Controllers

Controllers and final control elements act on the process. Learn feedback control loops, controller action (direct/reverse), control valves (characteristics, sizing with Kv/Cv, fail-safe), on/off vs modulating control, cascade, ratio and feedforward strategies, and loop troubleshooting.

A control loop measures a process variable, compares it with the setpoint, and moves a final element — usually a control valve — to correct errors. Most plants run thousands of such loops.

Selecting the valve and the control strategy matters as much as tuning the controller.

An analogy

Driving a car at constant speed: your eyes read the speedometer (sensor), your brain compares with the target (controller), your foot moves the accelerator (valve).

The feedback loop

Sensor → transmitter → controller (in a DCS/PLC) → valve actuator → process. The controller's action must be set correctly: if increasing output raises PV, the controller must be reverse-acting.

Control valves

Globe, ball, butterfly and rotary valves with pneumatic actuators and positioners. Characteristics: linear, equal-percentage (common for varying pressure drop), quick-opening. Fail-safe position (air-to-open fails closed) is chosen for safety — e.g. fuel valves fail closed, cooling water fails open.

Control strategies

On/off (heaters, simple levels), PID feedback (most loops), cascade (inner fast loop, e.g. flow, under outer slow loop, e.g. temperature), ratio (keep blend ratios), feedforward (measure disturbance and correct before it upsets PV), split-range and override/selector control.

Performance and advanced control

Valve stiction, oversizing and poor positioners cause oscillations — loop audits find them. Model predictive control (MPC) handles multivariable, constrained processes (distillation, FCC). Control performance monitoring tracks variability, which translates directly into money.

Formulas

• Error: e = SP − PV (reverse-acting) or PV − SP (direct)
• Valve sizing (liquid): Q = Kv √(ΔP/SG) [m³/h, bar]
• Conversion: Cv ≈ 1.156 Kv
• Ratio control: F_B,SP = R × F_A,measured

Worked example: Size a control valve

Given: Water 50 m³/h; ΔP across valve 1 bar, SG 1

Find: Required Kv and valve choice.

• Kv = 50/√1 = 50
• Choose valve so that design flow is at ≈ 70 % opening → rated Kv ≈ 71 (Cv ≈ 83)

Answer: Kv ≈ 50 needed; select Kv ≈ 70–80 valve

Worked example: Controller action

Given: Level controller moving an outlet valve (fail-open)

Find: Action.

• Opening outlet lowers level → increasing OP decreases PV
• So controller is direct-acting (output rises when PV rises)

Answer: Direct-acting

Common mistakes

• Wrong: Oversizing control valves. — Right: An oversized valve works near closed with poor resolution — size for 60–80 % opening.
• Wrong: Wrong controller action. — Right: Check sign: a wrong action drives the valve to the limit (runaway loop).

Applications

• Heat exchanger temperature: Cascade temperature-to-steam-flow control.
• Blending: Ratio control for fuels, additives and reactor feeds.
• Pressure & level: Split-range and override control on vessels and columns.

Key takeaways

• Sensor → controller → valve → process.
• Kv = Q/√(ΔP/SG); size for 60–80 % open.
• Fail-safe positions and correct action.
• Cascade, ratio, feedforward improve performance.

Going further

• Valve stiction: Detected from PV–OP plots; causes limit cycles.
• Model predictive control: Optimises multiple outputs within constraints each few minutes.
• Relative gain array: Chooses input–output pairings for multivariable plants.