What an automation system does
An industrial automation system is a set of equipment and software that runs part of a physical process with limited human involvement. It watches conditions, compares them with what is wanted, and acts to close the gap. The process might be filling bottles, holding a tank at a steady temperature, moving boxes along a line or sequencing a machine through its steps.
Nearly every automation system, from a single machine to a whole plant, repeats the same cycle: sense, decide, act, and report. Once you see that cycle, most of the vocabulary in this field starts to make sense.
The four building blocks
Sensors
Sensors turn a physical condition into a signal. Typical examples are devices that detect whether a part is present, measure temperature or pressure, read position or speed, or report the level of a liquid. A sensor does not decide anything. It only reports.
Controllers
The controller is where decisions are made. It reads signals from the sensors, runs a stored program, and produces output signals. The most common industrial controller is the programmable logic controller, often called a PLC. Others include controllers built for process loops, motion controllers, and ordinary industrial computers. What they share is a repeating cycle: read the inputs, run the logic, update the outputs, and start again, typically many times a second.
Actuators
Actuators are the devices that do the physical work the controller asks for. They include motors, valves, cylinders, heaters, pumps and relays. Some are simply on or off. Others can be positioned or driven at a variable speed, which gives finer control.
Software and interfaces
Software sits around the controller. At the machine, operators use screens and buttons to start, stop and adjust things. Above that, supervisory software gathers data from many controllers, shows the state of the process, records history and raises alarms. Engineers also use software to write and change the controller program itself.
How the loop closes
Consider a hypothetical tank that should stay about half full. A level sensor reports the current height. The controller compares it with the target. If the level is low, it opens a fill valve through an output. As the level rises the reading changes, and the controller closes the valve again. The sensor, controller and actuator form a loop, and the loop keeps correcting itself. This idea, called feedback, is explored further in the article on process variables.
Not every task is a feedback loop. Many machines mainly follow a sequence: wait for a part, clamp it, run a cycle, release it. This is often called discrete or sequential control, and it is built from the same four blocks.
Levels in a typical system
People often describe an automation system in layers, sometimes drawn as a pyramid:
- Field level: sensors and actuators out on the machine or process.
- Control level: controllers that run the logic.
- Supervisory level: operator displays and data collection, often handled by SCADA systems.
- Operations and business level: software that schedules production, tracks quality and connects to company systems.
Real plants vary, and modern designs blur some layers together, but the model is a useful map. The signals between layers travel over wiring and over industrial networking, which is its own topic.
Signals and communication
At the lowest level, devices speak in simple electrical signals. A switch is either open or closed. An analog sensor might vary a current signal smoothly across a range, with a common industrial convention using a current between 4 and 20 milliamps to represent the span of a measurement. Between controllers and higher-level software, digital communication carries values, commands and status. The details differ widely between plants and eras, which is one reason old and new equipment often coexist.
Why plants automate
The reasons are usually practical rather than dramatic. Automation can make a process more repeatable, reduce the number of repetitive or hazardous tasks done by hand, keep records, and allow operators to supervise more equipment. It also has costs: design, installation, training and ongoing maintenance, and results depend heavily on how well a system is designed and kept up. For how that plays out in budgets and results, see the discussion of automation budgeting.
Safety and reliability
Automation does not remove the need for safety thinking. Hazardous machines normally have separate protective devices such as emergency stop circuits that are designed independently of ordinary control logic. Servicing should follow the equipment manufacturer’s instructions and the plant’s own procedures, including lockout tagout practices, and should be done by trained, authorized personnel. Connected systems also create a need for industrial cybersecurity, since a controller reachable from a network is something to protect.
Where to go next
With the four blocks in mind, the natural next steps are the physical hardware, covered in the article on control system hardware, and the measurements that feed the loop. Details of programming, networking and specific industries build on the same foundation. This article is general information, not engineering advice, and specific systems should always be designed and maintained by qualified people.