Flow-Induced Vibrations: An Engineering Guide
By Eduard Naudascher and Donald Rockwell
5/5
()
About this ebook
Flow-induced vibrations are presented in terms of their basic elements: body oscillators, fluid oscillators, and sources of excitation. By stressing these basic elements, the authors provide a basis for the transfer of knowledge from one system to another, as well as from one engineering field to another. In this manner, well-known theories on cylinders in cross-flow or well-executed solutions from the field of wind engineering--to name just two examples--may be useful in other systems or fields on which information is scarce. The unified approach is broad enough to permit treatment of the major excitation mechanism, yet simple enough to be of practical use.
Related to Flow-Induced Vibrations
Related ebooks
Computational Fluid Dynamics with Moving Boundaries Rating: 0 out of 5 stars0 ratingsIntroduction to Fluid Dynamics Rating: 5 out of 5 stars5/5Theory of Elastic Stability Rating: 5 out of 5 stars5/5Rotating Fluids in Engineering and Science Rating: 0 out of 5 stars0 ratingsAdvanced Strength of Materials Rating: 3 out of 5 stars3/5The Laminar Boundary Layer Equations Rating: 0 out of 5 stars0 ratingsDesign for Thermal Stresses Rating: 5 out of 5 stars5/5Offshore Mechanics: Structural and Fluid Dynamics for Recent Applications Rating: 0 out of 5 stars0 ratingsMechanical Vibrations Rating: 3 out of 5 stars3/5Structural Analysis and Design of Process Equipment Rating: 0 out of 5 stars0 ratingsThe Finite Element Method: Linear Static and Dynamic Finite Element Analysis Rating: 5 out of 5 stars5/5Compressors and Modern Process Applications Rating: 0 out of 5 stars0 ratingsViscous Hypersonic Flow: Theory of Reacting and Hypersonic Boundary Layers Rating: 0 out of 5 stars0 ratingsStress in ASME Pressure Vessels, Boilers, and Nuclear Components Rating: 0 out of 5 stars0 ratingsTheory of Thermal Stresses Rating: 5 out of 5 stars5/5Fundamentals of Vibration Analysis Rating: 5 out of 5 stars5/5Control Valve Primer, Fourth Edition Rating: 3 out of 5 stars3/5Principles of Turbomachinery Rating: 0 out of 5 stars0 ratingsIntroduction to Mathematical Fluid Dynamics Rating: 5 out of 5 stars5/5Case Studies in Mechanical Engineering: Decision Making, Thermodynamics, Fluid Mechanics and Heat Transfer Rating: 5 out of 5 stars5/5Chemical Reactor Design and Control Rating: 0 out of 5 stars0 ratingsNon-Linear Elastic Deformations Rating: 4 out of 5 stars4/5Introduction to Finite Element Analysis and Design Rating: 0 out of 5 stars0 ratingsOffshore Piping Design: Technical Design Procedures & Mechanical Piping Methods Rating: 0 out of 5 stars0 ratingsStructural Reliability Analysis and Prediction Rating: 0 out of 5 stars0 ratingsHeat Conduction Rating: 5 out of 5 stars5/5An Introduction to Random Vibrations, Spectral & Wavelet Analysis: Third Edition Rating: 4 out of 5 stars4/5Case Histories in Vibration Analysis and Metal Fatigue for the Practicing Engineer Rating: 4 out of 5 stars4/5Programming the Finite Element Method Rating: 0 out of 5 stars0 ratingsCorrosion and Materials Selection: A Guide for the Chemical and Petroleum Industries Rating: 0 out of 5 stars0 ratings
Science & Mathematics For You
Sapiens: A Brief History of Humankind Rating: 4 out of 5 stars4/5Homo Deus: A Brief History of Tomorrow Rating: 4 out of 5 stars4/5Think in Systems: The Art of Strategic Planning, Effective Rating: 4 out of 5 stars4/5Ultralearning: Master Hard Skills, Outsmart the Competition, and Accelerate Your Career Rating: 4 out of 5 stars4/5Free Will Rating: 4 out of 5 stars4/5Gut: the new and revised Sunday Times bestseller Rating: 4 out of 5 stars4/5What If?: Serious Scientific Answers to Absurd Hypothetical Questions Rating: 5 out of 5 stars5/5The Science of Storytelling: Why Stories Make Us Human, and How to Tell Them Better Rating: 4 out of 5 stars4/5Feeling Good: The New Mood Therapy Rating: 4 out of 5 stars4/5Sapiens: A Graphic History, Volume 2: The Pillars of Civilization Rating: 4 out of 5 stars4/5Chaos: Making a New Science Rating: 4 out of 5 stars4/5How to Think Critically: Question, Analyze, Reflect, Debate. Rating: 4 out of 5 stars4/5The Trouble With Testosterone: And Other Essays On The Biology Of The Human Predi Rating: 4 out of 5 stars4/5The Gulag Archipelago [Volume 1]: An Experiment in Literary Investigation Rating: 4 out of 5 stars4/5Outsmart Your Brain: Why Learning is Hard and How You Can Make It Easy Rating: 4 out of 5 stars4/5The Gulag Archipelago: The Authorized Abridgement Rating: 4 out of 5 stars4/5How Emotions Are Made: The Secret Life of the Brain Rating: 4 out of 5 stars4/5Being Mortal: Illness, Medicine and What Matters in the End Rating: 4 out of 5 stars4/5This Explains Everything: 150 Deep, Beautiful, and Elegant Theories of How the World Works Rating: 4 out of 5 stars4/5The Structure of Scientific Revolutions Rating: 5 out of 5 stars5/5Deep Utopia: Life and Meaning in a Solved World Rating: 0 out of 5 stars0 ratingsThe Confidence Code: The Science and Art of Self-Assurance---What Women Should Know Rating: 4 out of 5 stars4/5Alchemy: The Dark Art and Curious Science of Creating Magic in Brands, Business, and Life Rating: 4 out of 5 stars4/5Activate Your Brain: How Understanding Your Brain Can Improve Your Work - and Your Life Rating: 4 out of 5 stars4/5The Psychology of Totalitarianism Rating: 4 out of 5 stars4/5The Systems Thinker: Essential Thinking Skills For Solving Problems, Managing Chaos, Rating: 4 out of 5 stars4/5The Illustrated Theory of Everything: The Origin and Fate of the Universe Rating: 5 out of 5 stars5/5No-Drama Discipline: the bestselling parenting guide to nurturing your child's developing mind Rating: 4 out of 5 stars4/5Healthy Brain, Happy Life: A Personal Program to to Activate Your Brain and Do Everything Better Rating: 4 out of 5 stars4/5Visual Intelligence: Sharpen Your Perception, Change Your Life Rating: 4 out of 5 stars4/5
Reviews for Flow-Induced Vibrations
2 ratings0 reviews
Book preview
Flow-Induced Vibrations - Eduard Naudascher
CHAPTER 1
Introduction
1.1 GOALS AND SCOPE OF MONOGRAPH
To this day, flow-induced structural vibrations and fluid oscillations or noise are being treated by different disciplines using different kinds of descriptions that depend generally on the type of structure or system involved. Moreover, methods of vibration and noise control are usually being obtained in an ad-hoc fashion as the problems arise. We are thus faced with a bewildering diversity of information and are tempted to reinvestigate a problem instead of applying results to resolve it.
The primary goal of this monograph is a synthesis of research results and practical experience from a variety of fields in the form of engineering guidelines. Guidance is intended, specifically, in assessing the possible sources of excitation in a flow system, in identifying the actual danger spots, and in finding appropriate remedial measures or cures. To this end, the material follows a unified approach, broad enough to permit treatment of the major excitation mechanisms, yet simple enough to be of practical use.
Flow-induced vibrations of structural or fluid masses can occur in so many different forms that several volumes would have been required had they been presented in terms of the variety of possible structures or systems affected. Since a given source of excitation can take a multitude of forms, moreover, this approach would still not have guaranteed reliable detection under all circumstances. Even well-understood vibration problems often come in disguise. Hardly any two gates, valves, or seals are alike in geometry, dynamic characteristics, and integration into an overall system; hence, they may be susceptible to quite different excitation mechanisms that call for different curative measures. On the other hand, completely different types of structures can be exposed to the same type of excitation; information and experience gathered from one of these structures, therefore, can well apply to the other, provided the basic ingredients of this excitation are recognized.
For these reasons, flow-induced vibrations are presented in terms of their basic elements: body oscillators, fluid oscillators, and sources of excitation. By stressing these basic elements, we wish to provide a basis for the transfer of knowledge from one system to another as well as from one engineering field to another. In this way, well-known theories on cylinders in cross-flow or well-executed solutions from the field of wind engineering, to name just two examples, may be useful in other systems or fields on which information is scarce.
In the area of flow-induced vibration, research is being performed so intensively and practical experiences are being gathered so abundantly that it is impossible to present today all material in a final form. All colleagues working in this field are therefore invited to report both criticism and findings from research and practical experience to the IAHR Editorial Committee (c/o IAHR Secretariat, Delft Hydraulics, Delft, the Netherlands) in order that corresponding corrections and supplementary material may be incorporated in the revised edition that is planned to appear in five to ten years.
Publication of the IAHR Hydraulic Structures Design Manual was undertaken with the intent to consolidate hydraulic-design information that is normally scattered among a vast number of journals and books. In spite of such consolidation, the manual will most likely be regarded as too extensive. The inevitable, and unfortunate, course of events may be that only a small part of the information will be extracted, simplified, and reduced to straightforward ‘design criteria’. The dangers inherent in such an outcome are obvious.
The material in this monograph does not codify existing knowledge or present binding design guidelines. The authors’ goal was simply to give the interested engineer an overview on a subject which is still little understood.
1.2 SOURCES AND ASSESSMENT OF FLOW-INDUCED VIBRATIONS
Flow-induced vibration phenomena have been treated by a variety of engineering disciplines, each having its particular terminology. In an attempt to provide a unified overview, we propose the following definition of basic elements of flow-induced vibrations:
Body oscillators;
Fluid oscillators; and
Sources of excitation.
Oscillators are defined herein as systems of structural or fluid mass that are acted upon by restoring forces if deflected from their equilibrium positions and undergo vibrations in conjunction with appropriate types of excitation. An engineering system will usually possess several potential oscillators and several sources of excitation. The first and most important task in the assessment of possible flow-induced vibrations is therefore to identify them.
A body oscillator consists of either a rigid structure or structural part that is elastically supported so that it can perform linear or angular movements (e.g., a rod or a sluice gate), or a structure or structural part that is elastic in itself so that it can perform flexural movements (e.g., a thin-walled pipe or the skinplate of a gate). No matter how simple the system may seem, it may contain a number of body oscillators, each of which can lead to vibration problems either by itself or in combination with other body or fluid oscillators. A brief review of the dynamics of basic body oscillators is given in Chapter 2, and the types of fluid loading and response of such oscillators are summarized in Chapter 3.
A fluid oscillator consists of a passive mass of fluid that can undergo oscillations usually governed either by fluid compressibility or by gravity. In both cases, the oscillating fluid mass can be discrete (e.g., the fluid in a pipe oscillating due to changes in fluid volume or changes in free-surface elevation in an adjacent chamber, Figures 4.3a, b), or it can be distributed (e.g., the fluid in a pipe or an open channel oscillating in the form of an acoustic or a gravity wave, respectively, Figures 4.6a, b). Fluid-flow systems may contain a number of fluid oscillators. They may give rise to undesirable fluid pulsations when excited (e.g., surging or wave oscillations); and they may amplify the vibration of a body oscillator if one of their natural frequencies coincides with the natural body-oscillator frequency. An overview of common fluid oscillators is presented in Chapter 4.
Sources of excitation for either body or fluid oscillators are numerous and may be difficult to detect. It is therefore useful to treat them within a basic framework. The following material distinguishes three types:
– Extraneously induced excitation (EIE);
– Instability-induced excitation (IIE); and
– Movement-induced excitation (MIE).
Extraneously induced excitation (EIE) is caused by fluctuations in flow velocities or pressures that are independent of any flow instability originating from the structure considered and independent of structural movements except for added-mass and fluid-damping effects. Examples are the bluff body in Figure 1.1a, being ‘buffeted’ by turbulence of the approach flow, and the pipe filled with compressible fluid in Figure 1.1b, being excited by a loudspeaker. The exciting force is mostly random in this category of excitation, but it may also be periodic. A case in point is a structure excited by vortices shed periodically from an upstream cylindrical structure. In either case, the vibration is sustained by an extraneous energy source. More information on sources of EIE and their characteristics and control in cases of body and fluid oscillators is found in Chapters 5 and 8, respectively.
Instability-induced excitation (IIE) is brought about by a flow instability. As a rule, this instability is intrinsic to the flow system. In other words, the flow instability is inherent to the flow created by the structure considered. Examples of this situation are the alternating vortex shedding from a cylindrical structure (Figure 1.1c) and the oscillations of an impinging free shear layer near the mouth of an organ pipe (Figure 1.1d). The exciting force is produced through a flow process (or flow instability) that takes the form of local flow oscillations even in cases where body or fluid oscillators are absent. The excitation mechanism can therefore be described in terms of a self-excited ‘flow oscillator’. (Note that the flow rather than the body or fluid oscillator is self-excited in this instance in contrast to cases of MIE, cf. Naudascher & Rockwell, 1980a).
Figure 1.1. Examples of body and fluid oscillators excited by (a, b) extraneously induced excitation (EIE), (c, d) instability-induced excitation (IIE), and (e, f) movement-induced excitation (MIE).
An important role regarding instability-induced vibrations of body or fluid oscillators is played by the type and strength of the control exerted on the flow instability. As shown in Sections 6.1 and 6.2, the nature of such control can be
– fluid-dynamic;
– fluid-elastic; or
– fluid-resonant.
In the fluid-dynamic case, the exciting force is a function of the flow conditions only. In the fluid-elastic and the fluid-resonant cases, this force depends on the dynamics of both the flow and the resonator in the system: a resonating body oscillator in the former case, and a resonating fluid oscillator in the latter. The main feature of fluid-elastic and fluid-resonant control is an amplification of the exciting force and a ‘locking-in’ of its frequency to that of the resonator within a certain range of flow velocities.
The main characteristics of five basic types of flow instability and methods for attenuating structural vibrations produced by these IIE are reviewed in Chapter 6. Examples of fluid oscillators affected by IIE are presented in Chapter 8.
Movement-induced excitation (MIE) is due to fluctuating forces that arise from movements of the vibrating body or fluid oscillator. Vibrations of the latter are thus self-excited. Examples are shown in Figures 1.1e and 1.1f. If the air- or hydrofoil depicted in Figure 1.1e is given an appropriate disturbance in both the transverse and torsional mode, the flow will induce a pressure field that tends to increase that disturbance. This situation can be described in terms of a dynamic instability of the body oscillator which gives rise to energy transfer from the main flow to the oscillator. Similar situations are possible with regard to fluid oscillators. As an example, Figure 1.1f shows an open pipe in supersonic flow in which a standing wave is sustained by MIE involving oscillatory movements of a shock front. In Chapter 7, different types of MIE are presented in terms of four basic groups. The main features of these MIE mechanisms are described so that they can be identified in any system. In addition, this chapter contains information on the onset of structural vibrations and methods of controlling them. Examples of MIE with fluid oscillators are given in Chapter 8.
If a system is susceptible to MIE, it is generally sufficient to ensure that the system operates below the thresholds of MIE. The latter are determined by the stability criteria presented in Chapter 7.
Frequently, the excitation of flow-induced vibration in a complex system is mixed in the sense that (a) both body and fluid oscillators are involved at the same time or (b) EIE, IIE, and MIE are present simultaneously. For example, the cylindrical structure of Figure 1.1c can be placed in a duct or channel in such a way that a standing wave is produced by the vortex shedding, in addition to structural vibration (Figure 8.14); or, it can be excited by turbulence buffeting, in addition to vortex excitation. Even in these mixed cases, however, it is advantageous to first identify all body oscillators, fluid oscillators, and sources of pure EIE, IIE, and MIE. Some examples of mixed excitation are presented in Chapter 9.
Whereas Chapters 5 to 8 focus on the basic mechanisms of flow-induced excitation, and their detection and evaluation, the focus of attention in Chapter 9 is the structure itself and the flow-induced vibrations it can undergo depending on its geometric and dynamic characteristics. Prismatic bodies and grids of prisms, on the one hand, and gate and gate components, on the other, are used as examples to illustrate the wide variety of mechanisms that can excite structural vibrations either individually or in combinations.
In rare cases, flow-induced vibrations are due to parametric excitation. These cases involve the variation with time of one or more parameters of the vibratory system such as mass, damping, and rigidity. As pointed out in Section 2.5, this variation may be of either the EIE or MIE variety. To safeguard against overlooking a potential source of excitation, one should search for these sources of vibration separately from the classification scheme presented in Figure 1.1.
In summary, then, the assessment of possible flow-induced vibrations in a system involves, first, a thorough search for
All body oscillators;
All fluid oscillators;
All sources of extraneously induced excitation;
All sources of instability-induced excitation;
All sources of movement-induced excitation; and
All sources of parametric excitation,
and, second, an assessment of all possible combinations of structural and fluid oscillations arising from (a) and (b) in conjunction with (c) through (f). What makes combinations of body and fluid oscillators dangerous is the coincidence of their natural frequencies. Estimates of these frequencies, as well as the dominant frequencies of possible EIE and IIE, should therefore be an integral part of the assessment.
A useful aid in preliminary investigations of ‘danger spots’ and dangerous operating conditions of a system is a global or lumped-parameter analysis as suggested, e.g., by Naudascher & Rockwell (1980a). If the magnitudes of loads on, and strains in, certain structures or structural parts are of interest, one has to resort to the methods of analysis described in Sections 3.3, 5.2 and 7.2 or to specific model tests. Nevertheless, prior assessment and global analysis will be important even in these cases in order that no essential element of the system and none of the possible types of excitation and their dangerous combinations is disregarded in the analytical or experimental investigation. In the case of model tests, such assessment will also help to exclude spurious excitations such as those caused by peculiar resonance conditions in the laboratory set-up that have no counterpart in the prototype.
CHAPTER 2
Body oscillators
2.1 OVERVIEW AND DEFINITIONS
Because the writers’ main goal is to identify the variety of mechanisms by which flow-induced vibrations are excited, the structural dynamics are presented in the simplest way possible throughout this monograph. In most cases, this means representing the vibrating structure or structural part as a discrete mass, free to oscillate with one degree of freedom, linearly damped, and supported by a linear spring (Figures 2.1 and 2.2). The following sections contain a brief review from the field of mechanical vibrations concerning these simple body oscillators that is sufficient for the understanding of the monograph. A method of generalization is presented, finally, by which simple-oscillator concepts become applicable to more complex systems with continuous or distributed masses such as beams, plates, and shells.
Any vibration is describable in terms of sinusoidal functions. The simplest vibration is a harmonic motion of the form
(2.1)
where x = body deflection from its time-mean position, xo = amplitude, t = time, ω = circular frequency, and ƒ = frequency in cycles per second or Hertz. One of the most useful ways of describing simple harmonic motion is obtained by regarding it as the projection on the horizontal axis of a vector of length xo rotating counterclockwise with uniform angular velocity ω (Figure 2.1). This rotating-vector description is commonly represented by the complex exponential function
(2.2)
for which Ox is the ‘real’ and Oy ). Thus, the real part of this expression may be considered the horizontal projection and the imaginary part the vertical projection; and again, it is the former which represents the harmonic motion or vibration. In Figure 2.1b, T = 1/ƒ denotes the period of vibration.
Figure 2.1. Definition sketch. (a) Simple undamped body oscillator. (b) Histogram of harmonic motion. (c) Vector respresentation of harmonic motion.
2.2 FREE VIBRATION
The simplest body oscillator consists of a discrete mass m free to vibrate in one direction (Figure 2.1a). In the absence of damping and exciting forces, an initial displacement xo (the dots denote derivatives with respect to time) is set equal to the sum of all forces
(2.3)
Fx = −Cx. The larger the spring constant C, the greater is the stiffness of the oscillator. The solution of the equation of motion
(2.4)
is
(2.5)
with
(2.6)
where ωn is the natural circular frequency and ƒn is the natural frequency of the discrete-mass system. For the particular initial condition cited above, the phase angle ϕ is zero.
For systems of greater complexity, determination of the natural frequency from the equation of motion becomes so complicated that it is advisable to use an energy approach. Since the spring force Cx in the middle of the stroke (x = ωxo) must be equal to the elastic energy in an extreme position (x = xo= 0), i.e.,
(2.7)
Figure 2.2. (a) Simple body oscillator with linear damping. (b, c) Histograms of responses for an underdamped (ζ < 1) and an overdamped (ζ ≥ 1) case.
C/m, independent of the amplitude xo. A method for evaluating ωn in cases of nonlinear systems is described in conjunction with Figure 2.10.
With linear damping (i.e., resistance proportional to velocity) included, Equation 2.3 takes the form
(2.8)
in which −Bdx/dt is the damping force (Figure 2.2a). The solution of this equation is
(2.9)
as long as the damping factor or damping ratio ζ is smaller than one (ζ < 1). The latter is defined as
(2.10)
and ωd is the frequency of the damped free vibration. (Note that for a given damper, B = const., and ζ decreases with increasing mass m and stiffness C.) In Figure 2.2b, this exponentially decaying response is represented for the initial condition t = 0, x = xo; thus, ϕ = 0. For ζ ≥ 1, the displaced body simply returns to its equilibrium position in an exponential fashion (Figure 2.2c). The damping for the limiting case (ζ = 1) is called critical damping.
In the underdamped case, 0 < ζ < 1, the ratio of any two consecutive amplitudes is obtained from Equation 2.9 as xn/xn . The logarithm of this ratio is called the logarithmic decrement
(2.11)
2πζ.
With the aid of Equations 2.6 and 2.10, the equation of motion (Equation 2.8) may also be written as
2.3 FORCED VIBRATION
2.3.1 Harmonic exciting force
If a simple oscillator is acted upon by a harmonic exciting force
(2.12)
(Figure 2.3), the equation of motion (Equation 2.3) takes the form
or
(2.13)
The solution of this equation is composed of the solution for the homogenous equation presented in Equation 2.9, and the particular solution given by
(2.14)
Since the former or transient solution dies out with time on account of damping (Figure 2.3b), only the latter or steady-state solution is of general interest. Its frequency is equal to the forcing frequency ƒs = ωs/2π; its amplitude xo is obtained as
(2.15)
(Figures 2.4a, b); and the phase angle ϕ by which the response x lags the exciting force F follows from
(2.16)
(Figure 2.4c); ωn is the natural frequency of the undamped system. The ratio of xo and the static deflection Fo/C of the body due to Fo, depicted in Figures 2.4a and b, is called the magnification factor. Note that for lightly damped cases, corresponding to values of ζ of the order of 0.05 or less, this ratio can be much larger than unity. The maxima of the response occur at
(2.17)
Figure 2.3. Definition sketch of forced vibration of a simple body oscillator.
Figure 2.4. Magnification factor and phase angle for forced vibration of a linear body oscillator.
The symmetric representation of the magnification factor in Figure 2.4b has advantages, for example, if the damping is to be determined from a few measured points on that diagram.
For a body with a torsional degree of freedom, Equation 2.13 takes the form
or
(2.18)
where Iθ = mass moment of inertia of the body, M(t) = damping moment, −Cθθ = restoring moment, ζθ = Bθ/(2Iθωn) = damping ratio, and ωθn = undamped circular natural frequency. The response to a harmonic exciting moment M(t) = Mo cos ωst is equivalent to the one just derived, i.e., where θo and ϕ follow from Figure 2.4.
Figure 2.5. Definition sketch for quality factor Q.
(2.19)
A definition most useful in applying the above results to diverse kinds of physical systems, both mechanical and nonmechanical, is the quality factor
(2.20)
For lightly damped system, ζ < 0.05, the oscillator approaches resonance near ω/ωn = 1 and the maximum amplitude may be approximated by
(2.21)
The points P1, and P2 in the amplification diagram, where the relative amplitude xo/(Fo/C) falls to Q/√2, are called half-power points, because the power absorbed by the damper in a system responding harmonically is proportional to the square of the amplitude (Figure 2.5). The increment Δω of the circular frequency ω associated with the half-power points, sometimes called resonance width, takes the form
(2.22)
for light damping. Consequently,
(2.23)
This relationship can be used as a convenient way of determining the damping ratio ζ.
2.3.2 Mechanical admittance and impedance
A useful representation of forced vibration is obtained with the aid of input-output functions such as: (a) mechanical admittance
(2.24)
or (b) mechanical impedance
(2.25)
Response and excitation in Equations 2.24 and 2.25 are commonly presented in complex form,
(2.26)
with the tacit understanding that the excitation is given by the real part of F(t), just as the response is given by the real part of x(t) in Equation 2.2. Introducing these expressions in Equation 2.13 and deriving the particular solution, one obtains
(2.27)
which gives information concerning both phase and amplitude. From complex algebra, the absolute value of χm(ω) is equal to the ratio of response amplitude xo to excitation amplitude Fo, i.e.,
(2.28)
This, evidently, is the magnification factor of Figure 2.4a divided by the spring constant C (Equation 2.15).
Since eiωt ≡ cos ωt + i sin ωt can be regarded as a unit vector rotating in the complex plane with angular velocity ω (Figure 2.1c), one can deduce from Equations 2.27 and 2.28
(2.29)
with ϕ as given by Equation 2.16. If one substitutes this and Equation 2.26 into Equation 2.27, one obtains
(2.30)
This illustrates clearly that the response lags the exciting force by the phase angle ϕ (Figure 2.3c).
2.3.3 Vector and force-displacement diagrams
The complex description introduced by Equations 2.2 and 2.26 permits an interesting geometric interpretation of the equation of motion (Equation 2.13) in the complex plane. Through differentiation, one arrives at
(2.31 )
lead the body displacement x by phase angles of π/2 and π, respectively. Therefore, the restoring force – Cx, the damping force —B , and the intertia force —m can be represented by vectors as shown in Figure 2.6a, all rotating with the same angular velocity ω, equal to the driving circular frequency ωs. The projections of these vectors onto one of the axes give the instantaneous values of the respective forces. If plotted as functions of the instantaneous values of the displacement xin Figure 2.6a.) An advantage of such diagrams is their applicability to nonlinear oscillators or non-harmonic excitations (Sections 2.3.5 and 2.3.6).
Figure 2.6. (a) Vector diagram for a harmonic exciting force F(t) = Fo cos ωst acting on a simple body oscillator (Equation 2.13). (b) Corresponding force-displacement diagram.
The following features of the force displacement diagram have special significance in this monograph:
a) The linear restoring force −Cx corresponds to a straight line with negative slope. A nonlinear restoring force would correspond to a curved line, and negative stiffness (C < 0) would produce a positive slope.
b) All non-conservative forces (like the exciting and damping forces) form a loop; the areas included in these loops correspond to the work done on the body during one cycle.
c) If a point on a force-displacement loop moves in the direction of ωt as time proceeds (i.e., counterclockwise), the work done on the body is negative and energy is dissipated.
The last of these statements is a result of Equation 2.32.
2.3.4 Energy consideration
integrated over one cycle:
(2.32)
where T is the vibration period. Applied to the exciting force (Equation 2.12 or 2.26), this equation yields
(2.33)
The result shows clearly that energy transfer to the body depends on the phase angle ϕ between exciting force and body displacement (Equation 2.16) and that this transfer is a maximum for ϕ = +π/2 and its multiples.
The work per cycle done by the damping force – B is
(2.34)
It is negative, which means that energy is being dissipated. From the plot of We and Wd for a given value of excitation frequency ω = ωs in the energy diagram one may conclude that steady-state force vibrations are obtained only at x = xo where We + Wd = 0. For any amplitude different from xo, the imbalance of We and −Wd is such that the amplitude will adjust with time to x → xo. Diagrams of this type facilitate illustration of nonlinear systems.
Even without a damper (B in Figure 2.3b), energy may be dissipated by what is called internal friction or structural damping. Any spring made of material which is not perfectly elastic shows some hysteretic deviations from Hooke’s linear stress-strain relationship. A typical force-displacement (or stress-strain) diagram for such a case is depicted in Figure 2.7a. During both loading and unloading of the spring, the force-displacement trace deviates from a straight line and forms a hysteresis loop as shown. The shape of this loop is nearly independent of amplitude xo and strain rate (Kimball, 1929). Consequently, the energy dissipated during one cycle, which is given by the area W. W* can be equated to the work per cycle done by an equivalent linear damping force −B,
(2.35)
Figure 2.7. Force-displacement diagrams for (a) a spring with internal friction and (b) an equivalent linear system.
, where ωn is the natural frequency of the undamped system (Figure 2.7b).
2.3.5 Nonharmonic exciting forces
Frequently, the exciting force is periodic but not harmonic, or it is nonperiodic or random. In the former case, it can be represented by a convergent series of harmonic functions whose frequencies are integer multiples of a fundamental frequency ωo. Such a series of harmonic functions is known as a Fourier series and can be written in the form
(2.36)
with the tacit understanding that the excitation is given by the real part of F(t) (Meirovitch, 1975, p. 61). If the system of the body oscillator is linear, its responses to the first, second, ... harmonics (N = 1, 2, ...) of Equation 2.36 can be deduced separately and then added. Thus, in analogy to Equation 2.30,
(2.37)
where |χm|N and ϕN are obtained from Equations 2.28 and 2.15, respectively, after substitution of ω = Nω0, i.e.
(2.38)
Clearly, the response x(t) is periodic just as is F(t). If the value of one of the harmonics Nωo, of the excitation is close to the natural frequency ωn of the system and if the system is lightly damped, this harmonic will provide a relatively larger contribution to the response because it is associated with the peak value of the mechanical admittance |χm| (Figure 2.4a).
If the period To = 2π/ωo of the excitation function in Equation 2.36 approaches infinity, this function becomes nonperiodic and must be represented by a Fourier integral instead of a Fourier series. In practice, such nonperiodic or random excitation is represented by means of a spectrum as shown in Figure 2.8a. If one treats the body oscillator as a linear, lightly-damped system, then the relationship between the spectra of the excitation and the response is given by
(2.39)
are the power spectral densities of the excit-ing force and the response; Ff and xf are the amplitudes of the exciting force and the body deflection associated with frequency ƒ, which are obtained after filtering out all frequencies outside a narrow band dƒ centered at ƒ; and the bars denote averaging (more details are contained in the IAHR Monograph on Hydrodynamic Forces, Naudascher, 1991, Figure 2.5). The equation implies that one has to simply multiply each ordinate value of the excitation spectrum with the corresponding value of |χm|² to obtain a point on the response spectrum (Figure 2.8). The mechanical admittance function (Figure 2.4a) is hence ideally suited to determine linear-system responses to nonharmonic excitation as well.
Figure 2.8. Evaluation of the response spectrum Sx(f) from the excitation spectrum SF(t) with the aid of the mechanical admittance function |χm|² in accordance with Equation 2.39.
The root-mean-square of the response or mean response amplitude is obtained by summing up the contributions to the spectrum from all frequencies as follows
(2.40)
(either positive or negative) as follows:
x > x′ or < −x′ 15.9% of the time each;
x > 2x′ or < −2x′ 2.3% of the time each;
x > 3x′ or < −3x′ 0.14% of the time each.
2.3.6 Nonlinear effects
The body-oscillator system in Figure 2.3a is called nonlinear if one or more of the coefficients m, B, or C depend on the displacement x or on its time derivatives. In mechanical cases, the most important nonlinearities occur in the damping or in the spring. Examples of nonlinear springs are shown in Figure 2.9. The most important consequence of nonlinearity in the spring is the fact that the natural frequency becomes a function of amplitude. With nonlinear damping, on the other hand, ωn but the rate of decrease in amplitude no longer follows the simple law of Equation 2.11.
A very simple and generally applicable method for investigating nonlinear systems consists in constructing phase diagrams(x) /dt /dx)dx/dt /dx, the equation m + Cx = 0 can be transformed into m d + Cx dx = 0. Integrated, this yields
(2.41)
/ωn plotted against x yields concentric circles in this case (Figure 2.10a).
/ωn versus x is
Figure 2.9. Systems with nonlinear stiffness. (a) Combination of linear springs and clearances. (b, c) Springs with more or less gradually increasing stiffness (hard springs).
Figure 2.10. Velocity-displacement or phase diagrams for (a) linear undamped free system, (b) free system sketched in Figure 2.9a. (c) Definition sketch.
easy to construct: To the right of O2 and to the left of O1, the system reacts like a linear C – m-system, whereas in between, the mass moves with constant speed as it is not acted upon by any force. The result is a periodic, nonharmonic vibration completely specified by the phase diagram in Figure 2.10b. Its natural frequency can be determined graphically by carrying out the second integration in this diagram. Considering the element of curve in Figure 2.10c, one has
= dx/dt or dt = dx
is the ordinate value. Thus, the time consumed by the mass progressing a distance dx is t = ∫dx. For the full period T, or one complete cycle of oscillation, hence
(2.42a)
or for a symmetrical spring
(4.42b)
This integral can be evaluated analytically or numerically. For the example shown in Figure 2.9a or 2.10b, the result is a period dependent on amplitude xo:
(2.43)
Typical effects of nonlinearity on the forced vibration of a body are illustrated in Figure 2.11 for a body oscillator as shown in Figure 2.9b. In case the system damping is linear and the spring cubic, the equation of motion becomes, after division by the mass m,
(2.44)
The diagrams for the magnification factor and the phase angle in this case deviate substantially from those of the linear case. Again, ωn is a function of amplitude xo (Magnus, 1965).
The significance of the bent-over resonance curves in Figure 2.11 a is illustrated in Figure 2.12 for a case with given damping ζ. Within a certain range of ωs/ωn, there are now two or three solutions, rather than one, for the amplitude xo at a given value of ωs/ωn. Physically, this means that there are ‘jumps’ from one to another amplitude as ωs/ωn is changed, and the system shows a hysteresis behavior in that different values of xo are obtained depending on whether ωs/ωn is increased or decreased (the relative amplitude jumps from A to B for increasing ωs/ωn and from C to D for decreasing ωs/ωn). Both of these jumps as well as the hysteresis effect are typical of nonlinear systems.
Figures 2.12b and 2.12c depict the energy and phase diagram, respectively, for a value of ωs/ωn = (ωs/ωn)* for which Figure 2.12a indicates three solutions, namely (xo)1, (xo)2, and (xo)3. Whenever a vibration is started within the shaded areas of the phase diagram (e.g., at point 4 and 6), more energy is supplied than dissipated (We > −Wd), and the amplitude increases until the trajectories reach what is called the limit cycles(x) curves corresponding to the solutions x = xo. Clearly, then, the shaded areas correspond to regions of amplification. If xo is below (xo)2 (say at point 5), on the other hand, dissipation will be larger than the supply of energy, and the amplitude will decay. The argument shows that only (xo)1 and (xo)3 are stable solutions, while (xo)2 marks the threshold amplitude. If xo is larger [or smaller] than (xo)2, for example due to an initial disturbance, the amplitude will change until it approaches (xo)3 [or (xo)1]. Which of the two stable solutions is obtained for ωs/ωn = (ωs/ωn)* has thus been determined.
Figure 2.11. Magnification factor and phase angle for forced vibration of a nonlinear oscillator (Figure 2.9b, Equation 2.44) with hard spring (α = + 0.04) (after Magnus, 1965).
Figure 2.12. Explanation of jumps and hysteresis for a nonlinear oscillator. (a) Resonance curve; (b, c) Energy and phase diagrams for ωs/ωn = (ωs/ ωn)*.
More detailed treatments of nonlinear oscillators, including the possibility of subharmonic responses at frequencies ωs/N (N = 1, 2, 3, ...), are presented in the relevant literature (e.g., Nayfeh & Mook, 1979).
2.4 SELF-EXCITED VIBRATION
Even without an external exciting force, a body oscillator may undergo sustained vibration if there is an energy source from which the oscillator can extract energy during each cycle of free movement. Vibrations of this type are called self-excited. An example is shown in Figure 2.13. Essential for the energy transfer in this example is the negative characteristic of the dry-friction force transmitted by the belt on the body (Figure 2.13b). For small values and small changes of the relative velocity between belt and body, ν , this friction force can be approximated by Fd = (Fd)o − B*(ν ) = const + B, and the equation of motion becomes m F = B− Cx + const. Redefining the equilibrium position, one may write
(2.45)
Comparison with Equation 2.8 shows that this equation describes a free vibration with negative damping. (or x), therefore, the response is as described by Equation 2.9 except that the exponent there becomes positive. Initially, hence, the amplitudes increase exponentially. With growing amplitudes, of course, the nonlinearity of the damping force becomes more and more pronounced, positive damping comes into play, and the vibration approaches asymptotically a steady state with constant amplitude xo (Figure 2.13c).
Probably the best known nonlinear equation describing a certain class of self-excited oscillators (e.g., electrical circuits containing vacuum tubes) is of the Van der Pol