Multi-degree-of-freedom systems and modal analysis
Mass and stiffness matrices, the eigenvalue problem, orthogonality, modal mass and stiffness, modal coordinates and superposition, Rayleigh damping and mode truncation.
Drafted with Aria, reviewed by the AiCanCode.org team. Spotted an error? Use Give Feedback at the bottom of the page.
Why it matters
A finite-element model of a wing or fuselage has thousands of degrees of freedom, yet its dynamic behaviour in flight is governed by a handful of low modes: first wing bending, first torsion, fuselage bending, control-surface rotation. Modal analysis turns the large coupled set of equations into independent single-degree-of-freedom equations, one per mode, so loads, gust response and flutter can be computed with a small, accurate model. It is also how test data from ground vibration tests are compared with the analysis.
Key ideas
Equations of motion. For n coordinates {x}, [M]·{ẍ} + [C]·{ẋ} + [K]·{x} = {F(t)}. [M] and [K] are n × n, symmetric; [M] is positive definite and [K] is positive definite for a restrained structure (positive semi-definite if rigid-body motion is possible). They come from lumped masses and springs, from influence coefficients (the flexibility matrix [a] = [K]⁻¹) or from finite elements.
Eigenvalue problem. Undamped free vibration gives ([K] − ω²·[M])·{φ} = {0}. There are n eigenvalues ω_r² and n eigenvectors (mode shapes) {φ_r}. Computer methods (subspace iteration, Lanczos) find only the lowest modes needed; by hand you solve det([K] − ω²·[M]) = 0 for 2 or 3 DOF.
Orthogonality and modal properties. For r ≠ s: {φ_r}ᵀ·[M]·{φ_s} = 0 and {φ_r}ᵀ·[K]·{φ_s} = 0. For r = s these products define the modal (generalised) mass M_r = {φ_r}ᵀ·[M]·{φ_r} and modal stiffness K_r = {φ_r}ᵀ·[K]·{φ_r}, with ω_r² = K_r/M_r. Scaling each mode so M_r = 1 gives mass-normalised modes, for which [Φ]ᵀ·[M]·[Φ] = [I] and [Φ]ᵀ·[K]·[Φ] = diag(ω_r²).
Modal coordinates. Write {x} = [Φ]·{q}, where [Φ] has the mode shapes as columns and {q} are modal coordinates. Pre-multiplying by [Φ]ᵀ decouples the equations:
M_r·q̈_r + C_r·q̇_r + K_r·q_r = f_r(t), with f_r = {φ_r}ᵀ·{F}
Each line is an SDOF oscillator, solved with the earlier methods (free vibration, harmonic forcing, Duhamel), and the physical response is rebuilt by superposition. The modal force f_r shows that a force applied at a node of mode r does not excite that mode.
Initial conditions. q_r(0) = {φ_r}ᵀ·[M]·{x(0)}/M_r and similarly for velocities.
Damping. The equations decouple only if [Φ]ᵀ·[C]·[Φ] is diagonal. Proportional (Rayleigh) damping [C] = α·[M] + β·[K] does this, and then ζ_r = α/(2ω_r) + β·ω_r/2. In practice a modal damping ratio for each mode is taken from test (typically 1–3 % for metallic airframes).
Truncation. Usually only the lowest few modes are kept, because loads in the low-frequency range excite them most. Keep enough modes to cover the frequency range of the excitation with margin; for stresses or local loads more modes, or a static correction, may be needed.
Links. This is the generalisation of the 2-DOF topic. Rayleigh–Ritz (next topic) gives approximate low modes cheaply. In aeroelasticity the structural modes are used as generalised coordinates and the aerodynamic forces are projected onto them as generalised aerodynamic forces.
Formulas
[M]·{ẍ} + [C]·{ẋ} + [K]·{x} = {F(t)} — MDOF equations; [M] (kg), [C] (N·s/m), [K] (N/m).
det([K] − ω²·[M]) = 0, ([K] − ω_r²·[M])·{φ_r} = {0} — eigenvalue problem.
M_r = {φ_r}ᵀ·[M]·{φ_r}, K_r = {φ_r}ᵀ·[K]·{φ_r}, ω_r² = K_r/M_r — modal mass, stiffness and frequency.
{x} = [Φ]·{q} — modal superposition; f_r = {φ_r}ᵀ·{F} — modal force (N, for unit-scaled modes).
M_r·q̈_r + 2ζ_r·ω_r·M_r·q̇_r + K_r·q_r = f_r — decoupled modal equation.
q_r(0) = {φ_r}ᵀ·[M]·{x(0)}/M_r — modal initial condition.
[C] = α·[M] + β·[K], ζ_r = α/(2ω_r) + β·ω_r/2 — Rayleigh damping; α (1/s), β (s).
Worked examples
Example 1 (standard). M = diag(2, 1) kg and K = [[6, −2], [−2, 4]] kN/m. Find ω₁, ω₂, the mode shapes and the modal masses and stiffnesses.
- det([K] − λ[M]) = (6000 − 2λ)(4000 − λ) − 2000² = 0 → 2λ² − 14 000λ + 20 × 10⁶ = 0 → λ² − 7000λ + 10⁷ = 0.
- λ = 2000 and 5000 (rad/s)², so ω₁ = 44.72 rad/s (7.12 Hz) and ω₂ = 70.71 rad/s (11.25 Hz).
- From row 1, φ₂/φ₁ = (6000 − 2λ)/2000: mode 1, {1, 1}; mode 2, {1, −2}.
- Modal masses: M₁ = 2 × 1 + 1 × 1 = 3 kg; M₂ = 2 × 1 + 1 × 4 = 6 kg.
- Modal stiffnesses: [K]{1, 1} = {4000, 2000}, so K₁ = 6000 N/m; [K]{1, −2} = {10 000, −10 000}, so K₂ = 10 000 + 20 000 = 30 000 N/m.
- Check: K₁/M₁ = 2000 and K₂/M₂ = 5000, matching λ. Cross term {1, 1}·[M]·{1, −2} = 2 − 2 = 0.
Answer: ω₁ = 44.7 rad/s, {1, 1}, M₁ = 3 kg; ω₂ = 70.7 rad/s, {1, −2}, M₂ = 6 kg.
Example 2 (GATE level). For the same system, mass 1 is displaced 1 mm and both masses are released from rest. Write x(t), and find α and β for 2 % damping in both modes.
q_r(0) = {φ_r}ᵀ·[M]·{x(0)}/M_rwith x(0) = {1, 0} mm: [M]{x(0)} = {2, 0}. q₁(0) = 2/3 = 0.667 mm; q₂(0) = 2/6 = 0.333 mm.- Released from rest, each modal coordinate is q_r(0)·cos(ω_r·t): x(t) = 0.667·{1, 1}·cos(44.72t) + 0.333·{1, −2}·cos(70.71t) mm.
- Check t = 0: x₁ = 0.667 + 0.333 = 1 mm, x₂ = 0.667 − 0.667 = 0. Correct.
- Rayleigh damping: solving ζ = α/(2ω) + β·ω/2 at both frequencies gives β = 2ζ/(ω₁ + ω₂) = 0.04/115.43 = 3.47 × 10⁻⁴ s and α = 2ζ·ω₁·ω₂/(ω₁ + ω₂) = 0.04 × 3162.3/115.43 = 1.096 s⁻¹.
Answer: x(t) as above (not periodic, since ω₂/ω₁ = √2.5 is irrational); α ≈ 1.10 s⁻¹, β ≈ 3.47 × 10⁻⁴ s.
Common mistakes
- Using {φ_r}ᵀ·{x(0)} instead of {φ_r}ᵀ·[M]·{x(0)}/M_r for modal initial conditions.
- Forgetting that mode shapes are only defined up to a scale factor; modal mass and modal force depend on the scaling, but the physical response does not.
- Assuming any damping matrix decouples in modal coordinates; only proportional (or classical) damping does.
- Mixing kN/m with kg, which shifts every frequency by √1000.
- Truncating to too few modes for a load with high-frequency content.
- Labelling frequencies out of order; ω₁ is always the lowest.
For GATE AE
Expect 2- or 3-DOF frequency calculations from given [M] and [K], mode shapes and orthogonality checks, modal mass and stiffness, response from initial conditions by modal superposition, and Rayleigh damping coefficients. Conceptual questions test why modal coordinates decouple the equations and when they do not.
Quick check
- What property of mode shapes makes modal decoupling possible?
- A mode {1, 2} with M = diag(1, 3) kg: what is its modal mass?
- If a force acts at a node of mode r, what is f_r?
- Write ζ_r for Rayleigh damping.
- For M = diag(2, 1) kg and K = [[4, 0], [0, 3]] N/m, what is ω₁?
Answers: 1. orthogonality with respect to [M] and [K]; 2. 1 + 12 = 13 kg; 3. zero; 4. ζ_r = α/(2ω_r) + β·ω_r/2; 5. √2 = 1.41 rad/s.
Interview questions
All Structural Dynamics and Aeroelasticity interview questionsTry answering each one aloud before you open it.
1.What is a multi-degree-of-freedom (MDOF) system in structural dynamics?Concept
A multi-degree-of-freedom (MDOF) system in structural dynamics refers to a system that has multiple independent movements or displacements. Each degree of freedom corresponds to a possible independent motion, such as translation or rotation, in a particular direction. MDOF systems are used to model complex structures like buildings, bridges, and aircraft, where multiple parts can move independently.
2.Explain the concept of modal analysis in the context of MDOF systems.Concept
Modal analysis is a technique used to determine the natural frequencies, mode shapes, and damping ratios of a multi-degree-of-freedom system. It involves solving the eigenvalue problem for the system's mass and stiffness matrices. The results help in understanding how the system will respond to dynamic loads, such as vibrations or shocks, by identifying the modes in which the system naturally tends to oscillate.
3.Why is modal analysis important in aerospace engineering?Application
Modal analysis is crucial in aerospace engineering because it helps predict how aircraft structures will respond to dynamic loads, such as turbulence, engine vibrations, or aerodynamic forces. By understanding the natural frequencies and mode shapes, engineers can design structures that avoid resonant conditions, which could lead to excessive vibrations and potential structural failure. This ensures the safety and reliability of the aircraft.
4.How can damping be incorporated into the analysis of MDOF systems?Concept
Damping can be incorporated into the analysis of MDOF systems by adding a damping matrix to the system's equations of motion. This matrix represents the energy dissipation mechanisms within the structure, such as material damping or friction. The inclusion of damping helps in predicting more realistic dynamic responses and reduces the amplitude of vibrations at resonance.
5.Explain the difference between mode shapes and natural frequencies in modal analysis.Concept
Natural frequencies are the frequencies at which a system tends to oscillate in the absence of external forces or damping. Mode shapes, on the other hand, describe the pattern of deformation or displacement that the system undergoes at each natural frequency. While natural frequencies indicate how fast the system oscillates, mode shapes show how different parts of the system move relative to each other during oscillation.
6.Why is it important to consider multiple modes in the analysis of MDOF systems?Application
Considering multiple modes in the analysis of MDOF systems is important because real-world structures can respond to dynamic loads in various ways. Each mode represents a different pattern of motion, and neglecting higher modes can lead to inaccurate predictions of the system's response. By analyzing multiple modes, engineers can ensure that all significant dynamic behaviors are accounted for, leading to safer and more efficient designs.
7.What is the role of the mass matrix in the modal analysis of MDOF systems?Concept
The mass matrix in the modal analysis of MDOF systems represents the distribution of mass within the structure. It is used in conjunction with the stiffness matrix to solve the eigenvalue problem, which determines the natural frequencies and mode shapes. The mass matrix is crucial for accurately predicting how the structure will respond to dynamic loads, as it influences the inertia of the system.
8.Calculate the natural frequencies of a two-degree-of-freedom system with mass matrix M = [[2, 0], [0, 1]] kg and stiffness matrix K = [[6, -2], [-2, 4]] N/m.Numerical
Solve det(K − ω²M) = 0: (6 − 2ω²)(4 − ω²) − 4 = 0, which expands to 2ω⁴ − 14ω² + 20 = 0, i.e. ω⁴ − 7ω² + 10 = 0, so ω² = 2 or 5. The lower (first) natural frequency is ω₁ = √2 = 1.414 rad/s and the second is ω₂ = √5 = 2.236 rad/s. Substituting back into the first row, (6 − 2ω²)φ₁ = 2φ₂, gives mode shapes {1, 1} and {1, −2}, which are M-orthogonal (2 − 2 = 0).
9.For a three-degree-of-freedom system, explain how you would determine the mode shapes.Concept
To determine the mode shapes of a three-degree-of-freedom system, follow these steps:
- Formulate the mass and stiffness matrices for the system.
- Solve the eigenvalue problem (K - ω²M)Φ = 0, where Φ is the mode shape vector.
- For each natural frequency (eigenvalue), solve the resulting set of linear equations to find the corresponding mode shape (eigenvector).
- Normalize the mode shapes for easier interpretation and comparison.
- Verify that the mode shapes are orthogonal with respect to the mass matrix.
Finished this topic? Mark it so your progress, study plan and readiness keep up.
Stuck on something here?