Root-Locus Oscillatory Stability Analysis

• Root-locus-based oscillatory stability analysis is a technique for quantifying stability margins and pole sensitivity in LTI systems as feedback gain is varied.
• The method leverages residue-based sensitivity and continuation algorithms to efficiently trace pole trajectories and determine oscillatory crossing points.
• Practical applications span classical SISO feedback, time-delay systems, and networked power grids, offering actionable insights for controller design and stability certification.

Root-locus-based oscillatory stability analysis is a foundational approach in control theory for characterizing and quantifying the stability margins, oscillatory modes, and sensitivity properties of closed-loop linear time-invariant (LTI) systems as a designated system parameter—most commonly feedback gain or delay—is varied. Its practical significance spans classical SISO feedback, infinite-dimensional systems, time-delay processes, and advanced applications to large-scale networks such as weakly connected power systems. Recent advances leverage analytical residue-based sensitivity, continuation methods for dead-time systems, and decomposed modal analysis in networked systems to yield efficient, quantitative criteria and computational pipelines for oscillatory stability certification.

1. Analytical Foundations: Partial Fraction Decomposition, Residues, and Pole Sensitivity

The classical root locus framework is built upon the partial-fraction decomposition of rational transfer functions: $$G(s) = \frac{N(s)}{D(s)} = \sum_{i=1}^n \frac{r_i}{s-p_i}$$ where $p_i$ are the simple roots of $D(s)=0$, and residues $r_i$ are defined as $$r_i = \lim_{s\to p_i} (s-p_i)\,G(s) = N(p_i)/D'(p_i)$$. In closed-loop, for unity feedback with gain $K$, the characteristic equation becomes

$A(s,K) = D(s) + K N(s) = 0$

and the roots $p_i(K)$ as functions of $K$ form the root-locus branches.

A critical theoretical insight, as established by Tebaldi and Zanasi, is that the sensitivity of each closed-loop pole to gain,

$\frac{dp_i}{dK} = -r_i^{(0)}$

where $r_i^{(0)}$ is the residue of the closed-loop transfer function at $p_i$, provides an exact analytic expression for the instantaneous velocity of the pole in the complex plane as $K$ varies. This residue-based sensitivity result underpins both fine-grained analytic stability margin assessments and efficient root-locus construction algorithms (Tebaldi et al., 2024).

2. Residue-Based Algorithmic Construction of Root Locus Curves

Traditional root locus plotting requires numerical root-finding for each parameter value, which becomes computationally intensive for high-order systems. Tebaldi & Zanasi present an efficient update rule exploiting residue structure: $$p_i(K+\Delta K) = p_i(K) - \frac{\Delta K\, N\bigl(p_i(K)\bigr)} {A\bigl(p_i(K),K\bigr)\prod_{h\ne i}[p_i(K) - p_h(K)]}$$ This difference-equation update iteratively propagates each pole along its root-locus branch, leveraging only evaluations of $N$, $D$, and current pole locations—thus avoiding expensive polynomial root solves at each step. Empirical results demonstrate substantial computational gains over standard tools (e.g., MATLAB's rlocus), with output equivalence confirmed (Tebaldi et al., 2024). Sensitivity |$r_i$| near the imaginary axis directly reflects the local root-locus "speed", quantifying closed-loop pole rapidity and sensitivity.

3. Oscillatory Stability Margins: Criteria and Calculation

Oscillatory (Hopf-type) instabilities arise when a complex-conjugate pair of poles crosses the imaginary axis as a system parameter is increased. The crossing conditions can be reduced to joint amplitude-phase equations: $$|D(j\omega)| = K|N(j\omega)|,\quad \arg D(j\omega) - \arg N(j\omega) = (2m+1)\pi$$ for minimal positive $\omega$ and integer $m$. The corresponding critical gain $K_{\text{crit}}$ yields the gain margin, and the crossing frequency specifies the emergent oscillation. Along the locus, the damping ratio $\zeta = -\Re(p_i)/|p_i|$ and natural frequency $\omega_n = |p_i|$ are readily extracted branch-wise. The loci of |r_i| serve as a "pole velocity" diagnostic: high velocities near the imaginary axis signal pronounced parameter sensitivity and reduced robustness (Tebaldi et al., 2024).

4. Extension to Infinite-Dimensional, Delay, and Distributed Systems

The root-locus concept extends rigorously to infinite-dimensional systems, including those governed by PDEs and delay equations. The transfer function $G(s)$ becomes meromorphic, with poles prescribed by the spectrum of the system's generator operator on a Hilbert space. The root locus is here defined as the set of closed-loop eigenvalues as a scalar feedback gain $K$ is swept: $G(\lambda) = -\frac{1}{K}$ Global properties include branch continuity, non-intersection, and (for self-adjoint or skew-adjoint generators) strict interlacing of poles and zeros, ensuring LHP branch confinement and monotonic motion under increasing K (Jacob et al., 2014). In PDEs or systems with delays, root-locus analysis allows for direct computation of stability margins, with branch behaviors such as infinite escape, accumulation at zeros, or axis interlacing characterized by operator-theoretic structure.

For SISO time-delay (dead-time) systems, closed-loop characteristic roots satisfy transcendental equations. Here, predictor–corrector continuation algorithms parameterized by arclength in (Re(s), Im(s), K)-space provide robust, globally convergent tracing of root-locus branches, robust to high-order and branch point-induced ill-conditioning. Initial seeds include open-loop poles, boundary crossings, and infinity (asymptotes), with crossing criteria for marginal stability formulated in terms of magnitude and phase equations evaluated on lines of prescribed real part (Gumussoy et al., 2020, Gumussoy et al., 2020).

5. Application to Networked Systems: Modal Decomposition and Frequency-Shaping Tuning

In large-scale networked systems, such as weakly connected power grids, oscillatory stability is governed by the inter-area modal structure exposed under linearization around operating points. Through coordinate diagonalization, the system decomposes into a center-of-inertia (COI) mode and inter-area oscillatory modes, indexed by Laplacian eigenvalues $\lambda_k$: $$\hat z_k(s) = \frac{g_o(s)}{1 + g_o(s)(\lambda_k/s - c_o(s))}$$ Oscillatory stability for all modes is reduced to a pole-placement task for a tractable scalar subsystem. Under Frequency-Shaping (FS) control, the main-mode loop gain is bi-quadratic, yielding explicit closed-form pole loci. The minimum damping ratio and decay rate are

$$\zeta_{\min} = \min\left\{1,\; \frac{d + d_b + d_t}{2\sqrt{m\lambda_n}}\right\},\quad \alpha_{\min} = \min\left\{ \frac{d+d_b+d_t}{2m},\; \frac{(d+d_b+d_t)-\sqrt{(d + d_b + d_t)^2 - 4m\lambda_2}}{2m} \right\}$$

and are directly sight-readable from the root-locus geometry as $\lambda_2, \lambda_n$ are varied (Jiang et al., 27 Jan 2026). This structure supports analytic tuning rules: $$d_b = \max\left\{ \textrm{COI threshold, oscillatory stability lower bounds} \right\}$$ Contrasted with Virtual Inertia (VI) methods, where root-locus analysis becomes algebraically intractable, FS-based loci provide transparent design trade-offs among response time, damping, and robustness.

6. Worked Examples, Computational Aspects, and Practical Insights

Combining these concepts, analytic and algorithmic frameworks for oscillatory stability assessment emerge:

• For rational plants, explicit residue-based updates trace root-locus curves with lower complexity.
• In the example $G(s) = 4s/[s(s+4)((s+1)^2+4)]$, initial poles are found, residues computed, and root-locus branches constructed by difference updates. The critical gain $K_{\text{crit}}$ and oscillation frequency $\omega_c$ at marginal stability are evaluated by phase-norm conditions; damping ratios decrease along branches as poles approach the imaginary axis, directly quantifiable by residue magnitudes (Tebaldi et al., 2024).
• For delay systems $G(s)e^{-\tau s}$, continuation-based tracing with predictor-corrector steps (utilizing secant predictions, Newton corrections, and adaptive step size) robustly plots roots as gain or delay varies, finds crossing points, and identifies stability windows (Gumussoy et al., 2020, Gumussoy et al., 2020).
• In power grids under FS control, the locus underpins algebraic tuning ensuring all modal decay rates and damping ratios meet explicit thresholds, guaranteeing no frequency nadirs and quantifiable security (Jiang et al., 27 Jan 2026).

7. Current Directions and Generalization

Root-locus-based oscillatory stability analysis continues to generalize to increasingly complex scenarios: distributed-parameter systems, high-order and nonrational (dead-time) dynamics, and structured network-feedback architectures. The underlying residue-sensitivity connection provides analytic control over local and global system behavior; modal decomposition and continuation techniques address infinite-dimensionality and nonpolynomial feedback; algorithmic advances improve computational tractability. These developments enable rigorous, quantitative stability assessment and controller synthesis across diverse modern control applications (Tebaldi et al., 2024, Jacob et al., 2014, Gumussoy et al., 2020, Jiang et al., 27 Jan 2026).