Dominant Poles

In control systems, a transfer function can have multiple poles, but not all poles contribute equally to the transient response. The concept of dominant poles allows engineers to simplify higher-order systems into an approximate second-order model, making analysis and design far more tractable. This approximation is fundamental in GATE and university examinations.

Core Concept Explanation

Every pole of a closed-loop transfer function contributes a decaying exponential term to the time-domain response. A pole at s = -a + jb produces a response component of the form e^(-at) multiplied by a sinusoidal term. The rate at which this component decays depends directly on the real part magnitude, which is the distance of the pole from the imaginary axis.

Poles that are far to the left in the s-plane have a large negative real part. Their exponential components decay very rapidly and die out before the slower poles even begin to influence the response. As a result, the overall transient behavior of the system is shaped almost entirely by the poles that are closest to the imaginary axis, which are the dominant poles.

A pole is considered non-dominant if its real part magnitude is at least 5 to 10 times larger than the real part of the dominant poles. Under this condition, the contribution of the non-dominant pole decays so fast relative to the dominant poles that it can be safely neglected in the approximate analysis.

Approximation Criteria

The dominant pole approximation is valid when the following condition holds. If the dominant poles are at s = -ζωn ± jωd, and a non-dominant pole is at s = -a, then the approximation is acceptable when |a| >= 5 * ζωn. This ensures the faster pole decays within a time frame much shorter than the system's settling time.

When the approximation is applied, the higher-order transfer function is reduced to a standard second-order form. The standard second-order transfer function is written as G(s) = ωn² / (s² + 2ζωns + ωn²). All the well-known time-domain specifications such as rise time, peak time, settling time, and percent overshoot can then be directly computed using standard second-order formulas.

Mathematical Expression

Consider a third-order system with transfer function G(s) = K / [(s² + 2ζωns + ωn²)(s + a)]. The partial fraction expansion produces three response terms. The term corresponding to the pole at s = -a has a time constant of τ = 1/a. When a >> ζωn, this time constant is negligibly small, and the third pole does not appear in the approximate step response.

The percent overshoot formula %OS = exp(-πζ / sqrt(1-ζ²)) × 100 and settling time formula ts = 4 / (ζωn) are both derived assuming a pure second-order system. Dominant pole approximation justifies using these formulas for higher-order systems, making them central to GATE numerical problems.

Practical Understanding

In real control system design, most plants are higher-order due to mechanical flexibility, electrical parasitics, or thermal dynamics. Designing controllers by accounting for every pole is computationally expensive. The dominant pole concept allows a designer to focus compensator design on the one or two poles that actually determine the visible transient behavior.

However, dominant pole approximation can fail when a zero is present near the non-dominant pole, partially cancelling it. In such cases, the zero-pole pair may still leave a residue in the response, and the approximation may overestimate how much the higher-order pole can be ignored. Always verify that no zero exists near the pole being neglected.

Given:
Third-order system G(s) = 500 / [(s² + 4s + 5)(s + 20)]
Dominant poles: s² + 4s + 5 = 0 → s = -2 ± j1
Real part of dominant poles: ζωn = 2
Non-dominant pole: s = -20, |real part| = 20

Why this formula applies:
Check ratio: 20 / 2 = 10 ≥ 5, so dominant pole approximation is valid.

Formula:
%OS = exp(-πζ / sqrt(1 - ζ²)) × 100
ts = 4 / (ζωn)
ωn = sqrt(5) ≈ 2.236 rad/s, ζ = 2 / 2.236 ≈ 0.894

Substitution:
%OS = exp(-π × 0.894 / sqrt(1 - 0.894²)) × 100
    = exp(-π × 0.894 / 0.447) × 100
    = exp(-6.28) × 100

Calculation:
%OS ≈ 0.19% (heavily damped, negligible overshoot)
ts = 4 / 2 = 2 seconds

Final Answer:
Approximate settling time = 2 s, Overshoot ≈ 0.19%
Third-order system treated as second-order using dominant pole approximation.

Exam Tip: GATE often gives a third or fourth-order system and asks for settling time or overshoot. Always first check if the real part of the outermost pole is 5 or more times larger than the dominant poles. If yes, reduce to second-order and directly apply standard formulas. Missing this step leads to unnecessary computation.

• Dominant poles are closest to the imaginary axis in the s-plane and have the slowest decaying exponential components.
• Non-dominant poles are at least 5 times farther from the imaginary axis compared to dominant poles.
• The approximation replaces the higher-order system with an equivalent second-order system formed by the dominant poles.
• Standard second-order formulas for overshoot, rise time, and settling time can then be directly applied.
• Approximation breaks down if a zero is located near the pole being neglected, because partial cancellation may leave a non-negligible residue.

Quick Revision

• Dominant poles are the closed-loop poles closest to the imaginary axis, controlling the dominant transient behavior.
• Approximation criterion: non-dominant pole real part must be 5 to 10 times greater than dominant pole real part.
• Key formulas after reduction: %OS = exp(-πζ / sqrt(1-ζ²)) × 100, ts = 4/(ζωn), tp = π/ωd.
• Approximation fails when a system zero is near the non-dominant pole.
• GATE trap: do not apply second-order formulas to a higher-order system without first verifying the dominant pole condition.
• A zero near the dominant poles can increase overshoot beyond the standard second-order prediction.