Você está na página 1de 4

PAPER LIMIT STATE DESIGN OF STRUCTURAL PARTS AT AND IN THE GROUND

Some points that deserve or need a more general discussion are out-

II.imit state desic~n


lined in the conclusion.

Shallow and spread foundations


Consider a skeleton with a stiff basement upon a group of block or

oI'structural pai~ strip foudations (Figure 1). A statical calculation with characteristic
structural loads leads to reaction forces R,, R3 etc, assuming non-yield-
ing supports in a first approximation. Statical equilibrium for each
foundation (Figure 1 inset, subscripts omitted) requires identity of

atanc in the grouncI action and resistance, F = R. The design situation is obtained with par-
tial factors, named IIF and n for convenience instead of 7 as in the
Eurocodes:
Prof Dr-Ing G Gudehus, University of Karlsruhe,8 Prof A
Fd = nF Fn c Ra = Rn/IIR
Weigenbach, University of Dortmund, Germany.
The subscript n indicates nominal values of action and resistance, F
Wurocode 7 proposes ultimate limit states for the design of struc- and R, calculated with characteristic values and assumed dimensions
Wtural parts at and in the ground, abbreviated ULS 1B. They have of the structure. R is the statically required supporting force. It is
&been introduced in order to reconcile the structural Eurocodes, in increased by a load factor nF as for structural design and can be split
particular EC2 for concrete, with EC7 within the frame of EC1. The into permanent and temporary parts. nF = 1.4 is obtained via '/,1.35 +
resulting compromise has some shortcomings from a geotechnical /3 1.50 with dominant weight ( /3), which is typical for such cases, and
point of view: the factors 1.35and 1.50for permanent and temporary actions. This fac-
~ The partial factor 1.35for permanent actions is not tractable in many tor cannot be justified by fluctuations of weight only, but by uncertain-
cases; ties of 'actions't the foundations, ie of internal forces there, due to
~ contact pressures among structural parts and ground cannot gener- irregularities of stiffhess and mechanical history.
ally be classified as action or resistance; The nominal resistance R is calculated with an assumed breadth b
~ it is difficult to transfer well-established design procedures and glob- of the foundation using characteristic soil strength parameters, sim-
al factors. plified soil layers and a nominal water table. It is reduced by n in order
The German National Application Document (NAD), which to cover uncertainties of the soil model and parameters. With the value
appeared April 1996, is an attempt to avoid these shortcomings. The n= 1.4, Eq 1 leads to the familiar 'global'actor II = IIF II-2.0. This
German pre-standard DIN V 1054-100 referred to in this NAD, which is may be replaced by a higher value, eg II = 3.0, for stiff clay. A desirable
outlined and discussed in the sequel, is intended to: reduction of the factor for action, say to 3)F = 1.25, can easily be allowed
~ incorporate established global factors so that a correction of the said for by a higher II=II/IIF . The proposed design format is thus the same
1.35 can also be allowed for; as for structural parts off the ground, incorporates well established
~ better avoid arbitrary interpretations of the partly vague EC7 text; experience and avoids details of soil mechanics and probabilistics.
~ avoid over- or underdesign in cases of well-established practice. A I'rame structure on shallow block foundations is more complicated
Shallow and spread foundations are treated first. The forces acting as the latter obtain inclined and eccentric forces (Figure 2). Nominal
upon groups of block foundations and the resisting forces from the forces with inclinations and eccentricities (F, 5, e in Figure 2 inset,
ground are first calculated without partial safety factors. The forces are without subscripts for brevity) are obtained from a statical calculation
then changed into design values by partial factors as in structural which may include assumed subgrade springs for representing the
engineering and so that established global factors are reproduced. yielding ground. Acting and resisting design forces are calculated with
Hydrostatic uplift and horizontal sliding are similarly treated as addi- the same partial factors nF and nas above, assuming the same incli-
tional ultimate limit states (ULS 1A and B). nation and relative eccentricity, 5 and e/b, as in the nominal situation.
Groups of piles and anchors are treated in the same manner. The Established relations for the ground resistance R in case of base fail-
forces from the superstructure are calculated as for shallow ure can be used avoiding lengthy discussions on safe values of 5 and
foundations. The resistance of the ground against penetration or pull- e/b.
out is reduced by a partial factor depending on loading test results. Lateral forces at the foundation can also be allowed for. An active
Ultimate limit states of pile and anchor groups can also involve large earth pressure, at the left side in the case of Figure 2 inset, is included
zones of the ground, leading to 'global'plift or base failure (ULS 1A into F, 5 and e. A passive earth pressure at the right side can likewise
and C). be included; it is thus part of the action at the foundation base, which
The same concept can be used for retaining structures. The breadth is inevitable for internal forces. The force component ratios, 5 and e/ b,
of L-walls is determined from weight and earth pressure as for a strip are again estimated for the nominal situation. A sufficient safety
foundation. Bedding depth and struts of pile walls are determined by against sliding on a cohesionless base is obtained with
considering simple beams. Similar beams are considered with several
rows of anchors and with allowance for vertical force components. The tan 5~ tan 5, /I4 (2)
length of anchor tendons is determined by group failure (ULS 1C). Soil
nailing is a special case. wherein 5= 5 is the nominal force inclination, 5, the base friction
Uplift and seepage forces can be incorporated by taking effective angle, and I4 an associated safety factor. The familiar safety level is
forces in the soil skeleton and in structural parts in the ground. obtained with I4 = 1.5. We thus consider the horizontal and vertical
Instationary pore water pressures related with volume changes of the components of F as coupled due to the interaction of structure and
soil are left aside. Possible redistributions of structural forces are ground.
briefly discussed. They are covered by partial factors in the proposed Hydrostatic uplift is considered as a special ultimate limite state
concept which requires a certain degree of ductility. (ULS 1A for spread foundations (Figure 3). A trough is lifted if its nom-

Rttsso t. Skeleton wNb ~~~~ basement watts npon tNoobfoenteeens aint


(tnS& sotiatnbnt bteeb tonwtatton.

tn, tn,

GROUND ENGINEERING SEPTEMBER 1996


PAPER LIMIT STATE DESIGN OF STRUCTURAL PARTS AT AND IN THE GROUND

out partial factors, safety against


ULS 1B is assessed again with Eq
1. The partial safety factor IIF can
remain the same as before, and
W/2 for the same reasons. The nomi-
nal resistance R is normally
determined from loading tests
and at best supported or extrapo-
lated with mechanical calcula-
tions including soil parameters.
The associated partial factor qR
F/2 depends on the number of load-
ing tests, the relevance of ground
investigation and the reliability of the mechanical model. A wider
spectrum of IIR than for block and step foundations can be necessary
therefore, and q = nR nR holds again for the global safety factor.
A closer inspection shows that both nR and qs should better be adapt-
ed to the type of building, ground and piles. A larger qs should be cho-
sen for an end bearing pile in sand, eg than for a friction pile in stiff
clay. One should realise that details of ground-structure interaction
cannot be covered by a few fixed partial safety factors.
Ai-'B, A counterpart of Figure 2 is an abutment of a bridge upon piles pass-
ing through soft ground (Figure 5). Nominal pile forces are calculated,
with an assumed array of the piles, as statically required reaction
tR, forces, eg F, and F, . Their component ratios, eg F,/F, and inclination
5, in the considered example, depend on the chosen mechanical model
TOP:Rgure 3. and parameters and is assumed again to remain the same for the
Hrdrostaeo uplift design situation. Acting and resisting forces are then treated for the
ot tnmgb (ten), assessment of safety against ULS 1B as before. F,/F, and 5, replace e/b
bottom ol an and 5 for a shallow foundation. The partial factor qR is justified by
suemndbm uncertainties of force redistribution as before. Eq 1 is used with the
(oontra), and same IIR as for the previous case.
In)ected layer Other force components and mechanisms can also be included.
(itgbt). Negative shaft friction from the soft ground and the top layer can be
F, Fa added to the action upon a pile which is then its axial force at the level
AII Rgmu d. of the bearing layer (likewise for vertical piles). Lateral pile forces,
gbetetsn wllb caused by laterally flowing soft soil, lead to bending moments as in
basement walls retaining structures. The whole structure and part of the surrounding
upon phe, and ground may undergo sliding towards a neighboured slope which has to
(Inset) separated be safeguarded by considering a ULS 1C (detailed in the next section).
pile. A group of vertical anchors can be needed against hydrostatic uplift
of a bottom plate (Figure 6). The force acting upon each anchor is
Rlgltr'gma g. F, = (F W)/N, with the resulting uplift force F and weight W (usually
Itrtdge abubmmt without contribution from the walls). Dividing through the number N,
on pile Ipeup of anchors means that a uniform distribution of anchor forces is
assumed. Multiplying the nominal value F, with a force factor IIF is to
cover uncertainties of redistribution as before. The nominal resistance
R is usually determined from pull-out tests, possibly supported by
mechanical calculations. The associated partial factor nR depends on
the number of tests and other factors as outlined above for piles, and
with the same advantages and shortcomings of partial factors.
inal weight and uplift forces are equal, W = F (Figure 3, left). The nom- Other ultimate limit states of the same system can be investigated
inal action F is calculated with a safe characteristic height of the with other safety factors. Hydrostatic uplift involves the anchor group
water table. The nominal resting weight W is obtained with a safe frac- with the lifted part of the ground and is covered by
tile of specific weight and structural dimensions. Uplift is sutTiciently
avoided with II/s W+ N/ /IIR (4)

Fs W /II (3)
Herein Fdenotes the resulting hydrostatic uplift force, II the associ-
wherein the partial factor is lower than in the case outlined above, typ- ated partial factor as in Eq (3), W the nominal weight of the slab plus
ically II-1.05 to 1.1.This is sufficient as the nominal forces Fand W the buoyant weight of the lifted part of the ground, and F a pull-out
do not depend on statical redistributions and are less uncertain there- resistance of the anchor sections below the assumed separating sur-
fore. face. Eq (4) is thus a combination of Eqs (1) and (3) combining ULS 1A
This format can be transferred to the bottom of a drained excavation and B. (Following EC7, one has to consider this case as ULS 1C, but the
with an injected sealing layer (Figure 3). Acting and resisting forces are partial factors for shear strength of the soil would lead to inconsistent
determined by nominal total and water pressures, a and u, of the layer. results.)
F= Wwould express hovering (II= 1), meaning zero effective stress,
a' a u = 0, at the bottom. Eq (3) can again be used, so that mean pres- Retalninl structures
sures are represented by forces. An L-wall can be treated similarly as a shallow foundation with
inclined and eccentric force (Figure 7). A gravity structure with total
Groups of plies or anchors weight W and earth pressure E is obtained for calculation by means of
The building of Figure 1 may now stand on a group of vertical piles a vertical separating surface behind the wall. A resulting force F with
passing through soft ground (Figure 4). the nominal supporting forces inclination 5 and eccentricity e is obtained from Wand E with positions
R,, Rr, etc are calculated as before (possibly including a force redistrib- b and hE . The nominal ground resistance R is calculated with the
ution which can be achieved by attributing subgrade moduli to the sup- nominal 5 and e/b, the characteristic soil parameters, nominal layers
ports). The statically required supporting force R of each pile is trans- and water table. The breadth b is sufficient if Eq (1) is satisfied with the
ferred to the bearing layer (Figure 4 inset). With the nominal acting same factors IIP and IIR as for any other shallow foundation.
and resisting forces, Fand R, determined with characteristic values, Dimensioning the bottom plate with ULS 1B thus includes its breadth b.
assumed dimensions (eg d, and d2 for the fixed pile section) and with- Other failure mechanisms have to be considered with generally

GROUND ENGINEERING sEPTEMBER 1996


PAPER LIMIT STATE DESIGN OF STRUCTURAL PARTS AT AND IN THE GROUND

Rgure L Bottom Rgureg.


slab wgh anchors A P Pge wall wgh
against anchors and earth
hydreshgc upglt III support
(hdt), separalad
anchor(rtght)

Rgme 7. L~pe
iehlnlng wall.

Rgme'tg.
Pge wall wgh two
P rews el anchors.
ha 1ai
rttttt
Az
a,
/
/
/
/
/

other partial factors. Safety against sliding without adhesion is soil parameters. With IIR 1.4 the familiar global factor = -
q IIPIIR 2.0 is
assessed with Eq (2). A passive earth pressure in front of the founda- obtained. Adaptations of Rz are adequate for elevated active earth pres-
tion may be incorporated into F if a ditch can be reliably avoided there. sures and for exceptional design situations.
The global failure can be analysed with a slip circle, as indicated in As stated above for groups of piles, the designs of more complicated
Figure 7, using partial factors for shear strength parameters as pro- retaining structures with a few fixed partial factors can be misleading.
posed for ULS 1C. Bending moments for the design of cross sections of The ratios of supporting forces and heights, such as A/B and hz /h,
the wall and the bottom plate can be determined by estimating parts of may be fixed in empirically well-established cases (comparable to 6 and
the earth pressure and the soil weight above the slab. e/b for shallow foundations). Redistributions can only be followed up
A concrete or steel sheet pile wall with a row of struts and further by detailed observation and calculation in the sense of the observa-
lateral support by the lower ground is treated as a vertical beam with tional method; partial factors are not helpful then.
two supports (Figure Sa). The statically required supporting forces A Resting or streaming ground water can be allowed for by using effec-
and B are obtained from the nominal earth pressure E with two equi- tive forces, ie total forces minus resulting water pressures (Figure Sb).
librium conditions and the heights hz and hz (only horizontal compo- Seepage forces f, at the soil skeleton and net water pressures au can be
nents are considered here). E is the characteristic active earth pressure estimated from nominal water tables. The effective earth pressure E's
due to soil weight and surcharge p if the wall is allowed to yield suffi- calculated with the resulting skeleton volume force y '+ fand E'ith
ciently. hz and hz can be estimated for the nominal situation with estab- y
'
f,, both for the flooded soil fraction, and with total specific weight
lished methods of earth pressure calculation, so representing assumed y for the soil above. The heights hz and hz have to be estimated as
pressure distributions. For the assessment of safety the bedding depth before. (The seepage force may be substituted by an additional net
d has to be assumed, and corrected by iteration. water pressure as indicated by the dotted line in Figure Sb.) As without
Eq (1) is now written for the ULS 1B of the struts and the earth sup- water, the chosen bedding depth d is sufficient if Eq (1) is satisfied. The
port. The nominal actions are the statically required supporting forces struts are calculated with the partial factors prescribed for other struc-
A and B. The associated partial factor IIF is the same as for the previ- tural parts. The total structure is thus treated similarly as the soil
-
ously considered cases. A rather high factor, Rz 1.40, is justified by skeleton as far as the actions of water are concerned.
uncertainties of the calculation model and the parameters although For a concrete or steel sheet pile wall with anchors, vertical compo-
the soil weight is nearly fixed. The nominal resistance Rof the earth nents and pull-out resistance have to be allowed for (Figure 9).
support is the passive earth pressure E calculated with characteristic Inclination e and head depth hof the anchors are chosen. The earth
pressure inclinations, 6, and 6, have to be estimated and must not
exceed the wall friction angle. The statically required supporting
forces A and B are obtained from the earth pressure E (normally
active) and its assumed position. (In addition, the equilibrium of ver-
tical forces can be considered with lower and upper bounds of the ver-
tical force S at the bottom.) The bearing capacity of the earth support
is assessed as before, using a cautious estimate of 6 for E . Cross sec-
tions and pull-out resistance of the anchors are calculatecf with Eq (1).
The partial factor qz may be different from the one for vertical
anchors. The factor qz for the statically required supporting forces A
and B is the same as before, which is again justified by uncertainties
fst of the mechanical model and the parameters. The bending moment for
ULS 1B of the wall cross section is estimated as before. An ULS 1C is
E'gmu
analysed with slip surfaces (dashed lines in Figure 9) which are not
constrained by the anchors. The partial pull-out resistance of the
anchors in the non-sliding part of the ground has to be taken then with
8. Pge wall supporhdhy slnds andhwersog (hdt), same syshmwgh the same partial factor IIR . This case is similar as the one of Figure 6,
groundwahr (right). but this time we have a hybrid of ULS 1B and 1C.

GROUND ENGINEERING SEPTEMBER 1996


PAPER LIMIT STATE DESIGN OF STRUCTURAL PARTS DISCUSSION

.
Rttwo 11 Cut
with soil nol lint)
gelt), ottulllbrlutn
of toraoo at a
T ~~
Rt+82+Rs
Consolidation of
alluvium resulting
olhNnj3 unnttto
w
(rhtht).

from the impounding


A similar approach is proposed for a pile wall with two rows of
anchors (Figure 10). The nominal ratio of the two anchor forces,
AI/As, has to be estimated with the aid of observations, mechanical cal-
culations and engineering judgment. The statically required anchor
of tidal estuaries
forces can then be calculated as in the case of Figure 9. The pull-out by MCR Davies & TA Newson (paper published Ground Engineering
resistance R, and R, are calculated with the same ratio, May, Vol29, No4). Discussion from Alan K Parkln, Dept of Civil
ie R,/R2 = AI/As . The lower earth support is calculated as before. The Engineering, Monash University, Clayton, Victoria 31 68,Australia.
pull-out resistance for an ULS 1C has to be referred to the sections of
the anchors beyond the slip surface, with different lengths for the two ~ Ihe authors rightly anticipate rejuvenated settlements to occur in
anchor rows. the Cardiff Bay sediments as a result of decreased salinity, follow-
A steep cut with a thin wall and several rows of soil nails can be ing construction of the proposed barrage - depending, of course, on
analysed in a similar way (Figure 11). The pull-out resistance of the how far the leaching might extend. They do not, however, make refer-
rows is determined by the section lengths, which is practically equal to ence to a similar experiment by Lun & Parkin (1985), performed on a
the maximal axial nail forces needed for the design of the nail cross saline silty clay from South Melbourne, Australia, in which the oedome-
sections. The resulting pull-out resistance is in equilibrium with the ter cell water was changed f'rom saline to fresh, and I'resh to saline
total weight W, the normal force 1V on the slip surface and the shear (Samples A and B respectively), some time after the conclusion of the
resistance T. As the latter is reduced with partial factors for shear primary phase (Figure 1).
strength we have again a hybrid of ULS 1B and C. Other actions and These results accord with those
mechanisms can be incorporated for the same case. For instance, uplift of the authors insofar as the
and seepage forces can be allowed for by considering effective forces as change to a distilled water envi-
for the case of Figure 8b. Bending moments and transversal forces of ronment leads to a significantly
the thin wall have to be estimated for the design of its cross sections. increased creep component (with
respect to the control sample), and
Condttslons in that there is an appreciable
The use of nominal and design forces in a unified way has some advan- delay before this effect manifests.
tages as against Eurocode 7. Consideration of the real system appeals Predicting these movements is,
to the familiar statical design. The proposed partial factors can easily however, another matter but
be related with established global factors so that ample experience is important details emerge if such
not lost suddenly with a new format. Engineering judgement can be results are presented in velocity
allowed for by adaptation of nominal situations and partial factors. format, as in Figure 1 (noting that
The bulk of experience is preserved with the proposed format, and con- this is not consolidation rate as
fusion with sophisticated mechanics and statistics is avoided. The referred to by the authors). Here it
effective stress principle can easily be incorporated. is seen that Sample B showed only
Limitations are mainly due to the assumed ductility of ground and a small transient increase in set- ' " '~ " "~ ""7
structure. Force distributions and redistributions have to be estimated tlement rate (presumably as salt ie
which is legitimate as long as the lower bound theorem of plasticity depleted water near the exterior
theory is justified. This is not theoretically correct for frictional forces, was drawn into the water bath),
but empirically justified as long as sufficient freedom for the consid- while Sample A showed a more substantial increase that was preserved
ered mechanism is given. in a permanent and parallel translation of the creep line. The merit of
Pre-stressing of structural parts is not considered and not relevant this approach is to show, first, that the continuing movement is creep
for ULS 1B due to the assumed ductility. A certain degree of brittleness and not consolidation (as it is referred by the authors), and, second,
can be allowed by restricting the deformation capacity as for concrete once the creep line has been established, that there exists a basis for
structures. Temporary negative or positive excess pore water pres- predicting continuing settlement with some confidence. There is also a
sures play a similar role as structural pre-stressing: they change due to valuable examination of these effects in Mitchell et al, (1973).
relaxation and dissipation and can lead to a more brittle soil behaviour.
Inertial effects have been excluded. They may be allowed for by a fur-
ther increase of actions and decrease of resisting forces in a quasi-sta- Lun, PTW; & Parkin, AK (1985). Consolidation behaviour determined by the velocity
method. Canadian Geotechnical Journal 22, 158-165.
tic manner. A more detailed analysis, including shakedown and lique- Mitcheu, JK; Greenberg, JA; & Witherspoon, PA (1973). Chemico-osmotic etrects in fine-
faction, should not be made with a few fixed partial factors. grained soils. J Soil Mech & Found Div, ASCE, 99:SM4:307-22, Proc Paper 9678.
Coming back to more conventional structural parts and ground con-
ditions, we summarise that main objectives are achieved by the pro- Reply by MCR Davies fl, TA Newson.
posed concept:
~ The adaptation to well-established global factors is secured; en conducting the study reported in the paper, the authors
~ over- or underdesign can thus be avoided; were not aware of the similar, but not identical, experiment
~ uncertainties of the present EC7, going back to concessions towards conducted by the writers which appears as one aspect of the
structural engineering, are avoided. work reported by Lun & Parkin (1985).With regard to the writer's first
We stress in closing that our concept is not the solution once and for- point, the authors agree that the term 'creep's more correct techni-
ever. The present German codes and standards can and will not serve cally to describe the apparent continued consolidation of the samples.
as the only reference for the coming European standards. Our common In the second point, the writer indicates that the experimentally
interest in the geotechnical community should be: to defend the pecu- obtained creep line provides a basis for predicting settlements. While
liarities of ground engineering as against structural engineering; to agreeing with this, it is important to note that the processes involved
defend well-established methods of design; and to avoid uncertainties in creep induced by chemical-osmosis are, as described by Mitchell et
where our understanding is already sufficient. al (1973), highly complex. In order to obtain accurate quantitative pre-
dictions it is very important that the boundary values of the problem
are correctly reproduced. One important parameter in controlling the
rate of creep is the salt concentration gradient; which is why great
care was taken to ensure that this was highly controlled in the exper-
iments described in the paper.

GROIIND ENGINEERING sEPTEMBER 1996

Você também pode gostar