By Terrence K. McMahon
This year, Distributed Control System (DCS) technology marks 30 years since the first TDC-2000 beta-types were installed at Exxons Sarnia, Ontario refinery, Kennecotts Garfield, Utah copper smelter, and other sites. During this period, the DCS has swept alternative process control technologies from the field (analog pneumatic or electronic with supervisory-level digital computers or dedicated direct digital control (DDC) systems with dual shared-file mini-computers). In the late 1970s, market analysts projected the decline of analog control but no one anticipated the speed or completeness of this transformation.
At least two generations of process control engineers have matured during the DCS era, which has endured much longer than earlier technologies. The pneumatic and electronic analog eras lasted 1520 years each while computer-based control systems, overlaying various primary control eras, spanned perhaps two decades from the late 1950s to the late 1970s.
DCS technology derives from DDC, the theory of which germinated very early in the digital epoch. The ENIAC (electronic numerical integrator and computer) machine at the University of Pennsylvania in 19461949 was the beginning of the stored-program, digital computer era. In 1952, Long and Holtzman of Shell Development discussed the feasibility of direct digital control in the Transactions of ASME. Monsanto Chemical first reduced DDC to practice in an ammonia plant in Luling, La., in 1962. The noted computer control pioneer, Ted Williams, an inductee of the CONTROL Process Automation Hall of Fame, was the intellectual driving force behind this installation. Exxon was another DDC pioneer at its Aruba Refinery using a Foxboro PCP-88 (dual DEC PDP-8s with a shared disc drive).
Honeywell had been working on a dedicated digital loop controller since 1969 but the problems of reliability and cost seemed insurmountable. The emergence of the integrated-circuit microprocessor in the early 1970s solved these problems. The Intel 8080, announced in early 1974, was the breakthrough event enabling multiple control loops to be handled with a single microprocessor. The Intel 4004 (1971) and 8008 (1972) were pointed in the right direction but were not quite up to the requirements dictated by process control loop. The 8080, fabricated with n-channel metal oxide on silicon (n-MOS) technology brought sufficient computing capability to allow process control loop requirements to be accommodated.
Honeywells focus on the digital loop controller was motivated by their need to leapfrog the competition. Foxboros analog controllers were generally considered the class of the industry. The SPEC 200, announced in 1972, was the high-water mark of analog electronic technology and ISAs 420mA standard (SP50) was approved in 1975 just as this technology was being rendered obsolete by digital controllers.
The TDC-2000 was an immediate success in the marketplace. Bailey Controls, under the strategic guidance of Bud Keyes, introduced the first significant competitive product (NETWORK 90) a year or two later. Fisher Controls and Fischer & Porter soon followed.
Foxboro and Taylor Instruments, hoping to preserve their analog franchises as long as possible, joined this trend in the early 1980s. At ISA/76 in Houston's Astrodome, Honeywell formally unveiled the TDC-2000. Only an aisle or two away, Taylor Instruments unveiled their MOD III centralized display and communication system for electronic analog controllers. Within a few years, MOD III had been displaced by the MOD 300 DCS.
Distributed control systems, because of their ability to integrate large numbers of measurements into the process control strategy, significantly increased the demand for transmitters and other peripherals. Shipments by U.S. manufacturers surpassed the $1 billion mark in 1989. Total shipments of all process control instrument products were just over $1 billion in 1972 at the high point of the analog era.
DCS shipments stayed strong and stable through 2001. During 2002-2003 they dropped precipitously reaching a level in 2003 not seen since 1986. Something is going on but the focus is not yet clear. It may be that the value-added from this activity is being reported as software or services in some other category. DCS technology has had a profound effect on our profession. The pioneers deserve recognition for this substantial contribution.
U.S. Manufacturers Shipments Electronic Systems Non-Unified Architecture
YEAR |
SHIPMENTS ($ in millions) |
YEAR |
SHIPMENTS ($ in millions) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2003 $689 1991 $954
2002 $835 1990 $1095
2001 $1400 1989 $1039
2000 $1492 1988 $860
1999 $1484 1987 $778
1998 $1225 1986 $690
1997 $1428 1985 $705
1996 $1546 1984 $652
1995 $1432 1983 $441
1994 $1352 1982 $519
1993 $1307 1981 $420
1992 $1236 (Classification did not exist prior to 1981)
Terrence K. McMahon, McMahon Technology Associates, [email protected]