|
Domain |
Explanation |
|
Vibration? |
- The physical oscillations (even or uneven) with respect to time, space or any desired variable of a well-defined structure due to either internal disturbances or external excitations or a combination of both
- The system is literally swinging to and fro with different amplitudes (max offset from rest position), frequencies (cycles in unit time) and inducing various engineering situations within the structural components
- A famous example is the student's vibrating ruler
|
|
What causes these vibrations? |
- Literally anything that directly or indirectly contact the structure
- To excite a structure externally, the requirements include:
- Direct contact
: impacts (hits, knocks, quakes, explosions or crashes) & traction / point
- Indirect contact
: force fields including gravitational, electromagnetic or nuclear origins
- Excitation energy/structure mass ratio
: the higher the more & intense the vibrations
- Location(s) of the excitations
: excitations at the dynamically weak spots of a structure induce heavier vibrations
- Intensity of the excitations
: certain intensity excites the structure much more than others given the same energy input => resonance (max disturbed response) occurs when excited intensity equals the structural dynamic frequency
- Phase difference between the excitation frequency and vibration frequency
: the greater the phase difference, the more the increase in induced vibrations => max when the "excitation wave" is out of phase with the "vibration wave" (wave interaction)
- To excite a structure internally, the requirements are the same as above - the Principle of Transmissibility (sources different, effects equivalent)
|
|
What are the types of excitations? |
- Static: invariant with time
- Periodic-harmonic: sin, cos, sinh, cosh loads
- Periodic-general: time domain, frequency domain
- Stochastic: fits known stochastic distribution(s)
- Random: stochastic & stationary Gaussian - mean, definite
- Arbitrary: earthquakes & contingencies - Duhamel integral with Dirac delta
- White noise: complete spectrum of all available disturbances - ideal, filtered/unfiltered, non-white, differential distribution
|
|
Modeling |
- The process of modeling is an oxymoron (irony) in itself
- On one hand, models are fundamental for engineers to understand a phenomenal reality and its attributes essential for engineering
- On the other, the mere representation of reality with models is inherently flawed - assumptions are made, reduction into a finite set of performance variables & relationship between these variables (laws, theories, theorems, principles, techniques)
- It is interesting to note that researchers are cracking their brains to form & improve models when their eyes should have been directed to the phenomenal reality itself - however, the structure is often yet to be built but there are always similar ones in existence whether artificial or natural
- The modeling process:
- Phenomenal reality
- Model representation: assumptions
- Mathematical model: continuous/distributed (exact towards model) & discrete (approximate towards model)
- Solution techniques: mixture of mathematics & engineering methodologies with issues of consistency, convergence, stability, etc.
- Results: interpretations & appropriate applications
|
|
Model representation |
- To formulate & understand the identified parameters that are significant to the phenomenon in question
- Overall structure, members, connections & loads
- Since continuum, use free-body diagram (FBD) of finite dimension with infinitesimal loads, forces and dimensions wrt. space & time
- Boundary-initial-value conditions
- Assumptions of relationships in terms of equilibrium, compatibility and trial solutions
- Governing equations
- Solution principles & techniques
- Approximations with truncations, rounding-off and boundedness concerns
|
|
Vibration aspects? |
- Free vibration: ODE problem - modal frequencies (wi) & mode shapes (y i)
- Axial
- Flexural
- Shear
- Torsional
- Orthogonality of mode shapes: [f r T][M][f s] = 0, [f r T][K][f s] = 0, r¹ s
|
|
Approximate method of determining natural frequencies? |
- Natural frequencies are the intrinsic structural frequencies where resonance occurs when equal to the excitation / disturbance frequencies
- Rayleigh method
: COE - KE + SE = constant, assuming v(x,t)=f (x).Y(t)
- Rayleigh principle
: with reasonable mode shape f (x) assumed, satisfying the slope & deflection BC, good approximation of natural frequencies (slightly higher - conservative side) w2 = K*/M*
- Rayleigh-Ritz method
: reduces into finite number of dof's, v(x)=NS j=1Aj.y j , with y j satisfying Rayleigh's principle
- Principle of Virtual Work
: WE = WI with Rayleigh-Ritz method
- Note the need for appropriate trial functions to be assumed (unnecessary in FEM, replaced by the polynomial shape interpolation function)
|
|
Uncoupling equations of motion |
- Process of mode superposition method:
- Assume mode shape: y j wrt. BC (slope & deflection), v(x,t)=NS j=1y j.Y(t)
- Stiffness: kij = ò EI.y i. y j.dx + k.q i. q j = (distributed) + (lumped)
- Mass: mij = ò r A.y i. y j.dx + M.y i. y j = (distributed) + (lumped)
- Rayleigh damping
: [C]=a .[M]+b .[K], thus orthogonality condition applies where V r = a /2wr + b .wr/2 with lightly damped: [C]=2.V r.wr.[M]
- Form governing equation of motion
- Det( [K]-w2[M] ) = 0 for modal frequencies wi & mode shapes y i
- Form the uncoupled N modal equations of motion
- Solve accordingly
- Note: earthquake analysis (higher energy input, more modes) & wind analysis (lower energy input, larger surface area, fewer modes needed)
|
|
Limitations |
- Modal superposition method only applies to linear systems due to the linear combination of trial solutions (mode shapes )
- Considerations of only the first N modes - truncation error
- Mode shapes form basis (linearly indept.) of displacement vector field, but incomplete basis
- Orthogonality of mode shapes applies to [M] & [K] only with assumption of energy-dissipating damping with Rayleigh damping [C]=a .[M]+b .[K]
|
|
Strengths |
- Computationally efficient
- Flexible & allows generalization of user-defined considered modes
- Enables Duhamel integral (for arbitrary excitations) and modal participation factor (G i=ò m(x).f i(x)dx/ò m(x).f i2(x)dx) to be used for analysis
|