Normal Modes Method --Mode shapes are used to decouple the equations of motion

MR KR

T T

--These operations diagonalize the stiffness and mass matrices by using the property of orthogonality of the matrices --The result is the problem written in normal coordinates transformed from our original coordinate system Example Given 2k m k m

x1 Equations of motion

x2

1 x m 2 x

3k k

k k

x1 x2

Results are
1 2

k m k 3.414 m 1 1 2.414 .414 .586

Rewrite the matrix equations on normal coordinates

1 q 2 q

q1 q2

where
T

M K

6.827 0 0 1.171 4 0 0 4

for k=m=1

1 6.827 0 q 0 1.171 2 q
in equation form
1 q 2 q 4 q1 0 6.827 4 q2 0 1.171

4 0 q1 0 4 q2

Thes solutions to these are already known

qi(t )
where

Ai cos it Bi sin

(1) T

qi qi

M{x (0)} MRi (1) T M{x (0)} MRi

For the system shown and x1(0)=1, x2(0)= xdot1(0)=0, xdot2(0)=0

6.827 1 {1 - .414} 1.171

A2

Then with Bi=0,

q1 .146" cos .586t q 2 .854" cos 3.414t

q1( t )

( 2)

q 2( t ) 1 .854" cos 3.414t .414

1 .146" cos .586t 2.414

Damped Systems --Damping in MDOF systems is problematic Example Example Given 2k m k m

2c x1 Equations of motion

c x2

m 0 1 x 0 m 2 x

3c c

c x1 c x2

3k k

k x1 k x2

Natural frequencies and mode shapes are obtained as before neglecting the damping matrix
1 2

k m k 3.414 m 1 1 2.414 .414 .586

Rewrite the matrix equations on normal coordinates

1 q 2 q

q1 q2

q1 q2

where
T

M C K

m c k

6.827 0

0 1.171

4 0 0 4 4 0 0 4

1 6.827 0 q 0 1.171 2 q

4 0 q1 0 4 q2

4 0 q1 0 4 q2

in equation form
1 q 2 q 4c 4k q1 q1 0 6.827m 6.827m 4c 4k q2 q2 0 1.171m 1.171m
qo
di

Thes solutions to these are already known

(qo cos

di

ni o

sin

di

t)

where
(1) T

qi qi

M{x (0)} MRi (1) T M{x (0)} MRi

For the system shown and x1(0)=1, x2(0)= xdot1(0)=1, xdot2(0)=0

{1 2.414} q1

1 1 0 1 1 0 .854" .146"

6.827 1 {1 - .414} 1.171

q2

q1( t ) qi( t )

e e

nit

(.146" cos (qo cos

di

nit

t) qo
di

ni o

sin

di

t)

Example Given 2k m k m

c x1 Equations of motion

c x2

m 0 1 x 0 m 2 x

2c c

c x1 c x2

3k k

k x1 k x2

Natural frequencies and mode shapes are obtained as before neglecting the damping matrix
1 2

k m k 3.414 m 1 1 2.414 .414 .586

Rewrite the matrix equations on normal coordinates

1 q 2 q

q1 q2

q1 q2

where
T

M C K

m c k

6.827 0 3 1 4 0 0 4 1 3

0 1.171

1 6.827 0 q 0 1.171 2 q

3 1

1 q1 3 q2

4 0 q1 0 4 q2

This is a more typical case. The damping matrix is not generally similar to the stiffness matrix. --In this case, we use the damping assumption, , and simply write it in the equations of motion. in equation form
1 2 q 2 2 q
1 2

q1 q2

q1 0
2

q2

0
qo
di

These solutions to these are already known

(qo cos

di

ni o

sin

di

t)

where
(1) T

qi qi

M{x (0)} MRi (1) T M{x (0)} MRi

For the system shown and x1(0)=1, x2(0)= xdot1(0)=0, xdot2(0)=0

{1 2.414} q1

1 1 0 1 1 0 .854" .146"

6.827 1 {1 - .414} 1.171

q2

q1( t ) e qi ( t ) e

nit

(.146" cos (qoi cos

di

nit

t) qoi
di

ni oi

sin

di

t)

q1( t ) q 2( t )

e e

k (.05 ) .586 t m

(.146" cos .586

k 1 .052 t ) m k 1 .052 t ) m

k (.05 ) 3.414 t m

(.854" cos 3.414

Transform back to the original coordinate system

x1( t ) x 2( t )

(1)

q1( t )

( 2)

q 2( t )

Forced Vibrations --Many types of forcing conditions may be handled using the normal modes method Example Given 2k m f1(t) k c x1 Equations of motion x2 m f2(t)

m 0 1 x 0 m 2 x

2c c

c x1 c x2

3k k

k x1 k x2

f 1( t ) f 2( t )

Natural frequencies and mode shapes are obtained as before neglecting the damping matrix and forces
1 2

k m k 3.414 m 1 1 2.414 .414 .586

Rewrite the matrix equations on normal coordinates

1 q 2 q

q1 q2

q1 q2

f 1( t ) f 2( t )

where
T

M C K F

m c k

6.827 0 3 1 4 0 0 4 1 3

0 1.171

1 2.414 1

f 1( t )

.414 f 2( t )

Resume here
m 6.827 0 1 q 1.171 2 q 0 c 3 1 1 q1 3 q2 k 4 0 q1 f 1( t ) 2.414f 2( t ) f 1 .414f 2 0 4 q2

This is a more typical case. The damping matrix is not generally similar to the stiffness matrix. --In this case, we use the damping assumption, , and simply write it in the equations of motion. in equation form
1 2 q 2 2 q
1 2

q1 q2

q1 f 1 2.414f 2
2

q2

f 1 .414f 2

The solutions to these are already known for each type of force Example: Use f 1(t ) 1# sin 4t , =.01 then
1 2(.586k / m)1 / 2 (.01)q1 .586 q 2 2(3.414k / m)1 / 2 (.01)q 2 q k 1# q1 m m6.827# s 2 / in k 1# 3.414 q 2 m m1.171# s 2 / in

the solutions are of the form

qi( t )

fi / ki [(1 ri ) (2 ri) 2 ]1 / 2
2 2

setting k=1, m=1

4 [(1 )2 2 (.766) q1( t ) .0095"sin 4t q 2( t ) ( .414)

4 2 1/ 2 (2(.01)( )) ] .766 1# / 4# / in
2

sin 4t

4 [(1 )2 2 (1.85) q 2( t ) .028"sin 4t

4 2 1/ 2 (2(.01)( )) ] 1.85

sin 4 t

TRANSFORM BACK
x1( t ) x 2( t ) x1( t ) x 2( t )
(1)

q1( t )

( 2)

q 2( t ) 1 q 2( t ) .414

1 q1( t ) 2.414

Example Same system, f2(t)=step defined below f2(t) 2# t

Given 2k m f1(t) k c x1 Equations of motion x2 m f2(t)

m 0 1 x 0 m 2 x

2c c

c x1 c x2

3k k

k x1 k x2

f 1( t ) f 2( t )

Natural frequencies and mode shapes are obtained as before neglecting the damping matrix and forces
1 2

k m k 3.414 m 1 1 2.414 .414 .586

Rewrite the matrix equations on normal coordinates

1 q 2 q

q1 q2

q1 q2

f 1( t ) f 2( t )

where
T

M C K F

m c k

6.827 0 3 1 4 0 0 4 1 3

0 1.171

1 2.414 1

f 1( t )

.414 f 2( t )

Resume here
m 6.827 0 1 q 1.171 2 q 0 c 3 1 1 q1 3 q2 k 4 0 q1 f 1( t ) f 2( t ) 2.414f 1 .414f 2 0 4 q2

This is a more typical case. The damping matrix is not generally similar to the stiffness matrix. --In this case, we use the damping assumption, , and simply write it in the equations of motion. in equation form
1 2 q 2 2 q
1 2

q1 q2

q1 f 1 f 2
2

q 2 2.414f 1 .414f 2

The solutions to these are already known for each type of force Use f 1(t ) 1# sin 4t , =0, f2(t)= step function then
1 2(.586k / m)1 / 2 (.01)q1 .586 q 2 q 2(3.414k / m)1 / 2 (.01)q 2 k f2 q1 m 6.827# s 2 / in k .414f 2 3.414 q 2 m 1.171# s 2 / in

the solutions are of the form

setting k=1, m=1, =0

If damping is a factor, numerically integrate the integral

TRANSFORM BACK
x1( t ) x 2( t ) x1( t ) x 2( t )
(1)

q1( t )

( 2)

q 2( t ) 1 q 2( t ) .414

1 q1( t ) 2.414

Base Excitation Given 2k y(t) c x1 Equations of motion c x2 m f1(t) k m f2(t)

m 0 1 x 0 m 2 x

2c c

c x1 c x2

3k k

k x1 k x2

f 1( t ) f 2( t )

Natural frequencies and mode shapes are obtained as before neglecting the damping matrix and forces
1 2

k m k 3.414 m 1 1 2.414 .414 .586

Rewrite the matrix equations on normal coordinates

1 q 2 q

q1 q2

q1 q2

ky1( t )

where
T

M C K Y

m c k

6.827 0 3 1 4 0 0 4 1 3

0 1.171

1 2.414 1

2kyo sin t

.414 2kyo sin t

Resume here
m 6.827 0 1 q 1.171 2 q 0 c 3 1 1 q1 3 q2 k 4 0 q1 2kyo sin t (2kyo sin t ) 0 4 q2

--In this case, we use the damping assumption, , and simply write it in the equations of motion. in equation form
1 2 q 2 2 q
1 2

q1 q2

q1 2kyo sin t
2

q2

(2kyo sin t )

The solutions to these are already known for each type of force Use y(t)=1sin4t, =0 then
1 2(.586k / m)1 / 2 (.01)q1 .586 q 2 2(3.414k / m)1 / 2 (.01)q 2 q k 2k (1" ) sin 4 t q1 m 6.827# s 2 / in k 2k (1" ) sin 4 t 3.414 q 2 m 1.171# s 2 / in

the solutions are of the form

setting k=1, m=1, =.01

1 (2(.01)( q1( t ) .586 2 2 [(1 ( ) ) 4 .586 2(.01) 4 ) 0 tan 1 ( .586 2 1 ( ) 4 2"[ 2"[

.586 2 )) 4 ]1 / 2 sin( 4 t .586 2 (2(.01)( ) ] 4

1 (2(.01)( q 2( t ) 3.414 2 2 [(1 ( ) ) 4 3.414 2(.01) 4 ) 0 tan 1 ( 3.414 2 1 ( ) 4

3.414 2 )) 4 ]1 / 2 sin( 4 t 3.414 2 (2(.01)( ) ] 4

TRANSFORM BACK
x1( t ) x 2( t ) x1( t ) x 2( t )
(1)

q1( t )

( 2)

q 2( t ) 1 q 2( t ) .414

1 q1( t ) 2.414