PI Controller

A PI controller combines proportional and integral actions to eliminate the steady-state error that a pure P controller leaves behind. It is the most widely used controller in industrial process control, and understanding its behavior is essential for GATE and semester exams on control systems.

Core Concept Explanation

The integral term in a PI controller accumulates the error over time. If a steady-state error persists even for a short duration, the integral term keeps growing until the controller output is large enough to drive the error to zero. This is the fundamental mechanism by which integral action eliminates offset. In mathematical terms, u(t) = Kp*e(t) + Ki * integral(e(t) dt), where Ki is the integral gain.

The addition of an integrator in the forward path increases the system type by one. A type-0 plant with a PI controller becomes a type-1 closed-loop system, which guarantees zero steady-state error for a step input. This is why PI controllers are preferred in process control applications such as temperature regulation, flow control, and pressure control, where a persistent offset is unacceptable.

The trade-off introduced by the integrator is a reduction in phase margin, since an integrator contributes -90 degrees of phase. This tends to slow the transient response and may increase overshoot compared to a pure P controller with the same proportional gain. The ratio Ki/Kp defines a zero in the controller transfer function at s = -Ki/Kp, and the designer chooses this zero location to balance error elimination speed against stability.

Mathematical Expression

The transfer function of the PI controller in the Laplace domain is Gc(s) = Kp + Ki/s = (Kp*s + Ki) / s. This shows a pole at the origin (the integrator) and a zero at s = -Ki/Kp. The integrator pole adds -20 dB/decade to the Bode magnitude at low frequencies and adds -90 degrees to the phase. The zero partially compensates for this phase loss, but the net effect at low frequencies is still a phase reduction.

An alternate parameterization uses the integral time constant Ti = Kp/Ki, giving Gc(s) = Kp * (1 + 1/(Ti*s)). In this form, Ti represents how long it takes the integral action to contribute an amount equal to the proportional action when the error is constant. Smaller Ti means faster integral action and faster error elimination, but also greater risk of overshoot and oscillation.

Practical Understanding

A well-tuned PI controller provides fast proportional correction during transients and gradual integral correction to eliminate offset. In industrial settings, the integral term also compensates for slow disturbances such as load changes or parameter drift. If Ki is too large, the system exhibits integral windup, where the integrator accumulates a very large value during saturation of the actuator, causing a long overshoot when the actuator unsaturates. Anti-windup schemes limit the integrator output during saturation.

Given:
Plant G(s) = 2 / (s + 4)
PI Controller: Kp = 5, Ki = 2
Input: unit step

Why this formula applies:
PI controller zero is at s = -Ki/Kp.
Closed-loop DC gain check for steady-state error.

Formula:
Gc(s) = (Kp*s + Ki) / s = (5s + 2) / s
System type after PI = type-1 -> ess = 0 for step

Substitution:
Open-loop = Gc(s)*G(s) = [(5s+2)/s] * [2/(s+4)]
= (10s + 4) / [s(s+4)]

Calculation:
This is type-1 system (one pure integrator in forward path).
For step input, ess = 0 by definition.
Velocity error constant Kv = lim s->0 s * open-loop
= lim s->0 s * (10s+4)/[s(s+4)] = 4/4 = 1

Final Answer:
Steady-state error for step = 0.
Kv = 1 rad/sec (ramp tracking error = 1/Kv = 1)

Exam Tip: A PI controller adds a pole at origin and a zero at s = -Ki/Kp. The additional pole increases system type by 1 (eliminates step error), but the added phase lag can reduce stability. Always check phase margin after adding integral action.

Quick Revision

• PI controller: Gc(s) = Kp + Ki/s = (Kp*s + Ki)/s. Has one pole at origin and one zero at s = -Ki/Kp.
• Integral action adds one to system type: type-0 plant becomes type-1 system with PI controller.
• Zero steady-state error for step input is guaranteed for type-1 or higher systems.
• Integral gain Ki controls speed of error elimination; large Ki causes faster correction but more overshoot.
• Integral windup occurs when actuator saturates and integrator accumulates excessively; needs anti-windup.
• Phase margin decreases with PI compared to P alone because the integrator adds -90 degrees of phase at low frequencies.
• Exam trap: PI controller does NOT add phase lead anywhere; it is a lag-type controller.