In the field of industrial automation, DCS represents the most widely adopted solution for large continuous plants, including refineries, energy production plants, paper mills, glassworks, and chemical plants. The characteristics of the DCS include the possibility of distributing modules for data acquisition, processing and control, and a communication network that is as efficient as possible between the various subsystems. Another important feature of DCS is its ability to change the network topology and add and remove modules on the fly with the system running. DCS can also perform functions that are both normally implemented on PLC and on SCADA in an integrated way.

The typical architecture of a DCS is made up of several levels. Starting from the bottom, we can find the interfaces to the field, consisting of the appropriate electronic acquisition boards (inputs) and commands (outputs). At this level, there are also the communication interfaces for the most common fieldbuses, through which the information is exchanged with the transmitters and actuators that support the same type of protocol.

Through the I/O bus, the values ​​of the inputs and outputs are conveyed to the controllers and processed according to the control strategies and the logic designed for the specific application. The software is developed through the implementation of function blocks, which therefore constitute the so-called system database. Each of these is labelled with a uniquely defined name, commonly known as a tag. It is therefore necessary to associate each database record to the relative tag and its fields to the different variables used in the function block, normally called items. A PID controller, for example, normally represents a database tag. Its items are the process variables, the SP, the control variable, the alarm thresholds, the proportional, integral, and derivative parameters, and so on.

Supervisors are above the controllers. Nowadays, they are usually implemented by commercial off-the-shelf (COTS) operating systems. They act as an interface for the control room operator by accessing the tags and the database items, and consist of numerous screens through which a graphic representation of the system and the process is developed. This is one of the main differences between a DCS and a PLC-SCADA system. PLC and SCADA are two separate systems, each making use of their own variables and data structures, and exchanging data through a suitable communication driver, sometimes making use of a proprietary protocol. In the DCS, however, the database is unique, shared, and distributed between controllers and supervisors, each of which takes control of its processing tasks.

In general, a PLC controls one or more machining processes, while the DCS controls the entire plant. Typically, a DCS is used when:

• The value of the product is very high.
• The production process is continuous.
• Failures in the system will result in a large money loss. The DCS normally has built-in redundancy, ensuring a higher level of system insurance.
• All upgrades are made online while the system runs continuously.

In any case, the differences between the PLC and DCS have been reducing over the years. The most common DCS vendors on the market are ABB, Yokogawa, Honeywell, Emerson, and Siemens.