You are on page 1of 17

Bridge Design to Eurocodes: UK Implementation

ISBN 978-0-7277-4150-9
ICE Publishing: All rights reserved
doi: 10.1680/bdte.41509.434

PD 6694-1: RECOMMENDATIONS FOR THE DESIGN OF


STRUCTURES SUBJECT TO TRAFFIC LOADING TO EN
1997-1
S Denton, Parsons Brinckerhoff, Bristol, UK
T Christie, Parsons Brinckerhoff, Bristol, UK
J Shave, Parsons Brinckerhoff, Bristol, UK
A Kidd, Highways Agency, Bedford, UK

Abstract
This paper gives the background to the development of the provisions of PD 6694-1. It gives
guidance on the application of PD 6694-1 where it is considered that further explanation may
be helpful and identifies recommendations in PD 6694-1 which involve design principles or
procedures significantly different from those used in past practice.
The paper covers the clauses in the PD 6694-1 relating to actions, spread foundations, buried
structures and earth pressure on gravity retaining structures and bridge abutments. Traffic
surcharge and integral bridges are covered in detail in companion papers, for which references
are provided.

Notation
The same notation is used as in the Eurocodes and PD 6694-1. Other symbols are defined
within the clause in which they occur.

The Clause numbers used in the headings of this paper are the Clause numbers in PD 6694-1
to which the text refers.

Introduction
The recommendations given in PD 6694-1 (hereafter referred to as the PD) apply to
structures that are subject to traffic surcharge and other traffic loading. The recommendations
therefore specifically relate to the rules and partial factors given for "bridges" as opposed to
"buildings" in the Eurocodes. Many of the principles described can however be applied to
earth retaining structures that are not subject to traffic loading.

BS EN 1997-1:2004 does not specifically cover aspects of design of some types of highway
and rail structures such as integral bridges and buried structures. Complementary design
recommendations and guidance is therefore included in the PD.

For highway structures, PD 6694-1 replaces BD 30/87[2] (Earth Retaining Structures), BD


31/01[3] Buried Structures, BA 42/96[1] Integral Bridges and BD 74/00[5] Foundations. The
design of reinforced earth structures is neither covered in BS EN 1997-1:2004 nor in the PD.

434

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
PD 6694-1: recommendations for the design of structures subject to traffic loading to EN 1997-1

Basis of Design (4)


Dispersion of vertical load through fill (4.4)
The justification for the 30q method of dispersing vertical loads is given later in this paper in
relation to buried structure (10.2.7). The use of the 30q dispersion method may however be
unsafe when the vertical pressures arising from it are favourable. For example, where sliding
resistance is dependent on the load on the base slab, it may be unsafe to assume that part of
the weight of the surcharge traffic behind the abutment is supported on the base slab because
other dispersion modes including vertical soil arching can occur which may result in the
vertical load being supported on the ground behind the base slab while the horizontal
surcharge effect is still applied to the wall.

When analysing the foundations for bearing pressure the vertical pressure on an abutment
base slab due to the traffic surcharge may be favourable or unfavourable. In some cases the
additional pressure may increase the toe pressures, but in other cases it may apply a large
enough restoring moment to reduce the toe pressure. If the effect of the vertical pressure from
traffic surcharge is favourable in respect of bearing pressure, it may be prudent to ignore it.

Model Factors on horizontal earth pressure at ULS (4.7)


Following the publication of BS EN 1997-1:2004, concern was expressed that the ULS partial
factors were significantly lower than those used in pre-Eurocode standards for bridge design.
In particular it was seen that the effective ULS partial factor JfL. Jf3 specified for horizontal
earth pressure in BD 37/01[4] equalled 1.5x1.1 = 1.65 compared with a JF of 1.35 for the
critical STR/GEO limit state, Design Approach 1, Combination 1 partial factors in the
Eurocode (i.e. Set A1 in BS EN 1997-1:2004 and Set B in BS EN 1990:2002). This would
mean that structures designed to BS EN 1997-1:2004 could be less robust than those designed
in the past.

To address this concern, the PD states that where it is required to maintain the same levels of
safety as were applied in the past, a model factor JSd;K may be applied to the horizontal earth
pressure (effectively to Ka or K0). The recommended value of the model factor was based on
the ratio of the pre-Eurocode factors to the STR/GEO Combination 1 factor, namely 1.65/1.35
= 1.22 (rounded down to 1.2), to give similar design values for earth pressures. Its effect was
examined for other ultimate limit states verifications.

For sliding and overturning, BD 30/87, 5.2.4.2[2] references CP 2[8] in which it says, in
relation to sliding: a factor of safety of approximately 2 should be applied and the
angle of friction below the base is equal to I the angle of friction of the soil beneath the
foundation.

On this basis, the required heel length Bheel for an abutment of height Z is given by:

Bheel = 2Ka;k{JZ2/(2tanI' `

For a Eurocode design using the model factor JS;dK:

Bheel = JSd;KKa;d{JZ2/(2tanI'cv)}

435

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
Bridge Design to Eurocodes: UK Implementation

From this is can be shown that, using the model factor and the relevant values of the partial
factors, the Eurocode value of Bheel will not be less than the pre-Eurocode value if tanIccv is
not greater than about 0.9tan I'. In practice tanI'cv is almost invariably less than 0.9tanIc. 

For sliding resistance of an undrained foundation CP 2[8] uses a similar method to the
Eurocode. For the CP 2[8] method with a factor of safety 2 on sliding: 2H = Bcu where B is
the base length, Z is the height of the wall and the horizontal action H = Ka;kJ= Thus,

B = 2Ka;k{JZ2/(2cu)}

In the Eurocode, for a retaining wall subject to permanent actions and the model factor,
JGJSd;kH = Bcu/JM where in Design Approach 1, Combination 2 JG = 1 and JM = 1.4. 

B = 1.4 JSd;KKa;d{JZ2/2cu}

From this it can be shown that based on Ick = 33q for the backfill and the relevant values of
the partial factors, the Eurocode base length will be approximately 5% longer than the CP 2[8]
base length if the model factor is included, and approximately 13% shorter if the model factor
is not applied.

The above comparisons apply to retaining walls subject to permanent earth pressure only.
When surcharge, braking and acceleration are applied, the pre-Eurocode base lengths will
theoretically be relatively longer. In practice though, the Eurocode surcharge action is so
much larger than the pre-Eurocode surcharge action that it is unlikely that base slabs subject
to the Eurocode surcharge will be shorter than base slabs designed in the past.

Bearing resistance is frequently governed by settlement requirements at SLS for which the
ULS model factor is irrelevant. For ultimate bearing resistance it is less easy to make a direct
comparison between Eurocode and pre-Eurocode designs because of the number of different
acceptable pre-Eurocode design methods available. Specimen comparative calculations have
however shown that if the model factor is applied to the horizontal earth pressure, the
Eurocode designs for bearing resistance will usually be comparable with pre-Eurocode
designs.

In relation to overturning, CP 2[8] says "...in gravity walls the resultant thrust should not fall
outside the middle third of the base, and for other types of wall a factor of safety of at least 2
against overturning is required". Overturning is not usually an issue with conventional
gravity walls and abutments because the bearing resistance under the toe will normally
become critical before the structure overturns and the length of heel required to provided
sliding resistance is usually sufficient to give an adequate restoring moment. Overturning
could however become an issue with a mass gravity wall seated on rock or a concrete slab and
propped or keyed into the slab to prevent sliding as shown in Figure 1.

436

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
PD 6694-1: recommendations for the design of structures subject to traffic loading to EN 1997-1

Figure 1

For the above structure, considering overturning about A at ULS, the Eurocode effectively
requires that the maximum design overturning moment should not be greater than the
minimum design restoring moment:

Y H JG;soil;supJSd;kJMK  X V JG;conc;inf

where H and V are characteristic actions and JMK = Ka;d/Ka;k | 1.11 at EQU and 1.25 at
STR/GEO combination 2 if I' is about 33o.

The overall factor of safety is XV/YH which equals (JG;soil;supJSd;KJMK )/(JG;conc;inf ).

This equals (1.05 x 1.2 x 1.11/0.95) = 1.47 at EQU and (1.35 x 1.2 x 1.0/0.95) = 1.70 at
STR/GEO Combination 1.

These values reduce to 1.23 and 1.42 respectively if the model factor JSd;K is not applied.
However, it can be shown that if this structure was designed to comply with the "middle
third" rule at SLS then the factor of safety would automatically be t 3.0.

From the above comparisons it can be seen that the 1.2 ULS model factor compensates for the
difference between the Eurocode and pre-Eurocode values of ULS partial factors in relation to
earth pressure, sliding resistance and ultimate bearing resistance, and it is irrelevant in regards
to settlement and overturning except in the unusual situation where a structure such as that
shown in Figure 1 is not designed to comply with the middle-third requirement at SLS.

The Eurocode surcharge loading for highway structures is substantially more onerous than the
HA and HB surcharge used in the past, and as this will result in stronger rather than weaker
structures, the 1.2 model factor is not required to be applied to the effects of traffic surcharge
loading.

The PD does not offer an opinion as to whether the pre-Eurocode standards were unduly
conservative. The option to use the model factor is for designers and clients who wish to
maintain past levels of safety in their earth retaining structures.

437

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
Bridge Design to Eurocodes: UK Implementation

Pile Foundations (6)


This clause is based on the recommendations in BD 74/00 Annex B[5].

Gravity Retaining Structures and Bridge Abutments (7)


Earth pressures on retaining walls and abutments with inclined backfill (7.2.3
and 7.2.4)

Active pressure on walls with long heels (D<\) (7.2.3)


Figure 2(a) shows a structure retaining fill with a surface inclined at an angle E. Figure 2(b)
shows a plot of the horizontal thrust on plane CE as its inclination, O increases from D to 90o.
As can be seen the thrust reaches a peak when O = \. In Appendix 1 to this paper it is shown
that this peak thrust on CE (and therefore the horizontal thrust on the whole structure) is the
same as the thrust on the vertical virtual face CD when G on that face equals E. This means
that provided D is less than \, the total horizontal thrust on the structure can be found by
calculating the horizontal thrust on CD taking G = E. Appendix 1 also provides a proof of the
equation for \ given in the PD (which is the same as the corrected expression given in
Clayton and Milititsky[9]). Values of \ based on this equation are shown in Figure 3. \ is not
however used in the pressure calculation except to determine the minimum length of BC,
(ABcot\), for which the G = E method is valid.

E
O \ D
E2
E
E1
Vertical virtual plane

Critical thrust on
structure JKa;E CD2/2
A
G=E Ka on CD = Ka;E
G=I
Horizontal thrust

O
CE1

CE2
CA

CD
CE

H = 90o + I \
\ when O = \
D H
B C D \ 900
Inclination of planes (O)
(a) Retaining wall with inclined fill (b) Variation of horizontal thrust on
virtual planes of differing
inclinations when Gon these planes
equals Ic
Figure 2. Horizontal thrust on inclined backfill planes

Example 1 (Long heel)


A 6m high wall with a 4m heel supports backfill inclined at 20o. Find the active characteristic
horizontal thrust applied to the structure if J = 18 kN/m3 and I'= 35o:

438

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
PD 6694-1: recommendations for the design of structures subject to traffic loading to EN 1997-1

When G = E = 20o, Ka from PD 6694-1, Table 4 equals 0.30


Height of virtual face CD = 6 + 4tan20o = 7.46m
Horizontal thrust = JKaCD/2 = 18x0.3x7.46/2 = 150.3kN/m width.

Figure 3. Values of \

Active pressure on walls with short heels D ! \ (7.2.4)


When D ! \, (that is when BC is less than ABcot\as in Figure 4(c)) the critical inclined
virtual face CE is interrupted by the back face of the wall at H as shown in Figure 4(c) and the
results described in the previous paragraph will no longer apply because AH in Figure 4(c) is
a soil-to-wall surface and G over that length will be Gw rather than I'.

When D = \, as in Figure 4(b) G on CD = E. When D = 90o, as in Figure 4(d), G on CD equals


Gw because CD is then coincident with BA and Gw is applied over the full height of the wall.
When D lies between \and 90o, the effective value of G on CD therefore lies between E and
Gw, and the effective value of Ka on CD (that is Ka;CD) will lie between Ka;E and Ka;Gw as
described in the PD.

To determine the value of Ka on CD for an intermediate position with \ < D < 90, the shaded
triangle AHG in Figure 4(c) was considered as a Coulomb wedge with G = Gw on AH, and
HCFG was considered as a four-sided wedge with G = I ' on all soil-to-soil faces. Using this
model for a number of values of  it was found that although theoretically as  increased from
\ to 90o the plot of Ka;CD followed a curve between Ka;E and Ka;Gw , in practice it was simpler
and marginally conservative to assume that Ka;CD increased linearly with D from Ka;E to
Ka;Gwin this range

Figure 4(e) plots the value of the thrust on plane CA and on the whole structure as the wall
BA moves towards CD and D increases from less than \ to 90o. The arc PQ shows the
theoretical increase in effective Ka;CD as D increases from \ to 90o and the chord PQ shows
the linear variation in Ka;CD assumed in the PD.

439

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
Bridge Design to Eurocodes: UK Implementation

F
E F
D A,D
D G
D E A
E E
A A

G Gw
G E

G Gw
G E
Gw GE
H
G I

\
\ D
D H
B C B,C
B C
B C
(a) D < \ (b) D = \ (c) \ < D <90o (d) D = 90o

Critical inclined planes as BA approaches CD

\ D  90o
D \ Ka;CD increases
Ka;CD = Ka;E from Ka;E to Ka;Gw
Horizontal thrust Linear approximation Theoretical
Q
on structure equals G on CD =Gw
JKa;CD CD2/2 P Gon CD =E

Horizontal Gon CD =Ic


thrust on CA
horizontal thrust

as D increases
(a) (b) (c) (d)

D\ \ D!\ 900

Values of D
(e) Horizontal thrusts as BA moves towards CD
Figure 4. Abutments supporting inclined backfill

In the above paragraphs it is assumed that Gw is less than E. If Gw is greater than E then Ka;E is
greater than Ka;Gw and Ka;CD theoretically reduces from Ka;E to Ka;Gw as D increases from \ to
90o. As this effect is small and only significant when D is very close to 90o, the PD
recommends that when Gw is greater than E, Ka on CD should be taken as Ka;E for all values of
D.

440

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
PD 6694-1: recommendations for the design of structures subject to traffic loading to EN 1997-1

Example 2 (Short heel)


A 6m high wall with a 1.5m heel supports backfill inclined at 20o. Find the characteristic
active pressure applied to the structure if J = 18 kN/m3 and I'= 35o and Gw is taken
conservatively as 10o:

D = tan-1(6/1.5) = 76o
\ = from Figure 3 (above) (for E = 20o and I'= 35o) = 70.8o. D> \
When G = E = 20o, Ka;E from PD 6694-1, Table 4 equals 0.302
When G = G w = 10o and E = 20q, Ka;Gw by wedge analysis or other means = 0.322
Increase in Ka;CD = (Ka;Gw - Ka;E){(D-\)/(90-\)} = (0.322-0.302){(76-70.8)/(90-70.8)} = 0.005
Ka;CD = 0.302+0.005 = 0.307
Height CD = 6 + 1.5tan20 =6.55m
Horizontal thrust = JKaCD/2 = JKa6.55/2 = 18x0.307x6.55/2 = 118.5kN/m width

Walls with both Dand \ greater than 80o


It can be seen from the equation for \ that when E approaches I', \ approaches 90o. It has
been found that in the unusual situation when both D and \ are greater than 80o (i.e. steep
backfill with a very small heel), the value of Ka;CD starts to increase towards Ka;GZ when  is
approximately 80o (i.e. less than \ ). The PD covers this situation by artificially limiting the
maximum value of \ to 80 o in 7.2.4. This limitation only affects the heel length at which the
short heel effect becomes critical. It does not affect the value Ka;E.

Compaction pressures (7.3.4)


The pressure distribution shown by Clayton and Milititsky[9], Ingold[10] and others is
reproduced in Figure 5(a). In practice the position of (A) usually occurs only a short way
below the surface and it is simplest and only slightly conservative to consider the combined
active and compaction pressure having the value Vtop given in the PD and being constant from
ground level to the level at which active pressure equals Vtop as shown in Figure 5(b).

441

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
Bridge Design to Eurocodes: UK Implementation

2P
Vtop 2VJ
zc Ka
SJ V top
0 S
Vh =JzKp
zc z
A z

2 PJ
1 2P V h V 'hrm Vtop
hc S
ka SJ hc V h V top
JKd

B
Active Vh =JzKa
pressure
Active Vh =JzKd
pressure

(a) Clayton[9] & Ingold[11] (b) PD 6694-1

Figure 5. Compaction pressures

Movement Required to Generate Passive Pressure (7.5)


BS EN 1997-1:2004, Annex C effectively defines the relationship between mobilised passive
pressure and wall movement by three points and one gradient (see BS EN 1997-1:2004,
Annex C, Figure C.4). The three points are Ko and zero movement, 0.5Kp and a movement of
v2/h and Kp and a movement of v/h. By the nature of passive pressure the gradient of the
pressure/movement curve at full passive pressure is horizontal. This is illustrated in Figure 6.

In devising a formula to replicate the curve illustrated in Figure C.4 in Annex C it was
considered adequate to assume that the relationship was linear between Ko and Kp/2 and the
curve between Kp/2 and Kp was a cubic curve passing through the Kp/2 and Kp points and
having a horizontal gradient at Kp. The equation given in the PD achieves this. Inevitably in
some cases the gradient of the curve at Kp/2 is not identical to the gradient of the Ko-Kp/2 line.
The errors resulting from this are considered to be small compared to the tolerance on vp/h
and v2/h given in BS EN 1997-1:2004, Annex C.

442

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
PD 6694-1: recommendations for the design of structures subject to traffic loading to EN 1997-1

Hambly[10] values (see Table 1)

Figure 6. Plots of K/Kp against v/h using the empirical equation in the PD clause 7.5

Plots A and B in Figure 6 are based on the maximum and minimum values of v/h for K=Kp
and half Kp given in section (a) of BS EN 1997-1:2004, Table C.2. It is clear from the range
of figures quoted In Table C.2 and Table1 below, that the relationship between movement
and earth pressure can be very variable. The values of the wall rotations required to mobilise
full passive and half passive pressure given in BS EN 1997-1:2004, Annex C are very large
compared to those given by Hambly [10]. For example, to develop a "conventional" half Kp
behind a 10m high wall the deflection at the top would be between 110mm and 200mm
according to Annex C compared with about 50mm according to Hambly[10].

Rotations required to mobilise Rotations required to mobilise


Half Kp Full Kp
Hambly[10] 0.5% (approx) 3% (approx)
Annex C to
1.1% - 2% 5% - 10%
BS EN1997-1:2004

Table 1 Comparison of rotation required to mobilise passive pressure

Traffic surcharge (7.6)


The background to the surcharge model is described by Shave et al [13].

Integral Bridges (9)


See accompanying paper Denton et al [14].

443

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
Bridge Design to Eurocodes: UK Implementation

Buried Concrete Structures (10)


This section (Clause 10) is based on BD 31/01[3] updated in line with the Eurocode
requirements.

Superimposed permanent load (10.2.2)


The model factors JSd;ec to be applied for superimposed permanent load are taken from Figure
3.1 in BD 31/01[3].

Dispersal of vertical loads (10.2.7)


The dispersal of vertical loads through fill is given in the PD as 30q compared to 26.6q (2:1)
in BD 31/01[3]. The 30q value is taken from BS EN 1991-2:2003, 4.9.1, NOTE 2.

The Boussinesq equation given in clause 3.2.1 (iii) of BD 31/01[3] has been omitted from the
PD as it underestimates the pressure when there is a rigid plane, such as the roof of a buried
structure, located a short distance below ground level. The method in Table 2.1 of Poulos and
Davis[12] gives pressures below a point load which are marginally lower than those found
using the 30q dispersion method, as can be seen from Figure 7.

Figure 7. Comparison of the 30o pressures with pressures based on the Poulos &
Davis[12] and Boussinesq equations for 1 m fill depth.

Longitudinal road traffic actions (10.2.8)

Traffic surcharge (10.2.8.1)


The treatment of traffic surcharge for buried structures using the simplified model given in
PD 6694-1, 7.6 is explained in Table 5, Note B of the PD. Its background is given by Shave
et al[13].

Braking and acceleration (10.2.8.2)


The reduction of the braking and acceleration actions with increasing depth of earth cover
given in PD 6694-1, 10.2.8.2 has been retained from BD 31/01[3] in the absence of data to
justify a revaluation.

Longitudinal joints (10.5)


The 0.15Hc limitation of deflection on segmental units is taken from BD 31/01[3].

444

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
PD 6694-1: recommendations for the design of structures subject to traffic loading to EN 1997-1

Soil Structure Interaction Analysis of Integral Bridges (Annex A to


PD 6694-1)
See accompanying paper Denton et al[14].

Tables of Earth Pressures for Buried Structure (Annex B)


Maximum pressure (B.2 and tables B.1 and B.2)
The values of the earth pressure coefficients given in the tables in Annex A are based on the
assumption that the maximum characteristic at rest earth pressure coefficient taking
temperature and strain ratcheting into account, is 0.6, as in BD 31/01[3]. This may be
considered as backfill with a characteristic I' of 30q subject to an enhancement factor, Fenh, of
1.2 to allow for temperature, strain ratcheting and other unfavourable effects. In addition, at
ULS, the pressure coefficients are subject to the model factor JSd;K (also equal to 1.2) described
in 4.7. Using the values of JM given in the UK National Annex to BS EN 1997-1:2004 the
values of the earth pressure coefficient Kmax given in Tables B.1 and B.2 in Annex B are derived
as follows:

Characteristic EQU STR/GEO STR/GEO


Comb 1 Comb 2
Icd (inc JM) 30o 27.7o 30o 24.8o
K0 = (1-sinIcd) 0.5 0.53 0.5 0.58
K0 x Fenh 0.6 0.64 0.60 0.70
Kmax = K0 Fenh Jsdk 0.6 0.77 0.72 0.84

As traffic surcharge is not considered to be affected by temperature and strain ratcheting, and
as JSd;k is not applied to traffic surcharge (see 7.6) the design value of the at rest pressure
coefficient for traffic surcharge given in PD 6694-1, Annex B Tables B.1 and B.2 is simply
the value of Ko given in the table above.

Minimum pressure (B.3 and Table B.3)


For members which are critical with minimum horizontal earth pressure, Table B3 is relevant.
The minimum characteristic earth pressure coefficient of 0.2 is as in BD 31/01[3]. This is the
characteristic value of Ka for Ick = 30q multiplied by a reduction factor, Fred = 0.6. The values
of Kmin;d given in the tables are based on the equation:

Kmin;d = Fred(1-sinIcd)/(1+sinIcd)

where Icd = tan-1{(tan 30q)/JM*} and JM* = 1/JM . This gives approximately the same result as
taking Fred = 1, Ick as 38q with Ka;k based on G/Ic = 0.66 from BS EN 1997-1:2004, Figure
C.1.1 and JM J0 as before.

Figure C.1.1 gives values of Ka as low as 0.13 for Ic = 45o and G = Ic and it would therefore
be prudent to ignore active pressure altogether if it was considered that backfills with high
values of I' were relevant.

445

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
Bridge Design to Eurocodes: UK Implementation

Active pressure (B.4 and Tables B.4, B.5 and B.6)


When braking or acceleration actions are applied, active pressure based on a characteristic I'
of 30q is applied on the active face and Kmax is applied to the passive face as in Tables B.1
and B.2, except that as the single source principle in not applied at EQU and the Kmax actions
are favourable in resisting longitudinal actions, the characteristic value of Kmax (= 0.6) is
applied at EQU in Tables B.4 and B.5.

Where it is necessary to increase the pressure on the passive face above the Kmax pressure to
resist the longitudinal traffic actions it should be noted that movements related to specific
values of K given in Table C.2 of Appendix C of BS EN 1997-1:2004 are substantially
greater than those given by Hambly[10] (see Table 1 above).

References
[1] BA 42/96 Amendment No. 1(2003) The design of Integral Bridges, The Stationary
Office, London
[2] BD 30/87 Backfilled retaining walls and bridge abutments, The Stationary Office,
London
[3] BD 31/01 The design of buried concrete box and portal frame structures, The
Stationary Office, London
[4] BD 37/01 Loads for highways bridges, The Stationary Office, London
[5] BD 74/00 Foundations, The Stationary Office, London
[6] BS EN 1991-2:2003 (incorporating Corrigenda December 2004 and February 2010),
Eurocode 1, Part 2, Traffic loads on bridges, BSi, London, UK
[7] BS EN 1997-1:2004 (incorporating corrigendum February 2009) Eurocode 7:
Geotechnical design Part 1: General rules, BSi, London, UK
[8] CP 2 (1951) Earth retaining structures, BSi, London, UK
[9] Clayton, C. R. I and Milititsky, J. (1986) Earth pressure and earth retaining structures
Surrey university Press, London
[10] Hambly, E.C. (1991) Bridge Deck behaviour, 2nd edition, London: E& FN Spon
[11] Ingold, T.S. (1979) The effects of compaction on retaining walls. Gothechnique
29(3), 265-283
[12] Poulous, H.G and Davis, E. H. (1974) Elastic solutions for soil and rock mechanics
John Wiley & Sons, Inc, London
[13] Shave, J, Christie, T. J. C, Denton, S. and Kidd, A. (2010) Development of traffic
surcharge models for highway structures, in Proceedings of Bridge Design to
Eurocodes UK Implementation, Ed. by S. Denton, Nov 2010, ICE, London.
[14] Denton, S, Riches, O, Christie, T. J. C. and Kidd, A. (2010) Developments in integral
bridge design, in Proceedings of Bridge Design to Eurocodes UK Implementation,
Ed. by S. Denton, Nov 2010, ICE, London.

446

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
PD 6694-1: recommendations for the design of structures subject to traffic loading to EN 1997-1

Appendix 1

Derivation of Earth Pressures on Walls With Long Heels and


Sloping Backfill

This appendix provides a derivation of the critical wedge angle \, the friction angle G to be
used on the vertical virtual face, and the resulting earth pressure coefficient Ka, for a retaining
structure such as that illustrated in Figure A1(a) where the backfill is sloping at an angle E
and the heel of the wall is long enough that \!D.

Figure A1(b) shows the critical wedge 'ECF and the forces acting on it at the boundaries, R1
and R2, and the self-weight W.

(a) (b)
Figure A1. Earth pressures with sloping backfill

Figure A2 comprises the Mohrs circle for the critical wedge 'ECF as it reaches a critical
state simultaneously along CE and CF (represented by points e and f). Point d represents the
stress at the vertical plane CD.

447

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
Bridge Design to Eurocodes: UK Implementation

Figure A2. Mohrs circle for the critical wedge

The angle between planes EC and ED is (90-\), so this transformation of axes requires a
rotation on the Mohrs circle from point e to point d of
eod 2(90  \ ) (A1)

Triangle 'geo is right angled, hence


eog 90  I ' (A2)

From (A1) and (A2),


doh 90  I '2(90 \ ) 2\  90  I ' (A3)

The external angle of triangle 'gdo is therefore:


o dj 2\  90  I ' G (A4)

By considering triangles 'djo and 'gjo and using the sine rule,
sin G
oj sin( djo) R R sin( 2\  90  I 'G ) (A5)
sin I '
Leading to:
sin G
sin( 2\  90  I 'G ) (A6)
sin I '
Rearranging (A6) gives
1 1 sin G
\ 90  I 'G  sin (A7)
2 sin I '

Now consider the equilibrium of the wedge component triangles 'EDC and 'CDF as shown
in Figure A3.

448

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
PD 6694-1: recommendations for the design of structures subject to traffic loading to EN 1997-1

Figure A3. Equilibrium of wedge component triangles

The forces R1 and R2 act at the 1/3 points of CE and CF, because there is a linear stress
distribution along these lines. These points are vertically below the centroids of triangles
'EDC and 'CDF. The force R3 acts at the interface CD at an angle G. For equilibrium the 3
forces for each triangle must intersect at a point. These intersection points must therefore be
the 1/3 points on the boundaries CE and CF, as shown in Figure A3. Using similar triangles,
the angle of the force R3 must be identical to the slope of the backfill1, or:
G E (A8)

From (A7) and (A8),


1 1 sin E
\ 90  I ' E  sin (A9)
2 sin I '

By summing the angles around point O in Figure A2 to 360 degrees it can also be
demonstrated that the angle H is related to \ by the expression in (A10):

1 1 sin E
H 90  I '\ 90  I ' E  sin (A10)
2 sin I '

Knowing the critical value of \ as given in (A9), the ratio of the horizontal to the vertical
earth pressures may be determined from the Mohrs circle in Figure A2:
1
 sin 2\  I '
V hc sin I ' 1  sin I ' sin 2\  I '
(A11)
V vc 1 1  sin I ' sin 2\  I '
 sin 2\  I '
sin I '

1
An alternative derivation of (A7) considers the triangles of forces for the component triangle 'EDC and the full
wedge 'ECF shown in Figure A3 and demonstrates that when G E the horizontal thrust is equal in each case;
however the derivation is more complex and so not included here for space reasons.

449

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.
Bridge Design to Eurocodes: UK Implementation

However, while the ratio of stresses defined in (A11) could be thought of as an earth pressure
coefficient, the values obtained from (A11) are not the same as Ka defined in the conventional
way as in (A12), based on the total horizontal force H acting on a vertical plane of height h
and assuming a vertical earth pressure of Jz, where J is the soil density and z is the distance
below ground level.

H
Ka (A12)
1 2
Jh
2
H may be determined by considering the equilibrium of the wedge component triangle 'CDF
in Figure A3, from which:

W2
H (A13)
1
tan G 
tan(H  I ' )
The weight W2 is calculated based on the area of the triangle 'CDF:

1 2 cos H cos E
W2 Jh (A14)
2 sin(H  E )

Combining (A8), (A12), (A13) and (A14) results in the following expression for Ka:

cos H cos E
Ka (A15)
1
sin(H  E ) tan E 
tan(H  I ' )

A comparison of expressions (A11) and (A15) shows that they give almost identical values up
to a slope angle E of about half I, but as the slope approaches I the values diverge, with
(A15) giving higher values. This difference is due to the way that Ka has been defined in
(A12), which is convenient for design purposes, but this definition of Ka is not strictly the
same as the ratio of horizontal and vertical pressures when the backfill is sloping.

The values for Ka presented in Table 4 of PD6694-1 have been calculated from (A10) and
(A15) for various values of E and I.
Equation (A9) may be used to check whether \!D . If this is not satisfied (i.e. the heel is
short) then the derivation above is not correct; the angle G will lie somewhere between 0 and
E, and the thrust on the wall will need to be increased as described in PD6694-1.

450

Downloaded by [ University of Liverpool] on [31/01/17]. Copyright ICE Publishing, all rights reserved.

You might also like