What a process variable is
A process variable is a condition of a process that can be measured and that matters to the result. In most plants the four most common are temperature, pressure, flow and level. Others, such as pH, conductivity, speed, weight and humidity, appear in particular industries.
The goal of control is usually to hold a variable near a target value called the setpoint, or to move it along a planned path. The measured value is called the process value or measurement. The difference between the two is the error.
Measuring the common variables
Temperature
Temperature is often measured with a probe whose electrical behavior changes with heat. Two broad families are thermocouples, which generate a small voltage related to temperature, and resistance-based sensors, whose electrical resistance changes in a known way. Each has strengths in range, accuracy and cost, and the choice depends on the application.
Pressure
Pressure instruments typically sense how much a diaphragm or similar element flexes under the process pressure and convert that into an electrical signal. Gauge readings are relative to the surrounding atmosphere, while absolute readings are relative to a vacuum. Knowing which is being used avoids confusion.
Flow
Flow can be measured in many ways: by the pressure drop across a restriction, by the speed of a spinning element, by electromagnetic or ultrasonic effects, or by directly measuring mass. The choice depends on the fluid, the pipe and the accuracy needed. See flow meters for how those methods compare.
Level
Level can be found by floats, by pressure at the bottom of a tank, by radar or ultrasonic echoes, or by simple point switches that only say whether a tank is above or below a mark. The topic gets its own treatment under tank level measurement.
From measurement to signal
The sensing element rarely sends its raw output to the controller. A transmitter conditions it and produces a standard signal. A well-known industrial convention uses a current that varies between 4 and 20 milliamps across the measuring range. A reading at the low end of the range gives 4 milliamps and at the high end gives 20. Because 4 milliamps rather than zero represents the bottom of the range, a broken wire shows up as an obviously wrong signal rather than a plausible low reading.
A short hypothetical example: suppose a pressure transmitter is set up to cover 0 to 10 units, with the output running linearly from 4 to 20 milliamps. At the midpoint, 5 units, the signal is 12 milliamps, since 4 plus half of the 16 milliamp span is 12. The controller then converts the signal back to the number operators see.
Closing the loop
A feedback loop has three jobs: measure, compare, act. A hypothetical heated tank illustrates it. A temperature sensor reports the liquid temperature. The controller compares it with the setpoint. If it is too cold, the controller increases the output to a heating element or steam valve. As the liquid warms, the error shrinks and the output is reduced.
The device that changes the process is called the final control element, commonly a control valve or a variable speed drive. These are the subject of control valves and valve actuators.
How controllers decide
The simplest approach is on/off control, like a household thermostat: heat is switched on below a lower limit and off above an upper limit. The variable swings back and forth within a band. It is simple and often adequate where tight accuracy is not required.
For smoother results, many loops use a PID algorithm, named for proportional, integral and derivative actions:
- Proportional: the output responds in proportion to the current error.
- Integral: the output also builds up while an error persists, which removes steady offsets.
- Derivative: the output reacts to how quickly the error is changing, which can damp overshoot.
Settings for these actions are called tuning. Poor tuning can make a loop sluggish or make it oscillate, and good tuning depends on how the particular process responds. Many loops need only proportional and integral action. The design questions are covered further under control loop design.
Why measurements go wrong
A loop can only be as good as its measurement. Common problems include sensors installed in a poor spot, fouled or damaged probes, electrical noise on signal wiring, and slow drift over time. The last is addressed by periodic checks against a known reference, often called calibration. Delays also matter: a sensor placed far from the point where the process changes will report late, and a late measurement makes control harder.
Alarms and limits
Besides control, measurements feed alarms. A pressure that exceeds a high limit may raise an alert for operators. Protection against dangerous conditions is normally handled by separate safety devices rather than by the ordinary control loop alone.
Summary of the idea
Process control is the habit of measuring what matters, comparing it with a target, and adjusting something physical to reduce the difference. Different equipment implements it, but the logic stays the same. This article is general information, not engineering advice, and any real instrument or loop should be set up by qualified people following manufacturer instructions.