor, as shown in figure 11. The Foundation fieldbus H1 network includes the converter, DP sensor, power supply, and Pentium II computer, as shown in figure 12. The two fieldbus devices (converter and DP sensor) are programmed and configured from the computer with a National Instruments AT-FBUS interface card and accompanying Configurator software. In addition, the National Instruments Lookout human machine interface software displays process and control variable values from each fieldbus device.

This trainer was designed to demonstrate how process flow system components and control software can be configured tb maintain a continuous fluid flow, even when a process disturbance is presented to the system. In the original system, the PID control algorithm was implemented in a stand-alone digital controller. In the fieldbus implementation, the fieldbus DP sensor includes the PID controller with internal electronics and software that monitors and controls the fluid flow in the system to a rate designated by the user. Control has been moved from the separate digital loop controller to the fieldbus sensor.

The Honeywell DP fieldbus sensor incorporates a factory-installed Link Active Scheduler (LAS) program. This program allows the computer, which is used to configure and monitor the system, to be disconnected from the process network without creating any disruption in the process control or disturbance in the flow control.

The network interface for the fieldbus system uses a Smar fieldbus-to-current converter so that the control interface (pump with adjustable speed drive) did not require modification when fieldbus components were incorporated. The 4-20 mA control variable for the Seco AC adjustable speed pump drive originates in the fieldbus DP sensor instead of in the single-loop controller, but the control results and laboratory performance are the same.

All of the laboratory exercises previously performed with the original DP unit are directly applicable with the fieldbus system. The primary variation is that students must configure the Honeywell DP transmitter for proper control parameters instead of tuning the digital controller. In addition, students learn to use the Foundation fieldbus configuration program for field devices and to display a much broader range of process data than was available with the standard DP unit.

Future Expansion Plans

Based on the success of the initial fieldbus installation at Penn State Altoona, three additional process trainers will be modified to incorporate Foundation fieldbus technology: a temperature trainer, a liquid level trainer, and a pressure trainer. All four trainers will be linked together utilizing a H2 high-speed (100 MHz) bus. Additionally, the main computer on the H2 bus can be located at an instructor’s station in the lab to monitor student progress on the fieldbus systems, and students can use this central computer to test the remote monitoring and control features. Another advantage of the H2 bus and central computer is the instructor’s ability to inject problems into the local networks, enabling students to troubleshoot real-time failures.

The implementation of Foundation fieldbus on the three additional trainers is similar to the efforts used to convert the flow process trainer. The primary difference is in selection of the appropriate fieldbus transmitters and actuators: the temperature trainer would use a Honeywell ST-3000 fieldbus temperature transmitter, the level trainer would use the same DP transmitter used by the flow trainer, and the pressure trainer would use a Honeywell ST-3000 absolute pressure transmitter. Presently, all technology is available to retrofit each trainer with an independent Foundation fieldbus control system and to link the systems with a H2 high-speed bus.

The Fieldbus Advantage

Laboratory exercises performed with both the new fieldbus technology and the older single-loop controller technology yielded the following comparisons:

Entering process control parameters into the fieldbus system (values are entered into a software dialog box using the computer keyboard) and into the single-loop controller (values are entered through a keypad on the controller) required about the same length of time. However, the fieldbus software allows the dialog box to be activated with one mouse click and displays all the parameters on one screen, while the single-loop controller requires a number of key strokes to get to the parameter edit command and displays only one parameter at a time. The fieldbus interface efficiencies allow students to configure the system and collect test data more quickly with an average time savings of 15% per experiment. As a result, both students and faculty who used the two systems indicated that they preferred the fieldbus approach.

The National Instrument fieldbus software used for system configuration and data display is similar to the LabView software that is used for other process exercises in the Altoona technology laboratories. As a result, students with LabView experience are able to use the fieldbus configuration software after one introductory laboratory. For students not familiar with LabView, the fieldbus software requires an additional laboratory period. Mastering process control with the limited keypad interface of the single-loop controller requires more instructional time. Students spend a minimum of three labs becoming familiar with the large command tree that requires multiple keystrokes to perform many of the standard parameter editing processes.

The fieldbus human machine interface software permits computer screen displays of a wide variety of process and control variables in real time. These screen displays also allow the user to configure the data presentation to fit the needs of the experiments and tests. The traditional singleloop controller configuration, using strip chart recorders for display of process data, does not offer the same level of versatility or resolution in analysis of process results.

The fieldbus devices offer measurement of associated parameters in addition to the primary control parameter. For example, a fieldbus device providing flow data for use in a control loop could also measure the temperature of the flow stream for use as needed in the control algorithm. This added capability provides additional experimental data for process control exercises that is not available from the standard process control instrument or device used in a single-loop controller system.

The ability to network all the process trainers with a fieldbus H2 network protocol makes integration of the trainers into a single process control exercise easier and lets the instructor view the activities of student groups working on each trainer. This capability is not commonly found or easily achieved in a single-loop controller system.

Fieldbus-ready control elements such as DP sensors or pneumatically-controlled flow control valves cost between $1800 and $2400, about two to three times the cost of conventional devices. Accompanying software and the interface card would add an additional $1200 to the system cost. New technology is never inexpensive, but the advantages gained from using this control approach should justify the investment. In addition, the cost of the technology should become lower as more systems are adopted in industry. While a fieldbus system adds some incremental cost compared to a traditional single-loop controller system, the advantages outlined above and the opportunity to deliver current technology to students makes the argument for the introduction of some fieldbus technology into the process control laboratory compelling.

Conclusion

Fieldbus is the next technology that will find broad use in manufacturing control for the following reasons: it is relatively easy to minimize implementation costs by allowing a user to start with a small system and expand as needed, field wiring is reduced, and troubleshooting of system problems is enhanced. Also, since Foundation fieldbus components follow well-defined standards, many instrumentation manufacturers are introducing components for use in fieldbus systems, thereby creating greater competition in the market.

Because of the anticipated benefits inherent in fieldbus systems, it is important to teach Foundation fieldbus in the educational environment. The education of future control engineers in this new technology is the key to moving process control from the distributed control system model to the networked model described by Foundation fieldbus. Most older process trainers used in college and university process laboratories can be converted easily by incorporating fieldbus-type transmitters and actuators in place of the current passive devices. The use of a fieldbus-to-current converter allows existing control variable interfaces to remain unchanged in the system trainer.

References

1. Berge, J. Addressing Benefits and FAQs of Fieldbused FCS Architecture. Research Triangle Park, N.C.: Instrument Society of America, 1999.

2. Studebaker, P. “Fieldbus: Reaching the Promised LAN.” Control Engineering 10, no. 4 (April 1997): 55-58.

3. Glanzer, D. Wiring and Installation 31.25 kbits/s, Voltage Mode, Wire Medium – Application Guide. Austin, Tex.: Fieldbus Foundation, 1996.

4. McDougall, S., I. Verhappen, and S. Wheatman. Fieldbus Testing-Putting an Alliance to Work. Research Triangle Park, N.C.: Instrument Society of America, 1998.

5. Glanzer, D. Foundation Fieldbus and its Role in the Plant Network Hierarchy. Research Triangle Park, N.C.: Instrument Society of America, 1998.

6. Hendricks, D. Available: http://www.barn.org/FILES/historyofplc.html.

7. Hodson, W. Fieldbus to Change DCS Role, but Death Reports Greatly Exaggerated. Philadelphia, Penn.: Honeywell, 1998.

8. Relcom, Inc. Fieldbus Wiring and Installation Guide. Available: http://www.relcominc.com.

9. Fieldbus Foundation. Wiring and Installation Application Guide. Available: http://www.fieldbus.org/information/.

James Rehg is an assistant professor and program coordinator of the B.S. program in Electro-mechanical Engineering Technology (BSEMET) at Penn State Altoona.

William H. Swain is an engineer with the General Electric Locomotive Division In Erie, Pennsylvania, and Brian P. Yangula is an engineer employed by US Steel In Pittsburgh, Pennsylvania. Both were senior students In the BSEMET program when they worked on the fieldbus project for their senior project.

Steven Wheatman is a principal engineer with Honeywell Industrial Automation & Control in Fort Washington, Penn.

Copyright American Society for Engineering Education Spring 2001

Provided by ProQuest Information and Learning Company. All rights Reserved