CENTRALIZED AND DISTRIBUTED CONTROL SYSTEMS
Until the beginning of 1980s, computer control was strictly centralized. Usually a single large computer or minicomputer (e.g. the DEC PDP11 series) was used to control the plant. The computer, associated power supplies, input–output, keyboard and display unit were all situated in a central location. The advantages of centralized control are as follows:
• It is easy to manage the computer.
• Only one computer is used.
• Less number of people are required.
In a centralized control system, the controller algorithm is implemented in a single central computer. Hence, all sensors, actuators, input units and output units must be connected directly to this central computer.
Today, distributed control is more widely used. A distributed control system (DCS) consists of a number of computers installed at different locations, each performing an independent control action. Distributed control has emerged as a result of the sharp decrease in price, and the consequent widespread use, of computers. Also, the development of computer networks has made it possible to interconnect computers in a local area network (LAN), as well as in a wide area network (WAN). The main advantages of DCSs are as follows:
• A higher performance is obtained from a distributed system than from a centralized control system.
• A distributed system is more reliable than a centralized system. In the case of a centralized system, if the computer fails, the whole plant becomes unusable. In a DCS, if one computer fails, only a small part of the plant will be affected and the load of the failed computer can usually be distributed among the other computers.
• A DCS can easily be expanded by adding more computers to the network. For example, if 10 computers are used to control the temperature of 10 ovens, then if the number of ovens is increased to 15, it is easy to add five more computers to the network.
• A DCS is more flexible than a centralized control system as it can be easily adjusted to plant requirements.
In a DCS the sensors and actuators can be connected to local computers which can execute localized controller algorithms. Thus, the local computers in a distributed control environment are usually used for direct digital control (DDC). In a DDC application the computer is used only to carry out the control action for the plant. It is also possible to add some level of supervisory control action to a DDC computer, such as displaying the values of sensors, inputs and outputs.
Distributed control systems are generally used as client–server systems. In such a system one computer (or more if necessary) is designated as the server and carries out the common control operations. Other computers in the system are called clients and they obey and implement the instructions they receive from the master computer. For example, the task of a client computer could be to receive and format analog data from a sensor, and then pass this data to the server computer every second.
Distributed control systems usually exist within finite boundaries, such as within a factory complex, and all the computers communicate with each other using a LAN cable. Wireless LAN systems are becoming popular, and there is no reason why a DCS cannot be constructed using wireless LAN technology. Using wireless, system reconfiguration is as easy as just adding or removing a computer.