01 November 2003
Restoring single-loop integrity to process
By Henry Marks
Moving control to the field and implementing distributed fieldbus linking devices on Ethernet works
Should process control loops remain in a centralized distributed control system (DCS) or programmable logic controller (PLC)?
Or can they take root in a field-based architecture with control residing in intelligent instrumentation?
Not surprisingly, the answer to this question may depend upon the preference, experience—and product offering—of the control system supplier or engineering firm handling your project.
Although some companies are reluctant to abandon the familiar DCS or PLC environment, an increasing number of users are now choosing to locate most, if not all, regulatory and logic control functions in field devices.
The advent of modern fieldbus technology has changed the course of process control. In particular, control in the field enables end users to restore single-loop integrity to their production processes, and in doing so, improve plant reliability and availability.
Interface equaled, exceeded
In the late 1960s, automation equipment suppliers seeking to move beyond the limitations of pneumatic controls developed the first distributed, single-loop, miniature analog (4–20 mA) control systems. These systems offered users the reliability of single-loop control, so that in the event of a control valve failure, only one control loop was affected.
With the simplicity of single-loop control, instrumentation technicians could go into the field and quickly verify control or indicating loop operation, including the primary element, transmitters, controllers, control valves and their electric current to gas pressure (I/P) or positioner, as well as any auxiliary devices.
Furthermore, the technician could tune the loop if necessary. Documentation in most cases was limited to a single-loop diagram indicating all devices and interconnections and a data sheet listing the values of the various parameters.
Following the introduction of the latest distributed single-loop control offering, manufacturers set out to develop a next generation distributed control platform with capabilities for handling larger system requirements.
Although work continued on the single-loop miniature analog control system design, there were few additional technological breakthroughs to support this implementation. The need to interface indicating and control loops with a central data handling or computer system was also becoming prevalent. With an analog approach, the cost of the interface equaled, or in some cases exceeded, that of the basic device.
The most logical step forward seemed to be a single-loop miniature digital system with multiplexed communications to the central data handling or computer system. At the time, however, commercial microprocessor technology could not meet the processing requirements for this application and was still too expensive for practical use.
Ultimately, cost/performance considerations caused system designers to settle on a shared, digital, eight-loop distributed control platform employing 16-bit microprocessors. The first DCS was born! These systems came to market in the mid-1970s, with installations in continuous process industries such as refining and petrochemicals.
|DCS||distributed control system|
|I/O||input and output|
|I/P||electric current to gas pressure|
|OPC||OLE for process control|
|PLC||programmable logic controller|
From the outset, availability was a key issue surrounding the DCS. Users wondered about the consequences of a failure affecting multiple control loops, rather than just one. Suppliers of the new technology also had to educate customers about mean time between failures and mean time to repair—concepts that were unheard of with single-loop control.
The industry's hesitancy to adopt shared, distributed control led to a strategy of graceful degradation, whereby suppliers designed DCS with analog displays resembling those of conventional single-loop controllers and provided four-loop standby units to facilitate the replacement of major components.
Concerns about DCS reliability and availability eventually led to the development of the one-on-eight backup controller. The main justification for this redundancy approach was to ensure a constant operator interface while allowing twenty-four-hours-a-day, seven-days-a-week operation with maintenance tasks carried out during the normal plant work hours.
Soon after the release of the first DCS, many of the leading automation vendors introduced their own variation of the new control platform. What started out as a simple shared, eight-loop system evolved into complex configurations for controlling hundreds of loops. This trend towards creeping elegance was primarily due to competitive pressures for lower cost per loop.
By the 1980s, companies based most DCS systems on a similar architecture. In a typical system, 4–20 mA junction boxes interfaced by point-to-point wiring to field transmitters (AI) and I/Ps or positioners (AO), and digital I/O junction boxes interfaced to field switches (DI) and solenoids (DO). Temperature inputs (thermocouple and/or resistance bulb elements) had their own junction boxes, as did analyzers. These field-distributed junction boxes connected to marshaling panels in the control room by multiconductor cables, with both analog and digital inputs and outputs in the same cable. The marshaling panels allowed cross wiring between multiconductor cables from the field and single-function system I/O card termination panels.
Distributed digital systems used common I/O cards (i.e., 8AI, 4AO, 16DI, 16DO), and all analog and digital inputs/outputs were timeshared to handle indicating and control loops, alarm monitoring, display, and historical archiving/retrieval. With this approach, individual loops could not commission, or no one could access them until the entire system was fully operational.
Whereas testing of single-loop control systems was limited to identifying the location of indicating or control stations on a panel board and checking connections to field I/O and auxiliary devices, the shared nature of the DCS and PLC systems created the need for a factory acceptance test (FAT). FAT required interconnecting all system components at the supplier's facility and conducting rigorous tests to verify their behavior under normal and abnormal conditions. The vendor then dismantled the system, shipped it to the user's site, and reassembled it to perform a site acceptance test.
By the early 1990s, a new breed of smart field instrumentation was changing the outlook for users of conventional distributed control systems. The rise of fieldbus technology profoundly impacted plant automation strategies, enabling users to move control functionality to field devices and freeing higher level resources for real-time production control.
A key factor in the industry's acceptance of control in the field was the completion of the Fieldbus Foundation's high-speed Ethernet (HSE) specifications. Had Foundation fieldbus' digital device and bidirectional communications technology been available thirty years earlier, designers might not have developed the first DCS and PLC systems.
The open, nonproprietary Foundation fieldbus protocol includes H1 (31.25 kilobits per second) fieldbus for continuous control and HSE (100 megabits per second) for high-performance control applications and plant information integration (see figure 3). Within this architecture, HSE functions as a high-speed Ethernet backbone for device, subsystem, and enterprise integration. It also supports high-performance control applications using the same open and interoperable function blocks as devices on the H1 network.
End user adoption of Foundation fieldbus has happened as a result of the efforts of the Fieldbus Foundation's 200 members worldwide. The members include most of the major automation equipment suppliers. Smar International has been one of the most active participants in developing and implementing the technology. Smar introduced the world's first commercial HSE product (DFI302 linking device) in 2000, and later that year, opened the door to commercial HSE applications when Merisol Chemicals implemented its SYSTEM302 solution at the Greens Bayou refinery near Houston, Texas.
Plants installing a Foundation fieldbus system are free to implement batch and logic control at the field level. The technology enables primary proportional, integral, derivative (PID) and secondary PID (cascade) functions, as well as feedforward and lead/lag, to reside in the fieldbus control valve positioner/controller.
Foundation fieldbus also allows control to distribute to the field in HSE linking devices. In the fieldbus network architecture, the HSE linking device holds the key to integrating H1 fieldbus segments into the high-speed Ethernet backbone network. It functions as a link between H1 channels and the HSE linking device, interconnecting field devices to each other and to other hosts, and is an essential component for integrating system communication with direct I/O access and advanced control applications.
A single HSE linking device can interconnect four H1 fieldbus segments and accommodate up to 16 devices per segment—for a total of 64 devices per linking device.
Single-loop control is the most common method for controlling a process.
No cyclic control information
A field-centric control strategy eliminates the complexities of the DCS environment and restores single-loop integrity. Benefits include:
Improved reliability—Locating regulatory control in field devices improves loop reliability and increases availability—the consequences of a single failure are limited to the failure of a single loop.
Reduced complexity—Control in the field eliminates the need for complex and costly control room hardware. It also minimizes external link requirements and improves loop performance—the result is true distributed control.
Lower cost—A field-based control scheme reduces costs associated with centralized processing capacity, power supplies, signal conditioning, panel space, and redundancy hardware—implementation is "one bite at a time."
Increased flexibility—Fieldbus instruments with built-in control functions provide utmost flexibility for configuring control loops—the devices' instantiation capability simplifies incorporating or changing control strategies over time.
Improved scalability—Installing spare equipment to accommodate future modifications is easier and less expensive with fieldbus, because the cost and complexity of hardware is reduced. Unlike a DCS where each device consumes finite resources, each fieldbus device adds resources, making the system ever more powerful.
Decreased traffic—Fieldbus eliminates the need to send cyclic control information to higher levels—only supervisory and operation-related data need pass to the central control system.
Choosing a system
Companies seeking the benefits of Foundation fieldbus, including a return to single-loop integrity, should understand that not all compliant control systems fully support distributing the control among field devices.
Some fieldbus-based solutions incorporate a proprietary PLC device that functions as a remote controller and links to the host via fast Ethernet communications. Unlike a full Foundation fieldbus HSE implementation, these distributed systems only utilize a portion of the open, integrated HSE architecture.
Even manufacturers with Fieldbus Foundation-registered HSE linking devices often substitute DCS controllers in their fieldbus systems.
In such cases, a proprietary protocol serves for communications between the distributed controllers or between different manufacturers' devices, with system data channeled through a common central master station—a server.
In addition, the constraints imposed by proprietary communication protocols keep end users from taking advantage of OPC compatibility; Foundation fieldbus users running OPC can communicate through any HSE linking device or standard field device.
End users should also pay close attention when evaluating Foundation fieldbus devices for their installation. The guidelines for choosing devices include:
- Devices and other fieldbus system components should be officially registered by the Fieldbus Foundation.
- Devices should be registered with Interoperability Test Kit 4.0 or higher.
- Devices that are bus powered should have a maximum current consumption not exceeding 15 mA (preferably 12 mA).
- Devices should incorporate up to 17 types of function blocks, including the 14 standard Foundation fieldbus function blocks.
- Devices should have an instantiation capability allowing the user to implement a single-function block as many times as necessary—up to 20 function blocks per device.
- Devices (logic) should employ flexible function blocks supporting complex batch, discrete, and hybrid-control applications.
- Devices should have a total cycle time contribution of less than 500 milliseconds—even if two PID blocks are used.
And thus, by moving control functionality to the field, and implementing distributed Foundation fieldbus linking devices connected by the HSE communication protocol, end users can restore single-loop integrity to their process and achieve a control system that is easier to define, engineer, procure, test, install, and maintain. P
Behind the byline
Henry Marks has five decades of experience in plant automation, forty of it with Honeywell. He currently is president of Marks and Associates, a global automation engineering and consulting firm. Marks is a charter member of the ISA Montreal Section. Write him at firstname.lastname@example.org .