1) The document presents the derivation of the partial differential equation governing the transverse vibration of an undamped beam.
2) The beam vibration is modeled using the principles of mechanics and separation of variables is used to solve the PDE.
3) Natural modes of vibration are determined by applying boundary conditions to the solution. The natural frequencies are determined to be ωn = n2π2L2EI/m.
1) The document presents the derivation of the partial differential equation governing the transverse vibration of an undamped beam.
2) The beam vibration is modeled using the principles of mechanics and separation of variables is used to solve the PDE.
3) Natural modes of vibration are determined by applying boundary conditions to the solution. The natural frequencies are determined to be ωn = n2π2L2EI/m.
1) The document presents the derivation of the partial differential equation governing the transverse vibration of an undamped beam.
2) The beam vibration is modeled using the principles of mechanics and separation of variables is used to solve the PDE.
3) Natural modes of vibration are determined by applying boundary conditions to the solution. The natural frequencies are determined to be ωn = n2π2L2EI/m.
1) The document presents the derivation of the partial differential equation governing the transverse vibration of an undamped beam.
2) The beam vibration is modeled using the principles of mechanics and separation of variables is used to solve the PDE.
3) Natural modes of vibration are determined by applying boundary conditions to the solution. The natural frequencies are determined to be ωn = n2π2L2EI/m.
Copyright:
Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online from Scribd
Download as pdf or txt
You are on page 1/ 16
Undamped Vibration of a Beam
Louie L. Yaw Walla Walla University Engineering Department
PDE Class Presentation June 5, 2009
Problem - Undamped Transverse Beam Vibration
+u p(x, t)
m(x), EI(x) L +x 0 V M pdx 1dx
dx
V + V dx x M + M dx x
o 2dx
FBD of Slice dx
2 u Inertial Force fI = mdx 2 by DAlemberts t Principle dx
Derivation of PDE Sum Forces Vertically, choosing + up V (V + V 2u dx) + p(x, t)dx m(x)dx 2 = 0 x t (1)
Sum Moments about o, choosing CCW as + rotation M 2u +M + dx = 0 2 x t (2) Simplifying (1), and in (2) ignoring higher order terms in the limit as dx 0 gives M Vdx + p(x, t)1 dx 2 + m(x)2 dx 2 2u V = p(x, t) m(x) 2 , x t and V = M x (3)
Derivation of PDE
From mechanics of materials class, moment curvature relation (given here to save time) M = EI(x) 2u x 2 (4)
Substituting equation two of (3) and equation (4) into equation one of (3) and rearranging yields m(x) 2 2u 2u + EI(x) 2 = p(x, t) t 2 x 2 x (5)
Equation (5) is the PDE governing the motion u(x, t), subject to the external forcing function p(x, t).
Solving the PDE
Analytical solution difcult or impossible to obtain due to m(x) and I(x). Numerical methods such as Finite Element Method or Finite Differences can solve the PDE. Can simplify the PDE to demonstrate analytical methods by the following assumptions: m(x) = m =constant along the beam length I(x) = I =constant along the beam length p(x, t) = 0, ie, no forcing function
Solving the PDE
After simplifying assumptions the governing PDE (5) becomes 4u 2u m 2 + EI 4 = 0 t x
(6)
To make things pretty at the end, dene a2 = EI/m, so that 2u 4u + a2 4 = 0 t 2 x (7)
Solving the PDE
Assume a solution of the following form u(x, t) = (x)q(t) Substitute (8) into the PDE (7) to get 4 2q + a2 q 4 = 0 t 2 x (9) (8)
By separation of variables, observe that l.h.s and r.h.s must equal a constant, 4 1 4 1 2q = = 4 x 4 a2 q t 2 (10)
Solving the PDE
From (10), two ODEs are obtained 4 4 = 0 x 4 2q + 4 a2 q = 0 t 2 The respective solutions are (11) (12)
(x) = A sinh x + B cosh x + C sin x + D cos x (13) q(t) = E sin a 2 t + F cos a 2 t solution of the PDE (6) is u(x, t) = (x)q(t). (14)
Solving a Boundary Value Problem (BVP)
To solve a realistic problem, boundary conditions must be specied The six boundary conditions (BCs) are 1 2 3 4 5 6
@ x = 0, u(0, t) = (0)q(t) = 0 @ x = L, u(L, t) = (L)q(t) = 0 @ x = 0, u (0, t) = (0)q(t) = 0 @ x = L, u (L, t) = (L)q(t) = 0 @ t = 0, u(x, 0) = (x)q(0) = 0 @ t = 0, u(x, 0) = Gx(L x), G specied constant
Applying the boundary conditions to (x)
Applying the rst four boundary conditions yield the following results 1 2 3 4
(0) = B + D = 0 (L) = A sinh L + B cosh L + C sin L + D cos L = 0 (0) = B 2 D 2 = 0 BD =0 (L) = A 2 sinh L + B 2 cosh L C 2 sin L D 2 cos L = 0
From BCs (1) and (3), B = 0 and hence D = 0
Applying the boundary conditions to (x)
From BCs (2) and (4) A sinh L + C sin L = 0 A sinh L C sin L = 0 The above results imply A sinh L = 0 and C sin L = 0 (15)
From the rst expression of (15), A = 0. If A = 0 is not chosen, = 0 is required and this leads to (x) = 0 for all x which is the at rest condition (not very interesting). Using the remaining case (since A = 0), either C = 0 or sin L = 0. Choosing C = 0 isnt an option since that leads to (x) = 0 for all x which is the uninteresting at rest condition.
Applying the boundary conditions to (x)
Therefore, must have sin L = 0, which implies L = n. After solving for , the n solutions (which satisfy the B.Cs) for (x) are nx n (x) = Cn sin (16) L This implies that the beam vibrates in the following natural mode shapes for n = 1, 2, 3, 4... 1(x) 1 0.5 0 0.5 1 0 0.5 x 1 1 0.5 0 0.5 1 0 0.5 x 1 2(x) 1 0.5 0 0.5 1 0 0.5 x 1 3(x) 1 0.5 0 0.5 1 0 0.5 x 1 4(x)
Applying boundary condition (5)
Applying BC (5) yields q(0) = a 2 E sin a 2 0 + a 2 F cos a 2 0 = 0 (17)
The sine term equals zero and hence F = 0. As a result q(t) = E cos a 2 t In light of the fact that = n/L qn (t) = En cos an2 2 t L2 (19) (18)
Applying boundary condition (6)
Combining (16) and (19) and dening bn = Cn En yields un (x, t) = n (x)qn (t) = bn sin an2 2 t nx cos L L2 (20)
Equation (20) satises the PDE and the rst 5 BCs for any value of n and arbitrary constants bn . As a result, any linear combination of (20) also satises the requirements so that nx an2 2 t bn sin u(x, t) = (21) cos L L2 n=1
Applying the boundary condition (6)
To satisfy BC (6) the following must be true
u(x, 0) = Gx(L x) = n=1
bn sin
nx L
(22)
Hence, the bn are the sine Fourier coefcients for Gx(L x). That is bn = nx 2 L Gx(L x) sin dx L 0 L 8GL2 = 3 3 for n odd n =0 for n even (23) (24) (25)
Final solution of the BVP
Using the results of (21) and (23) gives the nal solution of the BVP.
u(x, t) = n=1,3,5,...
8GL2 an2 2 t nx cos sin L n3 3 L2
(26)
Comments: Recall a2 = EI/m which is known G species initial amplitude at t = 0, hence is known By observing the cosine term of (26) it is concluded that the natural frequencies for the beam are n = n2 2 L2 EI m
