The Dynamic Principles of Machine Foundations and Ground

512
The Dynamic Principles of Machine Foundations and

Ground
By J. H. A. Crockett, B.Sc. (Eng.)*, and R. E. R. Hammondf.
Recent research on soil dynamics and machine foundations has revealed that there are correct design
principles for vibrating foundations which can be used without dependence on empirical methods.
The most important discovery has been that any ground can oscillate at its own particular natural
frequency, and machines and ground quite often vibrate in resonance. Stress waves, some with
similar properties to light and sound waves, are of considerable and often of paramount importance,
for their interaction can lead to structural failure.
The authors have considered only steadily running machines and have merely indicated the
mathematics, the designer being advised to consult the proper engineering text-books for a fuller
account. They hope that the problem has not been over-simplified and that their conclusions may
form a sound basis for future design practice.
P R I N C I P L E S OF BEHAVIOUR OF V I B R A T I N G Table 1 shows some measured self-frequencies and corre-

FOUNDATIONS sponding bearing values for various soils. All measurements were
There can be no “static equivalent” corresponding to oscilla- taken either by using variable speed oscillators or from oscilla-
tions or stress waves, and the authors therefore postulate the tions produced by striking with a heavy weight and distributing
following two principles :- the force over a big enough area of ground to prevent plastic
(I) Oscillations. The machine, foundation, and adjacent settlement :these areas varied from one or two square feet up to
ground all act together as a system of oscillating springs and nearly a hundred square feet. Andrews and Crockett (1945)s
weights; each weight acts as a damped rigid body and oscillates in have measured resonant frequencies from machine foundations
three translational and three rotational degrees of freedom. up to a maximum size of about 2,500 sq. ft., but in all cases about
(2) Stress Wawes. Where the imposed oscillations have the same self-frequencies for the same type of soil have been
wavelengths nearly similar to, or smaller than, the dimensions noted; these workers also found that a range of harmonics and
of the spring-weight system, such members do not act in even “sub-harmonics” could exist.
oscillation; stress waves are the dominant modes.
It is suggested that successful design of machines and founda- Spring-Weight Equivalent of Ground and Foundations. A
tions properly depend upon complete collaboration between weight W,mounted on top of a spring of stiffness K2,as in
civil and mechanical engineers. Fig. 1, can be oscillated by a sinusoidal impulse of Wrl sin 2?rNt,
ACTION O F T H E G R O U N D
Until recently, only limited study has been devoted to the
behaviour of ground under dynamic influences. Seismologists
and geophysicists are mainly concerned with ground movements
relatively remote from the source of disturbance; this paper deals
only with ground behaviour at the source.
From 1928 to 1939 Degebo and the University of Gottingen
in Germany, using controlled oscillators, examined extensively
the natural frequency of vibration peculiar to sites and various
types of soil; many of their results have since been substantiated
by Andrews and Crockett (1945)$ who independently discovered
and studied the same phenomena, through resonance between
industrial plant and adjacent ground.
Ground Self-fregguencies. The works of Degebo (1936), Lamb
(1904), Leet (1946), Rayleigh (1885), and Sezawa and Kanai
(1937) have all dealt with the subject of ground self-frequencies;
it has been found that the self-frequency of vibration of ground
is a function of the physical properties of the material, and
dependent on the thickness of the strata immediately adjacent
to the place of disturbance. The bearing strength of ground is Fig. 1. Spring-Weight System
also directly related to the self-frequency, and Bergstrom and
Linderholm (1946) showed that it provided a practical method as by a piston, where W lis the weight which will produce a
of measuring bearing values of soil for ordinary building sites. static elongation a in the spring. If is the natural frequency
of vibration of the spring-weight system, the steady state of
The MS. of this paper was originally received at the Institution on motion of the weight is
12th September 1947, and in its revised form, as accepted by the no2
Council for publication, on 2nd July 1948. For the Minutes of
Proceedings of the meeting in London on 8th April 1949,at which this
110,-N2asin2rrNt . . . . .
paper was presented, see Proc. I.Mech.E., 1949, vol. 160, p. 259. which is the vertical displacement at any time t in the cycle,
* Consulting Engineer. assuming that there is no eccentricity in the system. Resonance
t Technical writer! formerly Temporary Assistant Engineer,
*
Grade 1, Chief Scientlfic Adviser’s Division, Ministry of Works.
An alphabetical list of references is given in Appendix 11.
§ Andrews and Crockett were assisted in this work by J. Duf€, B.Sc.,
and G. E. Walker.
Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

THE DYNAMIC PRINCIPLES OF MACHINE FOUNDATIONS AND G R O U N D 513
TABU1. FREQUENCY
AND BEARING OF SOILS
VALUES occurs when n, equals N, although in practice, damping takes
place.
Type of soil ’requenq idmissiblc Velocity of The stiffness K2 is assumed to be linear for this spring, but
cycles per pressure, Rayleigh with materials like cork and rubber the load-deflexion relation
second tons per wave, ft. is non-linear (Minorsky 1943) ;for such cases the resonance peak
sq. ft. per sec. at loses its symmetry, and bends over, as indicated in Fig. 2, to
20-25 C.P.S. either side. This figure shows a resonance curve from a soil
Valuesfrom Degebo (1936,001.4) (Hertwig, Fruh, and Lorenz 1933) in which the peak is at 23-12
Marsh 10 feet thick on sand . 4 262 cycles per second; if the amplitude induced at resonance had
Marsh 5 feet thick on sand 13.1 - been much smaller, this resonance frequency would have been
~ m h stratum
y (5 feet) on mu;! 15* - higher-about 28 cycles per second-but if it had been much
Loose sand agglomeration . 17.4 -
- larger than that shown, the resonating frequency would have
Very wet softened day . 18.4 been lower. Although Degebo (1933) do not appear to have
Wet clay . . 19.2 - noticed this non-linear phenomenon, their published figures
Finestsand . 19.3 361
Clay with fine sand’
Moist clay .
:. .
20.7
208 -459
demonstrate it, and their results have been substantiated by
Crockett and Hammond (1948). This non-linearity is very
Coarse gravel, vibrator-cfm- important when the self-frequency of a site or soil is stated;
pacted- , 21.2 - the amplitude of induced movement during the measure-
Very wet sand above spring . 21.6 - ment should also be stated. As this has not been done for the
Moist tertiary clay. . 21.8 427 values shown in Table 1, the higher frequencies which are
Moist medium sand . 21.8 459 based on small amplitude measuremmts have been marked with
Dry medium sand . . 22 525 an asterisk; the other values are lower than they might have
Clayey sand over marl rubble . 22.6 558 been. In practice, frequencies do not vary by more than about
Loose sand . 22.8 -
Medium-fine sand, stamped . 23 - 25 per cent below their highest values for small amplitudes of
Dry Loess of diluvial origin . 23.5 854 movement; such variations may not amount to much from a
Gravel with stones . 23.5 59 1 practical standpoint, but they do stress the guiding principle
Moist clay . 23.5 624 that the designer should always consider the self-frequency of
Medium sand, not compacted . 23.7 - the ground for any site.
Marl rubble . 23.8 624
Elutriated, stamped,mediumsanc 24.3 -
Dry clay 24.6* -
Dry clay with broken limeston; 25.3 657
Firmsurfaceclay . . . 25.5* -
Marl . 25.7 722
Berlinsand . 27.5 - 2
Very firm clay 31. fhet de;p
Very firm sandy gravel, 16 feet
: 28*5* - 2.
-I
GROUND NON-LINEAR
(CURVED AXIS)
above spring water . 29 - t<
Fine tertiary sand, 31. feet deep 29.5* -
Tightly packed coarse gravel .
Sharp compact tertiary sand,
30 1,379 8 SPRiNG LINEAR
n
3) feet deep . 30.2 - z
(VERTICAL AXIS)
Weathered mottled sandstone . 32 1,641

Very uniform medium sand . 33.4* -
Values from Lormz (1934)
Peat 6 feet thick, overlying sand 12.5 -
Medium sand, with peat rem- I
nants, old, 6 feet thick . 19.1 - 0
I
5
I
10
I
15
f
20 25 ,b 40
Gravelly sand, with clay lenses 19.4 - INDUCING FREQUENCY-CYCLES PER SEC.
Dense old slag fill, well com-
pacted by traffic. . 21.3 - Fig. 2. Non-linear and Linear Resonance
Very old well compacted fill of
loamysand . 21.7 - The relation between phase and induced frequency is also
Tertiary clay, moist . 21.8 - important, because it demonstrates still further the analogy
Lias clay, moist . 23.8 - between a foundation on the ground and a single spring-weight
Fairly dense medium sand 24.1 - system. When N is less than no, the displacement is in phase
Fine, with 30% medium sand. 24.2 624
Uniform coarse sand . 26.2 - with or in the same direction as the inducing force; at higher
Very dense mixed grain sand . 26.7 - inducing frequencies than the force and the movement are
Quite dry tertiary clay . 27.5 - anti-phased or opposite, so that when the force is downwards
Dense pea gravel . 28.1 - the ground moves upwards, and vice versa.
Limestone, undisturbed . 30 - This change of phase means that extra power is needed to
Sandstone, undisturbed. . . 34 - overcome the effect of resonance with the ground. Curve (a) in
Values from Andrews and Fig. 3 represents the power required by a certain well-balanced
Crocken (1945-7) machine (Crockett and Hammond 1948, and Hertwig, Fruh,
. . - and Lorenz 1933) to keep it idling at various speeds; curve (b)
Peat
Waterlogged es&e silt : 7.5
10 -
-
gives the extra power required to overcome the additional eiction
Verylightsoftclay , . 12
- in the bearings when the machine is put out of balance at various
Lightwaterloggedsand
Merllumday . . .
.. 15
15 - speeds. Curve (c) shows the large amount of extra power needed
. - to induce oscillations in the ground in the zone of resonance,-and
Hard peat and sand layers
stiff clay . .
17
19 -
-
also at higher frequencies, to maintain the ground in oscillatlon;
this latter movement is in the opposite direction or anti-phased
Silt and sand‘mixei . 23.3
- to that of the machine. The machine has to waste energy when
Sand &rubbleloosely compacted 235
Limestone . - it forces the ground to oscillate in this upper range, but only
Granite .
30
40 - limited power is required below the resonance frequency. Energy
loss around resonance is directly related to induced amplitude,
Degebo’s frequencies and bearing pressures were measured with as indicated by the subsidiary diagram.
oscillators, but Andrews and Crockett’s with a vibrograph, by setting
the ground into oscillation by striking it with a hammer; their Relation Between the Active Wezght, the Self-frequency,
and the
admissible bearing pressures are “engineers’ estimates” for the sites. Spring Constwt of a Foundation System. Available information
33
514 T H E D Y N A M I C P R I N C I P L E S OF M A C H I N E F O U N D A T I O N S AND G R O U N D
is Scarce and inconclusive, but the authors give a brief summary bulb, because the system can oscillate at a specific self-frequency.
and suggest a practical working method within present industrial This is probably why the most stable self-frequency of all these
limitations : foundations for any particular type of ground is substantially
(a) Acrive Wkght. There is a portion of ground of weight similar, even in spite of considerable variation in overall size,
W, which is immediately below and acts with the foundation from a foot or two square on the surface to deep and heavy
weight Wf. Since this active ground weight is also springy, the foundations for industrial machinery. This action has been
system is not that shown in Fig. 4u, but rather that of the called (Crockett and Hammond 1948) the ‘‘oscillation of the
rigid weight Wfas depicted in Fig. 4b supported by a spring bulb of pressure”.
which has itselfa weight W,(Crockett and Hammond 1948, and The further assumption is that the inner compression zone
of the bulb of pressure can or may represent W’;this has given
I successful results in practice. Fig. 5c shows the necessary bulb
i POWERTO modifications when the foundation is partly below ground level

WHEN (Andrews and Crockett 1945b), the active ground weight being
b POWER TO BEARINGS
WHEN UNBALANCED
Et Lz POWER FOR IDLING

WHEN BALANCED
L
c
d
z INDICATION OF
HOOP STRESSING
NCIPAL LINES OF STRESS
I
r I ,
0 10 20 30 40
INDUCING FREQUENCY-CYCLES PER SEC.
Fig. 3. Loss of Machine Power to Ground, Particularly at

Resonance b
Terzaghi 1943). This active ground weight is some function of

the bulb of pressure, but the exact dynamic properties of this
are unknown.
Hertz and Fuchs (Love 1944)have shown that under a circular
area of contact between one body and another there is an ellip-
soidal zone, adjacent to the area, where all principal lines of stress
are in compression. Outside this zone, some of the lines change
to tension, while lines of “hoop stressing” also occur around the
14;
Fig. 5. Proposed Method of Obtaining the Active Ground

Weight from the Bulb of Pressure
/
4 /
L’//// /i= /
,/’//
,
,/////////
//considerably increased by friction around the sides. Total weight
n b is not entirely dependent upon external dimensions of the con-
Crete; the foundation replaces some of the weight of the soil, and
Fig. 4. Effect of Active Ground Weight the concrete base itself may be of lightweight cellular form.
Fig. 5d represents a large block type of foundation taken
outside of this inner ellipsoidal compressed portion, as indicated down to a “good bottom”, where at times the upper parts of
by Fig. 5u. the sides may have little or no adhesion with the ground. If a
Degebo’s (1933)attempts to measure W, with an oscillator very heavy foundation is required, a more economical solution
showed that average errors of 25,50, and even greater percentages than the usual massive block is to use a large area (by stiff raft
occurred in the course of an experiment, though the self- for example) whereby W,is very large (Crockett and Hammond
frequency remained about the same. This implies that any 1948), as in Fig. 5.2. An alternative method is to use a piled
change in Wgis accompanied by an equivalent balancing change foundation, as in Fig. 5f, to increase the size and weight of
in K2,the spring constant. There must also be a definite relation the active ground still further.
between the vertical dimension concerned and the horizontal (b) Self-frequency Control. If the bulb proportions are con-
dimensions of the oscillator or other base which produces the stant, and the ground around is homogeneous, the self-frequency

THE D Y N A M I C P R I N C I P L E S O F M A C H I N E F O U N D A T I O N S AND GROUND 515
is also constant, provided that the top stratum is at least three settlement does not seem to be so much as that which takes
times thicker thanthe depth of the inner bulb, in order that there place a t resonance in unrestricted freely acting ground.
may be a sufficiently developed mechanism. Another foundation within two or three diameters of the one
If the weights of the inner bulb, the foundation, and the considered must, by its obstructing effects, cause increased
machine, are respectively Ws, Wj,and W m , and the self-frequency stresses and thence settlement, on that side. In addition there
of the ground is n,, and, for the addition of the machine system may be other settlement, even or otherwise, from ground
upon it, it is n d , or n e , then, for negligible or low damping, resonance. The number of cycles causing each will depend on
and for small amplitudes, the linear relations are (Andrews and the manner the machine is started, run, and stopped, so these
Crockett 1945b, and Crockett and Hammond 1948) :- two modes of settlement will usually take place at different
speeds. Thus the foundation may settle first in one direction and
then in another, but reaching a limiting value (Hertwig, Fruh,
and Lorenz 1933).
Thus, Wg is always the dominant controlling weight, and the Fig. 7 indicates that the settlement of a foundation depends
self-frequency of the system cannot be altered very much. on the eccentric or out-of-balance inertia of the system originally
A further complication can be introduced to allow for the causing the oscillations, apart from resonance.
effect of a heavy spring instead of a weightless one, but results
do not justify this at present. I n practice, the several frequencies FORCE OF INERTIA
above should be measured for checking at each stage of con- 0 I0 20 30
I I
struction.
(c) Spring Constant. The spring constant K g f m follows from
+
E
5
The dynamic elasticity from this constant is usually several I-
_I
times greater than the static elasticity (Crockett and Hammond t;

VI
1947a, and King 1946).
’
;-\
‘1
1
’ \
! \ Fig. 7. Curve showing Ground,Settlement Proportional to
the Force of Inertia
Settlement is also a function of the inverse of the total active
weight, probably of WEfm. Degebo (1933) took two oscillators
with base areas of a quarter and one square metre respectively
on similar ground, the eccentric weights and their eccentricities
both being the same. Fig. 8 shows that the oscillator with the
INDUCING FREQUENCY-CYCLES PER SEC. smaller base, and incidentally the lighter ground weight, attained
the greater settlement; after a few thousand cycles at each
increment of increase in the frequency, settlement of both
oscillators tended to become constantly limited. Settlement of
this type and settlement from resonance caused the ground to
become compact as far as about a quarter of the diameter of the
foundation beyond, outside which it had no practical significance.
O TOTAL NUMBER xOF CYCLE5
20,000 25.000
I I
LARGE BASE AND BULB.

/ SMALL SETTLEMENT
Fig. 6. Ground Settlement by Induced Vibration $1

E
I-
-1
Dynamic Ground Settlement. There is a more or less rapid

settlement of the ground connected with the rapid rise in power ---- SMALL BASE AND BULB,
LARGE SETTLEMENT
output needed to carry the machine through the resonance zone
of the ground and foundation. This settlement consists of further
consolidation or packing down of the soil particles during a large
number of cycles, and is particularly rapid within the resonance Fig. 8. Curve showing Ground Settlement as a Function of
range, as in curve a, Fig. 6 . Degebo (1933) state that for elastic, the Total Number of Cycles
soft strata unconfined horizontally, but not, of course, for rock,
this range appears to lie between plus or minus fifty per cent of Damp& Coeficient and Loss of Energy Outwards. Energy
the most stable self-frequency. Andrews and Crockett (1945a from a single impulse imparted to the ground by the foundation
and b) and Crockett and Hammond (1948) have further expressed is mostly radiated outwards in a series of waves away from the
the view that for this action the ground must be free of obstruc- source. This energy, however, is not radiated outwards at once;
tion or confinement horizontally for a radius of at least three if it were, any resonant build-up would be impossible. Crockett
foundation diameters, and that the ground must be free to and Duff have shown* that the successive reductions of oscilla-
“heave” elastically in the vertical direction. Where freedom is tions are nearly 30 per cent for large amplitudes, while at s m a l l
restricted, stresses are inevitably higher; even at very low amplitudes they may be as low as 2 or 3 per cent. Thus, very
frequencies the ground settles substantially until restraint pre- large inducing amplitudes can only cause, at resonance, a build-
vents its becoming more compact (Hertwig, Fruh, and Lorenz
1933)-as shown in curve b, Fig. 6. In this case, however, total * Not yet published.

516 THE D Y N A M I C P R I N C I P L E S OF MACHINE FOUNDATIONS A N D GROUND
up to about five times, but small inducing amplitudes may build- or another of them have been in evidence ;they correspond to a
up many more times, though, of course, to a lesser peak ampli- “half harmonic”, and there may be others at still smaller
tude. There is a small remnant which is absorbed locally by intervals. This unstable harmonic response is probably due to
damping and changed into heat. During any compaction of the daily or even hourly change of moisture content, ambient tem-
soil this remainder, according to Degebo (1933), is 3 or 4 per perature, agd barometric pressure affecting the damping coeffi-
cent, and it is very much less when compaction has reached a cient, elastic and rigidity moduli, density, vapour pressure and
stable condition. The damping is approximately proportional to other physical properties of the complex mechanism.
the velocity and thus has a logarithmic coefficient.
Fig. 9 is a typical vibrogram of the movement of a machine Self-frequencies in Other Degrees of Freedom. Crockett and
foundation set into oscillation by a blow; energy radiation is Hammond (1947) gave several records of beat notes between
very great for the large early oscillations, but only slight for the machines and ground in rotational modes of oscillation in the
vertical plane. The two modes in horizontal translation and three
in rotation have so far not been studied, but their principles are
probably similar to those of vertical oscillation.
A
Frequencies of Several Underlying Strata. It often happens
that there are several strata within three times the depth of the
inner pressure part of the bulb, and the self-frequency is more
complicated. Each stratum can oscillate, and may oscillate
individually, particularly when the wavelength of the induced
oscillation is of the same order as, or greater than, the thickness
Fig. 9. Typical Simplified Vibrogram of Ground when of the strarum itself, and the purely bulb oscillation can also
Struck occur. Thus, a large foundation may have more significant self-
frequencies than a smaller one on the same piece of ground,
later ones; it also shows clearly the increase m frequency accom- because the larger bulb affects more and deeper strata.
panying the reduction of amplitude, which appears to be the
normal non-linear behaviour. True damping is very small.
ACTION OF SPRING-WEIGHT SYSTEMS (OSCILLATIONS)
Unstable Harmonic Frequencies. Crockett and Duff (Crockett Double Spring-Weight System Damped. The action of a
and Hammond 1948) measured frequencies on one site during double spring-weight system, as in Fig. 1I, is that any oscillation
ten weeks ;typical daily responses were :-
23.3 and 70.0 C.P.S.
23.3 and 46.7 C.P.S.
10.5 and 46.7 and 93.3 C.P.S.
Fig. 11. Double Spring-Weight
is shared according to the respective weights and spring

constants. Moreover, oscillation of both weights is practically
synchronous, but may be in opposite phase-so that both may
move together, or while one goes up the other may move down-
always provided that they are well out of the resonance zone
with the inducing frequency. Fig. 12 shows that with one phase
the intermediate spring is compressed more than with the other.
The system comprises the top spring-weight and also the lower
spring-weight spring, so that there are two extra resonance
peaks (Karman and Biot 1940). In practice there is also internal
damping of the springs, particularly for cork, rubber, etc.
(Crockett and Hammond 1947, and King 1938).
A Traple Spring-Weight Industrial Installation. An oppor-
tunity occurred in 1944 (Andrews and Crockett 1945a and b) to
SELF FREQUENCY-CYCLES PER SEC.
observe the movements of a machine and of its foundation com-
prising three successive weights and springs, the top weight
Fig. 10. Possible Displacement-Frequency Response of being 250 tons, the middle 220 tons, and the bottom weight
Ground 1,100 tons of concrete in addition to the active ground, which
This curve shows resonances at octaves and major fifths. probably amounted to a further 1,000 tons. The upper spring
was an oak pad 3 feet thick, the middle spring was 6 inches of
As 23.3 C.P.S. occurred on most occasions they named it the cork, and the lowest was provided by the medium-stiff clay on
“most stable self-frequencyyy,and it corresponds with those the site.
frequencies measured by other workers using OsciUators, as in A downward impulse to the top weight of the system set all
Table 1. This harmonic range is not h e a r ; 105 is less than the parts oscillating at their own natural frequencies, and with
half of 23.3, but that is explained by the non-linear larger sufficiently small amplitudes to be reasonably linear. The vibro-
amplitude at lower frequency. The authors infer that the grams in Fig. 13 show that they all shared the movement; every
ground can respond to a frequency series similar to that shown member oscillated with the same frequency, and in this instance
in Fig. 10,which is truly harmonic but non-linear. Intermediate with no phase difference, as the initial impulse was all in the
and smaller peaks have been included because on some sites one same direction. This agrees with the published theory of a triple

THE D Y N A M I C P R I N C I P L E S OF M A C H I N E F O U N D A T I O N S AND GROUND 517
spring-weight system; by using an oscillator a phase difference well as from the surrounding concrete retaining walls. In effect,
of up to 180 deg. might have been excited between the weights, the oscillating system comprises five spring-weights mounted on
and that would not have been so simple to interpret. a lower single large spring-weight. The intermediate springs .of
cork between the upper and lower weights have a very high
“Anti-wibration” Foundations. The main design problem is spring constant or stiffness, and it is inevitable that a high pro-
to select convenient size and stiffness for the weights and springs, portion of the energy of the lack of balance of the machine should
thereby ensuring that the movement of the lowest weight is small be shared with the lower weight, or main foundation, and so with
enough to prevent troublesome ground vibrations. Furthermore, the ground. This causes a large amount of the vibrational energy
resonance must not be allowed between die normally running to be transmitted to the ground outside the active zone of the
machine and the spring-weight system nor with the ground, system, and that is undesirable. An air space all round the sides
while the self-frequency of any adjacent structure and its chief of the blocks would be better than cork, thereby eliminating all
side spring connexion.
In addition an attempt to reduce the bottom weight-movement
had been made by providing a large foundation block, and
employing bulb piles in order greatly to increase the active
ground weight. Although this appears to be sound practice, two
possible difficulties arise, namely : the ground immediately
under the main block may settle by the vibration, causing a
cavity or discontinuity, and hence a reduction in the weight of
the active ground bulb; and also the piles are susceptible to
lateral instability and bending to an extent which it is impossible
to estimate accurately. Andrews and Crockett (1945a) have
measured bearing piles buckling dynamically at their centres,
and Rausch (1943) refers to top bending; in both cases the un-
1 satisfactory behaviour is due to the disproportionate horizontal

TOP MACHINE
I
-
n.
Ln
5
- FOUNDATION AND GROUND
1 I 1
Fig. 12. Displacement of a Double Spring-Weight System

Both weights have the same frequency, but the distance between varies
according to phase.
harmonics can be nearly as important as those in the ground.
None of these can be substantially altered, but fortunately the
top spring and weight can be varied considerably. Although it is
possible to increase LVd as already described, so many complica-
tions are involved that this method may be justified only where
limited reductions of vibration are required.
An interesting plant recentiy installed (Crockett and Hammond
1947) comprises five large air compressors, some balanced in part
horizontally and others vertically, each mounted rigidly on a
large concrete block. These blocks were poured into a previously
prepared lining of slab cork, each concrete block and compressor
unit thus forming a complete sprung entity in itself; each unit is
separated from its neighbour and from the main foundation as

518 THE DYNAMIC PRINCIPLES O F MACHINE FOUNDATIONS AND G R O U N D
systems, but in practice each weight may oscillate in translation for the ground, and one of them is lower than the vertical self-
in each of the three dimensions and also in three rotational frequency, then those of the spring-weight should be lowered
degrees. Thus, for the simplest possible arrangement of a single accordingly.
spring-weight system there are six degrees of freedom, and for If the whole weight, as in Fig. 15u, can be sprung level with
a triple spring-weight there are eighteen, but there are very and symmetrically about its mass centre, the system has one
many more real roots to the equations of frequency. At present, rotational frequency round its mass centre, independent of its
neither spring constants of ground nor compressional, torsional, translational frequencies; but for a lower position of the springs
shear, or damping properties of spring pad materials such as there are two rotaanal and two translational frequencies inter-
asbestos, cork, felt, and timber are known. None of these materials related, different phasing producing the effects as in Figs. 1%
obeys Hooke's law. So far rubber is apparently the only resilient and c, to second order terms. If the supports are unsymmetrical
material which has been adequately examined.
The fundamental and harmonic oscillations of a simple linear
system initiated by an impulse are not complex compared
to those induced by continuous oscillations. Fig. 14, Plate 1,
shows the two-dimensional traces of the bottom of a compound
pendulum and of an epicyclograph. These Lissajous figures show
b
Fig. 16. Suspension of the Weight from Springs by
Suspension Rods
a Horizontal-translation isolation.
b Horizontal-rotation isolation.
on the mass centre, there may also be two centres of rotation,

one between and one outside the supports. Horizontal move-
ment can also be controlled by suspending the weight from
springs by suspension rods which function as the length com-
ponent of pendulums, as in Fig. 16a. This method also controls
the rotational oscillation in the horizontal plane, but if the bars
are of different lengths and situated at varying radii from the
mass centre, as often for reciprocating engines supported along
each side, there will be a number of possible natural frequencies,
/ A / CENTRES OF ROTATION and the system tends to become unstable, resulting in undesirable
liveliness which demands periodical accurate tuning or re-
balancing.
1 I
> '
//,%/,'*.'A'///
. d
C
Fig, 15. Rotational and Translational Freedom of a Sprung
Weight
a Rotation only.
b Right to left translation plus anti-clockwise rotation.
c Right to left translation plus clockwise rotation.
d Two centres of rotation for an unsymmetrical system.
simple harmonics, variation of damping, difference of phase,

variations of initial velocity in each dimension, and almost linear
conditions, but there are very few degrees of freedom. In a
machine foundation having twelve or eighteen degreesof freedom,
unsymmetricalproportions, harmonics, variable conditions, non-
linear spring constants, and damping, any accurate computation
of oscillational behaviour is impracticable ;if a single dimension Fig. 17. Symmetrical Foundation System for Unsymmetrical
be changed, the movement may be entirely altered. The authors Plant
consider that this evidence shows the limits within which the
designing engineer is constrained to work. It is probably only Modern Trend in Oscillation-controlled Foundations. All
practical to calculate, in all degrees, a single spring-weight oscillation-controlled foundations require a good base plate, a
system, whereby adequate control may be exercised over stress, concrete block or other type of balance weight arrangement, and
frequency, and amplitude of movement, using only springs or a main, pit-type foundation in contact with the ground. Helical or
resilients of which the static and the dynamic properties in all volute springs and spring pad materials can be used directly under
directions are adequately known. the machine or top weight, but leaf springs must be used with
In order that the dynamic load on to the ground should be suspension bars as they are unable to oscillate usefully in either
minimized by a single spring-weight, the natural frequency longitudinal or lateral directions. Springs and materials having
of this should preferably be not more than one-third of the a wide damping range are obtainable, but added damping by
lowest measured on the site in any of the six degrees. If any of rubber or oil dash pots is sometimes advisable. In all arrange-
the self-frequencies in the other five degrees can be determined ments it is probable that an air gap all round the weight is best;

T H E D Y N A M I C P R I N C I P L E S OF M A C H I N E F O U N D A T I O N S AND G R O U N D 519
this also allows for access to the resilients. The oscillational Thus the transmitted stress equals the algebraic s u m of the
movement must be so balanced that all controlling gear and original stress and the reflected stress; if d E X is greater than
machine services or connexions remain adequately effective for d E T z there w ill be a change of sign in the reflected stress, but
convenience and efficiency.
Fig. 17 shows an economical method of supporting and con- if z / E x i s the smaller, the reflected stress is unchanged in sign;
trolling an unsymmetrical plant, whereby the machine, the the transmitted stress is always of the same sign as the original
impulses imparted to the ground, and the total mass distribution, stress. Equation (4) can be rewritten as
are all symmetrically arranged; calculations are thereby
simplified. This method can be further simplified by supporting
at the level of the mass centre, when rotations and translations
remain separated.
Sometimes a badly balanced machine must be used on a site which shows the way in which the transmitted stress is affected
having about the same self-frequency; the solution may be to by the ratio of Elpl/Eg2. If E1p1/Ezp2is very small (E2pZvery
use soft springs with dampers of oil or other type, so that the large) the transmitted stress approaches twice the value of the
tendency to oscillation build-up may be avoided, although all original wave stress. If, on the other hand, B1pl/E2p~ is very
other conditions exist for resonance. The most practical answer large (E2p2 very small), practically no stress is transmitted; that
may be to have mass-produced controllable dampers, using oil is the case for air to steel conditions. A number of values have
for example, and working with soft springs, a light weight, and been worked out and are given in Tables 2 and 3 covering the
a light but large area main, pit-type foundation resting in the combinations often encountered in foundation practice. Table 3
ground. Rubber springs, large and small, and hydraulic dampers cannot apply for a tension wave if it separates the two materials.
now on the market, suitab!e for heavy machines, are illustrated
in Figs. 18, 19, and 20, Plate 2.
TABLE 2. VELOCITY OF P STRE~S WAVE(v* = dmp)
HIGH F R E Q U E N C I E S ( S T R E S S WAVES)
Simplified Theory. If the impressed frequencies are high and Material 3 modulus p density, v velocity Remarks
of elasti- Ib. per of wave,
their wavelengths, either of fundamentals or harmonics, are city, lb. cu. ft. t . per sec
similar to or shorter than the dimensions of individual com- ?er sq. 111.
ponents, the weights cannot be considered as rigid bodies without
Steel . 30 x106 489.6 16,860
Aluminium 10 xloa 170 16,440
Granite . 6.2 x 106 170 13,010
Concrete . 3.5 x 106 153 10,300 Code of practice :
1.1.2.
PI
Limestone . 2.7 X 106 170 8,584
El Concrete . 1.7 X 106 153 7,603 Code of practice :
1.2.4.
BOUNDARY
Sandstone . 9.0 x 105 150 5,274
Sand and
gravel . 4.0 x 105 115 4,032
Clay . 1.1 x 106 135 6,100
E2Pi Timber . 1.0 x 105 57.2 2,914 Thin pad, per-
pendicular to
@-?in.
Fig. 21. Wavelength and Refraction at a Boundary between Timber . 3.5 x 104 57.2 1,560 T h c k pad, per-
Two Different Materials pendicular to
grain.
flexural distortion; reflection, refraction, and diffraction of stress
Cork . 1,050 15 570 Compressed thin
layers.
waves become important phenomena. A compression wave Air at 60
originating in a machine bedplate obeys some of the airborne deg. F. . 20.32 0.0765 1,110
sound-wave laws; on reaching the dividing boundary with
another body (say, the foundation), part of the wave energy is * This formula is usually quoted without the term g, which is
reflected back and the rest is refracted into the second body. included here because it is necessary to use it with the normal engi-
Waves of shear are also important. neering constants.
The velocity of a P or simple “push-pull” wave (Rayleigh 1885,
and Thornton 1937) travelling along a long thin rod is d E 3 , The amplitude laws for the wave are similar, although they
where E is Young’s modulus, p is the density of the material, have not apparently been stated in any published literature; they
andg is the acceleration due to gravity. On reaching the boundary are important because the actual strain and compression of the
with another material, the wave crosses and proceeds at a new layers can be measured and estimated from them. When the
velocity, direction, and wavelength, dependent on the density and boundary has its maximum stress of f+fr (Elpl being very
Young’s modulus for the material, as shown in Fig. 21. The small and E2p2 very large) the movement or displacement
velocity of the wave is given by v = lw/2rr = lf, where I is the of the boundary is the least. Conversely, the least stress is
wavelength, f is the frequency in waves per unit time, and w accompanied by the greatest movement : Elpl is large and E g z
denotes the angular velocity in radians per unit time. If the is very small.
moduli of elasticities and densities are El, E2, and pl, p 2 respec- Tables 2 and 3 have been prepared for long thin bars having
tively, the stressft transmitted into the second bar, as compared free bulging or Poisson’s ratio effect, whereas in practice the
with the original stressf, is material is more extensive and cannot bulge sideways so much. In
addition, there are always waves of shear and waves of compres-
ft
J
2d\/E,Pz
dE2Pz+dB,pl
.. . . (4) sion travelling at an angle to the boundary, diffraction around
= -. - corners, and other complications which are not considered here,
and the ratio of the reflected stress fr and the original stress is but the presence of which should be borne in mind by the
engineer.
. Action of Sand, Gravel, and Other Insulating Layers. For
The difference between these is low-frequency oscillations, with relatively long wavelengths,
the spring-weight action is rather similar for sand or for clay,
and if one of these forms only a thin layer within a large mass
of the other, or only a small proportion of the active part of the

__-
m
Z 86 -1001 0 €81 €8 061 06 P61 P6 L61 L6 861 86 861 86 661 66 661 66 661 66 661 66 OOZ 001
6 0 OOI-ILI €8 -1 001 0
U 'OOOCSE = B
h
I -~
0 001-001 06
IL 62- - OOT 0 EZI EZ LPI LP 991 99 ELI EL 081 08 181 18 P81 P8 681 68 961 96 T'bsJad'qI
'OOO'001 = 9
~- J=qv.L
0 001-9 P6 - ZS 8P- LL €2- I 0 IEI I € IS1 IS 091 09 OLI OL ZLI ZL 6LI 6L €81 €8 P61 P6 M3
0 001-2 auoispueg
0 001- I 66 - 21 aoolsa~~
alpif)
-rur-w
*u!.bs iad '91 *u! .bs lad *q1
'd .Sap 09 'OOOCSE = B '000'001 = B lanes 1 auois p'y~ amis z.1.1 Umpp
It3 JW YJO3 JJP!L i XIqKUIL K E I ~ puF: pUeS I -p"S 'J33.73KIO3 -9-7 '3lJJ3U03 alFIf) -nm 1331s
c
-
T H E D Y N A M I C P R I N C I P L E S O F M A C H I N E F O U N D A T I O N S A N D GROUND Plate I
2
2
k
[I.Mech.E., 19491
Rate 2 THE DYNAMIC PRINCIPLES OF MACHINE FOUNDATIONS AND G R O U N D
Fig. 18. Rubber Fig. 23. Cork Lining inside the Tank of an Insulated Foundation before the Inner Concrete had been Cast
Spring
Fig. 19. Rubber Springs Fig. 20. Hydraulic Dampers

[I.Mech.E., 19491
THE DYNAMIC PRINCIPLES OF MACHINE FOUNDATIONS AND GROUND 521
bulb of pressure, there can be very little effect from such a out of phase with one another. Thus there may be building up
layer. or partial cancelling out of parts of the waves, with resulting
Fig. 22 illustrates the action of stress waves travelling through standing waves and troughs, so the sand may be a moderate or
severalmedia :consider a high-frequency oscillationwith a wave- poor insulator for a specific frequency. Most machines have
length of about 1 foot, travelling in concrete, passing into 3 feet several frequencies and it is probably impossible to ensure that
of sand and finally out into clay. This short wave, originating all will have the right phase relationship to the insulating bed
in the concrete, has about 37 per cent of its stress (according to thickness.
Table 3) reflected back from the boundary with the sand, and The authors do not consider that sand beds and similar
63 per cent carries on into the sand. On reaching the clay insulating arrangements are reliable anti-vibration methods,
boundary, the latter portion has about 31 per cent reflected back either for low- or high-frequency oscillations j many observations
in tension and 69 per cent, or 43+ per cent of the original wave, and tests show remarkable agreement with the theory. It is hoped
carries on. If there had been no sand layer, 40 per cent of the that high-frequency amplitude methods of calculation and a
original wave stress would have been transmitted directly into table of percentages similar to those for stresses in Table 3 will
the clay-a very small difference. Amplitude of movement is shortly be produced.
here approximately of the same order as the stress. If the A large foundation (Crockett and Hammond 1947) of the
frequency of the waves had been less, and the wavelength had single spring-weight type was recently built for an air compressor
been greater, the amount transmitted would have been greater, in a pit excavated within a protettive double ring of sheet steel
thereby reducing still further the influence of the sand bed piling, which provided an annular space about 2 feet wide filled
insulation. with sand in order to form an anti-vibration cushion. It now
If the tensile stress-wave reflected back from the clay has a seems clear that such a device will never work satisfactorily,
greater stress than the normal compressive stress between the either for low- or high-frequency vibrations; it would have been
better to have left the annular space empty. Air spaces of this
CONCRETE type have been in use for a long time and have given results
1 . 1 -1 I SAND
CLAY
superior to other spaces which have been filled.
Selfinduced Stress Wave Build-up. Another recent founda-
tion for supporting a gear hobber, which it w a s desired to insulate
from vibrations in the ground caused by adjacent machinery,
is shown in Fig. 23, Plate 2 j the large concrete block is insulated
by a layer of cork all round and underneath. Vibrational energy
with wavelengths shorter than the dimensions of the block,
emanating from the gear hobber, cannot “leak out” very fast
through the cork. As damping in the concrete is low, the
amplitudes of these waves must inevitably build up, perhaps to
an undesirable extent. This action is similar to that of a room
with not very porous walls wherein any sustained sound must
build up by successive internal reflections, or by reverberation,
to a high intensity (Fleming and Allen 1945).
Thus, a machine which has been insulated from external
disturbances may cause in itself greater trouble than if no such
precautions had been taken; a similar result can also occur with
low-frequency oscillations, and one remedy appears to be the
provision of large damping for the high as well as for the low
frequencies, apart from suitable springing (Haffield, Stanfield,
and Rotherham 1942, Hatfield, Rotherham, and Harvey 1944,
King 1938, and Zener 1940).
Badly Fitting Surfaces. A particular complication arises at
the boundary of two media which are not completely in contact
over the whole area, as in the case of a badly grouted baseplate
on the concrete block. The amount of wave reflected and re-
fracted from such an interrupted boundary cannot be satis-
factorily calculated: each small indiHdual area of contact will
often increase in size with greater pressure by elastic deformation,
and therefore any calculation would be of a non-linear character.
Fig. 22. Successive Refractions and Reflections at Two A rational theory and formulae have been evolved concerning
Boundaries Between Three Materials a wave passing through a row of parallel and equally spaced
cylinders (Rayleigh 1926), as in the case of sound travelling out-
sand grains due to gravity and to any live load which may be wards between the front show pipes of an organ case. This shows
carried by the sand at the moment, the energy is partly dissipated that waves considerably longer than the width of the space be-
by the grains “shaking down” into a more compact mass, and tween the pipes pass through with little reflection backwards;
partly in being reflected back again in the direction whence it waves shorter than this width are largely reflected back by the
came, once more being changed in sign into compression. In pipes. These long waves coalesce by diffraction on the other side
practice, after a comparatively short period of consolidation, the of the pipes. The short waves are not so strongly affected by
f i s t reflection in tension will be able to travel right through and diffraction; instead of joining up almost at once, they assume
back into the original concrete. Here both density and Young’s the form of a directional beam for some distance before gradually
modulus are higher than those of the sand, so that there is no coalescing. When such action is applied to two-dimensional areas
change in sign, and 37 per cent of the tension wave is reflected of contact, rather than to long narrow gaps, and with large areas
back in tension. of discontinuity between them, the effects are more marked; a
Fig. 22 also shows how multiple reflections and refractions lower proportion of the waves are reflected back and more pass
build up, causing a series of alternate compressions and tensions through.
of progressively decreasing stress to penetrate into the clay. For badly fitting surfaces, wave energy approaching the
Most of the high-frequency oscillational energy will “leak out” surface will not be largely reflected back from the discontinuous
in the same manner as heat, although the concrete block may areas, but will concentrate around and pass through the relatively
be surrounded by sand, but it may take a number of internal small areas of contact. Stress will be very high at such places,
rehctions and reflections to do this. and when the additional effects of different materials on either
Waves of compression and tension will traverse in phase or side of the boundary are taken into account, high tensilostresses

522 THE D Y N A M I C PRINCIPLES O F M A C H I N E F O U N D A T I O N S AND GROUND
can be induced. This may cause chattering or even lead to direct the Second International Conference on Soil Mechanics and
compression, successive impact, and grinding-action failures. Foundation Engineering, Rotterdam.
Barkan of the Russian V.I.O.S. organization in Moscow has
Stress Wuves Through “Plums”. When a stress wave passes made controlled experiments on large foundations on the ground
through and around a c‘plum” of different density and elastic surface, and found they have had natural frequencies and har-
modulus from that of the surrounding material (Rayleigh 1926), monics. Tschebotarioff has measured the natural frequency
it changes its velocity inside the “plum” and thereby produces a and other mechanical properties of soils under small model
shearing action around the boundary surface. A further result is foundations, and also investigated resonance troubles in some
to set the plum oscillating at its own self-frequency, which then large machine foundations. Bendel has investigated the dynamic
initiates a new set of stress waves radiating outwards in all bulb of pressure. The results of all these workers appear to
directions (Rayleigh 1926). This combination of stress waves can corroborate the basic ideas and their arrangement by the authors.
induce local stressing sufficient to cause failure. Andrews and The worlc of Ramspeck and Schulze (1938) has influenced the
Crockett (1945) examined a spectacular failure of this type, authors, but is not specifically mentioned in this paper.
wherein plums of dense concrete, originally caused by uneven
needle vibrating during the foundation casting, remained whole,
when the rest of the concrete was shattered by a wave reflected
in tension, particularly from a curved surface towards a focus.
APPENDIX I1
CONCLUSIONS
BIBLIOGRAPHY
It is now possible to lay down principles of design for the
foundations of reciprocating machinery, logically based on the A much fuller bibliography has been prepared by the authors,
twin principles of oscillation and stress waves :- and copies have been deposited in the Libraries of the Institu-
(a) Any machine-foundation system can be considered as a tions of Civil and Mechanical Engineers.
series of weights and springs, susceptible to reasonably simple
vibration analysis. The size and proportion of each part may be REFERENCES
suitably designed for specific requirements, and a satisfactory
type of springing and damping can be selected; the spring con- ANDREWS, W. C., and CROCKETT, J. H. A. 1945a J1. Inst.
stants of the springs or of the spring pads may not necessarily Structural Eng., vol. 23, p. 453, “Large Hammers and
be linear in their action, nor will they all have simple damping Their Foundations”.
coefficients. 1945b Trans. Inst. Eng. and Shipbuilders in Scotland,
(b) Every component of a spring-weight system may con- vol. 89, p. 53, “Theory of Forging Hammers and Their
tribute to oscillation in six degrees of freedom, and this must be Foundations”.
taken into account at every stage of design. BERGSTROM, SVENG., and LINDERHOLM, SVBN 1946 Swedish
(c) The ground is in itself a spring-weight system, of which Cement and Concrete Research Institute at the Royal
the self-frequency or the range of self-frequencies for each site Technical University, Stockholm, “A Dynamic Method
must be determined and checked periodically during the con- of Determining Average Properties of Surface Soil
struction period. These frequencies are not likely to be susceptible Layers”.
to much change by the foundation. CROCKBTT, J. H. A., and HAMMOND, R. E. R. 1947 Inst. C.E.,
(d) The ground, and any foundation in or upon it, can most Structural Paper No. 18, “Reduction of Ground Vibra-
probably oscillate at determinable frequencies in six degrees of tions into Structures”.
freedom. 1948 Proc. Second International Conference on Soil
(e) Resonance between ground and machine should be pre- Mechanics and Foundation Engineering, vol. 3, pap.er
vented if large amplitudes of oscillation and loss of power are lb.2, “The Natural Oscillationsof Ground and Industrial
undesirable in the machine. Foundations”.
(j)Resonance between the ground and the machine should DEGEBO, (Deutsche Gesellschaft fur Bodenmechanik) See works
be prevented if settlement is undesirable, particularly where byFoundation L.; Hertwig, A., Truh, A. and Lorenz, H.;
large amplitudes are present. and Ramspeck, A. and Schulze, C. A.
(g) Resonance between ground and machine should be pre- English translations are available the Ministry of Works
vented if interferences with nearby machines, instruments, Library London
structures and property are to be minimized. ROTHERHAM!L. 1936 Deutsche Gesellschaft fiir Boden-
(h) If resonance between ground and machine cannot be mechanik (Degebo) with the University of Gottingen,
avoided, the basic methods commonly used in other spring- “Die Anwendung Dynamischer Baugrundunter-
weight systems can be adopted involving the use of dampers, suchungen” (Julius Springer, Berlin).
springs, etc. FLEMING, N., and ALLEN,W. A. 1945 Inst. C.E., Structural
(i) Simple design is the best; a single rigid body with six Paper No. 12, “Modern Theory and Practice in Building
degrees of freedom can be calculated as in the standard text- Acoustics”.
books. Double- and treble-body systems can be used only as a HATFIELD, W. H., ROTHERHAM, L., and HARVEY, E. M. A. 1944
guide to possible action, but multiple-body foundations are apt Trans. North East Coast Inst. Encr. and Shbbuilders.
to become unstable, and hence their movements unpredictable. vol. 60, p. 227, “Further ExperimGts on thebampGg
0’) Stress waves must be considered at all stages of design, to Capacity of Metals”.
ensure that their reflection, refraction, and diffraction do not HATFIELD, W. H., STANFIELD, G., and ROTHJBHAM,. L. 1942
build up locally, particularly to focal points. Stress waves occur Trans. North East Coast Inst. Eng. and Shipbuilders,
in three dimensions, and are usually compressive-tensile, and vol. 58, p. 273, “The Damping Capacity of Engineering
shear. Materials”.
(k) Boundary surfaces should be as accurate as possible to HERTWIG,A., FRUH,A., and LORENZ, H. 1933 Deutsche
avoid local stress concentrations and chattering. Gesellschaft fiir Bodenmechanik (Degebo), “Die Ermitt-
Acknowkdgement . This paper is Crown copyright reserved, lung de fiir das Bauwesen: wichtigsten Eigenschaften des
and is published with the consent of the Controller, His Majesty’s Bodens durch Erzwuzgene Schwingungen” (Julius
Stationery Office. Springer, Berlin).
KARMAN, T., and BIOT, M. A. 1940 “Mathematical Methods
in Engineering” (McGraw-Hill, New York).
APPENDIX I KING,A. J. 1938 Engineering, vol. 146, pp. 124 and 198, “The
ADDITIONAL EVIDENCE Reduction of Structure Borne Noise by Vibration-
Attenuating Supports”:
Since this paper was written, further sources of information
have been revealed to the authors through the Proceedings of
1946 11. I.E.E.. vol. 93, Dart 2,- D
Vibiation Problems”.
__
- . 435. “The Analvsis of

THE D Y N A M I C P R I N C I P L E S O F M A C H I N E F O U N D A T I O N S AND GROUND 523
LAMB,H. 1904 Phil. Trans. Roy. SOC., Series A, vol. 203, RAYLEIGH, LORD 1885 Proc. Lond. Mathematical Soc., vol. 17,
p. 1, “The Propagation of Tremors over the Surface of p. 4, “On Waves Propagated Along the Plane Surface
an Elastic Solid”. of an Elastic Solid”.
L m , L. D. 1946 American Sn‘enrist, vol. 34, “Earth Motion 1926 “Theory of Sound”, Parts 1 and 2 (Macmillan and
from the Atomic Bomb Test”. Co., Ltd., London).
LORENZ, H. 1934 Z.V.D.I., vol. 78, p. 79, “Neue Ergebnisse SEZAWA,K., and KANAI,K. 1937 Bulletin of the Earthquake
der Dynamischen Baugrundforschung”. Research Institute, Tokio, vol. 15, part 1, “Resonance
LOVE, A. E. H. 1944 “A Treatise on the Mathematical Theory Phenomena and Dissipation Waves in the Stationary
of Elasticity” (Cambridge University Press). Vibrations of a Semi-Infinite Body”.
MINORSKY, N. 1943 “Introduction to Non-Linear Mechanics”, TERZAGHI,K. 1943 “Theoretical Soil Mechanics” (Chapman
Parts 1-4 (Washington: The David W. Taylor Model and Hall, Ltd., London).
Basin, U.S. Navy).
RAMSPECK, A., and SCHULZE, C. A. 1938 Deutsche Gesellschaft THORNTON, D. L. 1939 “Mechanics Applied to Vibration and
fiir Bodenmechanik(Degebo), “Die Dispersion Elastischer Balancing” (Chapman and Hall, Ltd., London).
Wellen in Boden”. (Julius Springer, Berlin). WILLIAMS, C. C. 1932 Civil Engineering (New York), vol. 2,
RAUSCH,E. 1943 “Maschinenfundamente und Andere Dyna- p. 675, “Small Earth Vibrations AfFect Foundations”.
mische Bauaufgaben”, vols. 1, 2, and 3 (Vertriebe ZENER,C. 1940 Proc. Phys. SOC., vol. 52, part 1, p. 152,
V.D .I.-Verlag, Berlin). “Internal Friction in Solids”.
Discussion
Mr. F. S. SNOW, M.I.Mech.E., said he had had very considerable living in houses 200 yards away lodged complaints with the
experience of the effects of vibration in buildings and various Council.
types of foundations. One building at Camden Town, a modern Another example was a building at the comer of London
block-some ten or twelve years old-adjoining the railway, Bridge where piles had been driven as close as 1 ft. 10 in. in
was on a comer site where there was a considerable amount of series along a wall quite satisfactorily, without any. difiidty;
loading and unloading on the premises, the major portion of the but 200 yards away another building had fallen down. Perhaps
vehicles being horse-drawn drays. the authors could explain how they had managed to do away
The architect and the engineer had endeavoured to minimize with wave effects 200 yards away, which was a matter’that
the vibration and noise by installing the whole of the building worried him at the present time.
on a cork foundation similar to the one shown in Fig. 23, though
much larger. The effect of the load over a number of years had Professor D. G. CHRISTOPHERSON, O.B.E., Ph.D., A.M.
been to compress the cork to such an extent as to render it I.Mech.E., said that the authors had stated in the summary of
virmally useless for the purpose originally intended, and the their paper that the most substantial advance which had taken
settlement of the whole building amounted to some 2+3 inches. place recently on the subject was the discovery that any particu-
It had also been found that where the cork was in contact with lar piece of ground had its own natural frequency; they had made
water, it had disintegrated and had gone into dust in parts, this it quite clear that the natural frequency was not intended simply
being a further cause of settlement, with the result that the to be a characteristic of a particular machine or a particular site
building, although of reinforced concrete, was considerably or foundation, but a characteristic of the ground at that place.
deformed and cracked in many places. As far as noise and It was impossible not to be struck by the queerness of that
vibration were concerned, the position was as it would have observation on purely theoretical grounds; by that he did not
been in the first place if there had been no expensive cork imply that the observation was not correct, for he was sure it
foundation. was. Similar observations had been made in a number of
Another problem which he had had to solve concerned a different ways.
series of “byrd” screens, used in the manufacture of pulp. Those It was equivalent to saying that, if the earth were considered
screens had been revolving out of unison, thus causing a con- as a very large elastic solid, then a natural frequency was being
siderable amount of vibration and sway on the floor. The produced which Mered from place to place ;the only physical
difficulty had been overcome by installing a series of struts quantities to determine this frequency appeared to be the elastic
coming down on to teak blocks at end-grain-the end-grain constants and the density, but a characteristic distance had to
rather than the cross-grain being an important point. No par- be introduced before the frequency could be determined.
ticular system fitted every machine. T o surmount the difficulty, gravity could be introduced as a
He had for many years experienced difficulty in piling work, controlling force, and the vibrations considered to be analogous
particularly when close to an adjoining building, to prevent the to gravity-controlled waves at sea, but a consideration of the
vibration from shaking the building down. At various times difference in frequency in the two cases showed that suggestion
cork mats had been used in the piling heads to prevent vibration, was most unlikely to be correct.
but those were very expensive. It had been found that packing The alternative explanation was that a characteristic distance
the hammer head of a pile with green elm was very much cheaper could be introduced by imagining the earth to be stratified in
and allowed about six piles to be driven, as against two when layers having Merent elastic constants. Oscillations could then
the cork mat was used. occur with a period equal to the time taken for an elastic wave
When piles were driven, or even when vibration was caused to go through a stratum and be reflected back again. There was
by a series of machines adjoining a wall, the worst vibration did no doubt that such oscillations did occur in appropriate cir-
not necessarily occur close to the machines. Some form of wave cumstances, but he believed that natural frequencies would also
was set up, which the authors had endeavoured to explain, but occur in places where no notable stratification could be discerned
which was still not very clear to him. On another site, where a for a very considerable depth from the surface, and where the
“Cecostamp” hammer had been installed, many complaints material appeared to be perfectly homogeneous.
were received from the occupiers of adjoining buildings. To He did not mean that in those places the earth was truly
minimize the disturbance Ole stamp was re-erected in a pit on homogeneous. He thought that at greater depths more con-
layers of cross timber (pitch pine), the whole being topped with solidation had occurred, and consequently the elastic constants
about 8 feet of concrete. As a result of those precautions, the were larger. Thus, in effect the earth might be considered not as
adjoining owners ceased to complain of the noise, but people a few thick strata of varying rigidity, but as an infinite number

524 DISCUSSION ON T H E D Y N A M I C P R I N C I P L E S OF M A C H I N E F O U N D A T I O N S
of very thin strata, each of which differed only infinitesimally surface (in the absence of internal damping) to keep it going
from the one above. He wondered whether it was possible to also had to be satisfied. Thus,
predict a natural frequency for such a system.
He had assumed that, in the region in which the vibrations
were occurring, Young’s modulus, E, increased linearly from
Eo at the surface to 2E0 at a depth b, as shown in Fig. 24, and which, if the two solutions occurred independently, reduced to
Thus the natural frequencies were :-

0.684 z
o from
n =
0.175
- b J 5.
, J>G . 7
a
0.432
J, b
-
1-
(12) . . .
Those frequencies formed a series in which the various
Fig. 24 members were related approximately by simple arithmetic
ratios.
that the foundation was standing on the top of an infinitely long If, alternatively, it was assumed that the ground surface was
column of material having the specified properties. The same constrained by having a mass m per unit area placed upon it,
theory would hold, with modifications, if the supporting column then in place of the frequency equation (1l), a condition
was regarded as increasing in area at increased depths, as shown au aZu
by the dotted lines in Fig. 25. E-
ax
=m-a t 2 . . . . . (13)
obtained, wgch could be solved by making use of tabulated
values of the Bessel functions. A series of modified frequehcies
could be obtained corresponding to the original frequencies
from equation (12), which, provided m was not too large, might
very well be expressed in an equation similar to equation (2),
the value of the parameter W,being appropriately selected.
Mr. D. M. CHADDOCK, M.Sc. (Eng.),M.I.Mech.E., saidthat

the subject of the paper was one which was not normally
associated with the design of artillery.
Artillery was a form of heat engine, a prime mover, and it
required a foundation to absorb its violent reactions. Statically
emplaced guns, coastal defence guns, and some anti-aircraft guns
were mounted on normal foundations of steel and heavy mass
reinforced concrete, and those problems were within the realm
of ordinary civil engineering, although due care had to be taken
Fig. 25 to distribute the very heavy loads and stresses through the
structure. The mobile gun, however, was in rather a different
category in that it had to carry its own foundations;they had to
In actual fact, all the material near the foundation vibrated, be light because they were mobile, and they had to be “emplaced”
but the effect of the masses I, I would both increase the mass on any type of Soil-sand, sat, or frozen earth.
and increase the spring support, so they might have only a The demand for minimum size and weight necessitated the
relatively small influence on the frequency. use of very high ground pressures, and the apparent measure
If x was measured from the point above the surface at which of success that had been achieved might be due to the impulsive
the line in Fig. 24 would reach Eo then nature of the loading.
X When a gun was fired, the reaction was very violent and of
E = Eiz very short duration. He asked the authors whether they had any
information on the resistance of soils to such high intensity
and the equation of motion giving the displacement u at time c loads of short duration-perhaps i L second-square
~ wave
for any cross-section of the rod was impulse applied to the soil. The total force was 2 tons per sq. fi.
or more on any sort of soil.
The designers were particularly anxious to find some form
of correlation between the performance of the soil to this type
where p was the density of the material. of loading and the more normal soil tests which were taken, such
The substitution of z = .\/x
in equation (8) gave as plate bearing tests and grading, to enable them to predict the
performance of the guns on a particular type of soil. Similar
problems of short duration and high intensity loading probabIy
also existed in other applications, such as in the design of pave-
Solutions of that equation were required to describe a vibration ments and on aircraft runways subject to high loads.
of frequency n. There were two possible solutions, namely In some of the mobile guns it was possible to allow a certain
amount of displacement of the foundations. If the gun moved,
the sights moved, and the axis of the bore pointed in another
* (10) direction; but the gun-layer could bring that back into line
again for the next shot by operating his controls. Some guns,
or notably anti-aircraft guns, were remotely controlled by radar
and predictor systems, and initially they had to be brought into
where Jo and Yowere Bessel functions of zero order and the line with their control systems, an operation similar to setting
first and second kind respectively. The condition that the vibra- up a theodolite; any subsequent movement of the foundation
tion was a “free” vibration, and did not require any force on the in relation to its controlling mechanism implied an error in the

DISCUSSION ON THE DYNAMIC PRINCIPLES O F MACHINE FOUNDATIONS 525
of them both, thus, demonstrating that the system necessarily
shooting of the gun, and as the range increased that angular
error was multiplied enormously. For example, one minute of approximated to the double spring-weight system described.
arc, which was equivalent to 1 inch at 100 yards approximately, That the ground itself would be damped seemed unlikely, and
was also equivalent to 3% inch in a structure which was 30 feet practice demonstrated that the application of the theory was
across. Thus, the problem was to produce a light-weight founda- sufficiently reliable for engineering purposes.
tion, considerably lighter than the load to be carried, which The dampers had been designed to provide any reasonable
would maintain its position to the required degree of precision amount of damping for machine foundations, where sufficient
while blows, which could only be likened to those of a pile movement was present and where the acceleration developed by
hammer, were rained upon it. a square-fronted wave form did not result in excessive loadings.
The authors had suggested, when they discussed the natural With a forging hammer the damper piston was pushed down
frequency of the soil in relation to the structure it was supporting, violently on the initial impact and the fluid was able to pass
that a one mass system should not have a natural frequency of
more than one-third of the natural frequency of the soil. As the
average value of the soil was about 20 cycles a second, that
meant the natural frequency of the structure on it should not be
more than about 6 cydes a second. That was contrary to current
theories of rigidity and stiffness in which high natural fre-
quencies were usually advocated. He asked whether it would be
advantageous to reduce the natural frequency of the gun
mounting structures by introducing resilience. The natural
frequencies were probably at present above 6 cycles a second and
it would be easy to reduce them.
The paper indicated that increasing the apparent mass of the
structure also lowered the natural frequency. He asked by what
practical ways the apparent mass could be increased in view of
the fact that the structure had to be kept to a minimum in size
and weight.
Finally, he raised a matter concerning a very large and unusual
type of machine which was being developed. Its purpose was
connected with guns, and it introduced very large unbalanced
forces which were repeated rhythmically. The forces were in a
horizontal plane, and they occurred a few feet above the ground.
An apparently adequate mass-concrete foundation had been
proposed, but it appeared that, if the machine were freely
supported, the ends of the foundation would tip up by about
2 inches. To reduce that deffexion to a reasonable proportion
by adding more mass woularequire the addition of an impossible
amount of concrete.
The machine was to be set up in a laboratory area, and he
asked whether the authors could give him any guidance as to
the form the foundation might take, there being no question of
considering balancing on the lines suggested in the paper.
Mr. C. B. BROW (London) said that the weight and spring
analogy as applied to subsoils was most interesting. I n Fig. 5
the bulb of soil acting as a weight was shown to depend very
largely upon the size and type of foundation. In addition, the
ground, whether it acted as a weight or as a spring, or as some-
thing in between the two, would obviously vary in its action
according to its nature and its state of compaction. A foundation
after having been in use for some considerable time might react
differently from the new foundation, owing to the consolidation
of the subsoil beneath it. That suggested the desirability of Fig. 26. Arrangement of Damper Unit
preventing excessive change in the subsoil by some form of soil
stabilization beneath the foundation. relatively freely from one side of the piston to the other. When
The stabilization of a large mass of subsoil by some cement the springs under the machine were compressed an inch or two
or chemical binder would obviously increase the weight in the by the movement of the mass, induced by the shock of the blow,
same manner as if the size of the concrete foundation were their resilience would normally not only return the machine to
increased. It might be possible to increase the spring effect of
the soil by stabilizing with a suitable binder which in itself had its position of rest, but also allow it to oscillate above and below
in that position in an uncontrolled fashion.
some degree of resilience. Certain bituminous or rubber com- The damper, introduced to work in a suitable position, could
pounds might fulfil that requirement. that excessive energy and allow the mass to return rela-
If that idea were at all practicable, it might be possible in the absorb
tively slowly to its original position. To control the movement,
case of granular materials to vary the soil frequencies over a
wide range merely by varying the quantity and nature of the the fluid passage from the top to the bottom of the damper
was restricted and adjustment provided so that the mass
stabilizingcompound. It would be interesting to hear the authors’ piston
could be dead beat in one half cycle.
views on that point. The civil engineer would determine the vibration frequency
of the ground and would design a suitable spring mounting
Mr. S. FOSTER (Leamington Spa) said that the damper unit (rubber in compression was proposed) to give the mass of the
(Fig. 26, and Fig. 20, Plate 2) had been developed entirely as a anvil and emplacement a vibration frequency sufficiently remote
result of the general theory described in the paper, and the results from the natural frequency of the local terrain to prevent a
of the first prototypes, applied to a forging hammer, had shown condition of resonance being set up between them. The approxi-
that the ground had acted as a large spring-weight system, the mately dead-beat action of the damper, which absorbed the
hammer and its timber pad acting as the upper spring-weight energy stored in the spring system, owing to the movement
of much smaller size. The original damper had been connected caused by the impact, should tend to reduce appreciably the
between the two weights and had damped out the oscillations effects experienced by the foundations.

526 DISCUSSION O N T H E DYNAMIC P R I N C I P L E S OF M A C H I N E F O U N D A T I O N S
The effect of the change from a continuous support to local from the oil companies and prospectors concerning explosions
spring pads under the machine should not be overlooked since and reflected waves; or whether those calculations took rock
the change in the effects caused by the shock loading might not formation too much into account for practical purposes. He also
be reliably predictable. asked whether the authors could give some information on the
In Fig. 22 no mention was made of shear waves in the con- nature of the indicator by which they measured the condition
crete ;the diagram also indicated the existence of tension waves of the clay or sand at a particular point, and whether the mass,
in sand which were difficult to conceive unless the sand were the depth, or the cross-sectional arrangement had made a great
wet. I n connexion with the tension and compressive waves in difference to the readings obtained.
the concrete he assumed that the waves were present owing to If two springs of different stiffness, each attached to a weight,
the impact at a point causing a more or less irregular distribu- were connected, and the system were put into oscillation, then
tion of load on the base of the machine, which together with the weights could be so chosen that the point connecting the two
varying effectiveness of the support for the foundations would springs remained in equilibrium. Perhaps that principle could
set up a complex and irregular form of stress in the concrete. be applied to machine vibration problems; he had not had time
to try it out.
Dr. R. JONES (Harmondsworth) said that the Road Research
Laboratory was actively interested in dynamic testing methods Mr. H. S. HVISTENDAHL, B.Sc. (Eng.), M.I.Mech.E., asked
and the data given in Table 1 would be of value in some experi- whether the curve shown for the settlement of the soil as a func-
ments that were being carried out. They had used small charges to tion of the impressed frequency represented a permanent settle-
produce a source of small transient stress, and by using sensitive ment after a theoretically infinite time. If the soil were considered
pick-ups they had recorded several cycles of ground oscillation as a spring, when a mass was mounted on it, it might be expected
at distances of up to 60 feet. that the frequency would drop with the increase in mass. He
They were also interested in the stress distribution under asked if that were the case or whether the fact that the mass
various loadings, and they were hoping in the future to carry was increased was offset by the greater volume of soil which
out comprehensive tests on the measurement of stress distribu- came into action.
tion under compacting machinery, some of which would include
the vibrating type. In a few prelimhary tests they had obtained Mr. S. S. WATTS,B.Sc. (Eng.), A.M.I.Mech.E., referring to
the distribution of vertical stress under a frog rammer impulsive Fig.
loading machine resting on homogeneous soil. The results were said it10, in which a number of natural frequencies were shown,
was not quite clear whether those were all due to a vertical
somewhat in agreement with the stress distribution shown in impulsive
the paper. Fox (1948)* had calculated stresses not only in a included. force or whether there were other modes of vibration
He asked whether other disturbances could have
homogeneous medium but also in a two-layer system represent- occurred when the disturbing force was applied which might
ing concrete resting on a soil foundation; the indication of hoop have caused other modes than the vertical mode.
stresses was present in both cases. In Table 1 the frequency was given in cycles per second, and
He had found the section of the paper dealing with high was not exactly implicit that those were all due to a vertical
frequency stress waves rather misleading, probably owing to it
simplification of the theory. The theory quoted referred to disturbance, although that was probably the case.
propagation along thin rods, so that propagation was mainly theIf low-frequency
the high frequency vibrations, which were superimposed on
oscillation, shown in Fig. 13 were due to
normal to the boundary. In practice the stress waves would be some high natural frequency in the machine, it was interesting
refracted as shown in Fig. 21 according to the relation to see that it repeated itself even in the disturbance of the ground
---
V1 02 and carried on for quite a considerable time, indicating that the
sin el - sin e2 ground was responding to that very high frequency. He asked
where 0, and e2 were the angles with the normal to the surface whether the authors had any explanation for that phenomenon.
in the two media and v1 and.v2 were the velocities in the mass
media. Mr. V. E. GOUGH,B.Sc. (Eng.), A.M.I.Mech.E., referred to
He did not know whether the authors had referred to the the rubber suspensions illustrated in the paper, and said that
theory for the propagation of waves across surfaces in extended it was possible to use sponge rubber, or a special mat develdped
media, which had been given by Boyle and Rawlinson (1928)t for the purpose, round the concrete block and to pour the con-
and which took into account the formation of standing waves in crete in. That had certain advantages compared with the indivi-
the second medium. dual mounting. Sponge rubber was an obvious replacement for
A method had been developed at the Road Research Labora- cork; a thin rubber mat with small cylinders on each of its faces
tory for determining the velocities of sound in mass media from could also be used.
tests on small samples (Jones 1949)$. Values for various con- He asked the authors whether the movements involved were
cretes, soils, and aggregates were available should the authors large enough to make the design of a direct damper on the actual
wish to extend their calculations to cover propagation of stress movements an easy problem, or whether it was a question of
waves in mass media. having a very small movement and requiring a very high damping
force. He felt that in the latter case it would be more practical
Mr. G. J. L. COWDELL, B.A., A.M.I.Mech.E., asked the to use a harmonic or a hysteresis type damper, and to use some
authors whether they had taken into account the data available of the foundation concrete as an auxiliary mass, with a hysteresis
* Fox, L. 1948 Road Research Technical Paper, No. 9, “Computa- andThe a spring element between the two foundations.
use of compaction by pile driving had been suggested as
tion of Traffic Stresses in a Simple Road Structure’’.
t BOYLE,R. W., and RAWLINSON, W. F. 1928 Roy. SOC.Can., a means of increasing the mass and therefore decreasing the
3rd series, vol. 22, p. 55. frequency; he thought that, with constant energy input, if the
$JONES, R. 1949 Magazine of Concrete Research, No. 2, p. 67, frequency were reduced the amplitude was automatically
“The Non-destructive Testing of Concrete”. increased in the case of a simple resonant system.

527
Communications
Professor R. N. ARNOLD,D.Sc., D.Eng., Ph.D., M.S., flotation, any of which might be capable of resonance. Such
M.I.Mech.E., wrote that he did not consider the paper a serious frequencies depended on the density and shape of the floating
contribution to the subject of foundation vibration. The authors’ object and objects of the same material and thickness would give
practical experience in suppressing such phenomena had ap- identical frequencies irrespective of size. The above was not
parently met with some success but it could not be inferred that intended to be a serious analogy of ground behaviour but the
they were thereby in a favourable position to expound the theory problems were not unrelated.
of ground vibration. If the authors had confined themselves to a Engineers would no doubt examine the conclusions to see if
description of their practical remedies, together with a sample they contairied any useful information regarding the suppression
of the main steps involved in their analysis, engineers would of excessive vibration in foundations. They were in general
have been able to assess the validity of their work. Instead they familiar with the phenomenon of resonance and did not require
presented pages of hypothetical theory, entirely unsupported to be reminded that it was something to be avoided. They would
by acceptable experimental evidence. If, however, the real prefer to know how it could be avoided and on that point the
function of the paper was to explain the “dynamic principles” authors’ paper seemed vague.
mentioned in the title, then that aspect must form the basis of
discussion. Mr. JAMES DUFF, BSc. (Sevenoaks), wrote that Timoshenko
The first part of the paper dealt almost exclusively with the (1934)t had shown that when a single impulse was applied to
experimental work of Degebo from which five diagrams were the ground, both the vertical and horizontal amplitudes fell off
reproduced. Those results had been misrepresented and the rapidly with depth. He himself had deduced that at a depth
authors had also perpetuated ideas already discredited. For greater than 100 feet those amplitudes became too s m a l l to
example, Fig. 3 was a complete travesty of the original diagram measure. He asked whether the self frequency increased
(Degebo 1933, Fig. 16). Precisely the same errors existed in indefinitely with depth or whether the figure of f25 per cent,
their previous paper (Crockett and Hammond 1948, Fig. 6). quoted by the authors, was an asymptotic limit.
Fig. 7, with the description given, was very misleading. The The authors had stated that a stratum depth of at least three
authors did not explain that the two curves were the result of times the bulb depth in homogeneous ground was necessary to
settlement during consecutive experiments on the same piece of ensure a reasonably constant self frequency. He did not know
ground under identical conditions. Moreover, the unit “Inertia what constituted homogeneity; but, presumably, the elastic
of unbalanced parts”, presumably indicating ft.2-lb. units, was constants should not show wide departures from average values.
in the original a function of mass x (frequency)*, namely, the He had examined the results given by Bergstrom and Linder-
holm (1946)$ and also some obtained at Sevenhampton. The
familiar ?u+ (lb.). Those curves had little scientific meaning Swedish tests had extended to a depth of 26 feet and the English
g
in the Degebo report and still less as presented by the authors. tests to 21 feet. Bergstrom and Linderholm had taken samples
The evidence for the existence of a ground weight Wg at various depths and determined values of E and G by various
required examination, since it occupied a key position in the static and dynamic methods. Of seventy-nine tests examined
authors’ theory. The obvious implication of Degebo’s inability only eleven gave acceptable values of Poisson’s ratio, and those
to measure W, was that it was non-existent. It was strange that values ranged between 0 and 0.49.
the spring force K 2 should be so accommodating as to vary In the English experiments, measurements of the horizontal
linearly with the fictitious W,. The statement “W, is always the and vertical displacements caused by a pile hammer were made
dominant controlling weight and the self-frequency of the system with a static weight vibrograph. It could be shown that the ratio
cannot be altered very much” implied (equation (2)) that WEwas of horizontal displacement/vertical displacement was a function
much greater than Wj+W,, but later a case was described in of Poisson’s ratio. A large number of values of Poisson’s ratio
which W’+ W, appeared to have a value of 1,570 tons while WE were obtained, but the variation was very great. He asked what
“probably amounted to a further 1,000 tons”. One could always effect the “homogeneity” of the ground would have on frequency.
choose a weight W, and spring force F2 to vibrate at a given The authors had mentioned that the unstable harmonic
frequency, but that did not mean that another complex elastic frequencies were non-linear, but he thought it was conceivable
system vibrating at the same frequency necessarily had any other that at some depth, or range of depths, the harmonic frequencies
factor in common with it. Simplicity of analogy was desirable might be linear.
but was only legitimate if the fundamental dynamics of the
problem remained identical. Mr. V. E. GOUGH, B.Sc. (Eng.), A.M.I.Mech.E., wrote that
He reserved judgement on whether the ground had or had it was clear from the paper that a simple mass-spring analogy
not a natural frequency of its own, for he had seen no experi- was inadequate for the solution of ground vibration problems,
mental evidence leading to such a conclusion. By assuming the and it appeared that the authors considered a closer approxi-
ground to be a semi-infinite homogeneous elastic solid, Reissner mation would be obtained if the system were assumed to consist
(1936)* showed that no discreet natural frequencies existed in of a very heavy spring element of the “active ground weight”
such a system. Reissner’s mathematical model might be open to Their argument was difficult to follow closely at that point.
criticism,but it was certainly nearer the truth than the assump- No method of making a practical estimate of W ‘ had been
tion that the static stress distribution in elastic spheres could given, save a few sketches in Fig. 5 which would permit wide
be used to study dynamic actions of the ground, subject to the variations in estimates by various designers, even in examples
introduction of a dubious volume called the “bulb of pressure”. illustrated. The fact that “Degebo’s attempts to measure W8
That the ground might possess no natural frequency of its own with an oscillator showed that average errors of 25,50, and even
was suggested by the action of a large stretch of water. A stone greater percentages occurred in the course of an experiment,
dropped in the centre of the lake sent out a train of surface though the self frequency remained about the same” challenged
ripples, If a cork floating nearby were used as a rudimentary the suggested theory and supported another, more likely, theory.
vibrograph its vertical motion would indicate a damped vibration The ground was capable of transmitting vibrations from a
of some frequency. That would not be the natural frequency of machine or to a machine, and was thus more nearly analogous
the water but would be a function of the size of the disturbance, to a transmission line or wave filter or a radiating field than
a larger stone producing a lower frequency. The cork, however, t TIMOSHENRO, S. 1934 “Theory of Elasticity,” p. 400 (McGraw
would possess natural frequencies due to its conditions of Hill, New York and London).
* RBISSNER,E. 1936 Inpnieure Archiv., vol. 7, p. 381. $ See list of references, p. 522.

528 COMMUNICATIONS ON T H E DYNAMIC P R I N C I P L E S OF MACHINE FOUNDATIONS
to a simple resonant system. The foundation and machine the ground had low frequencies if they were high and thin,’and
suspension device was then the coupling unit between machine higher frequencies if they were flat and broad.
and ground, and the design of a foundation resolved into the The ground was a spring with its own weight as shown in
problem of matching or “mis-matching” the characteristics at Fig. 4b; that subject had been treated thoroughly by Ehlers
various vibration frequencies of the source (the machine) and (1942)t,who had concluded that the ground (without damping)
the line (the ground). In the case of isolating a machine from behaved as a spring subjected to damping. A double spring-
ground vibration, the coupling unit (the foundation) design weight system had also been considered. Ehlers had applied his
again rested on the problem of matching or mis-matching the method to the foundation of a steam turbine and he had found
ground and the machine over a range of frequency. a good correlation between measurements and calculation.
In the latter case the foundation and machine mounting would He had previously pointed out that cork was a relatively rigid
have to be designed either to attenuate or to reflect back into the spring which was only suitable for use under a foundation of a
ground the frequencies present in the ground, especially those machine rotating at high speeds and situated on a rigid ground.
which would cause most response in the machine to be isolated.
In the former case the requirement to be attained was not quite Mr. J. H. A. CROCKETT, in reply, wrote that the trouble caused
so obvious. For a given machine, the most satisfactory results in nearby buildings and machines by the vibrations set up in the
might well be obtained if the energy of vibration was passed ground during pile driving lay outside the scope of the paper. A
into the ground and radiated away, if it were at a suitable steadily oscillating machine caused a steady train of waves,
frequency. That, however, would not be the solution for the mainly Rayleigh together probably with coupled* and gravity
neighbourhood. Alternatively, the design could be chosen to waves*, to move outwards from the source of disturbance. These
minimize the vibration power passed to the ground but that were all radial in their direction of movement, and somewhat
might have other undesired results, such as the machine like those occurring on the surface of water as a result of a stone
structure itself vibrating in resonance with the energy dissipation being dropped vertically into it; they were due to vertical move-
taking place by internal hysteresis of the machine members. ment at the middle. If a horizontal movement occurred simul-
The reason why the “self frequency remained about the same” tbeously, another wave, the Love wave, acting purely in
then appeared to be clear: the natural self frequency of the horizontal shear, also moved radially outwards but with a side-
ground was, in fact, a characteristic of the input impedance- ways action, rather like a wriggle. Those waves all took place in
frequency curve of the ground, and because it depended on the the surface layers only, within a depth of, say, 50-200 feet for
ground through which the vibrations passed, even if the relevant industrial machines; They formed a train of waves having, when
parts of the ground were some distance away, it remained un- initiated by a steadily running machine, exactly the same
altered. frequency as the machine ;such a train of waves was impressed
The use of piling or other means of consolidation of the upon the ground mechanism by the machine, while any natural
ground merely changed its characteristics at the coupling to oscillation of the ground and foundation mechanism itself
the vibrator. Tapered lines were well known in communication remained quiescent. That behaviour in itself was logical for any
circuits and could be used to obtain desired matching charac- complex oscillatory system, and an excellent demonstration that
teristics. Tapered lines and consolidated ground would appear the ground behaviour was not an incomprehensible muddle,
to have some analogy with each other. but obeyed proper laws. Any buildings or parts of buildings
A possible approach to a practical case would be to h d having the same natural frequency as the train of waves and
(a) the input impedance-frequency c u v e of the ground at the standing in its path would “build up” in impressed oscillation
proposed site (it would bear some relationship to Fig. 10) and into resonance, and it was that action which nearly always seemed
(b) the output-frequency cullre of the machine, and then to to be responsible for vibration troubles arising at a considerable
decide the overall characteristic of the system that had to be distance from the source. All loose parts like panels and window
met. The machine suspension and foundation could then be panes also built up clearly into a noisy rattle, alarming the
designed as a coupling unit to meet the foregoing requirements inhabitants, but often without serious effect upon the building.
taken as a complete system. Nevertheless, annoyance and discomfort, alarm, adverse effects
The resulting design might necessitate the foundation block on precision machinery or instruments, and possibly fatigue and
being fitted with hysteresis dampers with auxiliary masses, as other structural defects, could only be satisfactorily reduced by
well as the main foundation block. The design might well be balancing reciprocating parts of the trouble-causing machines
impracticable on cost or other grounds, in which case either more accurately (an expensive and often impossible method) or
the machine would have to be altered or the required charac- by an oscillation-controlled foundation (often but not always
teristic of performance would have to be relaxed, or in extreme expensive) or by mounting all buildings or their relevant parts, or
cases a new site, having suitable ground characteristics, chosen. all affected machines, etc., on oscillation-controlled foundations
That the above suggestion was worth investigation followed or resilients. Since the third method might involve several
from the great improvements which had been brought about in hundred buildings and machines in a populous district, it could
gramophone design when filter theory was applied to the be very expensive indeed. The crux of the problem was the steady
acoustical elements. frequency condition of the waves and oscillations.
The general theory that the ground was a form of transmission For pile driving, forging hammers, and other sudden impulse
line, wave filter, or a radiating field and that the foundation was cases, another factor-shock-was of equal importance, and
a coupling unit between machine and ground, should, he had an extra effect. A sudden shock given to any mechanism set
thought, rank with the authors’ principles stated at the beginning it into oscillation at its own natural frequency. Thus, any
of the paper. The two principles stated were the two extremes machine-foundation-ground system, or pile-ground system,
of the general case: under certain conditions a line might act when given a sudden impulse, was set oscillating at its natural
as a resonant circuit, and under other circumstances, when the frequency; some of the energy remained in the central system
frequencies were very high and the wave-lengths s m a l l compared during four or five cycles of oscillation, flowing outwards cycle
with the oscillating bodies, the behaviour of the line corresponded by cycle into the surrounding terrain, in the form of a train of
to the transmission of stress waves in bulk media. four or five waves, having the same frequency, though, of course,
each cycle became successively less in amplitude. There was an
Dr. E. RAUSCH(Essen) wrote that he considered that the additional action: on the blow being given, the whole mass of
ground under the machine foundation should be treated as a the system w a s accelerated to its maximum initial velocity in an
spring on which the foundation and the machine could oscillate extremely short time (a “shock” was experienced), after which
with self frequencies in six degrees. That had been confirmed it was free to oscillate naturally. Two important stress waves,
by measurements of ground oscillations made by Degebo on known as the P and S waves (pressure and shear respectively),
building machine foundations. From those results, the spring moved outwards in the ground, having thus a huge acceleration.
values of different grounds with different pressures had been
calculated. He had shown (Rausch 1943)* how those oscillations t EHLERS 1942 Beton und Eisen, p. 197, “Der Baugrund als Federung
could be treated in a simple manner. Foundations situated on in $schwingenden Systemen”.
. LEET,L. D. 1946 American Scientist, vol. 34, “Earth Motion
* See list of references, p. 522. from Atomic Bomb Test”.

COMMUNICATIONS O N THE D Y N A M I C P R I N C I P L E S OF M A C H I N E F O U N D A T I O N S 529
However, the duration of the blow was so brief that the during pile-driving operations it might, then the shocks could
material in the stress waves did not have time to move very far cause a “build up” swaying to a damaging extent. A good pile-
under its force before the wave had moved on elsewhere, so the driving engineer instructed the driver foreman to break the
wave could not carry with it much energy. Those shock waves timing sequence of the blows, as a marching column broke step
were said to be square-fronted, and their effect on any structure when crossing a bridge. That action was less important when
in their path was that of a sudden kick. They moved away from the structure was some distance from the centre. The third
the foundation or pile in all directions, being thus called “body method of failure occurring in structures from shock initiation
waves” and decreased in intensity with the square of the was rather similar to those from steadily reciprocating machines.
distance from their source, unlike the other mainly Rayleigh The two simultaneous trains of waves stretching outwards with
waves travelling within the “surface” layers, which decreased an aperiodic frequency, nevertheless were steady enough in
much less quickly. There was another complication arising from frequency to cause a resonant build up over several cycles.
the shock: the P and S waves gave rise to a train of the usual Usually near the centre there were up to six or eight cycles in the
(STEADY MACHINE AT
71 CYCLES PER SEC.)
m
SOME COMPONENTS IN
gESONANCE AT 7i CYCLES PER SEC.
GROUND FREQUENCY 15
N CYCLES PER SEC.
11
CYCLES PER SEC.
f--
N CYCLES PER SEC.)

(MOSTLY RAYLEIGH SURFACE WAVES
‘3
S O M E COMPONENTS IN
RESONANCE AT-! CYCLES PER SEC. HAMMER
N CYCLES PER SEC.)

(LITTLE SHOCK)
(5-6 RAYLEICH WAVES

N CYCLES PER SEC.1
:i i RAYLEIGH WAV
N CYCLES PER SEC.)
Fig. 27. The Two Kinds (Steady and Shock) of Industrial Ground Vibration Wave Trains,and their Effect
kinds of surface waves by means of Waction, reflection, and train with a more or less even single frequency; at a distance
refraction, described elsewhere*, which had a frequency much there were up to about twenty, though of lesser amplitude.
the same as the frequency of the central system. Structural failures brought about by the building up of those
There were, therefore, near the source of disturbance an initial few cycles could not amount to much, especially since the train
kick followed by two simultaneous trains of waves, but remote of waves only had an approximate and not a steady frequency,
from the centre there was only the train left. Farther away again but under constant repetition, over long periods, such failures
the train became too weak to be of much importance, and still might be expected, particularly by fatigue in low-quality work.
farther away it could be considered dead. This sequence and Two complications to this shock action were worth mentioning :
its ef€ects was shown well in Fig. 27. the initial shock or kick set the structure into motion at its
Damage to neighbouring structures could result from such a natural frequency, and if that frequency happened to coincide
ground movement in several ways, the first being from forces with that of the train of waves which followed the initial shock,
set up by the high initial acceleration kick, which had been found a much bigger build up was likely; secondly, all tall structures
to give a trigger action, initiating failures where there already acted as inverted pendulums, and so had a greater oscillatory
existed high internal stresses.* In another kind of failure where movement at the top than lower down, and thus making the
the internal stresses were low, the shock accelerations existed for accelerations greater at height, with quicker fatigue and other
such a short time that one, or even several, such shocks were not effects, like loose rattles, noise, and disturbance.
likely to cause any noticeable permanent movement of one small In short, he suggested that steadily oscillating machines could
part of a structure relative to another, as with brick to brick, or produce simple resonance up to a considerable distance around,
a bolt in its hole ;if the shock was repeated thousands of times, and possibly with fatigue effects; heavy impulse installations
the accumulation of minute movements could develop into a could give shock conditions leading to failures in nearby struc-
serious distortion. One of the authors when blasting had slowly tures, but also resonance at more remote places during a small
dislocated the top of a church spire over a period of several number of cycles, and there were intermediate effects in between.
months. Such failures occurred locally to the source of disturb- Without an extensive dynamic survey of a neighbourhood,
ance because the shock waves died out so rapidly. Another one could not predict in which of 400 or 500 buildings, or in
important type of damage could result from shock waves; the which part of a building, resonance troubles would occur. One
natural sideways or swaying frequency of a structure was often had to wait till they occurred and then deal with them, knowing
a b u t two or less a second, and if it coincided with the number a little of the physical mechanism, either by altering the
of shocks per second, as with very large forging hammers and frequencies impressed on to the ground, or by mounting the
* C R ~and ,HAMMOND, 1947 Inst. C.E. Structural Paper No. machine on resilients in order to decrease the impressed forces,
18, “Reduction of Ground Vibrations into S.tructures”. or by reducing the shock action-as with more resilient packing
j-CRANDELL, F. J. 1949 Jl. Boston SOC. Civ. Eng., U.S.A. in the pile helmet or a softer spring pad under an anvil.
34
530 C O M M U N I C A T I O N S ON T H E DYNAMIC P R I N C I P L E S O F M A C H I N E F O U N D A T I O N S
A possible explanation of Mr. Snow’s experience, in which a frequency of the train of Rayleigh waves was mostly measured ;
building had been damaged 200 yards away from pile driving, they were sufficiently similar to be taken as the same for engi-
while other nearer buildings remained unaffected, was that both neering purposes if the ground did not vary much.
the building and the pile driving were upon the same soft The oscillation principle of a double spring-weight system
stratum, while the intervening houses rested on a rather harder was mentioned extensively in engineering vibration text-books
stratum overlying it ;a house was said to have been demolished under the name of the “dynamic vibration absorber”, though
in that way in 1941 when a large forging hammer first started only two degrees of freedom were therein considered, where the
working. system had only two modes of oscillation and two natural fre-
Large forging hammers often produced a shock wave-or quencies when taken in the purely vertical sense (as usually
original acceleration-of about 0.001 second duration; Fig. 13 drawn). He was working upon that system in all its twelve degrees
illustrated such a case, though if a hammer was forging a large, of freedom where there were twenty natural frequencies and
soft, hot billet, the initial shock might easily become as long as modes, and the results were already applied and giving good
0-05 second. Work on the correlation of live load, dead load, and results. At least for large steadily running reciprocating machines,
soil characteristics, especially under oscillation conditions was it seemed to be a very powerful method of absorbing vibration,
being particularly examined by Bergstrom and others in Sweden and one which could be applied at least to some existing machines
for practical building and engineering purposes. without closing them down and completely rebuilding the
Mr. Chaddock’s gun base and the load applied to the ground foundation.
was interesting, but without more knowledge of gun-mounting Mr. Hvistendahl had asked whether any natural ground
design, the frequencies and the impulses given, he was unable frequencies existed within the important range of turbine fre-
to make suggestions. The paper stated that the natural frequency quencies. They had obtained apparent harmonics up to 120 cycles
of a single spring-weight system ought to be about one-third per second, and f i s t harmonics not infrequently had the same
that of the ground-foundation system in order (a) that there amplitude as the fundamental. Fig. 6, curve a, represented the
should be no resonance, with its accompanying soil failure, and settlement during a Degebo experiment, using the oscillator,
(b) that the live pressure on to the soil should be reduced; but and indicated the settlement after a large number of oscilla-
those required that the amplitude of the single spring-weight tions at rising increments of impressed frequency. The curve
should be large. If that amplitude had in any particular case to was probably not very accurate but was useful to engineers as an
be much smaller, either the weight had to be increased markedly, “indication” of what was likely to happen. When the weight of
or the natural frequency had to be increased to about three times a foundation was increased without its dimensions being altered
higher than that of the soil-foundation system, which, of course, in any way, as when the machine was erected upon it, the natural
increased the live load and the ground vibration again. The choice frequency was lowered ; that could easily be derected without
had to be made between two undesirable effects. However, since elaborate instruments. If the foundation weight were enlarged
a gun barrel moved already a considerable distance, and was fitted by increasing its dimensions below ground level, the natural
with controllable dampers, perhaps the gun mounting structures frequency could be expected’to alter, but to an extent it was
could have their natural frequencies reduced to say three or four probably impossible to predict, since not enough of the theory
cycles per second without uncontrollable liveliness. was known. Hence they insisted on repeated instrumental
Mr. Brown’s suggestions on soil stabilization around machine readings for all installations where accuracy was needed.
foundations ought to be borne in mind. Tschebotarioff had T o Mr. Watts they replied that the values of ground
quoted a remarkable case of resonance between a number of natural frequency in Table 1 were vertical. Andrews and
large water pumps and the ground ;the enormous build up had Crockett had found beat actions with some horizontal ones.
been much reduced by lowering the high water-table using a The high frequencies of Fig. 13 were not all the same (shown
continuously working well-point system all round the pump by counting them), but were all local natural frequencies inside
house. Perhaps soil stabilizing methods might vary a soil’s each of the three masses, those small oscillations being set into
natural frequency enough for some purposes, though there did motion by the initial shock wave used in that instance to promote
not seem to have been any research on the subject. the natural oscillation of the whole system.
Mr. S. Foster drew attention to a point not quite clear: that They agreed with Mr. Gough that sponge rubber, or
of a tension wave existing in a non-adhesive material like sand. mats of rubber with little rubber cylinders on each side, were
The sand was always precompressed by gravity. A tension wave good materials upon which a concrete block could be cast,
could exist in such a material up to the intensity of precompres- provided that the limitations of high frequency response was
sion. The algebraic s u m of the stresses of all stress waves passing remembered for such thin materials ;they were totally unsuited
at any moment was the actual stress. Cohesive soils could take for mounting machines having low frequencies. For some low-
more tension. frequency machines, like forging and sand-moulding equipment,
Mr. Hammond and he were indebted to Dr. Jones for the movement at high enough velocity and amplitude could often
reference to Boyle and Robinson’s work on refraction through be arranged so that an hydraulic damper like those illustrated
extensive boundaries rather than through the end of one thin in Fig. 20 could be used, but for higher-frequency oscillations
rod into another. The equivalent works for displacements were of small amplitudes, hysteresis damping seemed to be the only
still needed. Theory in the paper was necessarily very simplified practical solution, and there did not appear to be any engineer-
and short through lack of space; there was no menuon in it of ing material with a higher damping coefficientthan that of cork,
standing wave effects, nor of interference and distortion near with 15 per cent, which was rather low. A large concrete block
boundaries, which were important. certainly acted as a damper itself through dispersing stress
They had considered the extensive geophysical knowledge waves within its mass. Mr. Gough’s electrical analogy was
and experience of oil and other prospectors, referred to by excellent and seemed well worth further study, particularly the
h4r. Cowdell, but what happened at the centre of disturbance idea of the input impedence-frequency curve for the ground.
differed greatly from what happened remote from it ;their paper Electrical analogy had from time to time been applied to the
dealt almost exclusively with the former, which was not of subject; it was a most useful tool.
much interest to the geophysicist. All types of soil seemed to They were indebted to Dr. Rausch for his remarks, which
obey the general dynamic actions mentioned. They had used a confirmed their own experience, and for the reference to Ehlers’
simple vibrograph to measure the natural frequency of the work of 1942. Some variations of natural frequencies for tall
ground or of a foundation-ground system, and they usually and thin foundations and for low and broad ones appeared in
dropped a weight of several hundred pounds from 2 or 3 feet Dr. Rausch’s book on the subject, and answered Mr. Hvisten-
height to start the motion. It was necessary to impart the dahl’s query about loading a foundation, which was much the
energy in the weight very quickly to the ground, so some form same action as adding height to it.
of striking plate was needed, well bedded into the ground so as The variation of ground self-frequency of f25 per cent
to give a sharp and short impulse ;the bare ground was mostly mentioned was a rough judgement; they had no idea whether
much too soft. The natural ground frequency could be measured there were asymptotic limits in any way produced by the non-
close to the point of impact, but there was less damage to the linearity, but stated, in reply to Mr. Duff, that they were
instrument if it was placed some distance away, in which case the beginning to think that that subject could not be treated as

COMMUNICATIONS O N T H E DYNAMIC P R I N C I P L E S OF MACHINE FOUNDATIONS 531
though the ground was homogeneous with constant physical of natural frequencies might exist. They suggested that the
properties of elasticity, density, etc. other two constants also ought to be varied. If an impulse were
Professor Arnold had not realized that the object of the paper applied to the surface of a semi-infinite elastic homogeneous
was to put before engineers, for the fist time, a general picture solid at a dimensionless point, that place of application had no
of the subject, with plenty of references and an extensive though horizontal dimensions, and no variation existed vertically of any
incomplete bibliography of about a thousand works. The kind which might provide a vertical dimension :in that case there
majority of those works were not generally available to practicing did not exist the necessary conditions of three dimensions to
engineers;their examination by any designer would take far too form a mechanical system. As mentioned in the paper, no
much time; a general guide was needed through the great mass equations could be formulated to show any osciliatory action.
of material, and a judgement was required to help interpret it in A single vertical impulse produced in the semi-infinite solid a
the light of practical experience of many machine foundations pressure wave and a shear wave (both body waves) which then
designed according to those ideas. The paper was also written radiated outwards in all directions in straight lines through the
to show research workers where practical engineers need further body, from the initial point, provided the solid was homogeneous
theory from people skilled in producing theory, to help in the with constant physical properties. But in the ground those
practical job of producing good vibration-controlled foundations. waves were diffracted round and some even returned to the
Professor Arnold’s complaint that the paper was not a “serious surface again at some distance from their source, that action
contribution to the literature” of the subject of foundations was being due solely to increase in density, elastic and rigidity
thus unfounded. The greatest difficulty had always been the modulii with depth. The waves could also be returned to the
enormous amount of theorizing during the last century, and the surface by reflection and refraction at boundaries with other
small degree to which it bore relation to known and observable strata. Thus, that variation of properties might prove to indicate
phenomena. That state of affairs was of no use to the engineer, how P and S waves could induce specific oscillations.
who still had to design foundations for his client, and economical There was, however, a much larger type of wave initiated by
“doingyywas the most important object before the engineer. an impulse to the surface-the group known as surface waves,
The ideas, not theories, and particularly those about the active travelling radially outwards within a fairly short distance of the
ground weight, put forward by Ah.Hammond and himself were surface. The largest of them was the Rayleigh wave, rather like
those found to give a good code of practice ;no other code existed, the ripple or wave which occurred on the surface of water. Since
and no theorist had put forward anything covering the designer’s it had a vertical limiting dimension, and provided the impulse
field or even a substantial part of it. The paper was written to fill was taken to be applied over a finite surface area, the necessary
the gap, and not to present large amounts of experimental conditions with three dimensions existed from which oscillatory
evidence, for which reference could be made to the bibliography. equations could be produced. Sezawa and Kanai (1938) had
The two principles expounded by themselves did not seem shown that a Rayleigh wave could in certain circumstances
to have been mentioned before in this connexion and seemed to occur as a standing wave, with purely vertical motion. If there
be unassailable. The work by Degebo over a decade was still the could be a central portion having finite area or horizontal
leading and most practical research on the subject, and every dimensions, there would exist those three dimensions needed for
other leading research worker in the subject in the world, with an oscillatory mechanism, and Sezawa and Kanai had shown that
whom they had discussed the subject, had held it in high regard, the standing Rayleigh wave could exhibit both natural frequency
and considered much of the evidence useful. Fig. 6 was not taken and resonance. They themselves were strongly of the opinion
from Fig. 27, Degebo, vol. 1, 1933; it was in part taken from that the standing Rayleigh wave was most likely to be the
Fig. 26 and from other information, in order to give engineers dominant one in the natural frequency of the ground, and it was
an idea of what was likely to happen. Professor Arnold‘s remarks for those reasons that they had included the speed of travel of
about Pig. 7 were based on its form in the uncorrected proof. the Rayleigh wave in Table 1, showing that it was closely con-
They said as little as possible about that Degebo experiment nected with the natural frequency of the ground, and the allow-
owing to its difficulties, but included it because it seemed to be able bearing pressure. They would very much like to see Sezawa
important. Better information on the point would be accepted and Kanai’s work extended to include the effects of physical
gratefully. properties which increased in value with depth.
They did not agree that, because Degebo could only measure
WE inaccurately, W, could not exist; Degebo did measure Mr. ROLTHAMMOND wrote in reply that he was very glad to
it, very often. They were sure of its existence and so was have the criticism from Professor Arnold, whose familiarity with
every other practical engineer with experience of the subject. the work of Degebo gave him a great advantage over those who
The “1,000 tons” of W, under a machine and foundation had not studied it. He could only say that he had ample oppor-
W m +Wfof 1,570 tons was of the type shown in Fig. 5d, wherein tunim of seeing the practical results of Mr. Crockett’s work quite
much of W, was replaced by the deep foundation itself. The independently, and he could not believe that such excellent
bulb of pressure static distribution of stress idea was not in any results would have been achieved unless the working hypothesis
sense intended as a basis for studying dynamic actions in the adopted had been sound.
ground; that would be absurd. It was put forward as part of No soil was truly homogeneous, but every soil and stratum
the “practical working method within present industrial limita- appeared to have a dominant frequency, the true nature of which
tions”, as part of a code of practice, until a better method had had to be explained fully. That did not alter the fact that
proved itself. Engineers knew that resonance was to be avoided, Mr. Crockett had found a satisfactory practical and economical
but their attention had to be drawn to its possibility with the method of designing machine foundations.
ground. Since the fundamental dynamics of the problem were He himself had recently been studying the paper by S.
still unknown, no analogy could be very sound, but design Bergstrom and S. Lmderholm (,‘A Dynamic Method for
engineers certainly could use a code of practice based on tried Determining Average Elastic Properties of Surface Soil Layers”,
and proved methods which worked in practice. Royal Technical University, Stockholm, 1946), from which he
Professor Christopherson’s idea was most interesting. In 1936 noted the following :-
Reissner had shown that a “body wave” produced by an impact
on the surface of a semi-infinite homogeneous elastic solid could .
<‘. . control tests have demonstrated that the resonance
not set any particular natural frequency oscillations into motion, frequency was changed when the vibrator was supported on
but he had started with the assumptions of constant density, and a different soil. This shows that the resonance frequencies are
unvarying elastic and rigidity modulii. If that system were con- not determined by the natural vibration of the base plate of
sidered to represent the ground, not one of the assumptions was the vibrator, but rather that they are dependent on the
correct, and the result was completely at variance with what properties of the soil. There is no doubt that considerable
seemed to happen when the bare ground was given a direct further research is needed into the exact nature of ground
sudden impulse. The value of Professor Christopherson’s idea vibrations, because this will probably throw much new light
was that, when one of the constants was varied more nearly as on such important matters as the bearing strength of piles and
in the ground, it was seen at once that not only one but a series of concrete aircraft runways.”

The Dynamic Principles of Machine Foundations and Ground

Uploaded by

Copyright:

Available Formats

The Dynamic Principles of Machine Foundations and Ground

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Dynamic Principles of Machine Foundations and Ground

Uploaded by

Copyright:

Available Formats

512

The Dynamic Principles of Machine Foundations and

P R I N C I P L E S OF BEHAVIOUR OF V I B R A T I N G Table 1 shows some measured self-frequencies and corre-

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Weathered mottled sandstone . 32 1,641

i POWERTO modifications when the foundation is partly below ground level

Et Lz POWER FOR IDLING

Fig. 3. Loss of Machine Power to Ground, Particularly at

Terzaghi 1943). This active ground weight is some function of

Fig. 5. Proposed Method of Obtaining the Active Ground

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

times greater than the static elasticity (Crockett and Hammond t;

LARGE BASE AND BULB.

Fig. 6. Ground Settlement by Induced Vibration $1

Dynamic Ground Settlement. There is a more or less rapid

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Fig. 11. Double Spring-Weight

is shared according to the respective weights and spring

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

1 satisfactory behaviour is due to the disproportionate horizontal

Fig. 12. Displacement of a Double Spring-Weight System

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

on the mass centre, there may also be two centres of rotation,

simple harmonics, variation of damping, difference of phase,

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Fig. 19. Rubber Springs Fig. 20. Hydraulic Dampers

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Thus the natural frequencies were :-

Mr. D. M. CHADDOCK, M.Sc. (Eng.),M.I.Mech.E., saidthat

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

N CYCLES PER SEC.)

N CYCLES PER SEC.)

(5-6 RAYLEICH WAVES

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

Downloaded from pme.sagepub.com at PENNSYLVANIA STATE UNIV on May 10, 2016

You might also like