What is DCS? (Distributed Control System)
A distributed control system (DCS) is a process plant control system that uses independent controllers scattered throughout the system.
In the mid-1970s, the world of industrial control was introduced to a completely new concept: distributed digital control. During that time, direct digital control had a significant flaw: the risk of catastrophic failure if a single digital computer performing numerous PID control functions came to a standstill.
Although digital control has numerous benefits, it isn’t worth the risk if the entire operation will shut down (or catastrophically collapse!) as a result of a hardware or software failure on that one computer.
Distributed control solved this issue directly by dispersing and networking many control computers, each of which is accountable for a small number of PID loops.
There would be less concentration of liability with individual process control “nodes” dispersed throughout the campus, each dedicated to controlling just a few loops, than there would be with a single-computer DDC system. Because the hundreds or thousands of analogue field instrument connections only had to extend as far as the computing hardware, the analogue signal wiring was also shortened.
Only the networking wire needed to go that far, resulting in a significant reduction in wiring requirements.
Furthermore, distributed control brought the concept of redundancy to industrial control systems, with “spare” digital signal acquisition and processing hardware units designed to take over all important functions automatically in the case of a primary failure.
Distributed Control Systems (DCS)
A typical distributed control system (DCS) architecture is depicted in the diagram below:
Each “rack” has a microprocessor that handles all of the control operations, as well as separate I/O (input/output) “cards” that convert analog field instrument signals to digital format and vice versa. The possibility of component failure is addressed via redundant CPUs, redundant network cables, and even redundant I/O devices.
Routine self-checks on redundant system components are typically programmed into DCS processors to assure the availability of spare components in the case of a failure. If one of the “control racks” fails completely and the redundancy is insufficient to handle the fault(s), just the PID loops in that rack will be affected, not any of the other loops across the system.
Similarly, if the network cables are severed or otherwise damaged, just the data flow between those two places will be affected; the remainder of the system would continue to function normally.
As a result, one of the “hallmark” aspects of a DCS is its tolerance for serious errors: the impact on process management is minimized by design, even in the face of severe hardware or software faults.
The Honeywell TDC2000 system (Note 1), introduced in 1975, was one of the world’s first distributed control systems. The technology was primitive by today’s standards, but the concept was innovative.
Note 1: To be fair, the CENTUM distributed control system from Yokogawa Electric Corporation of Japan was unveiled the same year as Honeywell’s.
Each rack (Honeywell referred to it as a “box”) was made up of an aluminium frame that held numerous huge printed circuit boards with card-edge connectors. In the left-hand photo, you can see a “basic controller” box.
The termination board, seen on the right, is where the field wiring (420 mA) connections were established. Each termination board was connected to its appropriate controller box by a thick cable:
The TDC2000 DCS provided controller redundancy in the form of a “spare” controller box that could be used as a backup for up to eight other controller boxes. All analog signals were routed over thick cables to this spare controller, allowing it to take over in the case of a failing controller.
As a result, this redundancy solution protected against both processor and I/O failures. The “Data Hiway,” a dual coaxial cable network that connected all TDC2000 controllers, was used to communicate digitally. The dual wires enabled network communications redundancy.
The following image shows a typical TDC2000 operator workstation:
The Honeywell system became more sophisticated over time, with quicker networks (the “Local Control Network” or LCN), more capable controller racks (the “Process Manager” or PM series), and improved operator workstations.
Many of these enhancements were gradual, consisting of add-on components that could be used in conjunction with existing TDC2000 components to ensure that the overall system could be improved.
Other control equipment manufacturers responded to Honeywell and Yokogawa’s DCS revolution by introducing their own distributed control systems.
Examples include the Bailey Network 90 (Net90) DCS, Bailey Infi90 DCS, and Fisher Provox systems. Foxboro, already a leader in the control system field with their SPEC 200 analog system, first added digital capabilities to the SPEC 200 (the VIDEOSPEC workstation consoles, FOX I/A computer, INTERSPEC and FOXNET data networks), then developed an entirely digital distributed control system (the VIDEOSPEC workstation consoles, FOX I/A computer, INTERSPEC and FOXNET data networks).
The following are some examples of current distributed control systems:
- ABB 800xA
- Emerson DeltaV and Ovation
- Foxboro (Invensys) I/A
- Honeywell Experion PKS
- Yokogawa CENTUM VP and CENTUM CS
DCS IO Modules
Examine the following photograph of an Emerson Delta V DCS rack with the processor and various I/O modules for visual comparison with the Honeywell TDC2000 DCS:
Because of their ever-increasing speed, functionality, and low cost, programmable logic controllers (PLCs) are becoming more and more popular as PID control platforms, as previously noted in the Direct Digital Control (DDC) part.
With modern PLC hardware and networking capabilities, it is now possible to create a truly distributed control system with individual PLCs serving as processing nodes and redundancy built into each node to ensure that important control functions are not disrupted by a single failure. A system like this may be obtained for a fraction of the price of a full-fledged DCS.
However, the same level of hardware and software integration required to produce a viable distributed control system that is as ready-to-use as a system pre-built by a DCS vendor is currently lacking in the PLC sector.
To put it another way, if a company decides to construct its own distributed control system utilizing programmable logic controllers, it must be prepared to put in a lot of effort.
Any engineer or technician who has used a modern DCS, with its self-diagnostic, “smart” instrument management, event auditing, advanced control strategy, pre-engineered redundancy, data collection and analysis, and alarm management capabilities, understands that these features are neither frills nor easy to implement. Woe betides anyone who thinks these crucial traits aren’t important.