C.P. No.

905

MINISTRY OF AVIATION

AERONAUT/CAL RESEARCH COUNClL

CURRENT PAPERS

Stiffness, Damping and Creep

Properties of a Polyurethane
Foam Including the Effects of
Temperature and Humidity
by
D. 6. Poyen, H.N D

LONDON. HER MAJESTY’S STATIONERY OFFICE

1967

PRICE 7s 6d NET
 U.D.C. Noo. 678,664 : 621-496 : 539.531 : 534.372 : 539.434

C.P, No.9053
. ,b+pst 1965

STImSS, DAWING AND CREEP PROPERTIES OF A

FoLYURE!lXANE FOAM INCLUDING 'E-IX ?JFFECTSOF
!lTWXRA!IURF AND HUldIDITY

D. B. Payen, H.N.D.

Experiments have been made to establish the physical propsrtjei of a

flexible polyurethane fosm plastic used in the oonstruckion of aeroelastio
models f3r use in low-speed wind tunnels. Simple statio and dynnnio beniL1ng
tests mere made on a composite cantilever of plastic foam and a light alloy
bar. It is shown that dynatio stiffness and damping capacity measurements
must take account of temperature and absolute humic?ity, end that the dynsmic
stiffness was nppmxinatcly double the statlo value over the very limited
frequenoy rang0 consi&m3d. Other phenomena such as creep and 'do effect
of 'working' are discussed.

* Replaces R.A.C. Technical Report No.65198 - n.s.c.27537

 c01mNTs w

INTRODUCTION 3
SYNOPSIS OF TESTS 4
2.1 Creep test 4
2.2 Static stiffness tests 5
2.3 Dynamic stiffness tests 5
2.4 Damping tests 6

TEST EQUIPMENT 6

3.1 Specimen 6

3.2 Static aeflection measurement apparatus 7

3.3 Dynamic test apparatus 7
3.4 Humidity chambers 7
FXPZUMENTAL PROCETRJ3 AND hWl'HODS OF ANAtYSIS 8

4.1 Static stiffness analysis a

4.2 Dynamic stiffness analysis a

4.3 Damping tests 9

PRESENTATION OF RESULTS II

5.1 Creep 11

5.2 The effect of 'ww,rldng' the foam 12

5.3 Dynamic stiffness and the effect of humidity 12
5.4 Damping chsracterlstics of the foan and the effects of
humidity and temperature 13
AN EXPLANATION OF THE MFZHANICAL iWMJIOUR OF TIiX FOAM IN
TEFXS OF ITS COMPOSITION AND STRUCTURE 13
CONCLUDING FUMRKS 45
ACKNOWLEDGEMENCS 16

Appendix A Derivation of formula used to calculate the dynsmlc

stiffness of the composite cantilever q7
Table 1 Dynamic stiffness tests on a duralumln bar 19
Table 2 Recovery effect 20
Table 3 Effect of 'working' 20
Table 4 Dyne&c stiffness tests on composite beam 21
Table 5 Form of damping in the composite beam 22
Symbols 23
References 24
Illustrations Figures 1-14
Detachable abstract cards
 3

1 INTRODUC~ON

The advent of high speed aircraft has crested a nced for fresh
techniques in aeroelastic modelling. In particular the chordwise flexibility
of low-aspect-ratio wings cannot be represented adequately by the single-spar
or torsion-box type of construction usad extensively in the past.

A form of flexible construction for low-speed wind-tunnel models is to

make a plane network of interconnected metal becms with weights attached,
which represents the required stiffness and mass distribution. Around this
skeleton, a light expanded polyuretiiane plastic, hereafter called foam, is
cast to provide the desired aerodynamic profile. Ideally the foam should
make a negligible oontribution'to.the elastic properties of the completed -
model, but in practical applications it is found& be signifioant.
_ ..
NC relevant information regarding the stiffness properties of foam is
available, but Redshaw' has shove that the static and dynamic stiffnesses of
solid xylonita plastio differ, and that the static value is affected by
humidity. These peculiarities of-a related material indicated tho need to
examine the stiffness oharaoteristics of the foam,
The investigaXcn comprised static stiffness, dynamic stiffness s.nc
damping teats, made on a composite cantilever of low density foam reinforced
with a central duralumin bar. The static tests were simple bending
and were extended to examine qualitatively th e creep properties of the foam
and the effect of systematic squeesmg or 'working'. The dynsmio experiments
were more involved and provided a convenient means to examine the effcot of
humidity and temperature on the stiffness and damping in the foam, An
oscillatory technique was used where the cantilever was vibrated and its
response monitored with a proximity transducer. For the stiffness tests the
oscillation was forced and the resonant condition established. Damping was
measured by the direct comparison of the res~nse in free vibration with a
known electrical signal of controllable exponential aeosy. The results were
analysed on the assumption that the cantilever could be treated by engineering
beam theory and its stiffness is used to determine an effective Young's modulus
of the foam.

In Seoticn 5 the detailed observations of the experiments are presented.

It is shown that there is a large variation of d.ynamio stiffness and damping
with normal day to day changes in temperature and humidity. The dynamic
stiffness is almost double the stetio stiffness in states of comparable
humidity, and the initial static stiffness may be reduced considerably by
 'working' the foam without causing appreciable permanent deformation. The
damping in ths foam is found to be hysteretic. An explsnation of these
findings in terms of the chemical oonposition and structure of the foam, is
given in Section 6.
2 SYNOPSIS OF TESTS
A broad outline of the investigations is presented in the table below
followed by sunnnaries of test procedure3 which show the state and environment
of the specimen at each stage of the experiment,

Table of tests

Environment Remarks
Type Breakdown of specimen
1
I I Creep w open Numerous checks'
laboratory on H & T

Static '-Effect of working u + w Open

laboratory
2 Stiffness
bi-Values measured W Open H&T
laboratory measured
Controlled
chamber H&T

Test completed

Key:- W = worked; U = unxorked; H = humidity; T = temperature

2.1 creep test

The oreep* investigation was qualitative and simply entailed loading the
oomposite cantilever specimen and noting its deflection at set interval3 there-
after until sll motion had virtually ceased. The load was then removed and
deflection3 measured as before.
* The oreep property of the foam is probably anelastd, i.e. response to stress
is diffusion-controlled and molooular orientation is mduced by shear. The
configurational entropy, or molecular disorder, is thus decreased 30 that release
of the stress,reaults in an attempt by the molecule3 to restore maldmum entropy
and hence produces a definlte,though low, restoring force, with no permanent set.
2.2 Static stiffness tests
(a) Effect of 'worlang

Simple experiments indicated that the stiffness of the foam could be

reduced by squeezing or manipulating it between the fingers and thumb. This
process covering tha whole of its surface is herein referred to as 'working'.

A stiffness value for the cantilever was found by evaluating the average
slope of a load agcunat deflection graph. This plot was obtained by loading
the beam in equal increments and measuring the deflection 30 seconds after
each increment. In this way, the stiffness of 'unworked' foam and foam after
a number of successive 'workings' were compared.

Although allowances for humidity and temperature ware not included at

this stage, it is clear that the effeot of working is substantial.

(b) Measurement of stiffness

It was difficult lo assess a true value for the static stiffness

because of the rapid weep of the foam. However, the folloning procedure
was adopted. A single load was applzed to the beam and the deflcotion was
measured after thirty minutes - a period long enough to allow most of the
creep to take place. The response was then assumed to be lineer* with the
0
load applied.

2.3 Dynamic stiffness tests

In experiments to determine the dynamic stiffness of the foam the beam

was mounted ver'.ically with the mass above the support and harmonically
forced by an electromagnetic exciter. By adJustine the excitation frequency
until the displacement response under constant input power zas a maximum,
the resonant frequency of the bcem was established and its rigidity deduced.

(a) Effect of humidity end temperature

The composite cantilever was mounted in a closed chamber described in

Section 3.4(i), and its stiffness variation with temperature examined while
maintarning a constant relative humidity. This was repeated for a number
of humidity values covering a wide range of conditions.

(b) Frequenoy effect

An attempt was made to examine the variation of stiffness

_ with
frequency by varying the concentrated mass attached to the end of the beam.
* Plastics and polymers are materials that obey a generalised Hooke's La& i,e.
the total effect on a strain-time cirve of doubling the stress throughout the
test is to double the &rain throughout without changing the time scale.
6

The range of frequency was however, limited to 2.54 c/s to 6.07 c/s because of
the so called pendulum stiffness incurred by the mass.

2.4 Damping tests

Damping measurement was made by csmparing the response 'envelope' of the

beam in free vibration with the curve produced by the exponential decay of
potential across a capacitor discharging through a resistance. The procedure
used is detailed in Section 4.3.

(a) Nature of damping

The form of damping present in the foam was found by examination of the
variation of damping force n%th frequency in conditions of constant temperature
and humidity.

(b) Effect of temperature and humidity

For this test the cantilever was mounted inths humidity box described
in Section 3.4(ii), and the enolosed air dried as much as possible with
sulphuric acid. Damping and humidity measurements were then made at regular
intervals as the humidity slowly varied with amblent temperature and gradual
dilution of the acid. The moisture content in the ai- olowly rose over
seyersl days since water was always being introduced via the wet bulb of a
psychrometer. A smell number of tests were made at higher temperatures with
the aid of an electrically heated wire.

3 TESTEQUIPMZYT

3.1 Specimen
Two main considerations influenced the des?gn of the specimen cantilever.
Some form of reinforcement was necessary to minimise static droop, since the
strength-weight ratio of the foam is very lam and the length of the beam had
to be such that deflection due to shear and end fixing effects wore negligible.
Secondly the stiffness contribution of the foam had to be sufficiently large
compared to that of the reinforcement to ensure maximum accuracy of results.
A stiffness ratio of foam to reinforcement of about 3-1 was achieved.

The composite cantilever is illustrated in Fig.1, and was made by glueing

two strips of foam of 9 lb/ft3 density, on either side of a flat duralumin
bar. These strips were cut from a block about 7/B inch thick and stuck to
the metal by roughening the surfaces to be glued with glass cloth and applying
black Bostik compound. A good bond was produced which lasted throughout the
tests. The metal bar was clamped securely at one end between two steel blocks,
 7

and one end of each focus strip glued to its respective block rrhilst the beam
WBSunloaded. Provision for loading and adding mass to the bar was mada at
the free end.

3.2 ' Static deflection measurement apparatus

-
The end deflection of the cantilever was measured using sn optical
system that enabled a reading to be taken by simply viewing through a -.
telescope the reflection in a small mirror of a concave illurmnatod scale.
The'mirrcr ws.s mounted on a tripod, one log being located in a dimple
drilled into the head of a small bolt attached to the loading point on the
duralumin bar, and the other two legs on a flat horizontal platform held
nearby. Any deflection of the bsr thereby caused the mirror to rotate.

3.3 Dynsmio test apparatus

To test the plastic under dynsniic loading the composite cantilever was
oscillated by an electromagnetic exciter s-d its response measured rJlth e
Wayne Kerr vibration meter employing a capacity pick-up. The light weight
coil of the exciter was attached tc the free end of the beam and a small
disc of tin foil glued to the plastic surface provided an earth Plato for the
pick-up. The arrangement is shovn in Fig.2.

3.4 --Humidity chambers

(i) Humidity chamber for stiffness experiments

Stiffness experiments in humidity and tempcrnture controlled conditions
were made with the specsmen mounted in a closed chamber, which was of massive
construction and heavily lagged. A fen heater snd refrigeretion unit were
incorporated together with a number of thcmocouples sited to provide n
check on the uniformity of temperature in the Iworking section. In order to
measure relative humidity a hair hygrometer was used which was later
calibrated against an aspirated Ret and dry bulb psychrometer. NC direct
humidity,oontrcl was available and the use of saturated salt solutions proved
to be too slow for convenience. However, by careful mcnipulation of the
heater fan and refrigerator the required conditions were obtained and since
the insulation was efficient the conditions lasted long enough for tie'
specimen to tso&' and a test to be conducted. For control at extremes of
hwsid.ity,the introduction of steam in one case,or the use of a drying agent
at the other,was found to be helpful.
 (ii) Humidity box for damping experiments

A wooden box lxxd. with heavy gauge polythene and measuring approximately
20 inch x 14 inch x 14 inch was used to obtain the required atmospherio
condition. A small fan was incorporated, and the moisture content of the
enclosed space was controlled using concentrated sulphuric acid in a glass
container, the lid of which could be raised at will.

Humdity measurement was made with an aspIrated met and dry bulb
psyohrometcr that employed thermooouples. This instrument, whxh was found
to be sensitive snd simple to operate, had been developed by A.O.R.E.,
Farnborough.

4 EXPERIMENTAL PROCEDUREAND XETHODS OF ANALYSIS

The aim in all the stiffness tests vws to establish the rigidity of the
composite beam (EI), and hence an effective Young's modulus for the foam in a
known condition:-

(EI)c - @IId
Ef =
If

The suffixes 'f' and 'd' refer to fosm and dursiumin respectively,
The rigidity of the durslumin bar was determined by static and dynsmic
measurements before it was built into the specimen.

The experimental procedwes employed to obtain the stiffness of the

specimen are described in Section 2.

4.1 Static stiffness anal.ysis

From the measured stiffness of the cantilever the rxgidity eras calculated
using simple beam theory, viz:-

where S is the direct cantdever stiffness for tip loading.

It can be shown that, for small deflections, the effect of antiolastic

curvature restraint in the duralumin bar can be neglected,

4.2 Dynamio 3tdYness anal.ysis

The dynamic rigidity of the specimen wa8 found from the f'ollowlng
relationship which is derived in Appendix A for a cantilever held vertically
with its mass below the support.
 9

w is assumed equal to the measured resonant frequency of the beam, and the
sign of the terms in g is reversed when it is hold with its mass above the
support.
This formula was derived by assuming that the defleotion mode of the
vibrating beam is the same as the static deflection curve of a weightless
uniform cantilever with a concentrated load at its end. The tqrm Oe7502b$eK?
T
is an allowance made for the rotary inertia of the exciter coil. An exact t
solution for the fundamental frequency of a horizontal cantilever of uniformly
distributed mass with a'load at its end has been given by Timoshenko3. The
frequency given by the above formula in the absence of gravity (beam in
horizontal position) is only 1.5% higher than the exact solution for the
extreme case of the end load equal to zero. As a check on the method, the
above procedure was performed on a durelumin bar,and the results, shown in
Tables l(a) and l(b),were very‘oonsistent for the beam held in all attitudes
and at all tested frequenoios. Furthermore th8 average EI given by this
technique was only 1-G lower than that obtained by statio measurement.
4.3 Damping tests
Measurements of damping of the composite cantilaver vibrating freely in
the fundamental mode were accomplished by the procedure outlined in Section 2.4.
The beam was inilially displaced onto a stop by passing direct current through
the coil of the electromagnetic exciter attached to its free end, and its
deflection and subsequent oscillation monitored with the vibration meter.
(See Section 3.3.) Air damping in the exciter was minimised by moving the
magnetic core just out of the aoil. (See Fig.2.)
The circuit used to set the beam in motion and enable a measurement to
be made is shown in Fig.J(a). On olosing the hand-oporated micro-switch the
capacitor charged up, the ooil was energised and the beam displaced. On
opening it the capacitor discharged through th8 resistance B and the beam ma8
released and performed damped oscillations about the datim position; both
processes beginning at the some time. The voltage appearing across the
capacitor ma resistance R was applied to one Y amplifier of a high input
impedance double beam oscilloscope, while the attenuated signal from the
vibration meter was applied to the other. By adjusting the controls on the
IO

o3oillosoope .3.nd potentiometer F the tao traces wore made colnciaent, both

when the beam was at rest in the datum positlon and the capacitor discharged,
end when the beam was filly displaced and the capacitor charged. Thus tiile
the beem was osoillatlng the pattern produced on the face of the tube was
similar to those of Fig.j(b). Damping measurement was made by varywg the
resistance R and shaping the Y2 curve until it touched the wavy Y, curve Eft
points m,, m2, m3, . . . . and reading the final value of R off a calibrated
scde.

The damping coefficients were evaluated by identifying the expression

for the decay envelope of a single-degree-of-freedom system with that governing
the decay of potential of a capacitor discharging through a resistance on the
assumption that damping may be represented by velocity Seependent forces, i.e.

T denotes the time constant of the capacitance resistance circuit and is.
equal to the product of R and the capacitance, and b 1s the ratio of the actual
danping force coefficient C and the critical value for the system given by,

C = 2 sk
crit II-

where a-and k are the generalised mass and stiffness respcctlwly. A0 andE
are arbitrary constants indicating inltlal quantities and must be equal. Hei&

b=&. (1)

Because in the cases consi&red the difference between damped and undamped
resonant frequencies is negligible, we have, by definition,

so that,

If
it is found that the damping is hysteretic rather than viscous, the
concept of equivalent viscous danping may be used, The energy dissipated per
cycle of hysteresis for most engineering mxtcrijLs is proportional. to the
square of the amplituae;
so that,

(E.D~)~ 7 me* . (4)

The energy dissipated'per cycle by viscous damping at resonance is,

@.D& = xwce2. (5)

Equating (4) and (5) the equivalent visccus damping coefficient is deduced as,

Cequiv =-=- K h
(6)
m u

nhere h is the hysteretic constant for the material. By substitution for C

frcm (3) in (6) we get:-

h=y. (7)

-In tests where the concentrated mass at the end of the beam was varied
while the stiffness was kept constant it was ecnvenient to expass equations ~
(3) and (7) in the following forma. '

c+ (8)

5 PRESENTATION OFRESULTS
--
5.1 Crsep
A typical delayed strain curve is shown in Fig.4, nhich demonstrates
that even though a state of quasi-steady equilibrium is reached after thirty
minutes, movement continues slowly. Growth of deflection was never observed
to have ceased despite waiting for several hours on one occasion. There is
evidence that the creep property is anelastic in character since a high degree
of symmetry exists between the load on and load off cases and there is
ultimately very little permanent set. (See footnote of Section 2.1.)
12

52 The effect of 'working' the foam

The procedure dessribed briefly in Section 2.2 yielded an interesting

result, nsmely that slxost inmediately after the foam was 'worked' It
recovered its
overall dimensions, but took several hours to reach a steady
3tifie33. This is shown in Table 2. There is however, a permanent effect
of successive 'workings' indioated by Table 3 where the foam is gradually
weakened until a. minimum stiffness is reached that is approximately 65$ p?
the 'unworked' state.
No obvious deterioration of the foam was observed as a consequence of
this harsh treatment.

In Fig.5 a typical load-deflection curve for the beam is dra.rm which \

exhibits considerable hysteresis. Nevertheless tne curves are sufficiently
linear over their major Fortions to give a relisble average stiffness value.
5.3 Dvnamio stiffness and the effect of humidity
The variation of the dynamic stiffness of the foam with temperature and
relative humidity is plotted as a family of curves in Fig.6. For a given
relative humitity there is a 32&l amount of scatter, but the variation of
stiffness with humidity and temperature is so great that this becoines
insignificant. The possibility of the actual amount of water vapovur U-I the
air affecting the stiffness was investigated by replotting Fig.6 in terms of
absolute humidity in Fig.70. The coalesccnoe of the several curves of Pig.6
into a single characteristic indicate3 that the prevailing absolute humidity
is a major faotor m determining the stiffness of the foam. In Gt. Britain
the normal. variation of dbsolute hunidiiy is from P5 x IO -4 lb/ft' to
8-O x IO -' lb/ft3, so that a 3C?<change in stiffness could easily occur.
Fig.8 is also of interest since it shows that the variation of stiffness with
the reciprocal of absolute humidity is fairly linear. When large changes in
the air condition were made during the tests, the bearr, assumed a constant '
resonant frequency in 15 to 30 minutes, thus illustrating the rapid response
of the foam to changing climate.

Static and *nsmio stiffness measurements are compared in Figs.8 and 9.

In spite of some scatter the curves show that the effective Young's modulus
of foam under dynamic loading is nearly double that obtained by static
measurement. The range of humidity'covered was limited by laboratory
conditions, but the indication is that moisture has less effect on the static
than on the dynamic stiffness.
 Table 4 shows that ths variation of dynamio stiffness with fYequency is
insignificant for the limited range considered.
5.4 Damping characteristics of the foam and the effects of
humidity and temperature
The damping capacity of the composite beam and hence-the foaml'ws found
to be predominantly hysteretic. The results of three wpsrato tests sre
shown in Table 5, where the variation in hjrsteretic constant with frequency
is negligible in steady conditions of humidity and temperature. In using
equation (9) the generalised stiffness was assumed to remcin conatsnt snd
assigned U=I arbitrary value.
The effects of absolute humidity and temperature on the equivalent
viscous danping ratio rind hysteretic constant are illustrated in Figs.10 and
II, where each symbol represents moa~urenents made within the temperature
range indicated. Although there is scme overlap of symbols tha influence of
temperature on the damping is evident espeoially at the drier end of the scale.
The points marked by asterisks were obtained in an additional test lvhere the
temperature was maintained at 31.5'C. These have far less scsttcr snd
confirm the trend. The curves are limited at the higher humidities by the
onset of condensation in the foam.
The essential differance in chsracter of Figs.10 and 11 is due to the
variation in stiffness of the foam dth humidity and the accompanying change
in resonant frequency of the boem. From Fig.11 the hysteretic constant could
change by a factor of two in normal laboratory conditions, and because the air
tends to be &Le.-t when it 1s coldest a largo proportion of this change could
easily occur in * dey. This figure has an arbitrary scale since the
generalised mass in equation (7) was assumed to remain constant and given s
convenient value.

6 JN EXPLANATION OF THE hfECEANIC~ EEHAVIOUR OF THE FOAMIN l'E?X.i OF

ITS COMPOSITIONAND STRUCTURE
A simplified description of the chemistry of flexible polyurethane6 Nil1
facilitate further discussion. Basioslly the material is made by the reabtion
of sn alkyd resin with a liquid diacyanate, in which the evolution of
carbon dioxide and cross-linking of polymer chains occur simultaneously,
generating innumerable bubbles of gas within a gelling mass, to produce a
foamed structure.. Normally this structure is amorphous, but whwa it
crystalline alkyd is used as one of tha ingredients a tendency toward
crystallisation may appear in the final polyurethane. The moleculsr
14

composition is one of long polymer chains having a carbon-carbon backbone with

considerable freedom of movement at each individusl carbon-carbon bond, these'
chzLns being tied to each other at infrequent intervals by primary valence bonds
or cross links.

The microscopic structure of the foam is sh&vn by a photograph 1.1 Fig.12.

It consists of small flat-sided cells approximately one millimetre across.
Most of the oell walls, however, are punctured so that cell cavities inter-
oomnunicate; this being typicel of low density flexible polyurethanes.
Clearly moisture vapour can penetrate to all parts of the foam, where it will
exercise a plasticising action, that is it reduces the Van der Wads molecular
forces between adjacent polymer chains, with the net result thet some stiffness
is lost. The effect of plssticization has been demonstrated5 in a study of
the widely used 66 nylon &ich is chemically similar tc the foam. A diagram
taken from this work is shown in Fig.13 and indicates that a S$ water addition
reduces the stiffness by about a quarter.

The load carrying members in polyurethane foam are essentially the edges
of each cell,and these-normally termed fibril columns - can buckle and yield in
much the ssme way as conventionsl struts and ties. During the investigation
into the creep characteristic,s of the foam it was observed that the growth of
deflection with time was not continuous, but occurred in sudden Jerks which in
some cases caused the beam to oscillate momentarily. This implies that there
wa3 rapid yielding of parts of the structure at different time intervals.
Now if we suppose, that even under smsll loads there sre many compressive
elements in the plastic at or near the point of buckling, then a small
reduction in their rigidity would lead to a far greater reduction In stiffness
of the overall structure. Furthermore this stiffness reduction would depend
on the degree of plasticization in the plastic, which in steady corditions
will be proportional to the vapour concentration of the air passing through it.
The stiffness under prolonged loading however, may no+ be greatly affected by
humidity because of the slow inevitable buckling and yielding of flbril columns
8s stress relaxation occurs mthepolymer; a process that will happen in a
random fashion depending on &he geometry of the columns within the cellular
structure.

Tests showed that when the plastic was 'worked' there was recovery of
stiffness overnight, and the initial stiffness was never fully regained.
Moreover after several treatments a steady stiffness value was achieved. One
may conclude that successive 'workings' lead to permanent damage of the cell
 15

walls or very weak flbrils, tith consequential loss of efficlenoy. The

recovery effect however, is probably due to the temporary destruotion of the
stiffer crystalline regions withln the polymer, which arc restored over a rG3t'
period of several hours.

Damping or the dissipation of energy on displaocmcnt is associated with

molecular activity in the foam; the greater the activity the less the energy
loss. Above a so called glass transition temperature Tg i&ividual etoms
in the polymar chains rotate andtranslate continuously, this motion increasxng
with rising temperature. At Tg molecular motion is very small, and rapidly
deoreases firther 8s the temperature is lowered and the Van der Wads forces
between adjacent chains become dominant.

In the case of polyurethane foams at temperatures between 15" and 3O'C

the plastic is above its Tg of about -l5'C, but veil below the temperature GZ
which rapid and complete elastic behaviour is obtained, this normally being
80°C or more. Thus slow elastic response with oonsidarable hysteresis can be
expected.
Water vapour, aoting as G plasticizer has the effect of lowering Tg, or
conversely increasing molecular mobility at 8 given temperature.

7 CONCLUDING REMARKS

The dynamio stiffness of the foam is greatly affected by variation of

absolute humidity, and is approximately double the static value. The stiff-
ness may be considerably reduced in the first instance by 'working' or squceeing
it without causing obvious damage.
Temperature and humdity markedly influence tho danping oapacity,whxch
is hysteretic.

In view of the above properties it 1s desirable that, in eeroelastio

models of the metal frame - foon type of construction, the fosm's contribution
to the overall stiffness and danping should be kept small. This may be
di.fficult to achieve howvex-, beoause large stresses con arise in the foam
due to its interaction with support members and other components in the model,
especially where higher modes of deformation occur. An dternntive nould be
to seal the foam from the atmosphere by coating it with a durable film or skin.
No orgsnw skin of this nature oould stop the flow of air entirely, but it may
slow it down to such an extent that the variation in stiffhess of the mcdel
during a test would be acceptable,
 The qualitative properties of the foam described in this Report
are broadly applicable to foams of various densities, but precise values
relate to foam of 9 Ib/ft3 density.

a ACKIQUDGEKSNTS

The author would like to acknowledge the help gzven by Hr. L.N. Phillips
of CPM Department, R.A.E. with the chemical aspects of the foam's behaviour.
 Appendix A

DERIVATION OF FORMULA USED TO CALCULATE T!B DYIUMIC

STIFFNXiS OF 'IHE COMPOSITE CANTILEVER

Taking axes as shown in Fig.14, and assuming that the deflection mode
of the vibrating beam is the seme as the statiqdefleotlon curve of a weightless
cantilever with a concentrated load at its end, we have

EIx =
Me2
--- M z3
2 6
or

modex = (Jr12- q3) es,

Potential enera

Length of bar from 0 to q& at max displacement is

4-
e = dz = e dr,.
11
0 0

Assuming small oscillations

e = q e + (6q3 - 4+q4+ o.9q5) q2 4..

rl
Length of bar from OtoqC at zero displacement is q.5 so that height that
element of mass e dqp is raised is

h = 8(61-,~ - 4.5q4 + o.9Tj5) q2 .

rl

Let &-,o be distance of 0 to M along z axis. Then pdtentlel energy of

distributed mass is
'10
(P.E.),, = pe’gq2 (671~ - 434 + o.gq5) aq
J
0
18 Appendix A

Strain energy in mode 1s

Work done by concentrated mass is

(P.E.)td = Mg8 (6$ - 4.511: + o.vq;) q2 .

For small q, ~)o cz 1, and then

Total P.E. = + 2*4Mgt. + 0-75p~~~

1 q2 .

‘10

J&T -2
(3q2 - q3)2 aq + 2Me2q2
2 q
J
0

= [P4715pC3 + Zd tT2] t2 .

Hence applymg Lagrange's equation we have

(0.4715pe3 + 2?d
e’) 3 + F + 2*4hIge
L-
+ o.75pe2g
1 q = 0

so that the undamped natural frequency is given by

=-I = 6~1 + 2.4htge2 + 0.75&

w2 9 0.4715&I e4 + 2M e3
putting h zM we get
Pe

EI = M';w2 [, +
3
 Table IA

DYNAMIC STIFFNESS TESTS ON A lXJF?ATAJE~IN

BAR

1 Mass below support

Mass attached to Resonant frequency Rigidity of b$m
bean grams a cyo le S/SC 0 EDID lb in
. .
3.925 . - . 6.43 IO*25
3.673 5.48 IO.28
4-w 5.26 IO-31
5.129 4.95 IO-29
8.747 4-13 -10*,27
14.578 3.41 IO.26
24.416 2.80 40.26

3
44.085 2.25 10.25

Table IB

Mass abova support

/
Mass attached tc Resonant frequency ?.igsity of bea
beam gram.2 cycle s/se c EDID lb in2

1.925 6.18‘ IO*21

3.670 5.21 40.27
4.265 4.97 IO.30

1
5.516 4.54 IO.28
8-894 3-76 IO.30
45.284 1 2.93 10.27
24.699 2.27 lo.27
44.178 1.57 IO-26

Measured static rlgimty EDID = IO-42 lb in2

20

Table 2
RECOVSRYEFFECT

Worked and tested

after:-
I -
10 minute5 37.6
42.4
One night 50.5
t

Table 3
EFFECT.OF 'P~ORKING'

Initial working 50.5

2nd - . -'I_ 49’8
3rd v 43-a
4th It 44.4
 Table 4
DYNAMIC STIFFNESS TESTS ON COMPOSITEBEAM
A SET OF TYPICAL RESULTS
FOAMIN THOROUGHLY
VORKEDCONDIl'TON

Mass below support

Mass attached to Resonant frequency Yo!x-@'s mo@jl!k?
beam

5’241
a-617
grms cyclerr/sec

6.07
5.53
I Ef lb/in

63.5
64.1.
18.181 4-69 66.0

I 47.573
66.761
28.094
57.2~6
37-a 3.02
3.43
4'12
3.21
3.72 66.5
66.4
66.7
65.5

Ma33 above support

5.241 5.85 64-9
8.617 5.33 65-O
18.181 4.38 66-2
28.094 3.77 67.1
37.846 3.33 66.4
47-573 3.01 66.3
57.2% 2.75 65.9
66.761 2.54 65.2
22

Table 5

FORM OF DAMPING IN THE COMPOSI!IX BEAM

6*3? 21.5 22.50 ' 3.96 0.612

6-63 21-a 12-7-l 4.52 Q-620
6.54 21.6 2-35 5.45 o-643
6.54 21.6 0 5.80 o-632

5.94 19.0 0 6.0~1 O-636

5.86 19-o 2.35 5.60 O-631
5.86 19'0 12.71 4.61 O*CL;l
5-90 19.2 22.50 4.02 o-658
5-w 19.2 42.13 3.27 ! 0.637
 23

a generalised inertia
b ratio'of damp'ing to critioal damping
t3 Napierian log base
acceleration dU8 to gravity
hysteretic .damping constant
generd.i.sed stiffne?s-
length of beam
m ~4 (total mass of beam)
9 gener&sed co-ordinate
t time
C dscous damping coefficient
En Young's modulus for the plastic foam
61) a flexural rigidity of duritlumin bar
(No flexural ri&idit$ of'compdsite beem
If second moment of area of foam about neutral axle of beam
constant of proportio~sli~y (equation (4))
k radius ofL. gyration of exciter coil
M concentrated mass at free end of beam
mass of exciter coil
%
!c time constant of capacitance resistance circuit
x ratio of M to m
e amplitude of vibration
P mass per anit length of beam
w fundamentel resonant frequency
 REFERENCES

&A Author Title, etc.

1 S. C. Redshaw The construction ana testing of a xylonite

P. J. palmer model of a delta au-craft.
Aero. Quart. &83, 1951

2 I. Finnie Creep of engineering materials.

Vi. R. Heller U.S.A., McGraw-Hill Book Co. Inc.'l959,
pp.36 & 37

3 S. Timoshenko Vibretion pmblem.in engineering.

Second edition.
New York, D,Vm Nostrand Co, Ino. 1937
pp.Jfl-345.

4 Meteorological Office Handbook of Meteorological Instruments

Part I, London, H.N.s.o., 1956 p.155

5 R. H. Birtles New nylon polymers ana compositions.

Plastics Progress.
London, flihfe Book Ltd, 1955, p-729

6 L. N. Phillips Polyurethcnes. (Chemistry technology and

D. B. V. Parker properties).
London, Iliffe Book Ltd for the Flastics
Institute, 1964
CLAMPING
I BLOCKS
I
OURALUMIN
0.024” THICK
GLUE JOtNT
FOAM

FIG I GENERAL VIEW OF COMPOSITE BEAM

~ /WEIGHTS,
COIL FIXED
.
TO BEAM \

M A...,&$$$
CORE.

CAPACITY
INPUT FROM/ I PROBE.
WI---,
OSCILLATOR -

FIG 2 FREE END OF BEAM SHOWING

ARRANGEMENT FOR DYNAMIC TESTS
-1
r
-7E-
I-
+ *
0
WAYNE

0
OUTPUT
KERR

%
MICRO-SWITCH
a
-L
7

FIG 3 a&b DAMPING MEASUREMENT

I- n.% I ! I I I !II! I ! I ! ! j ! 1 2O:I’(

-2 TEM PERANRE

2 3 4 5 20 30 40 50 70 100
TlME mln

FIG. 4 CREEP CURVE FOR LOAD OF 12 GRAMS

 I I I \
0 IJl
UJJ~ avoi
 REL’ HUMlDllY 78% Q

REL’ HUMlOlTY 59% X

110
REL’ HUMIDITY 465% q
REL’ HUMIDITY 28% +

I(

90

60

70 II

b0

0 5 IO 15 20 25 30 35
TEMPERATURE ‘C

FIG.6 EFFECT OF TEMPERATURE AND RELATIVE HUMIDITY

ON THE STIFFNESSOF THE ‘FOAM
 IREC HUMIDITY 78% 0

loo-jy-j--
IREL; HUMIDITY 59 O/e X
RE+ HUMIDITY 46.5 % l3
REL HUMIDITY 28 % +

t
.
I
I
, -
I
,
I,

0
0
0

I i
0 25 r! 0 * o- 1; 5 15-o x 10-4
ABSOLUTE HUMIDITY lb / ft =

FIG 7 THE EFFECT OF TEMPERATURE AND RELATIVE

HUMIDITY RESOLVED TO ONE OF ABSOLUTE HUMIDITY
 DYNAM!C ‘E’ X
STATIC ‘E’ l

6C
x

SC

o-2 o-3 o-4 0-s

,& Ct’/lb

FIG 8 VARIATION OF THE STIFFNESS OF THE FOAM

WITH THE RECIPROCAL OF ABSOLUTE HUMIDITY
 9c

8C

4 DYNAMIC ‘E’ 0
lb/ In2 STATIC ‘E’ X-

70

2.0 3.0 4.0

ABSOLUTE HUMIDITY lb/ ft =

FIG 9 A COMPARISON OF DYNAMIC AND STATIC STIFFNESS

 12
TEMP RANGE t

16 - I8 0
x 18-21 x
* i
: 21-24 A
2b- 26 .
II Y x
26-29

2A xA X
315f02

x *
.
-x
,‘O

AA .h
%ftr
% A
.
9

-I-
A
A
x ;
7
I
x

l-
s 6 7 8 9 IO = lo--
ABSOLUTE HUMlOlTY lb/f t’

FIG. IO VARIATION OF DAMPING RATIO WITH HUMIDITY AND TEMPERATURE

 z x ;: x
(AWYLIEIMV SLINll) ,,INVLSN03,, 311lti3.LSAH
Fig.12. The microscopic structure of polyurethane foam magnified 20 times
 66 NYLON DRY . -

QQOO -

3000 -

LOO0 -

a00 -

!OOO

~
/
2 3 5 6
EXTENSlON %

FIC.13 THE EFFECT, OF WATER AND PLASTICISATION

ON THE STIFFNESS 0F 66 NYLON
 i

FIG 14 AXES OF CO-ORDINATES FOR VERTICAL BEAM

Printed in ak&znd’for Re+ Majesty’s Stot&mr+y Office by
the Royal Aircraft EstoblMnent, Pombormqh. Dd.126875.6.~.
A.R.C. C.P.No.935 A.R.C. C.P.No.905
AWUSC 1%5 670.664: Awust ,565 678.664:
payen. D. a. 621-4%: Pays”. D. B. 621-W
B’IWFNGS, DAWIHIG AJD CREEP PROPERTIW OF A POLWREiWAI~ FWlI 539.531: STIFRjES::. ZIpIt@ A:D CREEP PROPERTIE? OF A POLWRFIHANE FOAI, 539.531:
IliCWDIffi THE EFFECTS OF lTXPFXATLRE MD “LMIDIW 53b.372: IiCI,,DI:E ME EFFECTS OF Tt.PP.ARTtE AND kiIOITY 534.372
59.434 59.434

Experiments IBW bee” made to establish the physIcal pmwrtles or a Experiments lave bee” mde to establish the physical pmpertles of a
flexible ml~vrechane roam plastic used in the consL~ct.lon or aemelastlc flexible polyurethane roam plastic used 1” Che w”structlo” of aemelastlc
models for use 1” low-speed wlnd tunnels. .ilmple static and d~~rnlc models io,- use 1” low-speed tind t”nnels. Simple static and dynamic
bending tests ;vere mxde on a crmposlte Cantilever of pLastlC roam and a bendlw tests were made on a comwsite ca”tlle”er or plastic foam and a
11eht alloy bar. It Is shorn t&t dynamic stlfI"eSS and dampIng c.?.~clly 1lRhC alloy bar. It 18 .?Jlowr that dy”amlC 5tifr”eSs and denlPl”g capc1t.y
measuremnts must take account or temperature and absolute humldlty. and ueaswenents rrmst take account of tempemtuFc and ebaolute humldlty, and
Clat the dynamic stiffness was appmXl,“xtely double C218 static value on,. tbac the dyn¶DlC stirrne5s was approXlmztely double th3 static valve “Icr
the very limited Weque”cy range considered. Other pheno~sna s,,&, as the “WY llmlted frequency range considered. Other phenoeena such as
Mep and the effect of ‘ml’klngl m dlsc”ssed. CFBeP and the efrect of Mwklngl ale discussed.

A.R.C. C.P.No.935
*ueust ,955 67&6at
Paye”. D. B. &I i5fi
STIFMESS, DAWINC AND CRFXP PRCPERTIES (r A POLTWWHANE FM 5Y.531:
IMXXIIG Tr(E SFTFECTS OF -‘IUPS AND KiMIDITY %3?2:
539.04

Exp?rlnents fave bee” r;ade to establish the physical pmpertles of a

rlexlble po&reth3”e r0Qm phstlc Lsed ln the mnstnctlon 0r aemelastlc
models ror “se In low-speed .rl”d t”““els. Simple static and dynamic
bending tests "iere mde 0" a ~O~Poslte e."tlh?ver Or PhStk f-and a
11gJlt alloy bar. IL 1s show that d!JnzlC stirmess and damping cap3c1ty
measurerrents ,XISC. tap2 acco”“t or temperature and absolute huldlty, and
that the dyMmlC Stlrr”eSS WI.? apVXl&ely double the statle value oIw
the wry llmlted IrequencY range considered. Other phenomena such as
creep a”d the efrecL or ‘wrklng~ are discussed.
 C.P. No. 905

Q C~onvz Cop),rght 1967

Pubhshed by
HER MAJESTY’S STATIONERY OFFICE
To be purchased from
49 High Holborn. London w c 1
423 Oxford Street, London w.1
13.4 Castle Street. Edmburgh 2
109 St. Mary Street. CarddI
Brazennose Street. Manchester 2
50 Fairfax Street, Bristol 1
35 Smallbrook. Ringway, Birmingham 5
80 Chichester Street. Belfast 1
or through any bookseller

C.P. No. 905

S.O. CODE No. 23-9017-S