NVH
Noise, Vibrations & Harshness in vehicles
Lesson 04 – MDOF and Modal Analysis
𝑐1 ∗ 𝑥1ሶ
𝑘1 ∗ 𝑥1 𝑘2 ∗ ( 𝑥2 − 𝑥1 )
𝑘2 ∗ (𝑥1 − 𝑥2 ) 𝑘3 ∗ 𝑥2
01 Basis 2
A 2 DOFs System
A 9 DOFs System
Ys Z s Z u1 Z u3 Z u3 Z u4 φ θ ψ
A 4 DOFs System
A 2 DOFs System
“Quarter car”
01 Basis 3
[ m ] is the mass matrix, real symmetric and positive defined,
[ c ] is the matrix of viscous damping, real symmetric and positive defined (or positive semidefined)
[ k ] is the stiffness matrix, real symmetric and positive defined (or positive semidefined)
{ x } is the vector of generalized coordinates and
{ f } is the vector of forces
Remember that a matrix is
positive-definite if the number {z}T[M]{z} is positive for every nonzero real column vector {z}
01 Basis 4
Free response has a synchronous solution of this type (homogeneous equation for undamped system)
{ x (t) } = { X0 } e iωt
For a n degrees of freedom system, it gives origin to the algebraic equation of 2n degree in the variable
(degree n in the ω2 variable) which takes the name of characteristic equation or characteristic polynomial
Are the solutions (eigenvalues) of characteristic equation
By inserting, one by one, the eigenvalues in equation the following is obtained
contr. to node 1
contr. to node 2
eigenvalue matrix modal matrix
Total response can be obtained as the sum of the contributions contribution on node 1 contribution on node 1
by the modal freq. w1 by the modal freq. w2
node 1
𝑛
𝑥1 𝑡 = 𝐴1 cos( 𝜔1 𝑡 + 𝜗1 )𝜓11 + 𝐴2 cos( 𝜔2 𝑡 + 𝜗2 )𝜓12
𝑥 𝑡 = 𝐴𝑟 cos( 𝜔𝑟 𝑡 + 𝜗𝑟 ) 𝜓𝑟
𝑟=1 𝑥2 𝑡 = 𝐴1 cos( 𝜔1 𝑡 + 𝜗1 )𝜓21 + 𝐴2 cos( 𝜔2 𝑡 + 𝜗2 )𝜓22
Each ωr corresponds to a single ψr
Vectors orthogonality
By taking the r-th and the s-th eigenvalues
Transposing the second
Also considering
2
Bringing 1 and 2
If
Ψ Vectors are orthogonal with respect to
01 Basis the mass matrix 7
But we have defined it to be positive defined !
positive constants kr and mr take the name of r-th modal stiffness and r-th modal mass
modal mass matrix Matrices are diagonal and eigenvectors are said to be
modal stiffness matrix m-orthogonal and k-orthogonal
Scaling factors
Modal matrix is composed with column vectors of eigenvectors, each one can be scaled freely. Constant
are often chosen in the following ways
each first element settled to 1
each highest value is scaled to 1 m-normalized eigenvector
each lowest value is scaled to –1
each modal mass is scaled to 1.
01 Basis 8
𝑥 𝑡 = 𝜓 𝜂 𝑡 Application of the direct modal transform
Pre multiplication by transpose of eigenvector matrix
Reduction to n independent equations, each in the ηr coordinate
Free response for the modal coordinate ηr is given by the already seen SDOF
Which can be used to build up the total response in physical coordinates by using the direct modal
transform coordinate ηr
9
Two methods to determine initial conditions
• Initial conditions can be computed by inverting the eigenvector matrix
−1 −1 −1
𝜂0 𝑡 = 𝜓 𝑥0 𝑡 𝜂0ሶ 𝑡 = 𝜓 𝑥0ሶ 𝑡 = 𝜓 𝑣0 𝑡
• Or by using the orthogonal properties
And becomes the following for the initial velocity vector
01 Basis
10
Example for 2 DOFS system
01 Basis 11
(4 – 5* 2/5 ) x1 - 2 x2 = 0
– 2 x1 + (6 – 10 *2/5) x2 = 0
01 Basis 12
if we establish this kind of initial
condition, the system vibrates at
the natural frequency relative to
the mode n (omega_n)
in REAL NODES the oscillation
of the nodes are either in phase
or out of phase (no delay!,
phase=0, 180...)
in COMPLEX NODES we have a non-zero phase!
the maximum displacements of
the nodes are delayed one to
the other
01 Basis 13
01 Basis 14
Initial conditions
01 Basis 15
Case 2 𝑥0 = 1 0 𝑇 𝑣 = 0 0 𝑇 Different Initial conditions
Both modes are excited due to these initial conditions
𝑥0 = 1 0 𝑇 𝑣 = 0 0 𝑇
Only first mode is excited due to these initial
conditions
𝑥0 = 0 0 𝑇 𝑣 = 2 2 𝑇
01 Basis 16
Modes are numbered WRT increasing eigenvalues
01 Basis 17
