Flow Lab

Syllabus / Module 7: Systems of ODEs & Dynamical Systems

Lesson 7.3

Phase Portraits & the Trace–Determinant Plane

Classifying 2×2 linear systems: saddles, nodes, spirals, centers, degenerate cases; the trace–determinant map.

Written byClaude Sonnet

🌊
Why this matters for simulation: Critical points of a velocity field (saddles, vortices, sources) organize the whole flow topology.

By the end of this lesson you can…

  • Name the five portraits of $\mathbf x' = A\mathbf x$ (saddle, node, spiral, center, and the degenerate or star node) from the eigenvalues, and sketch each with its eigenlines.
  • Derive $\lambda = (T \pm \sqrt{T^2 - 4D})/2$ from the characteristic polynomial and use $T$, $D$ and $T^2 - 4D$ to classify a system without computing eigenvectors.
  • Decide stability from signs alone: asymptotically stable exactly when $T \lt 0$ and $D \gt 0$.
  • Read the rotation direction from the sign of $a_{21}$ and compute the amplitude factor per turn $e^{2\pi\,\mathrm{Re}\lambda/|\mathrm{Im}\lambda|}$.
  • Say which boundaries of the $(T, D)$ plane are structurally unstable, and show with numbers how a small perturbation destroys a center.
  • Show that explicit Euler turns a center into an outward spiral, measure the gain per step $\sqrt{1 + h^2\omega^2}$, and explain why an incompressible flow has only saddles and centers.
On this page
·
  1. Five pictures from the eigenvalues
  2. Degenerate and star nodes
  3. Trace and determinant: the whole classification in one plane
  4. Worked examples
  5. Moving across the plane: which pictures survive a perturbation?
  6. Numerics: a classifier, and Euler on a center
  7. Computing it
  8. Concept checks
  9. Problems
  10. Key Takeaways

In 7.2 Linear Systems & the Eigenvalue Method we solved x′=Ax\mathbf x' = A\mathbf x when the eigenvalues were real and distinct. This lesson asks a simpler question: what does the picture look like, and can we know it before solving anything? For a 2×22\times2 matrix the answer fits on one page of axes, because two numbers, the trace TT and the determinant DD, decide the whole picture. The payoff for simulation is concrete: the particle paths of a steady 2D flow near a critical point are exactly such a system, and whether a vortex stays a vortex is a statement about the line T=0T = 0. We also meet the case this course cares about most: explicit Euler on a pure rotation does not rotate, it spirals out.

▶ Video Morphing a phase portrait as the point (T, D) moves across the planecaptions only
Transcript
  1. For x prime equals A x, the picture of all trajectories, with arrows for the direction of time, is the phase portrait. Two numbers decide it: the trace T and the determinant D. We move one point around the T D plane and watch the portrait change.
  2. Start with the matrix with rows 1, 2 and 3, 2. Its eigenvalues are 4 and minus 1, real with opposite signs, which a negative determinant guarantees: a saddle. Only the stable line approaches the origin; every other curve is bent away along the unstable one.
  3. Same-sign real eigenvalues give a node. Rows minus 4, 2 and 1, minus 3 have T minus 7, D 10, T squared minus 4 D 9, so lambda is minus 2 and minus 5: stable. Curves enter tangent to the slow line. Rows 4, minus 2 and 1, 1 give 3 and 2: unstable, time reversed.
  4. A complex pair with nonzero real part spirals. Rows minus 0.2, minus 1 and 1, minus 0.2 give lambda equal to minus 0.2 plus or minus i. Each lap multiplies the radius by 0.28461: 28.5 percent is left. A 21 is positive, so it turns counterclockwise.
  5. With the real part zero the spiral closes into ellipses: a center, lambda plus or minus 2 i, period pi, factor exactly 1.00000 per turn. Nearby starts stay nearby: stable, but not asymptotically. Here a 21 is minus 4, so it turns clockwise.
  6. Repeated eigenvalues sit on the parabola. Minus I is a star node: every vector is an eigenvector, so every trajectory is a straight ray. Rows minus 2, 1 and 0, minus 2 have one eigenline only: a degenerate node, every curve tangent to it. Why is the next lesson.
  7. Expanding the determinant gives lambda squared minus T lambda plus D equals 0, so lambda is T plus or minus the square root of T squared minus 4 D, over 2. Negative D is a saddle; positive D gives nodes below the parabola, spirals above, a center on T equals 0. Stable exactly when T is negative and D positive.
  8. The oscillator y double prime plus c y prime plus k y equals 0, with k equal to 4, has T equal to minus c and D equal to 4. Increase c. At 0, a center. At 2, a spiral, minus 1 plus or minus 1.7321 i. At 4, on the parabola, minus 2 twice. At 5, a node.
  9. A center needs T exactly 0. Add 0.01 to a 11: T is 0.01, the real part 0.005, and each turn multiplies the radius by 1.01583, a slow outward spiral. Minus 4.01 in a 21 keeps T at 0, lambda plus or minus 2.0025 i: still a center, but a thin slice of perturbations.
  10. The parabola is a hinge. Perturb rows minus 2, 1 and 0, minus 2 by epsilon in a 21: T squared minus 4 D is 4 epsilon. Epsilon plus 0.01 gives minus 2.1 and minus 1.9, a node; minus 0.01 gives minus 2 plus or minus 0.1 i, a spiral.
  11. Explicit Euler multiplies a center's amplitude by the square root of 1 plus h squared omega squared each step. At h 0.1, 31 steps, the gain is 1.8366; at 0.05, 63 steps, 1.3681; ten laps, 22.7463. RK4 stays near 0.99998629. Euler expands area: det of I plus h A is 1.04.
  12. Near a critical point of a steady 2D flow, A is the Jacobian, and its trace is the divergence. Incompressible means T equals 0: only saddles, like u equals x, minus y, and centers, like minus 2 y, 2 x. From memory, sources and sinks need divergence.
  13. To sum up. D and T squared minus 4 D give the picture; stable means T negative, D positive. The center and the parabola are fragile. Euler turns a center into an outward spiral. Incompressible flows have only saddles and centers. Next: complex and repeated eigenvalues.

Five pictures from the eigenvalues

A phase portrait is the plane of states (x1,x2)(x_1, x_2) drawn with many trajectories, each with an arrow for the direction of time. The origin is the only equilibrium when AA is invertible. The eigenvalues (from 7.2 Linear Systems & the Eigenvalue Method) decide everything. Concrete matrices, with eigenvalues computed in block A of the lab cell below and shown by the sim:

TypeMatrixλ\lambda
saddle(1232)\begin{pmatrix}1&2\\3&2\end{pmatrix}4, −14,\ -1
stable node(−421−3)\begin{pmatrix}-4&2\\1&-3\end{pmatrix}−2, −5-2,\ -5
unstable node(4−211)\begin{pmatrix}4&-2\\1&1\end{pmatrix}3, 23,\ 2
stable spiral(−0.2−11−0.2)\begin{pmatrix}-0.2&-1\\1&-0.2\end{pmatrix}−0.2±i-0.2 \pm i
center(01−40)\begin{pmatrix}0&1\\-4&0\end{pmatrix}±2i\pm 2i

Saddle (real, opposite signs). The two eigenlines carry the two straight-line solutions e4tv1e^{4t}\mathbf v_1 and e−tv2e^{-t}\mathbf v_2. One is pulled in, the other pushed out, so every trajectory off the stable line is bent away along the unstable one. Only the stable line approaches the origin; a point on it that is nudged off drifts away, which is the knife-edge of the last lesson.

Node (real, same sign). Both modes grow, or both decay. In the stable node the −5-5 mode dies first, so every curve bends onto the slow eigenline (λ=−2\lambda = -2, direction (1,1)(1,1)) and enters the origin tangent to it. In the unstable node, running time backwards does the same, so curves leave tangent to the slow (λ=2\lambda = 2) line, and later turn to follow the fast one.

Spiral or focus (complex pair, Reλ≠0\mathrm{Re}\lambda \neq 0). A solution is eαte^{\alpha t} times something that rotates at angular rate ω=∣Imλ∣\omega = |\mathrm{Im}\lambda|. We take that on trust here, since turning the complex solution into a real one is the work of 7.4 Complex & Repeated Eigenvalues, and we check it by measurement. If α<0\alpha \lt 0 the radius shrinks while it turns, so the curve spirals in; if α>0\alpha \gt 0, out.

Center (pure imaginary). α=0\alpha = 0: the amplitude never changes, so trajectories are closed loops around the origin, ellipses for our matrix. It is stable (nearby starts stay nearby) but not asymptotically stable.

Which way does it turn? At the point (1,0)(1, 0) the velocity is A(1,0)T=(a11,a21)A(1,0)^T = (a_{11}, a_{21}). A complex pair needs T2−4D=(a11−a22)2+4a12a21<0T^2 - 4D = (a_{11}-a_{22})^2 + 4a_{12}a_{21} \lt 0, which forces a12a21<0a_{12}a_{21} \lt 0, so the point at (1,0)(1,0) must be moving sideways around the origin, and the sign of a21a_{21} says whether it moves up or down. So a21>0a_{21} \gt 0 means counterclockwise. The stable spiral above has a21=1a_{21} = 1, counterclockwise; the center has a21=−4a_{21} = -4, clockwise. The sim’s readout prints the sign of a21a_{21} and, as a check, the sign of x0×Ax0\mathbf x_0 \times A\mathbf x_0.

How fast does a spiral close? After one full turn, of duration P=2π/ωP = 2\pi/\omega, the rotating part is back where it started and only the factor eαPe^{\alpha P} is left:

ρ=e2π Reλ/∣Imλ∣.\rho = e^{2\pi\,\mathrm{Re}\lambda/|\mathrm{Im}\lambda|}.

For λ=−0.2±i\lambda = -0.2 \pm i this is e−0.4π=0.28461e^{-0.4\pi} = 0.28461: each lap around the origin leaves 28.5 percent of the radius. The measured value is 0.2846100.284610 (block B, RK4 over one period from (1,0)(1,0), and the sim’s readout). Because ρ\rho depends only on the ratio Re/Im\mathrm{Re}/\mathrm{Im}, a small ratio means ρ\rho close to 1 and many laps before the radius shrinks much.

Degenerate and star nodes

When the eigenvalue is repeated, λ1=λ2=λ\lambda_1 = \lambda_2 = \lambda, there are two possibilities. A=λIA = \lambda I (the sim’s (−100−1)\begin{pmatrix}-1&0\\0&-1\end{pmatrix}) is a star node: every vector is an eigenvector, so every trajectory is a straight ray into the origin. (−210−2)\begin{pmatrix}-2&1\\0&-2\end{pmatrix} has the same repeated eigenvalue −2-2 but only one eigenline, and is a degenerate (improper) node: every curve is tangent to that single line. These sit on the parabola in the next section. Why there is only one eigenvector, and the extra solution that replaces the missing one, is the generalised eigenvector story of 7.4 Complex & Repeated Eigenvalues; here we only classify and look.

Trace and determinant: the whole classification in one plane

Expand det⁡(A−λI)\det(A - \lambda I) for A=(abcd)A = \begin{pmatrix}a&b\\c&d\end{pmatrix}:

λ2−(a+d)λ+(ad−bc)=λ2−Tλ+D=0.\lambda^2 - (a + d)\lambda + (ad - bc) = \lambda^2 - T\lambda + D = 0 .

So the eigenvalues are

λ=T±T2−4D2,\lambda = \frac{T \pm \sqrt{T^2 - 4D}}{2},

and, comparing with (λ−λ1)(λ−λ2)(\lambda - \lambda_1)(\lambda - \lambda_2), λ1+λ2=T\lambda_1 + \lambda_2 = T and λ1λ2=D\lambda_1\lambda_2 = D. Now read the pictures off the sign of DD and of T2−4DT^2 - 4D.

  • D<0D \lt 0: the product is negative, so the roots are real with opposite signs: saddle (and T2−4D>0T^2 - 4D \gt 0 is automatic).
  • D>0D \gt 0 and T2>4DT^2 \gt 4D: real roots of the same sign, the sign of TT: node, stable if T<0T \lt 0.
  • D>0D \gt 0 and T2<4DT^2 \lt 4D: a complex pair with real part T/2T/2: spiral, stable if T<0T \lt 0.
  • D>0D \gt 0 and T=0T = 0: λ=±iD\lambda = \pm i\sqrt D: center.
  • T2=4DT^2 = 4D: repeated root T/2T/2: the parabola, with star or degenerate nodes on it.
  • D=0D = 0: a root λ=0\lambda = 0, a whole line of equilibria rather than an isolated one.

Stability. The origin is asymptotically stable when both Reλ<0\mathrm{Re}\lambda \lt 0. For real roots that means the same sign, D>0D \gt 0, and a negative sum, T<0T \lt 0. For a complex pair Reλ=T/2<0\mathrm{Re}\lambda = T/2 \lt 0 and D=∣λ∣2>0D = |\lambda|^2 \gt 0 holds automatically. So in both cases: stable if and only if T<0T \lt 0 and D>0D \gt 0. The test needs no eigenvalues at all, which is why it appears in every control textbook.

Worked examples

Example 1: a full sketch from TT and DD. Take A=(−421−3)A = \begin{pmatrix}-4&2\\1&-3\end{pmatrix}. Then T=−7T = -7, D=(−4)(−3)−(2)(1)=10D = (-4)(-3) - (2)(1) = 10 and T2−4D=49−40=9T^2 - 4D = 49 - 40 = 9. Since D>0D \gt 0 and 9>09 \gt 0 it is a node, and T<0T \lt 0 makes it stable. The eigenvalues are (−7±3)/2=−2(-7 \pm 3)/2 = -2 and −5-5. Only now do we spend effort on eigenvectors, because they fix the shape. For λ=−2\lambda = -2: A+2I=(−221−1)A + 2I = \begin{pmatrix}-2&2\\1&-1\end{pmatrix} gives v=(1,1)\mathbf v = (1, 1); check A(1,1)=(−2,−2)A(1,1) = (-2, -2). For λ=−5\lambda = -5: A+5I=(1212)A + 5I = \begin{pmatrix}1&2\\1&2\end{pmatrix} gives v=(2,−1)\mathbf v = (2, -1). To sketch: draw both lines, put inward arrows on them, then draw curves that run parallel to the fast line (2,−1)(2,-1) far from the origin (running time backwards, the fast mode grows fastest) and bend to become tangent to the slow line (1,1)(1,1) at the origin (the fast mode has died). The sim’s preset draws exactly this.

Example 2: damping is a walk in the (T,D)(T, D) plane. The oscillator y′′+cy′+ky=0y'' + cy' + ky = 0 has companion matrix (01−k−c)\begin{pmatrix}0&1\\-k&-c\end{pmatrix}, so T=−cT = -c and D=kD = k, and the sim’s draggable point is literally the damping-stiffness plane (5.2 Constant Coefficients & the Characteristic Equation found the same cases from the characteristic roots). Fix k=4k = 4 and increase cc. At c=0c = 0: T=0T = 0, a center, λ=±2i\lambda = \pm 2i (the undamped oscillator). At c=2c = 2: T2−4D=4−16<0T^2 - 4D = 4 - 16 \lt 0, a stable spiral with λ=−1±1.7321i\lambda = -1 \pm 1.7321i (underdamped). At c=4c = 4: T2=16=4DT^2 = 16 = 4D, on the parabola, λ=−2\lambda = -2 twice (critical damping, a degenerate node). At c=5c = 5: 25−16=925 - 16 = 9, a stable node, λ=−1,−4\lambda = -1, -4 (overdamped). Stability is c>0c \gt 0 and k>0k \gt 0 throughout, from T<0T \lt 0 and D>0D \gt 0. Block E reproduces the four rows; to see them in the sim, drag the point to (T,D)=(−2,4)(T, D) = (-2, 4) or (−4,4)(-4, 4), where the drag snaps to the parabola.

◆ Interactive Drag the point in the (T, D) plane; the portrait, eigenlines and readouts update live

Loading simulation…

What to try. (1) Load the center preset: the readout gives a factor of 1.000001.00000 per turn and closed ellipses; switch to center perturbed by +0.01 and the factor becomes 1.015831.01583 while the ellipse opens into a spiral. (2) Back on the center, tick the Euler overlay: at h=0.1h = 0.1 the readout reports an amplitude gain of 1.83661.8366 after about one period, at h=0.05h = 0.05 it reports 1.36811.3681, and RK4 with the same steps reports 0.9999860.999986 and 1.0000001.000000 (to six digits). (3) Pick the saddle and look at the dashed eigenlines: the curves hug them. (4) Tick the mirror box on a spiral and watch the spin reverse while TT, DD and the eigenvalues do not change.

Time reversal sends AA to −A-A, so T→−TT \to -T and D→DD \to D: the picture is mirrored across the DD-axis and every stable node becomes an unstable node, with the same curves traversed backwards. Mirroring the plane (x1↔x2x_1 \leftrightarrow x_2) keeps TT and DD and reverses the spin, which is why a21a_{21} and not (T,D)(T, D) carries the direction. The sim’s checkbox does exactly that.

Moving across the plane: which pictures survive a perturbation?

Entries of a real matrix are never known exactly (measurement noise, rounding, model error), so ask which pictures survive a small change A→A+EA \to A + E. TT and DD are continuous in the entries, so a point strictly inside a region stays inside: saddles, nodes and spirals are robust. Boundaries are different, because each is an equality:

  • The center needs T=0T = 0 exactly. T(A+E)=T+tr⁡ET(A + E) = T + \operatorname{tr}E, so any perturbation with tr⁡E≠0\operatorname{tr}E \neq 0 destroys it. Take the center (01−40)\begin{pmatrix}0&1\\-4&0\end{pmatrix} and add 0.010.01 to a11a_{11}: now T=0.01T = 0.01, Reλ=0.005\mathrm{Re}\lambda = 0.005, and the spiral factor is e2π(0.005)/1.999994=1.01583e^{2\pi(0.005)/1.999994} = 1.01583 per turn (block B, where RK4 agrees to six digits, and the sim’s perturbed-center preset). A tiny outward spiral replaces closed orbits, and the sign of the perturbation chooses in or out. Honest caveat: some perturbations leave a center alone, namely those with tr⁡E=0\operatorname{tr}E = 0; putting −4.01-4.01 in a21a_{21} gives T=0T = 0, D=4.01D = 4.01, λ=±2.0025i\lambda = \pm 2.0025i, still a center. But those are a thin slice of all perturbations.
  • The parabola needs T2=4DT^2 = 4D. Perturb (−210−2)\begin{pmatrix}-2&1\\0&-2\end{pmatrix} by ε\varepsilon in a21a_{21}: T=−4T = -4, D=4−εD = 4 - \varepsilon, so T2−4D=4εT^2 - 4D = 4\varepsilon. For ε=+0.01\varepsilon = +0.01 the eigenvalues are −2.1-2.1 and −1.9-1.9 (a node), for ε=−0.01\varepsilon = -0.01 they are −2±0.1i-2 \pm 0.1i (a spiral); block D and two sim presets. The sign of ε\varepsilon decides, so the degenerate node is a fragile hinge between two robust types.
  • D=0D = 0 (an eigenvalue exactly zero) similarly splits into a saddle on one side and a node on the other.

This matters for modelling. A center in a model is a statement that the physics is exactly conservative; any unmodelled friction (or numerical error) moves the point off T=0T = 0, and the long-term behaviour changes from “orbits forever” to “decays” or “blows up”. Which side you land on is decided by the sign of the error, as the next section shows for Euler.

Numerics: a classifier, and Euler on a center

Everything above is a function of the entries. The decision tree for the exercise is: ∣D∣≤tol|D| \le \text{tol} means a line of equilibria; D<0D \lt 0 saddle; else look at Δ=T2−4D\Delta = T^2 - 4D: ∣Δ∣≤tol|\Delta| \le \text{tol} repeated (star if AA is a multiple of II, otherwise degenerate); Δ>0\Delta \gt 0 node; Δ<0\Delta \lt 0 spiral, and a center when ∣T∣≤tol|T| \le \text{tol}. The tolerance is a modelling decision: floating-point arithmetic almost never returns T=0T = 0 exactly, and you choose how close counts as “on the boundary”.

The test case the course cares about is the center. Apply explicit Euler to x′=yx' = y, y′=−4xy' = -4x (so ω=2\omega = 2). One step is x↦(I+hA)x\mathbf x \mapsto (I + hA)\mathbf x, and the quantity Q=4x2+y2Q = 4x^2 + y^2 that the true flow conserves (Q′=8xy−8xy=0Q' = 8xy - 8xy = 0) is multiplied by

4(x+hy)2+(y−4hx)2=(1+4h2)(4x2+y2),4(x + hy)^2 + (y - 4hx)^2 = (1 + 4h^2)(4x^2 + y^2),

so the amplitude grows by 1+h2ω2\sqrt{1 + h^2\omega^2} every step, for every h>0h \gt 0. In terms of the eigenvalue: the mode zz obeys zn+1=(1+hλ)znz_{n+1} = (1 + h\lambda)z_n and ∣1+ihω∣2=1+h2ω2|1 + i h\omega|^2 = 1 + h^2\omega^2. Equivalently det⁡(I+hA)=1+hT+h2D=1.04\det(I + hA) = 1 + hT + h^2D = 1.04 at h=0.1h = 0.1, T=0T = 0, D=4D = 4: the Euler map expands area, though the true flow (T=0T = 0, 7.5 The Matrix Exponential & Fundamental Matrices will show det⁡eAt=eTt\det e^{At} = e^{Tt}) preserves it. Block C measures it over about one period P=πP = \pi:

hhstepsEuler amplitudeformula (1+4h2)N/2(1+4h^2)^{N/2}RK4
0.1311.83661.83660.99998629
0.05631.36811.36810.99999956
0.05628 (10 periods)22.746322.74630.99999564

The per-step factors are 1.019801.01980 and 1.004991.00499. Halving hh roughly halves the logarithm of the gain per lap (0.60790.6079 to 0.31340.3134): the error is first order, consistent with Euler. After ten laps at h=0.05h = 0.05 the “orbit” has grown by a factor 22.722.7. RK4 at h=0.05h = 0.05 is within 4.4×10−64.4\times10^{-6} of a closed loop even after ten laps, because its amplification factor on the imaginary axis is 1−O((hω)6)1 - O((h\omega)^6) rather than 1+O((hω)2)1 + O((h\omega)^2) (from 4.3 Runge–Kutta Methods & Butcher Tableaux, the degree-4 Taylor polynomial of eze^{z}). The sim’s Euler overlay draws the dotted outward spiral on the center and reports the same 1.83661.8366 and 0.999986290.99998629 for h=0.1h = 0.1.

Computing it

🐍 PythonLab: T, D and eigenvalues of every example; spiral factor against RK4; Euler against RK4 on the center; the fragile parabola; damping (blocks A to E)
⌘/Ctrl + Enter

Now build the classifier yourself. It must work from TT, DD and the entries only, and the tolerance is part of the specification.

🧪 Coding exerciseExercise: classify a 2x2 matrix by trace and determinant
⌘/Ctrl + Enter

Concept checks

✅ Concept check

A system has $T = -3$ and $D = 5$. Which description is right?

✅ Concept check

A model has a center at the origin. The entries are only known to within $\pm0.01$. What is the honest conclusion?

✅ Concept check

A stable spiral has the factor $0.28461$ per turn (the sim's preset). By what factor does the radius shrink after three full turns? (Give a number.)

✅ Concept check

Why can a 2D incompressible flow have a vortex or a saddle at a critical point, but never a spiral sink?

Problems

✏️ Pen & paperClassify without eigenvectors●○○

Classify x′=Ax\mathbf x' = A\mathbf x for A=(1−23−1)A = \begin{pmatrix}1&-2\\3&-1\end{pmatrix}. State the spin direction and the period.

Show worked solution

T=1+(−1)=0T = 1 + (-1) = 0, D=(1)(−1)−(−2)(3)=5D = (1)(-1) - (-2)(3) = 5. With T=0T = 0 and D>0D \gt 0 it is a center, λ=±i5=±2.2361i\lambda = \pm i\sqrt5 = \pm 2.2361i. Spin: a21=3>0a_{21} = 3 \gt 0, so counterclockwise (the velocity at (1,0)(1,0) is (1,3)(1, 3), upward). Period P=2π/5=2.8099P = 2\pi/\sqrt5 = 2.8099. (These numbers are from the formula, not the sim.)

✏️ Pen & paperA one-parameter family●●○

For Ak=(k1−10)A_k = \begin{pmatrix}k&1\\-1&0\end{pmatrix}, find TT, DD and classify for every real kk. Where are the structurally unstable values?

Show worked solution

T=kT = k, D=1D = 1, so T2−4D=k2−4T^2 - 4D = k^2 - 4 and the point travels along the horizontal line D=1D = 1. For k<−2k \lt -2: stable node (e.g. k=−3k = -3: λ=−2.618,−0.382\lambda = -2.618, -0.382). At k=−2k = -2: on the parabola, repeated λ=−1\lambda = -1 (a degenerate node, since the matrix is not a multiple of II). For −2<k<0-2 \lt k \lt 0: stable spiral. At k=0k = 0: center (±i\pm i). For 0<k<20 \lt k \lt 2: unstable spiral; at k=2k = 2 degenerate; for k>2k \gt 2 unstable node. The values k=−2,0,2k = -2, 0, 2 are structurally unstable, the three crossings of the line with the parabola and the DD-axis. Stability: T<0T \lt 0 and D>0D \gt 0 means k<0k \lt 0.

✏️ Pen & paperIncompressible flows near a critical point●●●

For u=(ax+by, cx−ay)\mathbf u = (ax + by,\ cx - ay) show that the flow is incompressible and find the types for (a,b,c)=(1,2,3)(a, b, c) = (1, 2, 3) and (1,−2,3)(1, -2, 3).

Show worked solution

∇⋅u=a−a=0\nabla\cdot\mathbf u = a - a = 0, so T=0T = 0. D=−a2−bcD = -a^2 - bc, so λ2=−D=a2+bc\lambda^2 = -D = a^2 + bc. For (1,2,3)(1, 2, 3): λ2=7\lambda^2 = 7, D=−7<0D = -7 \lt 0: a saddle with λ=±2.6458\lambda = \pm2.6458. For (1,−2,3)(1, -2, 3): λ2=1−6=−5\lambda^2 = 1 - 6 = -5, D=5>0D = 5 \gt 0: a center, λ=±2.2361i\lambda = \pm2.2361i. The sign of a2+bca^2 + bc decides: strain beats rotation for a saddle, rotation beats strain for a center. (Numbers from the formula, not from the sim.)

Key Takeaways

  • For x′=Ax\mathbf x' = A\mathbf x the characteristic polynomial is λ2−Tλ+D\lambda^2 - T\lambda + D, so λ=(T±T2−4D)/2\lambda = (T \pm \sqrt{T^2 - 4D})/2, λ1+λ2=T\lambda_1 + \lambda_2 = T, λ1λ2=D\lambda_1\lambda_2 = D.
  • D<0D \lt 0 saddle; D>0D \gt 0 and T2>4DT^2 \gt 4D node; T2<4DT^2 \lt 4D spiral; T=0T = 0 center; T2=4DT^2 = 4D degenerate or star node; D=0D = 0 a line of equilibria.
  • Asymptotically stable exactly when T<0T \lt 0 and D>0D \gt 0, with no eigenvalues needed.
  • A spiral’s turn direction is the sign of a21a_{21} (positive: counterclockwise), and its radius is multiplied by e2πReλ/∣Imλ∣e^{2\pi\mathrm{Re}\lambda/|\mathrm{Im}\lambda|} each turn (0.284610.28461 for −0.2±i-0.2 \pm i, measured 0.2846100.284610).
  • The center and the parabola are structurally unstable: adding 0.010.01 to a11a_{11} of a center gives Reλ=0.005\mathrm{Re}\lambda = 0.005 (1.01581.0158 per turn), and ±0.01\pm0.01 in a21a_{21} of a degenerate node gives a node or a spiral.
  • Explicit Euler multiplies the amplitude of a center by 1+h2ω2\sqrt{1 + h^2\omega^2} per step (gain 1.83661.8366 per lap at h=0.1h = 0.1, 22.722.7 over ten laps at h=0.05h = 0.05); RK4 at h=0.05h = 0.05 stays within 4.4×10−64.4\times10^{-6} of a closed loop after ten laps.
  • In a 2D incompressible flow the trace of the Jacobian is the divergence, so T=0T = 0 and critical points are only saddles and centers (vortices); sources and sinks need T≠0T \neq 0.
  • Next: complex and repeated eigenvalues solved exactly in 7.4 Complex & Repeated Eigenvalues, and the nonlinear case in 7.7 Nonlinear Systems & Linearization.
Finished this lesson?

Mark it complete to track your progress (saved in this browser). Reading every section does this for you.

Where this leads

These later lessons build directly on this one: