KBR Technical Portfolio

Infrastructure Course 3.1

3.1 General principles of dynamic design and specific applications

Contents
Introduction Units Sources of vibration Vibrational response of structures Design considerations Guides to specifications Design of pad foundations for machinery Spring mounted equipment Systems with two degrees of freedom Beats with multiple equipment Appendix A Use of general-purpose analysis software Appendix B Manual checking of dynamic response Appendix C Dynamic analysis and design for mining and other plant Appendix D Dynamic design issues for footbridges Appendix E Wind induced dynamics Appendix F Dynamic loads imposed by pneumatically-tyred vehicles Document revision history 1 2 3 4 7 11 12 17 18 19 20 21 25 28 29 31 36

Introduction
This document was originally published as Kinhills Technical Bulletin No 7/81, with some additional information separately published as addenda later. It has now been further supplemented with new material, and assembled as a complete document in a new form for use by KBR staff. The purpose of this document, Infrastructure Course 3.1, is to acquaint structural engineers with those aspects of vibration theory which are necessary for the design of structures subject to sources of excitation. Particular applications are addressed in Appendix Cfor mining and other machinery; Appendix Dfor footbridges or walkways; Appendix Efor wind sensitive structures; and Appendix Ffor the effect of pneumatic tyres.

General principles of dynamic design, Rev. E 1 September 2008

KBR Technical Portfolio

Infrastructure Course 3.1

Only the essence of the theory is presented, and the document explains the fundamental issues most relevant to day-to-day problems. It begins with an explanation of the sources of vibration and explains the difference between forced vibrations and resonance. The response of structures to vibration is then discussed, leading to a formulation of the principles of anti-vibration design and practical solutions. The acceptability of different levels of vibration is considered. The particular approaches to design of pad foundations for vibrating machinery are presented. This is a situation that is frequently encountered but poorly understood, and the preferred method of solution has been prepared as a spreadsheet. Practical advice is offered in terms of what should be included in specifications when issuing documents for purchase of machinery, with respect to information to be supplied and the boundary of responsibility for design of supporting structures. The question of what geotechnical information is required is also discussed. Appendix A offers some general guidance in the application of modern structural analysis software to the modelling of dynamics problems. Appendix B gives guidance on manual checking of dynamic response, useful for a first assessment or for checking the sense of a computer analysis. Appendix F discusses the forces likely to be imposed on supporting structures by vehicles with pneumatic tyres. It should be noted that wherever the term amplitude is used in this document it refers to the mean-to-extreme excursion.

Units
When evaluating formulae in structural dynamics, it is essential that a consistent set of units is used throughout. Particular attention must be paid to the distinction between force and mass, with forces always expressed in absolute units rather than gravitational units. Table 1 below presents some alternative sets of consistent units. Only the fundamental units and some key derived units are given: any other derived units can be expressed in terms of the fundamental units.
Table 1 Quantity Fundamental units: Length Mass Time Rotational displacement Key derived units: Force Stress / pressure / elastic modulus newton pascal kilonewton kilopascal newton megapascal metre kilogram second radian metre tonne second radian millimetre tonne second radian Possible sets of consistent units Set 1 Set 2 Set 3

General principles of dynamic design, Rev. E 1 September 2008

KBR Technical Portfolio

Infrastructure Course 3.1

Sources of vibration
Any elastic system will vibrate if displaced from its rest position by a force and then released. This is called natural vibration, and the frequency of the vibration is called the natural frequency. Easily understood examples from the physics of our schooldays are the tuning fork, or a weight on a spring, or a pendulum. Because some damping (i.e. absorption of energy) is always present, the amplitude of the vibration progressively dies down from its maximum value (the initial displacement) towards zero. In order to maintain an ongoing state of vibration, an input of energy is required to replace the energy dissipated by damping. Lightweight structures such as walkways can be excited merely by pedestrian traffic or the wind. However, the most common source of vibration with which engineers in industry are concerned is a pulsating force of regular frequency, usually caused by reciprocating machinery, out-of-balance parts of rotating machinery, or a combination of both. The types of machinery causing such pulsating forces include:
Motors and generators Turbines Compressors Vibrating screens, apron feeders or grizzlys Crushers (jaw and gyratory) Hammer mills Motor drives for drum screens, conveyors, etc.

Most rotating machinery is intended to be dynamically balanced. However, with large machines it will be necessary to allow for some eccentricity of the centre of gravity of the rotating parts due to manufacturing tolerances or wear. At the other extreme, vibrating screens have a mass which oscillates, and this motion is generated by a deliberately out-of-balance mass on a motor-driven flywheel. Jaw crushers and gyratory crushers also have deliberate out-of-balance forces. In rotating machinery, the centripetal force set up is F = me 2 m is the out-of-balance mass where e is its eccentricity is the rotational velocity (in radians/sec) The frequency in cycles per second (cps) is f =

It is normal, for structural design, to examine the structure or foundation for the effects of horizontal and vertical loads. Assuming the rotation is about a horizontal axis, the horizontal and vertical components of the dynamic forces vary with time as follows:

Fx = me 2 cos (t )
F y = me 2 sin (t )

General principles of dynamic design, Rev. E 1 September 2008

KBR Technical Portfolio

Infrastructure Course 3.1

Since the peak force is the same in each case, and equal to the centripetal force, the above equations are of interest mainly in further theoretical analysis. They do, however, show that the two forces follow a simple harmonic pattern, rising to a peak, dropping to zero, then reversing in sign. The above are the forces acting at the shaft of the motor or flywheel, and with a rigidly mounted machine these are transmitted to the supporting structure. If the system is further suspended on a spring frame, as is the case for a vibrating screen, or on rubber mounting pads, then the force transmitted to the supporting structure will be modified.

Vibrational response of structures

FORCED VIBRATIONS

A structure subjected to any fluctuating force will deflect in response to changes in that force, at the same frequency as the changes in the applied force. However, the amplitude of oscillation will differ from the static deflection which would have occurred under a steady application of the force. This condition is called forced vibration, and the designers primary concern is to ensure that the amplitude of the resulting vibrations will be at an acceptable level.
HARMONIC VIBRATIONS

Harmonic vibrations are a special case of forced vibrations. If all components of the fluctuating force are varying sinusoidally and at the same frequency, then the structure will settle into a steady state response in which every component of its displacement is varying sinusoidally. The frequency of these sinusoidally-varying displacements will be the same as that of the fluctuating force. This situation is known as harmonic vibration. It may take some time to establish itself, because any transient vibrations associated with start-up need to die away first.
NATURAL VIBRATIONS

As explained above, all elastic systems have one or more natural frequencies of vibration. For a concentrated mass attached to a mass-less elastic support without damping (a system with a single degree of freedom):

n = 2f n =
where

k M

n is the natural frequency in radian/sec,

f n is the natural frequency in cycles/sec, k is the spring stiffness of the elastic support (the static force required to
produce a unit deflection), M is the concentrated mass.

General principles of dynamic design, Rev. E 1 September 2008

KBR Technical Portfolio

Infrastructure Course 3.1

With damping, the natural frequency is modified slightly so that

f nd

fn 1 D2

where D is the damping ratio as defined below.

RESONANCE

As stated previously, the amplitude of vibration is not the same as the deflection that would occur if the applied load was static. As the frequency of the applied force approaches the natural frequency of the structural system the amplitude is greatly magnified, and the mathematical expressions show that without damping the amplitude would become infinite. There is always some energy loss or damping in the system, which prevents this situation from arising. However, for systems with low damping the amplitude can be magnified many-fold. This phenomenon is called resonance, and a primary aim of all dynamic design is to avoid it altogether by ensuring that the natural frequency of the system is well removed from the frequency of the oscillating applied forces.
DAMPING

Damping is dissipation of energy, from hysteresis within the material of the supporting structure itself, from frictional movement in bolted joints, slip between reinforcing rods and concrete, friction between soil particles, etc. An additional form of damping is applicable to pad foundations, where the vibrational energy is carried away to infinity by compression and shear waves in the soil: this is known as geometric or radiation damping. Critical damping is defined as the smallest amount of damping for which no oscillation occurs in the free response. The actual damping present in any system is then expressed as the ratio of the actual damping to the critical damping. This damping ratio ( D ) is useful in assessment of likely oscillation amplitudes, and if it is not known safe approximations can usually be made. Typical published values of damping ratios are given in Table 2 below, where only a broad range can be given for each type of structure since the damping is influenced by many factors and design details. It is useful to distinguish between damping in the bare structure and damping in non-structural elements (such as flooring, partition walls, furniture, mechanical services, equipment, etc.). This distinction underlies the Table's differences between general building structures (which usually have a variety of non structural elements), and pedestrian bridges (which usually have almost no non-structural elements).

General principles of dynamic design, Rev. E 1 September 2008

KBR Technical Portfolio

Infrastructure Course 3.1

Table 2 Structure type

Typical published values of damping ratio Damping ratio

General building structures: Continuous steel structure Bolted steel structure Prestressed concrete structure Reinforced concrete structure Lightweight pedestrian bridges: Steel Composite Reinforced concrete Prestressed concrete Miscellaneous: Small diameter piping system Large diameter pipes 0.010.02 0.020.03 0.0030.005 0.0040.008 0.0080.020 0.0050.017 0.020.04 0.040.07 0.020.05 0.040.07

MAGNIFICATION OF AMPLITUDE

As a starting point, it is necessary to calculate the deflection that would be caused by the out-of-balance force ( F ) if it were statically applied. The dynamic displacement amplitude varies from this static deflection as follows:

a =
where

1 (1 r 2 ) 2 + ( 2 Dr ) 2

static

= M d static

r is the ratio of the operating frequency to the undamped natural frequency, D is the damping ratio as already defined, Md is called the dynamic magnification factor.

At low operating frequency, the amplitude is the same as the static deflection, with the structural response being in phase with the applied force. As the frequency of the applied force increases towards the natural frequency of the system, the amplitude becomes greatly magnified. Because of the presence of damping, the peak amplitude occurs at a slightly lesser frequency than the natural frequency. As the frequency of the applied force increases beyond the resonant frequency, the magnification decreases. By about 1.4 times resonance the magnification has dropped below 1. The physical explanation of this is that the oscillating force cannot accelerate the mass quickly enough to achieve the static deflection before the force starts to act in the other direction. As a rough rule, undesirable amplification occurs if the frequency ratio r lies between 0.5 and about 1.3. The upper range requires more careful interpretation. We are not usually looking at the case of a machine whose operating speed is increasing. Usually we are concerned with designing for a fixed, or nominal, operating speed, and then we vary the design of the structure or foundation to change its natural frequency, aiming to move it further away from the operating frequency. If the frequency ratio is low, i.e. below 1, the natural frequency is higher than the applied frequency, so design modifications usually aim to make the structure stiffer. This decreases both the static deflection and the dynamic magnification that applies to

General principles of dynamic design, Rev. E 1 September 2008

KBR Technical Portfolio

Infrastructure Course 3.1

that static deflection. In addition, the machine will not pass through resonance during start-up and close-down. On the other hand, if the frequency ratio is high it means that the natural frequency is low relative to the operating frequency, probably because the stiffness of the structure is also low. Any further reduction in structural stiffness will increase the static deflection, and this increase in static deflection will usually more than offset any benefit obtained from a reduction in dynamic magnification. (It is generally difficult to change the mass significantly, except for foundations as discussed later in this document.) An added problem in the case of a high frequency ratio is that resonance can occur during start-up, or if the machinery is operated below its nominal speed: this factor is sometimes overlooked by a designer given a simple set of criteria to work to. Thus we can conclude that the best means of guarding against excessive amplitude of vibration is to design the supporting structure to have a natural frequency more than double the applied frequency (i.e. r < 0.5). If this cannot be achieved, we should aim for a frequency ratio of 1.5 to 2.0 or higher (1.3 is too close to resonance in view of possible errors, or changes in parameters, given the steepness of the magnification curve). If this latter approach is achieved by reducing stiffness, then the amplitude must be carefully checked: in this context, note that to halve the natural frequency means having to quarter the stiffness, which will lead to four times the static deflection. Although the amount of damping has a significant effect on the amplitude near resonance, it does not have as much effect for frequency ratios in the desirable ranges described above. Hence, whilst it was necessary to discuss its effect, damping can usually be ignored for the purposes of designing the supporting structure. Refer also to Section 8 for comments on how the magnitude of force transmitted to a supporting structure can be reduced by spring mountings and to Section 7 for transmissibility from pad footings to the supporting soil.

Design considerations
STRUCTURAL FORM AND STIFFNESS

As stated in the previous section, the approach to controlling vibrations should always involve first the elimination of resonance and then a check of the predicted amplitude. Since n = (k/M) it can be seen that large changes in stiffness and/or mass are required to shift the natural frequency by significant amounts. It is usually impossible to increase the natural frequency to any extent by reducing the mass, and the primary target is the stiffness. To double the natural frequency requires quadrupling the stiffness. Thus if it is found that resonance occurs with a 460UB67 supporting beam, then a 760UB173 will be required to give satisfactory performance (r < 0.5).

General principles of dynamic design, Rev. E 1 September 2008

KBR Technical Portfolio

Infrastructure Course 3.1

On the other hand, it should be noted that for a central point load = Pl 3/(48EI) and so can be halved by dividing the span by 3 2 = 1.26. Thus if the initial span was 4 m and resonance occurred, satisfactory performance would be achieved by reducing the span to 3.17 m with the same size of beam. Because of the layout of feeders, conveyors, etc, a pair of beams supporting a screen or small crusher are often supported in turn on two other beams running to the columns. This increases the overall deflection considerably, and requires all four to be stiffened. It is apparent from this that the most advantage can be gained at the layout stage, rather than giving the structural designer a committed layout to size up. If the span cannot be shortened and increased beam depth cannot be accommodated, another approach theoretically available is to drop the frequency by reducing stiffness. If this is done simply by reducing the inertia of the beam it is unlikely that it will have enough strength to carry the load. However, increasing the span will increase the deflection at a much faster rate than it increases the stress. This approach can be adopted if the span cannot be shortened, but is not the preferred approach for reasons outlined previously. The last approach available is to move the frequency as far as possible from resonance, and then to introduce sufficient damping in the mounting of the machine. This is a last resort, and requires expert advice. It is not covered in this document.
SECONDARY ELEMENTS AND EFFECTS

So far the discussion has been aimed primarily at the main structural elements supporting the vibrating machine. In mining applications, a vibrating machine is often contained within a larger complex, such as a screen building having several levels, and containing walkways and stairs. It is advisable to keep the structure that supports main items of equipment as separate as possible from other elements. However, sometimes other beams will be connected to the same columns. The forced oscillation of the main supporting beams can be transmitted into the columns as bending moments, particularly through end plate connections, and then back into other beams. In such a case, it is important to check these beams and avoid a resonant natural frequency. The further one goes from the source, the less energy from the vibration will be available to excite the various elements. However, a high energy source can affect quite remote areas, and resonant vibrations of even small amplitude can be very annoying in, for example, a control room where instruments have to be read. Despite attempts at isolation, it is likely that there will be directly connected bracing members which could be affected. Although these have low bending stiffness they also have low mass. This combination means that they can have a high natural frequency, and also vibrate at large amplitude in response to a low energy input. They should therefore be checked and designed to avoid resonance. Again, changes in configuration should be considered, for example -bracing instead of -bracing. If -bracing is added after the dynamic assessment has been performed, the apex should not be at a beam subjected to forced vibration unless the dynamic assessment is re-done. This is because the presence of the apex at the midspan of the beam will change its dynamic characteristics.

General principles of dynamic design, Rev. E 1 September 2008

KBR Technical Portfolio

Infrastructure Course 3.1

ACCEPTABILITY OF VIBRATIONS

The method of calculating vibration amplitudes has been discussed previously. Having arrived at a figure, it is necessary to decide whether it is acceptable or not. For some applications it will be necessary to establish the acceptability level as briefing data or from an equipment supplier. For example, very strict criteria apply to printing presses for colour printing where registration of three superimposed primary colours is important. Guidance on acceptable levels for machines is provided in Figure 1. Figure 2 gives acceptability levels for comfort of persons. However, more comprehensive information is contained in ISO 2631-1. Figure 1 is believed to be based on data to indicate vibrations caused by faults within the machine rather than the effect of external vibrations transmitted to the machine.

Figure 1 Acceptable amplitudes of vibration for rotating machines

General principles of dynamic design, Rev. E 1 September 2008

KBR Technical Portfolio

Infrastructure Course 3.1

Figure 2 General limits for vibrations

In all cases the acceptability of an amplitude is dependent upon frequency, since what we are really concerned about is acceleration and the inertia forces resulting from it. The figures show this relationship. Often criteria are expressed as a limit on velocity. This is an artificial device. One can travel at 1000 km an hour without discomfort. The velocity device is an attempt to express the criterion as a combination of acceleration and amplitude, having the units but not the true meaning of velocity, i.e. one can tolerate a higher acceleration at very low amplitude, or higher amplitude at low acceleration. For most machinery, it is the bearings that determine the level of vibration that can be tolerated. The sensitivity of persons to vibration is somewhat subjective, and depends not only on body-shape, but also on the level of expectation. Relatively large deflections will be accepted on a conveyor gallery without concern, and the acceptance level of vibration on the operating platforms of a crusher station will differ from that tolerated in a control room or sampling laboratory. Figure 2 sets out some guidelines, but judgement must be applied. For the types of structure we are concerned with (bare steel or reinforced concrete), the purely structural consideration is that of fatigue. Since the dynamic magnification factor will have been calculated to determine the vibration amplitude, it is an easy task to apply this magnification factor to the nominal static stresses to arrive at the peak stresses and stress ranges. The allowable stress range is determined by the fabrication details (refer Section 11 of the Steel Structures code AS41001998). Note that a

General principles of dynamic design, Rev. E 1 September 2008

10

KBR Technical Portfolio

Infrastructure Course 3.1

stress range allowable may be as low as 45 MPa. Foundations are a special case discussed later.
VIBRATION OF LIGHTWEIGHT STRUCTURES

Another source of annoying vibrations in addition to vibrations caused by machinery is the low frequency but high amplitude natural vibration of lightweight structures such as conveyor galleries, walkways, and handrails. People walking slowly make about 1.5 footfalls per second, a figure that rises to about 3 footfalls per second for moderate running. Therefore structures having a natural frequency in the range 1.53 cps can be excited accidentally simply by having people crossing them or deliberately shaking them. This can cause great concern. Furthermore, if a group of malfeasers is so inclined, deliberate excitations of up to about 4 cps can be achieved. The structure's natural frequency should therefore be checked, and the range 1.54 cps should be avoided by changing the span or otherwise stiffening (it being usually impractical to change the mass). Whilst the main dynamic effect of walking is a vertical oscillating force at a frequency of 1.53 cps, there is also a smaller lateral oscillating force at half that frequency. Lightweight pedestrian walkways should be checked for this as well. It has been observed that once a perceptible lateral oscillation has been established in a walkway people feel compelled firstly to fall into step with it and secondly to widen their gait, both of which will exacerbate the oscillation. Handrails sometimes give trouble due to inadequate fixity of the stanchions at their bases. Stanchions should be fixed to cross members and not to the longitudinal stringers. The latter are usually very flexible in torsion and will allow large deflections of the handrails to occur.

Guides to specifications
MACHINERY

If the design of the structure is undertaken as a separate assignment from the machinery supply, then it is important to obtain the correct information for the designers. The actual information required will depend upon the nature of the machinery and its mountings, and the enquiry documents must ask for it to be clearly spelled out. The information required by the designer includes:
The geometric layout and particular positioning of mountings. The mass of the machine and all attachments (mounting frame, etc.). Also, the

maximum and minimum additional mass that can be present, such as ore on a screen.
The possible range of operating speeds. Peak speed is not sufficient.

General principles of dynamic design, Rev. E 1 September 2008

11

KBR Technical Portfolio

Infrastructure Course 3.1

The out-of-balance dynamic forces transmitted to the structure (i.e. below the

mountings), and their centroid of action. These forces are required in the vertical and both horizontal directions if applicable, and any rocking effects should be stated (such as force alternating between front and rear mountings). If the machinery supplier cannot provide this force information, then more basic information is required to enable it to be calculated, viz:
The out-of-balance mass, the position of its axis of rotation, and the worst possible

eccentricity of the centroid of the out-of-balance mass from that axis.

The nature of the mountings, and if flexible then their spring constants both

vertically and horizontally (which may be quite different).

If the spring constants of the mountings are known, and also the maximum

amplitude of the oscillation that occurs, then this can serve as an alternative to the information on masses, eccentricities and axes (if that information is not available). Finally, the specification should state that none of the above parameters may be varied without informing the purchaser of the changes and their implications to the dynamic design.
STRUCTURES

If the design of the structure is made the responsibility of the machinery supplier, then the need for and submission of calculations for dynamic design should be specified. The acceptable limits on natural frequency should be specified, i.e. preferably more than twice the highest operating speed or failing that less than half of the lowest speed, also the limits on amplitude selected to suit the operating frequency. The extent to which suppliers are expected to check dynamic response should also be specified to set some limits on their responsibility, e.g. main supporting beams, columns, secondary beams attached thereto, bracing, foundations. The brief for the soil testing program should include reference to the information required for the dynamic design, if foundations for major rotating or vibrating machinery are required. This information is:
Classification of the soil type. The allowable static bearing pressure. The dynamic shear modulus

G (i.e. that which is applicable to short term

deformations).
Poissons ratio .

Design of pad foundations for machinery

The fundamental concepts of dynamic behaviour of foundations are the same as those already discussed for structures generally. The machine and foundation can be regarded as a mass supported on a series of coil springs, and the system has a natural

General principles of dynamic design, Rev. E 1 September 2008

12

KBR Technical Portfolio

Infrastructure Course 3.1

frequency and hence suffers resonance or response to forced vibrations. However, there are obvious differences between a mass supported on a large volume of soil and the case of a point load supported on a steel beam. In this section the discussion on foundations is limited to a stiff pad or block foundation, resting on or partially embedded in the ground. Friction piles and raft foundations are not discussed, and if these are involved specialist advice should be sought. Attempts to assess the dynamic performance of foundations by use of a Finite Element (FE) model are not recommended. This is because of the effects of boundary conditions and the inability of an FE model to include radiation damping which is of great significance. Even a near resonant situation can often be tolerated because the high level of damping available can limit amplitudes to acceptable levels. A spreadsheet has been developed to implement the recommended analysis method. To obtain access to this spreadsheet, see the details at the end of this section.
GENERAL APPROACH TO DYNAMICS OF PAD FOUNDATIONS

Whilst a pad foundation can be thought of as a mass supported on a series of coil springs, and the general expression n =

(k

) still applies, there are several

differences in interpretation between this simplified model and the actual situation. A number of investigators have modelled this system mathematically as a body supported on an elastic half-space. As a result of their investigations, workable solutions have been found in terms of the previously explained parameters (equivalent mass, equivalent spring constant and damping factor), thereby enabling natural frequency, resonance and amplitude magnification to be determined. A foundation supporting an unbalanced machine is potentially subjected to fluctuating loads in the vertical, horizontal and rocking modes, as the horizontal forces will act in a plane well above the foundation base. There could be a rocking (pitching) motion as well, from loads on mountings fluctuating to and fro. These rocking modes can be more significant than the translational modes, and must be handled properly. The text book solutions also deal with a yawing motion (oscillations about a vertical axis), but this is not considered necessary in the scope of the present discussion. A geotechnical engineers advice on soil parameters is needed for all but approximate assessments. These assessments are valuable to gain an understanding of the factors involved, and should be made at the layout stage whilst there is time to influence the general arrangements before they are frozen.
EQUIVALENT MASS

Although a substantial mass of soil would appear to be involved, D.D. Barkan, in his fundamental reference work Dynamics of Bases and Foundations (McGrawHill, 1962), has shown that in practical footings its equivalent effect never exceeds 23% of the mass of the footing, because the amplitude of its vibrations is rapidly dying down with increasing distance from the footing. Thus even to ignore it altogether will only cause an error never exceeding 11% in the calculation of the natural frequency.

General principles of dynamic design, Rev. E 1 September 2008

13

KBR Technical Portfolio

Infrastructure Course 3.1

SPRING CONSTANT

Methods for establishing equivalent spring constants have been developed, with the results differing for each of the modes of vibration. They depend upon the dimensions of the foundation, the elastic properties of the soil, and the length of effective embedment of the foundation in the soil. The procedure is spelled out in full in the book Design of Structures and Foundations for Vibrating Machines, by Suresh Arya, Michael ONeill and George Pincus (Gulf Publishing Company, Houston, 1979). It is quite a complicated procedure to carry out, involving jumps between graphs, tables and formulae. The spreadsheet developed for foundation design (mentioned above) automates it.
SOIL SHEAR MODULUS

The elastic shear modulus of the soil, G , is the value applicable to dynamic vibrations, so it is a short term figure applying to low values of shear strain, rather than the secant modulus that would be derived from large deformation tests. It can be measured from field tests (either a surface oscillator test or a cross-hole test), by laboratory tests (the resonant column test), or deduced from published equations for various types of soil. The method adopted will depend on circumstances and the relative importance of the application. This could be decided after undertaking a trial analysis using published generalised data for various soil types, such as that in Table 3 below. Such data must be regarded as only approximate, but could indicate whether dynamics are a factor.
Table 3 Soil type Soft clay Stiff clay Very stiff clay Medium dense sand* Dense sand* Medium dense gravel* Dense gravel* *at shallow depths Typical soil properties for dynamic calculations Shear modulus (MPa) 20 to 35 70 to 140 >150 35 to 100 70 to 140 100 to 170 140 to 270 0.4 to 0.5 0.3 to 0.4 0.4 to 0.5 0.35 to 0.45 Poissons ratio

For pads founded on sound rock, there is no theoretical dynamic issue. However, pad size must still be adequate to ensure that no tension will develop due to loads which could cause rocking motions.
SOIL POISSONS RATIO

If the Youngs modulus and shear modulus have been determined, Poissons ratio can be calculated from the standard elastic relationship

1 2G

Typical values are given in Table 3 above.

General principles of dynamic design, Rev. E 1 September 2008

14

KBR Technical Portfolio

Infrastructure Course 3.1

EMBEDMENT

Embedment has a significant effect on the damping by the soil, but care must be taken before relying upon its benefits. The footing should be poured directly against the sides of the excavation, which should be in good condition. Alternatively, backfill should be well compacted and not of a type that will shrink away from the footing. It is suggested that even when these conditions are met, the upper 300 mm of embedment should be ignored.
DAMPING

Elastic analysis shows that even if the soil is purely elastic and has no inbuilt material damping of any kind, considerable damping occurs. This is because the supporting soil is of infinite extent, and so vibrational energy is constantly dissipated by being carried away in elastic waves. This form of damping is referred to as geometric or radiation damping. A method for deriving it is described in the book by Arya et. al. (op.cit.), and is automated in the spreadsheet. Of course the soil also has inbuilt material damping. This is expressed as the ratio of the area of the hysteresis loop (the energy lost) to the energy input. For sands and clays, an average value of 0.03 can usually be used for the material damping ratio. The total damping ratio is the arithmetic sum of geometric damping ratio and the material damping ratio. For the translational vibration modes the geometric damping is the more important and the material damping can be disregarded. However, for the rotational modes the material damping is more significant. Note that if a hard stratum of soil or rock exists at a shallow depth below the footing, the value of geometric damping should be reduced: the spreadsheet does not do this automatically, and at the present stage the effect will have to be fiddled by using an appropriate negative value for the material damping.
RESONANCE

Total damping is much more significant for foundations than for steel structures, and so its effect must be considered. Firstly, it lowers the resonant frequency below the undamped natural frequency according to

fr

f n 1 2D 2

However, it also lowers the amplitude magnification factor to such an extent that, whilst the design should seek to avoid resonance, it is permissible to design to be much closer to it. The general rule is to check that the magnification factor is less than 1.5, and that the operating frequency is outside the range 0.81.2 times the resonant frequency (not the undamped natural frequencysee above). There will, however, be times when it is impossible to avoid even this narrowed range. Given the degree of abstraction in the idealised model adopted, and the inherent inaccuracies in the parameter values used, it is considered prudent that if the design lies within this range, the design should be performed as if full resonance had been predicted.

General principles of dynamic design, Rev. E 1 September 2008

15

KBR Technical Portfolio

Infrastructure Course 3.1

TRANSMISSIBILITY

Magnification of displacement amplitude has already been discussed. In a similar way, the dynamic force transmitted by the footing to the soil differs from the applied out-of-balance force.

Ft

F M d 1 + ( 2 Dr ) 2

where the term M d 1 + ( 2 Dr ) 2 is called the transmissibility factor. Here d is the magnification factor which has been previously defined by

Md

1 (1 r ) + ( 2 Dr ) 2
2 2

The soil must be checked for the increase in pressure due to this effect. Texts suggest that the total pressure from static loads plus the transmitted dynamic loads (from all the simultaneous modes of vibration) should not exceed 0.75 times the allowable static bearing pressure given by the soil consultant. The writer believes this can be a misleading concept however, as the allowable pressures are often quoted for small footings. For large footings, whilst the factor of safety against failure rises rapidly, it is the total load rather than pressure that determines settlements, and the allowable pressure recommendations should be made in this light.
RULES OF THUMB FOR TRIAL SIZES

Dynamic analysis can be performed only after a foundation size has been selected. It is appropriate, therefore, to use the following rules to make a trial selection:
Keep in mind these general considerations: A broad flat footing is usually cheaper and more effective than a heavy

deep one.
It is a mistake to believe that a heavier footing will automatically reduce the

displacement amplitude. Additional mass will lower the natural frequency, probably moving it closer to the operating frequency. This increases the magnification factor, and hence the amplitude (which is a function of soil stiffness and footing area, not footing mass). Thus, whilst there is some advantage in setting a footing deeper into the ground, it is the depth of embedment which helps, not the increased mass.
Calculate footing area so that the bearing pressure under the static loads is about

0.5 times the allowable bearing pressure.

Select width to be at least 1 to 1.5 times the height of any laterally applied forces,

measured from the footing base.

Arrange dimensions so that the centroid of the static loads and the centroid of the

dynamic vertical loads are both within 5% of the footing centroid (to avoid tilting due to settlement).
A recognised rule of thumb is to calculate footing mass on basis of 23 times mass

of machine for centrifugal machines, and 35 times mass of machine for reciprocating machines. However, this is frequently misused. Performance is improved by increasing the area of the footing, which rapidly increases its

General principles of dynamic design, Rev. E 1 September 2008

16

KBR Technical Portfolio

Infrastructure Course 3.1

rotational moment of inertia. Increasing the mass by increasing thickness can move the footing closer to resonance. A good design does not have to be so heavy.
Select footing thickness to achieve required mass, but keep it at least one-fifth of

least horizontal dimension or one-tenth long horizontal dimension.

PERFORMING THE DYNAMIC ANALYSIS

The calculations described in this section can be performed most easily by using the specially prepared Excel spreadsheet, which can be found on the Infrastructure/APAC section of the KBRconnect intranet. It is spreadsheet SSN003, in the structural engineering pages. The spreadsheet contains its own specific instructions. It is frequently found that the rocking modes due to fluctuating horizontal forces having a considerable lever arm above the base of the footing are the most critical, rather than vertical modes,

Spring mounted equipment

The principles and formulae set out in this document can be applied to the situation of equipment mounted on springs or special pads. The natural frequency of the upper system can be calculated from the mass of the vibrating machine and the spring constant. (This is the sum of the value of each spring for the vertical mode, and can easily be derived for a rocking mode in terms of restoring moment per radian of rotation.) The natural frequency is deliberately selected to be very low so that the frequency ratio r is high. For springs the damping is close to zero, and so the formula for the transmissibility factor as given above simplifies to

1 . 1 r2
Thus the force transmitted can be reduced to 2% by selecting r to be 7. This reduced force will still be transmitted to the supporting beam with a pulsation equal to the operating speed, and it is still advisable to avoid resonance in those beams since damping in steel beams is low. The system is no longer a one degree of freedom system: it has two, or perhaps even more, independent degrees of freedom, and in general it should be analysed accordingly (see below). However, provided the natural frequency of the supporting beam is well removed from the natural frequency of the spring system, the beam can be considered separately. For this calculation the mass applicable is that directly fastened to the beam, and the spring-supported mass is not included. In summary, the natural frequency of the spring system is selected to be a small fraction of the operating frequency. The spring constant is then calculated from

fn

1 2

k M

General principles of dynamic design, Rev. E 1 September 2008

17

KBR Technical Portfolio

Infrastructure Course 3.1

and the springs are chosen to give this spring constant. Then select the supporting structure to have a natural frequency of (say) double the operating frequency. Finally, check the amplitude of vibration of the supporting beam, i.e. the deflection caused by the out-of-balance force times 1 1 r 2 .

Another reason why the natural frequency of the spring system should be low is that during start-up the system will pass through resonance. The out-of-balance force is proportional to the square of the rotational velocity. Therefore if r is (say) 7, the outof-balance force at the time resonance occurs will be only 2% of its operating value. It is of course necessary to check the deflection of the springs caused by the out-ofbalance force at the operating speed, and ensure that this is acceptable in terms of clearances, connection of electrical cables, air lines, etc. Vibrating screens are usually excited by a pair of coupled but counter-revolving flywheels with out-of-balance weights that are opposite each other in the horizontal position. Thus they cancel in the horizontal direction while adding in the vertical. Nevertheless some horizontal and longitudinal bracing should be provided, and natural frequencies of members should be kept away from the operating frequency. The above approach reduces the force transmitted, but it may be necessary to check that the vibration amplitudes of the machine itself will not be detrimental to it. For example, fuel lines from a separate supply point may need to be flexible.

Systems with two degrees of freedom

The behaviour of systems with two degrees of freedom is not able to be described through explicit algebraic formulae. Such systems need to be modelled numerically, with numeric results being extracted from the model. To aid in this process, an Excel spreadsheet has been developed. This models a generalised two degree of freedom dynamic system under loadings that can include both applied forces and foundation motion, and can be used for:
natural frequencies and mode shapes; harmonic response to harmonic loadings; time-history response to more arbitrary loadings.

The spreadsheet is available on the Infrastructure/APAC section of the KBRconnect intranet. It is spreadsheet SSN002, in the structural engineering pages. The spreadsheet contains its own specific instructions. To get an approximate estimate of the lowest natural frequency for a 2-dof system, you can use Dunkerley's formula (also sometimes known as the Southwell-Dunkerley formula, or the "inverted squares" formula). This formula is

1 f2

1 1 1 + 2 +K + 2 2 f1 f2 fn

General principles of dynamic design, Rev. E 1 September 2008

18

KBR Technical Portfolio

Infrastructure Course 3.1

where f i is defined as the frequency of a modified form of the structure in which all the springs remain present, but all masses except Mi are set to zero Points to note with this formula:
It applies only to lumped mass systems. Thus the structure consists of a heap of

lumped masses, interconnected by a tangle of linear springs.

It always gives an underestimate of the true structural frequency. (This is not to

say it is necessarily inaccurate, just to alert you to the direction any error will take.)
You need to be careful in how you calculate your "component frequencies", and

adhere to the above definition.

Beats with multiple equipment

It should be noted that if a foundation or structure supports two pieces of equipment operating at different speeds, then beating will occur. The frequency of the beats is equal to the difference between the operating frequencies of the two sources, and the peak out-of-balance force is the sum of the two effects (but these will probably be separated by some substantial distance). This is an added requirement to be checked if the natural frequency of the supporting structure or foundation has been made lower than the higher of the two operating frequencies. This situation can arise with a large motor driving another piece of equipment through a gearbox, or if two or more items of variable speed equipment are used.

General principles of dynamic design, Rev. E 1 September 2008

19

KBR Technical Portfolio

Infrastructure Course 3.1

Appendix A Use of general-purpose analysis software

Many of the modern structural analysis programs include at least a limited ability to perform dynamic analysis. This, of course, in no way obviates the need for the designer to understand the general principles described in this document. Such understanding will help the designer to:
create better structural models more quickly; check the sensibility of any results produced from a model; move from the analysis process to the design process. LIMITATIONS OF SUCH PROGRAMS

These programs, as opposed to highly specialised dynamics programs such as Strand7, do not usually offer a full repertoire of options.
Types of analysis: nearly all the programs offer natural frequency and mode shape

extraction. However, not all will follow that up with either time-domain analysis (time history) and/or frequency-domain analysis (spectral analysis). Furthermore, most will not be able to include any non-linear effects, material or geometric, in their dynamic analysis.
Harmonic analysis:

only some programs offer harmonic analysis, where the program directly determines the harmonic response to some form of harmonic stimulus. If you require this but it is not available, then you will have to run a time-history analysis, for a simulated time that is long enough for the transients to have died away. for the structure as a whole described by a single damping ratio. For different damping types, and for the ability to concentrate the damping in discrete locations, specialised programs (and probably specialised advice) are required.

Damping: most programs allow only simple viscous damping, with the damping

CREATING AND ANALYSING YOUR DYNAMIC MODEL

Advice on the creation, running and verification of models for the dynamic behaviour of structures is provided in a companion document in the KBR Technical Portfolio. See "Infrastructure course 3.4: Advanced finite element analysis with the Strand7 software". This companion document, despite its title, has a section that contains a lot of advice that is applicable regardless of the analysis software being used.

General principles of dynamic design, Rev. E 1 September 2008

20

KBR Technical Portfolio

Infrastructure Course 3.1

Appendix B Manual checking of dynamic response

NATURAL FREQUENCY

As stated in Section 4, for a single degree of freedom system, which can be represented as a concentrated mass attached to a mass-less elastic support,

n = 2f n =

k 1 or f n = M 2

k M

It is surprisingly easy to represent a number of real situations by an equivalent simple model (a weight on a spring) with sufficient accuracy. Some examples:
A machine mounted on springs allowing motion in one direction.

The above equation applies directly, where M is the mass of the machine and its platform (spring mass is negligible), and k the total stiffness of the springs. elastically in response to a vertical force, they can be treated as a spring where k is the reciprocal of the deflection caused (to the beam pair) by a unit load. Thus

A machine mounted centrally on a pair of parallel beams. As the beams deflect

k=

48 l3

The mass of the beams is distributed, but it can be shown (see below) that it can be adequately treated as a concentrated mass, having half the value of the distributed mass, and added to the machine mass.
A machine supported on a beam grillage. In a similar manner to above, one can

calculate the deflection caused by a unit force at the machine position (not necessarily central) and apply an equivalent mass.
Tower like structures subject to wind loading. These can still be addressed by

calculating the horizontal deflection caused by a unit force and considering equivalent mass. An approximate calculation can show whether more precise analysis is warranted or not. The methodology for calculating equivalent mass is explained in the following section.
EQUIVALENT SYSTEMS

To consider the response of structures to vibration, it is first necessary to convert them to a mathematical model which will have similar characteristics but will be amenable to analysis. A simple example is illustrated below in Figure B1.

General principles of dynamic design, Rev. E 1 September 2008

21

KBR Technical Portfolio

Infrastructure Course 3.1

Figure B1 Example of an equivalent mathematical model

The system to be examined is a simply supported beam of uniform mass per unit length, carrying at its centre a massive machine subjected to out-of-balance forces. The equivalent model suitable for analysis is a concentrated mass hanging on a massless spring. In this system, the out-of-balance force acts at the concentrated mass, and the response is also required at this point. It is first necessary to determine the beams equivalent mass, modelled as all concentrated at one point. Since the central part of the beam will oscillate with the same amplitude as the machine whilst the other parts of the beam will oscillate with a lesser amplitude, it is obvious that the beams equivalent mass will be less than its total mass. The technique for quantifying this equivalent mass is to equate the kinetic energies of the real and equivalent systems. In the real system the concentrated mass of the machine oscillates with amplitude equal to the maximum deflection, but each element of the distributed load oscillates with an amplitude proportional to the deflected shape. It is sufficiently accurate, but not strictly correct, to assume that this shape is the same as that caused by a concentrated static force of magnitude in the middle of the beam. (It would also be valid to assume that the deflected shape was the same as that caused by a uniform distributed load: the results are surprisingly insensitive to the choice of deflected shape provided the shape is reasonable.) Under this assumption the deflected shape is defined by the following equation (with the origin located at the centre of the beam).

y max

(6lx 2 4 x 3 l 3 ) 48 l 3 = 48

The peak velocity at any point x along the beam will be V x = Cy where the constant C is defined by Vmax = Cy max .

General principles of dynamic design, Rev. E 1 September 2008

22

KBR Technical Portfolio

Infrastructure Course 3.1

By substitution we get

Vx

= Vmax
l/2

6lx 2 4 x 3 l 3 l3

Kinetic energy of the beam = 2

muV x2 dx 2
68 2 mu lVmax 280

With substitution and integration this becomes and by definition this must equal so equiv

1 2 equivVmax 2

17 mu l 35

The total concentrated effective mass e is then the mass of the machine ( ) plus 0.486 times the total uniformly distributed mass (mul). This type of calculation can be performed for different beam support conditions. It can be shown that for a fully fixed beam the equivalent mass is 0.371 times the uniformly distributed mass. For a partially fixed beam a value between these two could be selected, say 0.45 for a nominally simply supported beam and 0.40 for a nominally fixed beam such as an encastr concrete beam. The equivalent kinetic energy approach can also be used for masses concentrated at points other than the centre. Table B1 lists the results of some such calculations. It also gives the factor by which the total dynamic loads should be multiplied to make them equivalent to a single centrally applied force. As described above, for a point load on a simply supported beam (or pair of parallel beams) the equivalent spring constant is equal to the force required to produce unit deflection. Therefore, since

y max

48 l 3 = we get k e = = 3 y 48 l

This calculation can also be performed for other cases, or extracted in numerical terms from the deflection calculations for the supporting structure. For example, in the case of a beam supported on other beams, the relevant deflection is the sum of the deflection within the inner beam and the deflection of the supporting beam that occurs when a load is applied to the point under consideration. The spring constant is then the load that is required to cause a total deflection of one unit.

General principles of dynamic design, Rev. E 1 September 2008

23

KBR Technical Portfolio

Infrastructure Course 3.1

Table B1

Equivalent mass and force for various configurations Equivalent concentrated dynamic force 1.00 F Equivalent concentrated mass Lumped mass Uniformly at load points distributed mass 1.00 M 0.49 mul

Case Simply supported beam, uniformly distributed mass mu per unit length, dynamic force F applied at midpoint, lumped mass M at midpoint. Simply supported beam, uniformly distributed mass mu per unit length, dynamic force F/2 applied at thirdpoints, lumped mass M/2 at third-quarter points. Simply supported beam, uniformly distributed mass mu per unit length, uniformly distributed dynamic force fu per unit length. Fixed ended beam, uniformly distributed mass mu per unit length, dynamic force F applied at midpoint, lumped mass M at midpoint. Fixed ended beam, uniformly distributed mass mu per unit length, uniformly distributed dynamic force fu per unit length.

Spring constant 48EI/l3

0.87 F

0.76 M

0.52 mul

49.1EI/l3

0.64 ful

0.50 mul

49.2EI/l3

1.00 F

1.00 M

0.37 mul

192EI/l3

0.53 ful

0.41 mul

203.5EI/l3

General principles of dynamic design, Rev. E 1 September 2008

24

KBR Technical Portfolio

Infrastructure Course 3.1

Appendix C Dynamic analysis and design for mining and other plant
DISTINGUISHING CHARACTERISTICS

Whilst the approach to dynamic design of structures (including foundations) supporting machinery in various types of plant is common for each of a few generic machine types, there is a big difference in the quantum of the problem when addressing some types of plant in the mining industry. In most types of plant and in bridges or tall towers the dynamic loads are an unwanted or parasitic side effect and efforts are made to eliminate or reduce the loads. However, crushers and screens in mining plant require dynamic loads for their fundamental operationthe larger the dynamic load, the more effective the plant. Handling these dynamic loads becomes the principal structural consideration, not a mere side effect. The main categories of equipment, classed in accordance with the nature of the dynamic loads are:
rotating machinery (rotary crushers, motors, turbines, pumps, rotary engines) reciprocating machinery (diesel engines) combined action equipment (screens). BRIEFING CRITERIA Loading

For most types of rotating machines manufacturers strive to produce machinery as balanced as possible, i.e. so that the axis of rotationthe centre of the bearingsand the axis of rotational inertia, or centre of mass, coincide. The required briefing data is the mass of the rotating part and its possible eccentricity. This is the initial tolerance plus an allowance for later wear of the bearings or shafts. This information should be sought from the vendor. Some guidelines are provided in Figure 1. Most rotary machines have a horizontal axis, so the dynamic force produces a vertical component pulsating sinusoidally, and a horizontal component (at right angles to the axis of rotation). As the axis is above the level of the supports (or base of foundation) this horizontal component creates a rocking motion. This can be the critical case for foundation design. The situation with reciprocating machinery is more complex, and detailed vendor supplied data must be sought listing out of balance forces and their directions. At the very least, the designer must be aware that the motion of the reciprocator is not harmonic, and so there will be dynamic force applied at frequencies higher than the fundamental operating frequency. These forces must be considered. Screens in mining plant are vibrated as the result of an eccentric flywheel. They are spring mounted, so the forces transmitted to the supporting structure, both vertical and longitudinal, are not the same as the primary effect of the flywheel eccentricity which is acting on the equipment above the springs. These forces could be calculated

General principles of dynamic design, Rev. E 1 September 2008

25

KBR Technical Portfolio

Infrastructure Course 3.1

using the principles set out in the main document, but are usually supplied by the manufacturer. It is necessary to establish the normal or maximum operating speed and also the possible range of speeds, particularly if an under-tuned solution is adopted. The static mass of the machine as a whole is also needed for the natural frequency assessment.
ACCEPTABILITY CRITERIA

As stated in the reference document, the acceptability criteria are a composite of amplitude and acceleration, but sometimes given as a velocity. The acceptable level is dependant on where, in the plant as a whole, the vibrations are being felt. An international standard, ISO 2631-1, addresses the effect on persons as related to the time of exposure. A much lower level is acceptable at a continuously staffed operator position than the level acceptable for a point which may be visited only for short periods. The acceptability level for an item of equipment can be obtained from the manufacturers but this criterion is sometimes misunderstood or wrongly quoted. For example, if a motor exhibited vibration amplitudes of 0.1 mm, due to its own operation, this would be seen as a sign of deteriorating bearings or worn shaft and deemed unacceptable. But a motor mounted on a platform which is vibrating in response to a crusher or screen could be subject to this vibration amplitude without any damage being caused. To demonstrate the point; if amplitude is 0.1 mm at a frequency of 900 RPM the peak acceleration is a 2 , i.e. 0.9 m/s2, i.e. the load felt by the bearing would be 10% higher than the static weight of the rotating parts onlyan insignificant increase. Refer to the explanation in the main reference document Section 5.3.
MITIGATION

The methods of analysis of response to vibration sources are addressed in the main document, as are suggestions for mitigation.
Mining plant

In mining industry plant the main issues are:

Vertical vibration of beams directly supporting machines; these vibrations can also

be transmitted through connections to columns, other beams, platforms, etc. The objective is to not only ensure that resonance does not occur, but also to limit the dynamic magnification. The available methods are to increase the support stiffness to about double that corresponding to resonance. Stiffness is most effectively increased by reducing the span of beams and/or re-configuring supports to avoid a grillage situation. To reduce the beam stiffness and hence magnification is not usually practical, but for some types of machines the transmissibility of force can be reduced by the use of resilient mounts. This can be done for large screens as used in coal plants by double spring mounting as follows. A screen and its exciting eccentric flywheel

General principles of dynamic design, Rev. E 1 September 2008

26

KBR Technical Portfolio

Infrastructure Course 3.1

and motor are always spring mounted. It is possible to spring mount the screen on a frame and then provide further spring mounts between that frame and the main structure. See Section 8 of the main reference document.
Vibration of secondary elements:

platforms and handrails often vibrate in resonance (sometimes in higher modes); this is often due to flexible fixing details so ensure stanchions are as rigidly fixed as possible. Usually caused by incomplete bracing systems. However, do not ignore the possibility of providing an undertuned solution if additional stiffness is impractical.

Horizontal vibration of multilevel frames (particularly in screen buildings).

Vibration of foundations for crushers: the main issue will be the rocking motions

due to horizontal loads at a significant height above the underside of the foundation pad (which can be aggravated if the pulsating vertical dynamic load is off centre). These are best mitigated by increasing the moment of area of the footing/soil interface. Increasing the length by 26% (in the relevant direction) will double the stiffness in this mode.
TURBINE FOUNDATIONS

Gas turbines are very high speed machines. They are in consequence manufactured to high standards and well balanced, but with increasing wear significant forces can develop due to the squared effect of speed. It is usually impractical to provide a foundation with a natural frequency higher than the turbine operating speed, (except perhaps if founding on strong rock) so an undertuned solution is used which can reduce the amplitude below the static value. This is done by having a deep foundation block, i.e. large mass.
PRINTING PRESSES

Colour printing presses are very sensitive. Full colour is achieved by superimposing three separate layers of the primary colours. Vibrations cause lack of registration with poor definition and edges showing the primary colours. The out of balance forces are caused by large, rapidly rotating reels of paper which develop eccentricity relative to the shafts on which they relate (they oval), and the machines are relatively high so rocking motions can occur. It may be necessary to add piles under the foundation or bored piers if rock is at reasonable depth.
THE MYTH OF ISOLATION OF FOUNDATIONS

Engineers sometimes seek to isolate the foundation of a machine from an adjoining foundationeither another machine or an adjacent structure. This is based on a misconception. If a machines foundation pad is vibrating significantly the movement is not within the pad itself but in the ground below it. The zone of influence is significant and a purely nominal separation (e.g. a joint) will not prevent significant movements being caused in the adjacent foundation. However, if the pads were monolithic, it is likely that the overall stiffness would be greatly increased and soil movement reduced. This option should therefore be considered.

General principles of dynamic design, Rev. E 1 September 2008

27

KBR Technical Portfolio

Infrastructure Course 3.1

Appendix D Dynamic design issues for footbridges

The design of footbridges is the subject of its own document in the KBR Technical Portfolio. Material that was previously in this Appendix has been merged into that document. See "Infrastructure course 2.3: Elemental Reference Design for footbridges".

General principles of dynamic design, Rev. E 1 September 2008

28

KBR Technical Portfolio

Infrastructure Course 3.1

Appendix E Wind induced dynamics

DISTINGUISHING CHARACTERISTICS

The distinguishing issue in the analysis of structures for possible dynamic sensitivity to wind is the complexity of the loading and the range of wind speeds at which exciting forces can occur. For tall towerlike structures there are two basic, separately caused forms of excitationalong wind and across wind. The former is caused simply by the fluctuation of gust velocities. A simplified explanation is as follows. Suppose the wind was blowing at a fixed speed, causing your structure to deflect; then it stops blowing or blows at a slower speed; the structure will spring back, wholly or part-way, and oscillate at its natural frequency; while still doing so, a wind gust of increased speed hits it and the process goes on. Cross wind excitation is quite different. All structures create downstream eddies. In the case of some structures, particularly of circular shape in plan such as chimneys, eddies break off the structure alternating from side to side, setting up lateral exciting forces. If these coincide with, or are close to, the lateral natural frequency a resonant response occurs. The lateral oscillations of the structure reinforce the eddy shedding pattern. However, in practice both effects can combine, resulting in elliptical (and even figure-eight) motions. Free edged roofs of buildings like sports stadia and similar structures can be subject to flutter due to eddy effects. Expert advice and/or wind tunnel testing of aerodynamic aeroelastic models is needed.
BRIEFING AND CRITERIA

There are no issues to be provided by the client, other than possibly the amplitude of vibration that lights or delicate equipment (such as communications antennae or dishes) can tolerate. Usually the client cannot provide such criteria as part of the brief and the designer has to research it from equipment manufacturers and interpret it. For tower-like structures (including multistorey buildings) the structural issue is that along-wind oscillations magnify the base moments and shear forces. The normal structural design rules then apply to the magnified moments. However, this effect is not very great for the reason explained in Section E3, Analysis. Cross-wind excitation is often more of a nuisance than a cause of structural distress.
ANALYSIS

The principles of dynamic analysis are as set out in the main document, and some of its practicalities are discussed in the companion document "Infrastructure course 3.4: Advanced finite element analysis with the Strand7 software". AS 1170, Part 2 Section 6 deals with the dynamic loading effects in wind. The methods set out include equations whose derivation is far from obvious. It is essential

General principles of dynamic design, Rev. E 1 September 2008

29

KBR Technical Portfolio

Infrastructure Course 3.1

to read section C6 of the Commentary to AS 1170 Part 2 in parallel with the code methodology. The code states, in effect, that the overall load and bending moments will not be increased above the value obtained by the routine quasi static methods of Sections 1 to 5 of the code if the first mode natural frequency is greater then 1 Hz. It is not clear whether this statement applies to cross wind effects as well as along wind effects. Whilst the load and stress effects may not be significant, the statement should not be taken to mean that dynamic deflections should not be considered if a tower like structure is supporting sensitive equipment. If in doubt, seek advice from the Monash University Mechanical Engineering Department, which specialises in wind dynamics and will carry out model testing if required.
MITIGATION

For tower-like structures, if the assessments indicate unsatisfactory performance, the obvious solution is to increase the stiffness if practicable. However, this is not always possible and another option is to increase damping. This is effective for wind induced vibrations and quite modest increases in damping often suffice. Very tall buildings, observation towers and the like have sometimes been fitted with large mass dampers. Cross wind excitation of steel chimneys is frequently dealt with by the use of spiral strakes around the chimney which prevent the eddy pattern from forming. Cross wind excitation of tubular diagonal braces on a container crane were damped out by placing a substantial chain inside the tubethe friction between links and between the chain and the tube provided the necessary damping.

General principles of dynamic design, Rev. E 1 September 2008

30

KBR Technical Portfolio

Infrastructure Course 3.1

Appendix F Dynamic loads imposed by pneumatically-tyred vehicles

INTRODUCTION

This section is intended to provide guidance to engineers in assessing the likely dynamic loads imposed on supporting structures by the passage of vehicles with pneumatic tyres. Simplified formulae are presented for two situations:
a single discontinuity in the road surface; continuous sinusoidal corrugations in the road surface.

Whilst it is not directly applicable to the subject of short-term dynamic forces, it is worth pointing out that significant vehicle loadings occur when the vehicle brakes. Obviously, horizontal forces are generated, via friction between the tyres and the road surface. Less obviously, the front wheels impose additional vertical forces since the weight of the vehicle tends to pitch forward. Less obviously again, the effective stiffness of the front suspension might be increased by this increased force, since many suspension systems exhibit deliberate non-linear behaviour.
SINGLE DISCONTINUITY Basic model

The road surface is assumed to consist of two semi-infinite level portions whose vertical separation is H . The vehicle is travelling along the road at a steady speed such that a wheel will traverse the discontinuity in time T . The vehicle suspension system consists of an axle of total mass m supported on a number of tyres whose total stiffness is k t . The axle supports a portion M of the total vehicle mass and load, via a number of springs of total stiffness k s . In some instances there is also a hydraulic arrangement between the axle and the body whose effect is to impose a constant force Ps : the purpose of this is to provide a smoother ride by allowing softer springs to be used. applicable to the springs and tyres respectively. Damping ratios of Ds and Dt are

General principles of dynamic design, Rev. E 1 September 2008

31

KBR Technical Portfolio

Infrastructure Course 3.1

Diagrammatically we have

and

Analysis

The above system has two degrees of freedom. To reduce it to a single degree of freedom we assume that over the period of interest the vehicle body does not move vertically. This assumption is usually acceptable, because generally M is considerably greater than m , and k t is considerably greater than k s , so the axle responds to a disturbance much more rapidly than does the vehicle body. The damped natural frequency (in radian/sec) of the axle under these assumptions is

where is the undamped natural frequency.

k s + kt 1 ( Ds + Dt ) 2 m

= 1 ( Ds + Dt ) 2

General principles of dynamic design, Rev. E 1 September 2008

32

KBR Technical Portfolio

Infrastructure Course 3.1

If damping is ignored, it can be shown that the effect of the disturbance is to superimpose on the static force between the tyres and the road, a dynamic force that oscillates between the two values

kt H (k s k t ) k s + kt
where is a function of T , and is given by

2 sin ( d T 2 ) for d T < d T

1

and

2 for d T d T

0 0 1 2 3 4

dT/
Similarly, it can be shown that the (total) static force in the suspension springs has superimposed upon it a dynamic force that oscillates between the two values

k s kt H (1 ) ks + k
Discussion

Several key assumptions are made above, some explicitly and some implicitly.
The vehicle body responds relatively slowly to disturbances. The validity of this

can be gauged by calculating the natural frequency of the body, which is given by

k s kt (k s + k t )M
and comparing the result with the natural frequency of the axle.
The road surface, which might be a bridge deck, does not deflect significantly

under the action of either the static or dynamic loads. The likely effect of the dynamic loading on the supporting structure can be assessed using the methods described earlier in this document, and then the validity of this assumption can be assessed by comparing the support deflection with H .
The discontinuity is a single change in the elevation of the road surface. If the

surface subsequently returns to its original level, that constitutes a second

General principles of dynamic design, Rev. E 1 September 2008

33

KBR Technical Portfolio

Infrastructure Course 3.1

discontinuity whose effects will be additive to those of the first (having due regard for the sign of H in the formulae).
The axle traverses the discontinuity in time T . The estimate for the value of T is

crucial to the entire exercise, and it will be under-estimated if obtained merely by dividing the horizontal extent of the discontinuity by the horizontal speed of the vehicle. The vehicle has pneumatic tyres, which will tend to flow over the discontinuity (the caterpillar effect). Approximate allowance for this can be made by adding half the length of the tyre footprint to the horizontal extent of the discontinuity before dividing by the speed.
Damping is ignored except insofar as it affects the value of

d . This has the

theoretical effect that the independent vibrations set up at the start and the end of the discontinuity, do not die away, but continue forever. In practice they will die away within a few cycles. The greater the value of d T the more the above simplified approach will tend to over-estimate the dynamic forces. For small values of d T the error is insignificant, and this is the case that is usually of interest because it gives rise to the larger dynamic forces. (As an aside, for very large values of d T the simplified approach gives answers that are too large by a factor of two. The physical reason for this is that the vibrations set up at the start of the discontinuity will actually have died away completely when the end of the discontinuity is reached.)
CONTINUOUS CORRUGATIONS Basic model

The vehicle, its axle and its tyres are idealised in the same manner as before. The road surface elevation is assumed to be sinusoidal, with amplitude H and a wavelength such that the angular frequency of the disturbance imposed on the vehicle is (radian/sec).
Analysis

As in the case of the single discontinuity, the problem is reduced to a single degree of freedom by assuming that the vehicle body does not move vertically. Under this assumption, and with much algebraic manipulation, the following formulae can be derived. Let

=
r =

D = Ds + Dt

k s + kt m

(undamped natural frequency of axle) (frequency ratio) (total damping ratio)

General principles of dynamic design, Rev. E 1 September 2008

34

KBR Technical Portfolio

Infrastructure Course 3.1

Then:
The amplitude of vibration of the axle is given by

kt k +k t H s 2 1 r

2 + (2 Dt r ) 2 2 + (2 Dr )

The dynamic force applied to the road surface has an amplitude given by

Hk t

kt 2 1 r k + k s t

(1 r ) + (2Dr )
2 2

2 k s + kt 2 1 + + ( ) 2 D r s 2 Dt r k t 2

which, if the spring stiffness is very much less than the tyre stiffness, simplifies to

Hk t r 2

2 Ds 2 2 1 + (2 Dt r ) 1 + r 2 2 1 r 2 + (2 Dr )

The dynamic force applied by the suspension to the vehicle body has an amplitude

given by

Hk s

k t k +k t s

2 k s + kt 2 1 + ( ) + 2 D r t 2 Ds r k s

(1 r ) + (2Dr )
2 2

Discussion

Most of the assumptions made for the analysis of a single discontinuity apply in the analysis of continuous corrugations. However, damping has been included. A feature of corrugation-induced vibrations is that they are usually imposed at a frequency close to the natural frequency of the suspension system. This is because the corrugations are themselves caused by resonance in the suspensions of the larger vehicles which travel the road. The corrugation profile will probably not be sinusoidal. The actual profile could be measured, then analysed into its Fourier components. This would almost certainly show that the fundamental component was the predominant one, a predominance that is enhanced by the fact that this fundamental component will probably be exciting the system at close to its natural frequency. The caterpillar effect of the tyres footprint will serve to lessen the effective value of the corrugation amplitude. The likely magnitude of this can be estimated by drawing a full corrugation to scale, then superimposing a type footprint on that drawing. (When doing this, remember that the actual profile of the corrugations will not be sinusoidal.)

General principles of dynamic design, Rev. E 1 September 2008

35

KBR Technical Portfolio

Infrastructure Course 3.1

TYRE STIFFNESS

In cases where the tyre stiffness is not directly available, an estimate can be made from knowledge of the tyres radius and operating pressure. Consider a single tyre of radius R and internal air pressure p . The tyre has been flattened by an amount , acquiring in the process a footprint length L . Assume, in the absence of any better information, that the footprint width is L 2 .

The footprint area is 0.5 L2 , and so the load being supported is 0.5 pL2 . The values of L and are related through the geometry of a circle by (2 R ) = L2 4 , which for small deflections simplifies to = L2 (8R ) . Hence, the load being supported is equal to 4 pR , and so the stiffness of the single tyre is equal to 4 pR .

Document revision history

Version Description Initial paper-based document produced. Various bits and pieces added in loose-leaf form over the years. Converted to electronic format. More bits & pieces added over the years. D E Becomes part of KBR Technical Portfolio. Cosmetic improvements. Dunkerley's method described. 2006 Sep 2008 Tony Dawson Rob Niall Rob Niall ~1998 Rob Niall Release date 1981 Released by Tony Dawson

General principles of dynamic design, Rev. E 1 September 2008

36