80) corresponding to Aw < 60-80mm Based on A-logt plot, the failure load isthe load just before an obvious bend atk Sin Quick maintained Nil ? FDT (1999) av took load test. Factor of safety. ‘Smaller piles use Davisson criterion. Diameter > 600 mm, yield at the pile tip, the magnitude of the residual movement is strongly affected by ‘locked-in movement’ of the pile as a result of side shear acting downward against the pile upon unloading thus preventing it from rebounding to its original position. This locked-in effect is almost impossible to assess accurately and reliably. In addition, the residual movement is affected by any creep in the concrete pile. These two phenomena are likely to be more prominent in long and large-diameter bored piles and barrettes than short and small-diameter driven piles (GEO, 1996; Ng et al, 2001c). The residual pile movement criterion is therefore irrational, unreliable and unsuitable for use as an acceptance criterion, especially for long piles and barrettes, and it always leads to unnecessarily conservative design (GEO, 1996). 29SHORT COURSE IN SOIL-STRUCTURE ENGINEERING Table 1.4 Settlement/failure load criteria for pile load tests in Hong Kong Criteria Test method/pile Settlement Key assumptions Related type and size rate references Buildings Am < ae ‘Slow maintained Before Based on Davisson = GEO. Department, pray load Westin Zeyeles Inerements/ caterion (1072) (1906) HKSAR (BD) Mox.test load.” decrements: ‘which are based on PNAP.6, an < (Peed mm) 2sdesign working <00Smmin driven pile #Rwide BD (19973) vy load. 10 min. (0.3m), Davi where PL/ABis the where PL/AEIs the Unioading at end of Before hold: D/120is directly (1872) Sasi shortening, each cycle. <0.1mm/n. transformed from Disthepile |, yah hold at peak tin on it pu, Sramcter, is the oad of cycle 2 given in Davisson pile head sellement. pre type and size not th973, Agisthe residual ated but intended Reasoning behind for bored pies result setlement Exsterion nt stated PLD Housing A 24.P (inmm) HA specification efore Based on Yiu and HA (1998) ‘Author, A requited test carried increments/ Lam 1990; the GEO HKSAR ak|CHAPTER FOUR Establishment of a new failure load criterion for large-diameter bored piles and barrettes Overview: the problem of large-diameter, long bored piles Ng et al. (2001c) have observed that often in the literature, piles used to verify design criteria have been small-diameter driven piles; hence, there is uncertainty as to how suitable these criteria are for large-diameter long bored piles which are common in the Far East in general and in Hong Kong in particular. Because the load-carrying capacities of large-diameter bored piles are relatively large, high-capacity loading-reaction systems are required and the targeted failure load may not always be practical to achieve, A total of 38 large-diameter test bored (drilled) piles founded in weathered rocks and saprolites were critically examined and investi- gated. ‘Saprolites' refer to soils that retain various degrees of the original texture, fabric and structure of the parent rock from which they were formed by decomposition (GCO, 1988). The test piles under review range from 0.6 to 1.8m in diameter and from 12 to 66m in length. The rock sockets were generally around 2m in depth. Details of all of the tests are given in Table 1.5. Among the 38 piles, 31 tests were examined for toe resistance and shaft shortening and incorporated in an investigation of pile behaviour related to the development of a failure load criterion. Based on the study of the mobilization of toe resistance and on a consideration of shaft shortening, a new semi-empirical method was proposed for determining an approxi- mate moderately conservative failure load for large-diameter bored piles. The interpreted failure loads using the new criterion were compared for the five pile tests and further verified independently for two additional piles that were not used in the formation of the criterion. Moreover, pile movements at factored loads were reported and discussed using global and partial mobilization factors. A comparative study of failure loads using some existing criteria is shown in Table 1.6. Pile-head movements corresponding to failure are not reached according to the failure criteria of Brinch Hansen (1963) and Butler and Hoy (1977) and failure load as suggested by the ISSMFE (1985). However, criteria such as the modified Davisson criterion and [a]SHORT COURSE IN SOIL-STRUCTURE ENGINEERING (21002) 1m 12 BN v'D rs ss OOF TI su doy YOM LT CEL yOE (21002) 1 72 BN v'D as zs 000 €% aa qu YOM 8st OORT bz (1002) 1 72 BN wD 06 = 009 00S Fz Lt aoa YOM 709 OFT 48t (21002) 1 72 BN wD Tel eee 002 €1 su GY YOM 709 OKT lz (1002) 142 BN vd o0z «ost STO 1a GX WAM YOF OTT 49% (6661) ‘fo 79 uPUYsUAPTEg va é a é us Aq SINE] OCie | 09/0 4 (6661) 10 12 ueuysUe[eR va Q a i aS Mad SN OST «SLO (9661) 1p 79 ueL, Va @ oBr~ Ozer us @ sD 062 = 060 et (8661) 1p 79 UPL, v'd é o@~ oseL aS. 2 SN 007 190 % (8661) 70 39 wey v'a 2 oer~ —o00ST as 2 SW 0€9 OOL —TZ (e661) 70 32 UPL, vid é o0s~ osze8 us é SN Of SOT oz (8661) 7 39 EL, We é 4 ee us a SW 090 61 (01002) 1» 12 BN wD ge Fst 00s GS goo SN wet gt (21002) 1 72 BN w'D ge or oors us goo. SW ier at (2861) 12 79 NOH. v'D 6st Osh 00L8 ys qoYD so oor 91 (e661) O39 wD VIE 0% 00211 us MOD SO Cty OWT ST (e661) 08D o'D OTE FFE 009 TT as MOD sO 907 OFT as (e661) 08D w'D gol 96T = o0s TT Ys MOD sO r7s OFL et (1002) 1 12 BN vd o8t STE — 0zz0e To GON WAM ZTE OWT (1002) 1 12 BN vd 6L €€% OST TT YS MOD SA 9ST OTE (©1002) 1 72 BN wD 18 ol 001 6T Us MOD SA BTE —OZT or (01002) 1m 12 BN v sp 92 ooze YS MOD SA LE OT (1002) 1 72 BN w'o s9 se ogger ys MOD. SA LB8E OT 8 (9460) aaem vay Q O9E — 09Fe wu 2 UM OT Ok 1002) P12 BN BD 'a'¥ OFT 86L 00097 La ay YOM Ok ET (1002) P19 BN G'DAY se O19 ozzoT ee Cou MAME ce Oct) sc (4661) ‘10 19 WT viv t vis 0002 us a sar 0199 102.0) v (e661) Buc pue BueyD viv Q see osha us ebny so roe S90 £ (21007) 1272 BN. v'O'V ss Bsc 08h KT us dMo SIN coer 0c) z (g661) P12 PAHS aa BD 'a'V OZ 048 ~—-000 Fz as MO Soe Oey, 0a ay mg (ut) (NY) aouarsjoy s210N, eave amy! ™d WLIN OW ws (7 (wp and (21002) ‘18 18 BN Aq pamaraar spoLraDUIOab parayyDam ut sapd perog 19}@uIDEp-abiD] Jo asvqnI0G $I 14D, 32PART | DEEP FOUNDATIONS - PILES AND BARRETTES “dysuoneyor mowarou-peo; wos pareodeaxs sonyen 5166 Suypeo| wi paypes: Jou pow sshreue Aa PoUYEP peOL BIEU=V/N “HUH Es IM aT= ¥ BION OOCE © OSEE «OTE = OPT Z =O = OUSz©=—«NRT=COSPE = WN W/NOETY gOVOE GET = EZ.SC NOEL OOFSz —000SZ_OOZEZ —OODET HOREZ © GOTLT OBST HODEZ ©—GOTST WIN ,OUZFE ,0000F 0O0ZT. © W/N WN 00097 OOL0L —Q0ZOI_ 00S OOZAT_ OTS -—«GEGS_-—«GORE.—«OBOT.-« UNS. «AUOLTT GLLOT OFZ © 00LR_-«S WNL. 00861 §—OORLT 001 ozzI 00811 oGPEY © GOgLL_ © W/N 30000P 00L@2 «WN WN WN BPD ook = WIN 00 st agoz WN 008% “s9LeZ OOFSL_ «WN WN 000K ) a0) osood | AmoUTTSE (x0 aa) %06 oe son uouoyu “pue)pub sox xo sor uosmae “sor xu wOsTenT ost pao poor son posodorg WeNO AUS YH VK (PAMIOOIN SV SY Yau) MoUHg eq aq se at (21002 “Te 19 BN 421)0) 1 pun 9 “G ‘2 “I Sallq :DUAILID pun suojDoyFoads juDAayes UIO4y (NH) SOO] amnyrD) payoudsayUy 9°T A{D, ‘SHOU ft poyraUL 189) peo} paUTEIU mors - pompous so uMouyuN ~ y ajqetreaPun eye ~ 2 aU dus WEA and ~ 0PM LOM OS !AISAL~ SL up pur grub ~ M495 ‘BuIseo OMEN cou Pan Buyseo Arwiodussi yum 19}! zepun jasiyo pue qes6 — Cy ‘oyTordes AreyuouNDasEiOW - SW. eu = WALA “KouDenst0D jo 1b ysod pue Buyse> Arexodwoy wat 191 LMo ‘Busta Areiodusoy ‘ua r9yem tapun qe — ides qn ~ go sBurse> Aresoduray iu JoVm JopuN GIy PUP aeI6- CED iB - 99 ‘suontpL0o {Aap rapt orjeavoxs — Aaq]‘pasiyeue Buluayiots yous ~ 9 ‘pasdfoue sduysuonefOr YuoUTOAotH-aoUPIsIsOr 901 ~ | PassTeUD sdyysHONDIAr WoAsoUl-PeO| ~ y “SION (©1002) 112 BN wD Let 0'6 «= 0008T aL (Gon; AME cer cor ae (21002) 1 72 BN oD ror Fal oz 08 aL dou UNM F6E (OE ALE (91002) ‘12 72 BN o'D OL 907 00961 aL dou YOM 6h OST oe (1002) 1 32 BN v'D B0Z = 9ET 00008 UL dou WN 90F OTT 4SE (01002) ‘1 79 BN wD ost | set oor ge a qdou YAM i rn (01002) 1 79 BN wD or ors 005 sz a qou YAM TL ee (01002) 19 72 BN wD gst 0% (09z8T La qou 4M OO Ze (01002) 1 12 BN wo 6TL FLT — o00ez Li aay YOM oft gle 3BSHORT COURSE IN SOIL-STRUCTURE ENGINEERING 34 the failure threshold of Hirany and Kulhawy (1989) were achieved in all cases, The method of Hirany and Kulhawy (1989) seems to be attractive for use with large-diameter bored piles, as it was verified specifically by loading tests on bored piles, most of which were of large-diameter but founded at relatively shallow depths. The method does not explicitly include the shaft shortening, because the amount of shaft shortening for the short piles considered is small and so the pile-head movement is approximately equal to the pile-toe movement. However, inconsistencies in interpreted failure load could be expected for long piles of 30-70 m in length, if shaft shortening is not explicitly considered. The modified Davison criterion does include an estimation of the shaft shortening, but, from the comparisons, the criterion appears to be over- conservative. Possibly, the general form of the modified Davisson criterion could be adopted, but greater movement of the pile toe should be specified to provide less conservative results. Also, the estimation of shaft shortening in the modified Davisson criterion may not be accurate enough due to shaft resistance reducing the axial load in the shaft, and the estimation should be refined based on actual measured shaft shortening results from loading tests. Method for establishing a new failure load criterion Study of pile toe movement ‘Weltman (1980) and the ISSMFE (1985) suggest that a movement of 10% of the pile diameter corresponds to ultimate load, based on the suggestions of Terzaghi. However, Reese and O'Neill (1988) provide toe resistance— toe movement relationships of bored piles in clays suggesting that full mobilization of toe resistance capacity occurs ai a movement of around 45% of the pile diameter. For bored piles in sands, they report that toe resistance may continue to increase at movements exceeding 15% of the pile diameter. For the purpose of interpreting an approximate failure load for moderately conservative designs, a moderately conservative movement for the mobilization of capacity is preferred. For piles display- ing increasing toe resistance at movements greater than the moderately conservative movement, the interpreted failure load may not represent the actual total load-carrying capacity of the pile, but the interpreted fail- ure load will be comparable with the failure load determined for other piles, Furthermore, as suggested by Reese and O'Neill (1988) and Welt- man (1980), actual ultimate toe resistance for piles in granular materials effectively may never be achieved. Figure 1.5 shows the toe resistance-toe movement relationships for 11 of the piles having reliable data available. Details of the piles are given in Table 1.5. The unit toe resistance, ,, was generally calculated using the measured strain near the pile toe by strain gauges (or by extensometers)PART | DEEP FOUNDATIONS ~ PILES AND BARRETTES 000 2500 = 2000 5 1500 s ©- Pl: Gante soto Aa TE Pie to metnsertany sapote e TE Plo 2 Motsndmosay eoate z SP 21: atasederetnysaprote Pho 22 Metamora eprae 500 Topless: Grant sap TE phe 24 Metanctay sgote Tr fhe 28 Meteseomertay spre ° 012045 6789 0H 2 19 4 15 (a) Normalized toe movement, 5,/d (96) 16000 14.000 = 12000 © 10000 5 000 el a) 8 5 4000 -0- Pa: Weathered vlan ok ‘2000 45% —@ Pile 6: Weathered granitic rock “a Pio 7 Wenbered ok ° 01245 67 89H B14 15 (o) Normalized toe movement, dy/d (%) Fig. 1.5 Toe resistance-movement relationships for (a) piles embedded in saprolites; (b) piles socketed in and founded on weak rocks (aiter Ng et al., 2001¢) and by estimating the pile modulus of elasticity. The movement 4, is the ‘local’ toe movement, obtained by subtracting the shaft shortening (measured by extensometers or strain gauges) from the measured pile head movement and is normalized by d. Figure 1.5(a) shows the relation- ships for piles founded on saprolites. The figure indicates scatter in the rate of mobilization of resistance and varying magnitudes of r,, at which yielding began. The variation may be due to the soil originating from different rock types, the varying density and confining pressure. The 35SHORT COURSE IN SOIL-STRUCTURE ENGINEERING LE] construction of each pile, involving unknown levels of disturbance, must also be a factor. For most of the test piles, toe resistance capacity was not reached. An exception is Pile 20, which did achieve its capacity, and this occurred at a displacement of 4.5% of d. Figure 1.5(b) shows the relationships for three piles founded on weathered rocks. There is scatter in the magni- tudes of mobilized toe resistance, but all three relationships are consistent, in that significant yielding clearly occurs at a movement of 0.5% of d. However, resistance is observed to be still increasing after this point. The three tests were terminated at a toe movement of around 4.5% of d and in the case of Pile 5, the test was terminated due to the rate of pile movement exceeding the maximum allowed by the test requirement (0.5mm in 10mm) From Fig. 1.5, it appears likely that the full mobilization of toe resistance is difficult to achieve, as it may require extremely large toe movements. However, a moderately conservative movement for the mobilization of toe resistance capacity has been identified to be 4.5% of d, and this is con- sistent for both piles founded on saprolites and piles founded on weath- ered rocks. The results are also consistent with the suggestions of Reese and O'Neill (1988). This approximate movement could be adopted in a new method of interpretation of moderately conservative failure load. Study of shaft shortening The modified Davisson criterion includes an estimation of the pile shaft shortening equal to PL/AE. The estimation assumes that no frictional resistance acts along the whole pile length. The load transferred as shaft friction depends on many factors, such as soil type and construction details. Whether the pile toe is ‘floating’, i.e, founded in soil or bearing on rock, also affects the total axial load in the shaft. A comparison of measured pile shaft shortening and calculated pile shaft shortening using PL/AE at measured P,,.; obtained from 28 pile loading tests is shown in Fig. 1.6, Details of the 28 test piles are given in Table 1.5, All of the measurements in Fig. 1.6 lie on or above the line for calculated shortening equal to PL/AE, indicating that shortening is overestimated by the expression. The piles founded in soils generally have shortening measurements between values calculated by }PL/AE and }PL/AE. The second expression is consistent with the theoretical shortening for piles with zero toe resistance and for which the average shaft resistance is con- stant along the pile length. For piles with toe resistance, the theoretical shortening will be greater than }PL/AB. This explains the measured values in soil, which are as high as } PL/AE. Figure 1.6 also shows results for 'end-bearing piles’, piles with rock sockets and piles with slip liners. ‘End-bearing piles’ refer to piles founded on rocks and relying predomi-PART | DEEP FOUNDATIONS - PILES AND BARRETTES 3 PL PL 4 Ae/ 88 aE pam 2 ‘© Piles founded in soil ‘¢ End-bearingirock socket piles (without slip liner) ‘4 End-bearingfrock socket ples (with slip finer) terpreted shortening, PLIAE: mm 20 30 40 Measured shortening: mm Fig. 1.6 Pile shaft shortening (after Ng et al., 2001¢) nantly on toe resistance. ‘Slip liners’ refer to casings, often coated in a material such as bitumen and sometimes surrounded by a soft grout annu- lus. They are intended to allow minimal shaft resistance and thus allow maximum axial load to reach a test section below the liner. The piles with rock sockets and slip liners and the end-bearing piles may be expected to give similar results, due to the majority of the pile load being transferred at or close to the pile toe. The measured shortening would be expected to be higher than the piles founded in soils, due to shaft resistance being relatively lower, and this is supported by Fig. 1.6 with the results lying between PL/AE and $PL/AE. For use in interpreting failure load, a moderately conservative estimate of shaft shortening may be appropriate. From the results in Fig. 1.6, the moderately conservative estimates are }PL/AE and 3PL/AE for piles founded in soils and end-bearing piles, respectively. End-bearing piles are piles that rely predominantly on toe resistance, piles with rock sockets and piles with slip liners. Note that these estimations may also err on the conservative side for the reason that the maximum test load P;,,. may be less than failure load, as determined by the new criterion in which the esti- mation will be incorporated. New proposed method of interpretation From the previous discussions, 4.5% of the pile diameter (or 0,045d) seems to be a moderately conservative toe movement for the mobilization of toe resistance capacity of large-diameter bored piles founded on saprolites and weathered rocks. Combining this movement with the observations of shaft shortening, a new method for interpreting failure load is given 7SHORT COURSE IN SOIL-STRUCTURE ENGINEERING [=| as follows (in mm): 1 PL Am ¥ 0.045d + = = (17 mM 3 AE (1.7) for piles founded in soils, and An x 0.085d-+2 7 (1.8) 4 AE for piles founded on rocks and/or with short rock sockets and for piles with slip liners. In any case that the pile-head displacement under maximum test load is less than the value given by the proposed method, the maximum test load can be taken as the conservative failure load for design, Sensitivity analysis of global Factor of Safety For practising civil engineers, pile-head movements at the interpreted failure load scaled down by factors of safety are often of interest. Move- ment limits of piles will vary with structural requirements, the design methodology, and the choice of factor safety. These issues will not be addressed here. However, the effectiveness of some of the failure criteria introduced in this chapter in controlling pile-head movements can be examined by studying the pile-head movements measured at the inter- preted failure load and at the interpreted failure load divided by a factor of safety. Figure 1.7 shows the relationship between pile-head movement, Aawork {assumed working load movement), as measured during loading testing at 100 p> + Davison eas. + Proposed erterion ' Proposed criteria (max) Proposed crteria (min) AS (max) pec ET ee AS (min) 20 [Mod. Davissony \\ Mod. Davisson (max) (min) Pile-nead movement at assumed working load (ty ah ey ry re ER ee Nominal global factor of safety F Fig. 1.7 Pile movement at assumed working loads (after Ng et al., 2001¢)PART | DEEP FOUNDATIONS - PILES AND BARRETTES a load equal to the interpreted failure load divided by a nominal global factor of safety, F, defined as ‘working load’. Values of F from 1.0 to 4.0 were applied to the interpreted failure loads for Piles 1-7, calculated using the new proposed criterion, the modified Davisson criterion by BD, and the AS criteria, which is dominated by its residual movement cri- terion. The ‘max’ and ‘min’ in Fig. 1.7 refer to the maximum and minimum measured movements for each criterion or set of criteria, respectively. At F equal to 1.0 (.e, ultimate load) the scatter in A, yor is considerable, but the scatter decreases as F increases. The modified Davisson criterion pro- duces the lowest values of Aaworx Of the three criteria, with the maximum line for this criterion on the figure being less than 12mm for F equal to 1.5 and less than 8 mm for F equal to 2.0. For F equal to 2.0, the AS criteria and the new method provide similar results, with the maximum movement being around 15mm. Unlike the new method, the AS criteria are not dependent on pile dia- meter; hence, at greater pile sizes, the AS criteria and the new proposed criteria would be expected to produce a greater variation in interpreted results, The AS criteria should tend to limit the factored movements con- sistently to around the same values as those shown in Fig. 1.7, which could be considered to be an advantage by structural engineers. However, as mentioned in previous sections, the degree of mobilization at failure load for different pile diameters is likely to vary because the criteria are independent of diameter. Provided that structural tolerable movements are not exceeded, the new method will result in a more effective design. If structural requirements are demanding, for greater pile sizes the safety factor may need to be increased to a value greater than 2.0. Summary A new failure load criterion for large-diameter bored piles and barrettes is that the pile head displacement Ay, at failure load in weathered geo- materials may be given as follows (in mm): 1 PL An © 0.045d +5 Fe (1.7) for piles founded in soils, and 3 PL , An © 0.045d +7 Ze (1.8) for piles founded on rocks and/or with short rock sockets and for piles with slip liners. [>]CHAPTER FIVE Analysis, design and parameters Categories of analysis/design procedures As pointed out by Poulos (1989), analysis and design procedures can be divided into three broad categories, depending onthe level of sophistication and rigour. An extended classification system of these procedures has been proposed by Poulos and Hull (1989) and is shown in Table 1.7. Category 1 procedures probably account for most pile design done throughout the world, Category 2 procedures have a proper theoretical basis, albeit simplified, and are being increasingly used for pile deflection calculations. Such procedures involve the use of simple computational methods or design charts, and generally do not demand the use of a computer, Category 3 procedures involve the use of a site-specific analysis based on relatively advanced analytical or numerical techniques such as the finite-element method or the boundary-element method. In most cases, such procedures require the use of a computer. Category 3 procedures are frequently used to carry out the necessary parametric solutions and develop the design charts which can then be used as category 2 solutions. Typical examples of the various categories of procedures for axially loaded piles are shown in Table 1.8, In choosing an appropriate category of design for a practical problem, the following factors need to be considered: the significance and scale of the problem the available budget for foundation design the geotechnical data available the complexity of both the geotechnical profile and the design loading conditions © the stage of the design process (i.e. whether a feasibility, preliminary or final design is being carried out) © the experience of the designer with the methods being considered. Estimation of geotechnical parameters If soils can be idealised as elastic materials, the most significant para- meters required for many of the categories 2A, 2B, 3A and 3B analyses [|PART | DEEP FOUNDATIONS - PILES AND BARRETTES Table 1.7 Categories of analysis/design procedures (after Poulos, 1989) Category Subdivision Characteristics Method of parameter determination 1 = Empirical - not based on soil Simple in situ or mechanics principles laboratory tests, with correlations 2 28 Based on simplified theory or Routine relevant in situ charts — uses soil mechanics tests - may require some principles - amenable to hand —_correlations calculation. Theory is linear elastic (deformation) or rigid plastic (stability) 2B As for 2A, but theory is non- linear (deformation) or elasto- plastic (stability) 3 3A Based on theory using site- Careful laboratory and/or specific analysis, uses soil in situ tests which follow mechanics principles. Theory isthe appropriate stress linear elastic (deformation) or paths rigid plastic (stability) 3B As for 3A, but non-linearity is allowed for in a relatively simple manner 3c As for 3A, but non-linearity is allowed for by way of proper constitutive models of soil behaviour of pile behaviour under static loading are as follows: . . . the side resistance f, the end-bearing resistance fy, Young's modulus of the soil E, Poisson's ratio of the soil v,. For calculation of axial pile load capacity, f, and f, must be estimated as accurately as possible. For the calculation of settlement resulting from direct axial loading, the theoretical solutions reveal that the choice of an appropriate value of £, is generally crucial, unless the piles are long and compressible. For piles in soil subjected to external movement, the pile behaviour is generally much less dependent on E, and, provided that the soil movement is known,’ an approximate estimate of E, may be ® The soil movement is treated here as an independent variable, although it will be influenced to some extent by the soil modulus. atSHORT COURSE IN SOIL-STRUCTURE ENGINEERING Table 1.8 Examples of categorization of methods for evaluation of axial pile response (aiter Poulos, 1989) Category Axial pile capacity Settlement 1 Correlations with CPT Approximate correlations with pile (e.g. Schmertmann, 1975; De Ruiter diameter (Meyerhof, 1959; Frank, and Beringen, 1979) 1985) Correlations with SPT (Meyerhof, 1956; Column deflection multiplied by a Thorburn and McVicar, 1971) factor (Focht, 1967) Total stress (a) method (Tomlinson, 1957; Semple and Rigden, 1984) 24 Effective stress (3) method (Burland, _ Elastic solutions (Randolph and 1973; Meyerhof, 1976; Stas and Wroth, 1978a; Poulos and Davis, Kulhawy, 1984) 1980) 2B Effective stress method (Fleming ef al, Elastic solutions modified for slip 1985) (Poulos and Davis, 1980) 3A Plasticity solutions for end-bearing Elastic finite-element analysis capacity (Meyerhof, 1963; Giroud et al, (e.g. Valliappan et al., 1974) 1973) 3B Non-linear load transfer analysis (e.g. Coyle and Reese, 1966; Kraft et al, 1981) Non-linear boundary-element analysis (e.g. Poulos and Davis, 1980) Non-linear finite-element analysis (e.g. Desai, 1974; Jardine et al., 1986) 3c Finite-element analysis, including simulation of pile installation (e.g. Randolph et al., 1979; Withiam and Kulhawy, 1979; Nystrom, 1984) a2 adequate, although reasonable estimates of shaft and end-bearing resistance are desirable. Methods for determining parameters For evaluating the parameters for static pile response, a number of methods can be contemplated, including laboratory testing appropriate interpretation of field pile load tests empirical correlations with laboratory determined parameters empirical correlations with the results of in situ test data Conventional laboratory tests, such as triaxial or oedometer tests, are generally not suitable for direct measurement of the soil’'s Young's modulus as they do not follow, even approximately, the stress pathPART | DEEP FOUNDATIONS - PILES AND BARRETTES which the soil adjacent to the pile follows. Laboratory model pile tests may overcome this deficiency to some extent, but may not accurately reflect the behaviour of prototype piles because of the presence of scale effects, particularly for piles in sand. Some potential exists for more sophisticated tests such as the constant normal stiffness (CNS) direct shear test (Johnston and Lam, 1984; Ooi and Carter, 1987), and this type of test has been used in the design of grouted piles in offshore carbonate sediments (Johnston et al., 1988). However, the direct utilisation of laboratory tests for pile design is infrequent in practice, and still requires further research before it can be applied with confidence. The most reliable means of determining f, and E, is by back-analysing from the results of pile load tests. Methods for interpreting the pile load test data have been detailed by Poulos and Davis (1980) and Stewart and Kulhawy (1981), among others. Such methods are particularly effective if the pile is instrumented so that details of the load transfer along the pile shaft are available; it is then possible to determine detailed distributions of soil modulus and limiting pile-soil resistance along the pile shaft (Ng et al., 2001b). Side resistance f, Tables 1.9 and 1.10 summarize available methods for determining side resistance, f,, from laboratory strength data, for both driven and bored piles. Effective stress approaches can be used for all soil types, whereas a total stress approach is still adopted commonly for piles in clay, The parameters a and @ (or K and 6) are usually obtained from empirical correlations, despite the fact that the effective stress 9 approach is fundamentally sound and falls into category 2. A summary of some suggested correlations between f, and the standard penetration resistance N are given in Table 1.11. Considerable variations occur in these correlations, particularly for bored and cast-in-place piles. Figures 1.8 and 1.9 show values of f, correlated with the static cone resistance, q., These relationships have been developed by Poulos (1989) from the correlations suggested by Bustamante and Gianeselli (1982) and cover a wide range of pile types in both clay and silica sand. The classification of these pile types is shown in Table 1.12, It should be emphasized that several other correlations have been proposed and that wide variations exist between some of these. Figure 1.10 shows an example of this variability, for driven piles in silica sand. The potential inaccuracy of shaft capacity prediction using category 1 correlations, especially for loose sands, is clearly demonstrated. Schmertmann (1975; 1978) proposes a different approach to the utiliza- tion of cone data, whereby the pile side resistance is related to the measured sleeve resistance of the penetrometer. Corrections are applied,SHORT COURSE IN SOIL-STRUCTURE ENGINEERING Table 1.9 Shaft resistance {, for driven piles, determined from laboratory strength data {after Poulos, 1989) Soil type Equation Remarks Reference Clay f= acy, a= 1.0(¢, < 25kPa) API (1984) = 0.5(¢y > 70 kPa) Linear variation in between @ = 100(c, < 35kPa) ‘Semple and Rigden @ = 0.5(¢, 2 80kPa) (1984) Linear variation in between Length factor applies for L/d > 50 “@lGy ™( -QLE)" =(G Fleming et al. (1985) Bo, (1 sing’) tan g'(OCR)°* Burland (1973 Meyerhof (1976) Silica sand f, = J, 9 = 0.15-0.35 (compression) McClelland (1974) ( ¥ fim) 0.10-0.24 (tension) B=044 — for ¢! = 28° Meyerhof (1976) 075 forg’ 12 ford! =37° 8 = (K/K): Ko tan(- 5/4) Stas and Kulhawy (1984) 6/0 depends on interface materials (range 0.5-1.0); K/Ky depends on installation method (range 0.5-2.0) Kg = coefficient of earth pressure at rest, and is a function of OCR Ja, 9 = 0.05-0.1 Poulos (1988d) calcareous sand depending on soil type, pile type, relative pile length and depth below the surface. Robertson et al. (1985) have found that the method proposed by Schmertmann provides a more reliable prediction than the direct correlation to q. when applied to piles in a clayey silt. Correlations such as those outlined must always be employed with caution, as a number of other factors may also influence side resistance, e.g. the presence of overlying layers (Tomlinson, 1977). End-bearing resistance fi, Table 1.13 summarizes the two methods usually used for assessment of the end-bearing resistance of piles using laboratory data. A total stress approach is almost invariably used for piles in clay, whereas an effectivePART | DEEP FOUNDATIONS - PILES AND BARRETTES Table 1.10 Shaft resistance f, for bored piles, determined from laboratory strength data (after Poulos, 1989) Soil type Equation Remarks Reference Clay f= acy ‘= 0.45 (London clay) Skempton (1959) a= 0.7 times value for driven Fleming et al. (1985) displacement pile A Kis lesser of Ky or 0.5(1+Ky) Fleming et al. (1985) K/Ko = 2/3to 1; Ky is function of Stas and Kulhawy OCR; 6 depends on interface (1984) materials Silica sand f, = ol, $=01 — ford’ =33° Meyerhof (1976) 0.2 for ¢’ = 35° 035 for g! = 37 8 = Ftan(¢' — 5°) Kraft and Lyons (1974) Uncemented calcareous sand where F = 0.7 (compression) and 0.5 (tension) f, = Boy §=0.5 100.8 Poulos (198d) {fy # fetirn) fetim ~ 60 to 100 kPa stress approach is used for piles in sand. Two main problems arise in the latter case: * some experimental evidence suggests that a limiting value of f, may occur when the pile is embedded more than 10 to 20 diameters; no entirely satisfactory method of theoretical analysis has been devel- oped to take this into account and an empirical upper limit to fi, is usually specified ¢ the theoretical bearing capacity factor, Ng, is very sensitive to the angle of shearing resistance ¢! for values of ¢' in excess of about 35°, thus small changes in ¢ can theoretically lead to large changes in Nq although the effects of soil compressibility are then more important and may reduce the dependence of N, on ¢’. Table 1.14 shows some empirical correlations between f, and the standard penetration resistance in the vicinity of the pile tip. These correlations indicate that bored or cast-in-place piles develop a significantly smaller end-bearing resistance than do driven piles. Bustamante and Gianeselli (1982) suggested correlations between fy and the average cone penetration resistance value near the pile tip. The correlation factor to the latter value is between 0.3 and 0.55, depending on soil and pile type. These correlations contrast with procedures such as those proposed by Belcotec (1985) and De Ruiter and Beringen (1979), in which a factor of unity is applied to the average value (computed [+]SHORT COURSE IN SOIL-STRUCTURE ENGINEERING Table 1.11 Correlations between shait resistance f, and SPT value, with f, = 0 + NkPa (ater Poulos, 1989) Pile type Soil type a p Remarks, Reference Driven Cohesionless 0 2.0 f, =average value Meyerhof (1956) displacement over shaft Cast-in-place Bored N=average SPT —_Shioi and Fukui (1982) along shaft Halve f, for small- displacement pile Cohesionless 10 3.3. Pile type not Decourt (1982) and cohesive specified N23 i, # 170kPa Cohesive 0 100 Shioi and Fukui (1982) Cohesionless 302.0. f, ¥ 200kPa Yamashita et al. (1987) Shioi and Fukui (1982) 0 50 Cohesive 0 5.0 f, ¥150kPa Yamashita et al. (1987) Shioi and Fukui (1982) 0 100 Cohesionless 0 10 Findlay (1984) Shioi and Fukui (1982) 0 33 Wright and Reese (1979) Cohesive Oo 0) Shioi and Fukui (1982) Cohesive 10 3.3. Piles cast under _—_—Decourt (1982) bentonite 504NZ3 f, #170 kPa Chalk -125 12.5 30>N>15 After Fletcher and i, % 250kPa Mizon (1984) 46 differently than in the Bustamante and Gianeselli approach). However, the latter approaches are confined to driven piles, whereas the Busta- mante and Gianeselli approach is more general, simpler to apply, and probably more conservative. Young’s modulus E, of the soil Ideally, for piles in clay, a distinction should be made between the undrained Young's modulus, used for calculations of immediate or undrained settlement, and the drained Young's modulus, used for calcu- lations of total settlement of a pile. However, for many clays, the difference between the drained and undrained modulus values is not great and the approximate nature of most correlations makes such a distinction imprac- tical. It is therefore suggested that the correlations presented here should be considered to apply to the drained Young's modulus. For alll soil types,PART | DEEP FOUNDATIONS - PILES AND BARRETTES 160 Cure | Applicable, pile types | Note: Lower limit applies for No, (see Table 1.12) | unreliable construction contro! upper fmt applies for very sao ft | 1B, tower timit for fea coeseuant antl cece ‘careful construction control 2 | Upper limit for 18 320} 3 | INA, upper tit for 1A iA and 1B i) 100 Shalt resistance f; KPa 40 20 ers C0 Cone resistance a,: MPa Fig. 1.8 Design values of shait resistance for piles in clay (based on Bustamante and Gianeselli, 1982; after Poulos, 1989) a further distinction needs to be made, between the tangent value of the Young's modulus (if a non-linear interface model is being used) and the secant value of the Young's modulus (if a purely linear analysis is being used). Again, it is often difficult to make such distinctions with rough empirical correlations and, unless specified, the Young's modulus referred to here will be a secant value, relevant for normal working load levels of between one-third to one-half of the ultimate load capacity. For piles in clay, the Young's modulus has often been correlated with laboratory-measured undrained shear strength, ¢,. Some of these correla- tions are shown in Fig. 1.11, and a feature of this figure is the wide spread of correlations. Possible reasons for this spread might include differences in the method of determining c,, differences in the method of determining the modulus values, differences in the load level at which the modulus was determined, differences in the overconsolidation ratio of the clay between different tests, and differences between the clay types. Callanan and Kulhawy (1985) find that values of E,/c, generally range between 200 and 900, with an average value of about 500, These values apply to piles with a length-to-diameter ratio in excess of about 15. For shorter piles, the upper range of E,/c, may be greater because of the possible effects of fissuring, desiccation and overconsolidation of the clay near the surface. ]SHORT COURSE IN SOIL-STRUCTURE ENGINEERING 48 Fig. 1.9 Design values of shaft resistance for piles in sand (based on Shait resistance f,: KPa Cure | Apotcable pile tes | Note: Lower it appies for No] (s0e Table 1.12) unreitle constuction contol: 300 | th | Lower limit for 8 upper tnt apples for very 10 | Upgerurt fori catetul construction control 2 | ine 3. | Lower limit for 1A and A 3U_| Upper limit fort and iA 20 4 | mw 5 200 5 150 | 4 wy x 3 100 tu eae — ot 1 50 = ° ° 5 10 15 20 280 Cone resistance 9, MPa Bustamante and Gianeselli, 1982; after Poulos, 1989) Table 1.12 Classification of pile types (Bustamante and Gianeselli, 1982; aiter Poulos, 1989) Pile Type of pile category 14 Plain bored piles, mud bored piles, hollow auger bored piles, cast screwed piles ‘Type I micropiles, piers, barrettes 1B Cased bored piles Driven cast piles HA Driven precast piles Prestressed tubular piles Jacked concrete piles 1B Driven steel piles Jacked steel piles MA Driven grouted piles Driven rammed piles IIB High-pressure grouted piles (d > 0.25m) Type II micropilesPART | DEEP FOUNDATIONS - PILES AND BARRETTES 120 a 4 100 4 2| [as 3 80 € 2 60 5 3 3 3 12 2 4 _ 12 zo | | | ° (a) (b) © Number Source of correlation 1] “Bustamante and Gianesell (7082) 2 | Fleming and Troroum (1084) 3 | Verbrugge (1982) 4 | Vanimpe (1986) 5 | Poulos (1889) Fig. 1.10 Example of variations between correlations for shaft resistance against CPT-driven piles in sand: 6 MPa; (b) qe = 10 MPa; (c) dc = 20MPa (alter Poulos, 1989) Some correlations between the Young's modulus and the standard penetration test number are summarized in Fig. 1.12, and show alarming variability. At least some of this variability may be attributed to differences in the determination or definition of the SPT value, but it is clear that the potential for selection of inappropriate values of E, is great. Table 1.15 shows some suggested correlations between E, and the cone penetration resistance, qc, and, as with most of the other correlations, the range is large. Two correlations for initial tangent modulus, E,,, are shown, both being derived from dynamic triaxial tests but believed to be relevant for piles. Poisson's ratio v, ‘The Poisson's ratio of the soil is a necessary input parameter into analyses that involve elastic continuum theory, but its effect is generally quite minor when the solutions are expressed in terms of the Young's modulus [1 49}SHORT COURSE IN SOIL-STRUCTURE ENGINEERING Table 1,13 End-bearing capacity of pile tip {,: determination from laboratory data (after Poulos, 1989) Soil type Equation Remarks Reference Clay fy =Netyy Ne =9for L/d > 3 Skempton (1959) Gy = value of , in vicinity of pile tip Silica sand’ fy = Nqow Ny =40 API (1984) F losis’ N, plotted against ¢ Berezantzev ef al. (1961) Nj related to ¢', relative density Fleming et al. (1985) and mean effective stress 'N, from cavity expansion theory, Vesic (1972) as'a function of ¢! and volume compressibility Uncemented fy = Nya N, = 20 Datta et al. (1980) calcareous hum Typical range of Ng = 8-20 Poulos (1988d) sand [N, determined for reduced value Dutt and Ingram (1984) of 6! (eg. 18°) “For silica and caleareous sands, the above expressions apply for driven piles only "Typical limiting values fyys; range from 10-15 MPa for silica sand, and 3-5 MPa for calcareous sand; the latter value depends on soil compressibility (Nauroy et al, 1986) Table 1.14 Correlations between end-bearing resistance f, and SPT value, with 1, = KN MPa (after Poulos, 1989) Pile type Soil type K Remarks Reference Driven Sand 0.45 N =average SPT value in. Martin et al. (1987) displace- local failure zone ment Sand 0.40 Decourt (1982) Silt, sandy silt 0.35 Martin et al. (1987) Glacial coarse to fine 0.25 Thorburn and MacVicar silt deposits (1971) Residual sandy silts 0.25, Decourt (1982) Residual clayey silts 0.20 Decourt (1982) Clay 0.20 Martin ef al. (1987) Clay 0.12 Decourt (1982) All soils 0.30 For L/d>5 Shioi and Fukui (1982) ILjd <5; K=0.1+0.04L/d (closed-end piles) or K =0.06L/a (open-ended piles) Cast-in- _ Cohesionless fy = 3.0MPa Shioi and Fukui (1982) place 0.15 f, #7.5MPa ‘Yamashita et al. (1987) Cohesive = fy = 0.09(1 + 0.162) Yamashita et al. (1987) ‘where z = tip depth (m) Bored Sand oA Shioi and Fukui (1982) Clay 0.15 Shioi and Fukui (1982) Chalk 0.25 N<30 Hobbs (1977) 0.20 N>40 [>|PART | DEEP FOUNDATIONS -~ PILES AND BARRETTES 120 100 Young's modulus E,: MPa 20 ° ° 00 200 300 Undtained shear strength g, kPa Legend [Remarks Reterence 1 Driven piles Pouios (1968) 2 Bored piles Poulos (1968) 3 Driven piles ( ‘Aschendrenner and Olsen (1964) 4 Bored piles, lower bound | Callanen and Kulhawy (1965) (E=200¢) Fig. 1.11 Correlations for soil modulus for piles in clay (after Callanan and Kulhawy, 1985; Poulos, 1989) 600 Komorik (1974) D'Appotonia etal. (1970) 500} 3 Shiol and Fuku’ (1982) Denver (1982) CChristoulas and Pachakis (1987) Yamashita et al. (1987) MPa 0 20 40 60 SPT value Fig. 1,12 Comparison between correlations for soil modulus and SPT value for driven piles in sand (after Poulos, 1989) 51SHORT COURSE IN SOIL-STRUCTURE ENGINEERING Table 1.15 Correlations between soil’s Young’s modulus E, and CPT value for driven piles (aiter Poulos, 1989) Soil type Correlation Remarks Reference Clay and E,* =21.0qi Various pile types Christoulas (1988) silts E, and q. in MPa E, = 15qe Poulos (1988c) Silica sands E, ~ aq = 20-40 Milovic and Stevanovic (1982) a=5 Poulos (1988¢) (normally consolidated sands) a=75 (over-consolidated sands} Ey! = 53q2"! Bq, and q. in MPa Imai and Tonouchi Dynamic modulus value (1982) Unspecified Ey = aq. a= 24-30 Holeyman (1985) Dynamic modulus value E,=10.8+6.6q. E, and q. in MPa Verbrugge (1982) (for q. > 0.4 MPa) “8, = secant Young's modulus. * B =initial tangent Young's modulus of the soil. For saturated clays under undrained conditions, v, can be taken as 0.5. For clays under drained conditions, », generally lies within the range 0.35 + 0.05, whereas for silica sands, v, is usually within the range 0.30.1, Lower values, within the range 0.15 0.1, are applicable for many marine calcareous sediments. [=|CHAPTER SIX Dynamic formulae Overview As pointed out by Poulos and Davis (1980), perhaps the oldest and most frequently used method of estimating the load capacity of driven piles is to use a driving formula, or dynamic formula, All such formulae relate ultimate load capacity to pile set (the vertical movement per blow of the driving hammer) and assume that the driving resistance is equal to the load capacity of the pile under static loading. They are based on an idealised representation of the action of the hammer on the pile in the last stage of its embedment, There are a great number of driving formulae available, of varying degrees of reliability. Smith (1960) states that the editors of the Engineering News Record have on file 450 such formulae. The derivation of some of these formulae is discussed by Whitaker (1970), while details of some of the parameters required are available in Chellis (1961). The primary objectives in using a pile-driving formula are usually either to establish a safe working load for a pile by using the driving record of the pile, or to determine the driving requirements for a required working load. The working load is usually determined by applying a suitable safety factor to the ultimate load calculated by the formula. This safety factor, however, varies considerably depending on the formula used and the type of pile being driven. Also, because pile driving formulae take no account of the nature of the soil, the appropriate safety factor may vary from one site to another. An improvement in the estimation of load capacity by dynamic methods has resulted from the use of the wave equation to examine the trans- mission of compression waves down the pile, rather than assuming that a force is generated instantly throughout the pile, as is done in deriving driving formulae. The main objective in using the wave-equation approach is to obtain a better relationship between ultimate pile-load and pile-set than can be obtained from a simple driving formula. As well as providing a means of load capacity estimation, this relationship allows an assessment to be made of the driveability of a pile with a particular set of equipment. Moreover, this approach also enables a [>|SHORT COURSE IN SOIL-STRUCTURE ENGINEERING rational analysis to be made of the stresses in the pile during driving and can therefore be useful in the structural design of the pile. Pile-driving formulae Derivation of general formulae Pile-driving formulae attempt to relate the dynamic to the static resistance of a pile. They have been established on an empirical or a theoretical basis. Several of the latter are based on Newton's law of impact, modified in some cases for energy losses during impact and stress propagation, The assumed relationship between pile resistance and downward movement of the pile is shown in Fig. 1.13. The materials of the pile and the driving cushion are assumed to be perfectly elastic, and inertia forces in the soil and energy losses stemming from irreversible deforma- tions (except of the soil) are disregarded. Dynamic equations Dynamic equations can be categorized into three groups: theoretical equations, empirical equations and those which consist of a combination of the two. The theoretical equations are formulated around analyses which evaluate the tolal resistance of the pile, based on the work done by the pile during penetration. These formulations assume elasto-plastic force-displacement relations, as shown in Fig. 1.13. The total work is therefore: Q (s + at (1.9) where R, is the yield resistance, Q is the quake denoting the combined elastic deformation of the pile and the soil, and S is the set, denoting the Load Seitlement Fig. 1.13 Assumed load-settlement curve for pile (after Poulos and Davis, 1980)PART | DEEP FOUNDATIONS - PILES AND BARRETTES plastic deformation (permanent displacement ~ also called the ‘final set value’) under each blow. When the work of the resisting forces W is equated to the energy deliv- ered to the pile, that is, W = E,ior (Where Exe; is the energy transferred to the pile-soil system after impact), then: pee to) S4 2 2 The low reliability of dynamic equations has been reported (Housel, 1965). There are several reasons for this: © the parameters, such as the efficiency of energy transfer and the pile/soil quake, are assumed, and therefore may not reflect the high variability of the field conditions the theoretical analysis of the ‘rational’ pile formula (see for example Bowles, 1988) relates the energy transfer mechanism to a Newtonian analysis of ram-pile impact. This formulation is theoretically invalid for representing the ‘elastic’ wave propagation mechanism which actually takes place, and * no differentiation is made between static and dynamic soil resistances. A clear distinction is therefore required between the underlying valid energy analysis and additional estimations of the different parameters, many of which are either invalid theoretically or practically limited in their accuracy. If the anvil-pile cushion is included as part of the pile-soil system, energy will also be stored during elastic compression of the pile cushion. The elastic compression should therefore comprise three components, namely, Q=C,+C,+Cy where C., C, and C, are the temporary compression of pile cushion, pile and soil respectively. If the anvil-pile cushion is not considered as part of the pile-soil system, the energy transferred to the pile-soil system may be simplified to energy stored in the pile and soil only. In this case, Q = C, + Cy and E,ye, may be approximated by the total energy transferred to the pile head, Ey, where Eq can be measured directly using a PDA (Pile Driving Analyser). Eq can also be expressed in terms of the energy transfer ratio X, defined as E, = XWh, where W is the hammer weight and h is the drop height. The values of Ey and X are normally standard output in a PDA test report. In some pile-driving formulae, Ey is related to the kinetic energy of the hammer before impact F’. It is common to express £' in terms of the hammer efficiency a defined as F’ = Wh. The hammer efficiency measures the energy loss during descent of the hammer. Some hydraulic hammer systems impart energy to the hammer during descent of the 55SHORT COURSE IN SOIL-STRUCTURE ENGINEERING 56 hammer and the input energy of the hammer fer pile driving is actually higher than Wh. This may give a hammer efficiency higher than 1.0 even if there is energy loss before impact. As there is further energy loss after impact, Ey can be expressed in terms of the efficiency of hammer blow 7. This gives By = £’ = naWh. Hiley formula By considering the physics of impact of two spheres, Hiley (1922) derived the following expression to calculate 1 _W+(Wp+ Wie > '—"Wwa(We+w) ° where W, is the weight of pile, W, is the weight of anvil and e is the coefficient of restitution. The Hiley formula considers the anvil as part of the pile-soil system. By substituting Ey = naWh and Q = C, + Cy + Cy into Equation (1,10), then ___yaWh S+H(C.4+ CG, +C,) Ry Major deficiencies in using the Hiley formula for long piles have been pointed out and explained by Triantafyllidis (2001) and hence it should not be used blindly. Reliability of dynamic formulae Poulos and Davis (1980) point out that several investigations have been made to determine the reliability of the various pile-driving formulae by comparing the load capacity computed from the appropriate formula with the measured capacity from a pile loading test. Some of the most comprehensive investigations have been reported by Sorensen and Hansen (1957), Agerschou (1962), Flaate (1964), Housel (1966) and Olsen and Flaate (1967) Sorensen and Hansen used data from 78 load tests on concrete, stecl and wooden piles, most of these having their points bearing on sand (a few were founded on hard moraine clay). The results of their comparisons are shown in Fig, 1.14, in which the ratio, 1, of the measured to the computed load-capacity is plotted against the percentage of load tests smaller than js. This plot is a probability plot, and a straight line on this plot represents a normal or Gaussian distribution of results. Figure 1.14 shows that all the formulae considered, with the exception of the Eytelwein formula, follow approximately a Gaussian distribution. There is very little difference in the accuracy of the Danish, Hiley and Janbu formulae, and the theoretical curves derived from the wave equation, but the Eytelwein formula is very inaccurate.PART | DEEP FOUNDATIONS - PILES AND BARRETTES Janbu formula Hiley formula Eytetwein formula — Danish formula = —— Wave equation 4 loss than log u Fig. 1.14 Statistical distribution of load test results (after Sorensen and Hansen, 1957; Poulos and Davis, 1980) Agerschou’s investigation concentrated on the Engineering News formula but also broadly confirmed the conclusions of Sorensen and Hansen regarding the Danish, Hiley and Janbu formulae. Agerschou showed that despite its popularity, the Engineering News formula is unreliable. It has the highest standard deviation, and 96% of the allowable loads determined by this formula will have actual safety factors ranging between 1.1 and 30. Flaate (1964) investigated the accuracy of the Janbu, Hiley, and Engineering News formulae for 116 tests on timber, concrete and steel piles in sand. The conclusions reached by Agerschou regarding the unreliability of the Engineering News formula are reinforced by Flaate’s results. There is relatively little difference between the Janbu and Hiley formulae, although the former is perhaps the more reliable overall and gives good results for timber and concrete piles. Hiley’s formula is also reasonable for timber piles. Modified pile driving formula for long piles For relatively long piles the 7 factor in the Hiley formula suggests very low energy transfer coefficients from the hammer onto the pile, because the whole mass of the pile is taken into account (Triantafyllidis, 2001). If the pile is relatively long it is worth considering only that portion of the pile that is affected during the duration of the impact. Therefore a more realistic approach is to consider the impact between two rods instead of the impact between two spheres, Alternatively, due to the availability of field equipment for measuring the forces and accelerations in a pile during driving by using a PDA, a [=]SHORT COURSE IN SOIL-STRUCTURE ENGINEERING io different but simple approach can be adopted. Broms and Lim (1988) and Paikowsky and Chernauskas (1992) proposed an energy equation to determine the pile-driving resistance as follows: Ey R= od (1.13) “S45 (Cp + Cy) a where E, can be directly measured by a PDA In this modified formula the anvil-pile cushion is not considered as part of the pile-soil system and hence C is not included in the equation Accuracy of the modified pile-driving formula Figure 1.15 compares the ultimate capacity of piles predicted by the modified pile-driving formula of Equation (1.13) with that predicted by CAPWAP (Case Analysis with Wave Analysis Program) or pile loading tests, The data smaller than 3000KN are taken from Paikowsky and Chernauskas (1992) for which the failure loads were obtained from pile loading tests. The other data points are obtained from Li et al, (2003). A total of seven test piles of Grade 55C 305 x 305 x 224kg/m H-piles constructed in completely decomposed granite (CDG) were loaded to failure or 3.3 times the design pile capacity. The length of the piles ranged from 35 to 46m. Each test pile was initially driven to final set by hydraulic hammer followed by drop hammer. As shown in Fig. 1.15 it can be seen that the modified formula predicts the test loads for long piles reasonably well. However, further verifications of this modified formula are still needed. 19 000 10 000 5000 + Paikousky and Chomauskas (1962) * Drop hammer (eased to faire) * Drop hammer {rat leaded to flue) & Hycraule hammer adod to file) & Hydraulic hammer (not loaded ofa) Pile capacity measured by loading test: KN ° i — ° 5000 10 000 18 000 Pile capacity based on the modified pile driving formula: KN Fig. 1.15 Pile capacity based on the modified pile-driving formula correlated with measurements from loading testsPART | DEEP FOUNDATIONS - PILES AND BARRETTES The Wave Equation Stress propagation Stress propagation in a pile during driving can be described by the following equation of motion: &u Sp Fu Bonu Anu 2 1ie (1.14) where u(x,t) — longitudinal displacement of an infinitesimal segment; Ay, Sp = pile area and circumference, respectively; Ep, y = modulus of elasticity and unit density of the pile material, respectively. The friction stresses f, are activated by the pile movement and under free wave motion (/, = 0) Equation (1.14) becomes the familiar one-dimensional wave equation. Equation (1,14) may be solved for the appropriate initial and boundary conditions to determine the relationship among displacement, time and position in the pile, from which the stress variation in the pile may be determined. Because of the complications involved in practical piling problems, analytical solutions to Equation (1.14) generally are not feasi- ble, and resort must be made to numerical means of solution. A convenient numerical method has been described by Smith (1960). Smith (1960) proposed an idealization of the hammer-pile-soil system which would be capable of representing the passage of the stress wave down the pile (Fig. 1.16). The pile is modelled as a series of rigid masses connected by springs which can act in both compression and tension. The hammer ram and the anvil or pile cap are also modelled as rigid masses, but interfaces between ram and anvil and between anvil and top of pile are idealized by springs capable of sustaining compression but not tension. Side resistance and end-bearing actions of the soil into which the pile is being driven are represented by bi-linear springs acting on the embedded rigid masses. Side resistance springs can exert forces either upward or downward while the end-bearing spring can act only in compression. Soil springs act linearly up to a limiting displacement termed the ‘quake’. In addition, increased resistance of the soil due to viscous damping is represented by a series of dashpots, Smith expresses the instantaneous soil resistance force, R, acting on an adjacent rigid mass as R=R(I+JV) (1.15) where R, = static soil resistance, J = a damping constant, V = instanta- neous velocity of the adjacent mass. ‘Ascan be seen in Fig. 1,16, the basic Smith idealization is representative of a pile driven by a drop hammer or a single acting hammer. Diesel 59SHORT COURSE IN SOIL-STRUCTURE ENGINEERING 60 i — _ [-— batvicapotock x, Dspacement (+ armen Wh Form of sol spring =% gel Ky 1 It « eee EK «Fle ia Vit Sido springs to we lee Dass o represent represent skin Ke fete on ple = amping fects soll oe Ht «ete ge le a tl it im er Ww, _ tl} it ZK, Kae Mele —~ See —~ Ene-bearing —— Fig. 1.16 Smith (1960) idealization for Wave Equation analysis of pile during driving (after Tomlinson, 2001) hammers have to be considered in a rather different manner since the actual energy generated by such a hammer will vary with the resistance of the pile that is being driven. For low resistance there will be low ener- gies per blow at a high number of blows per minute, while for high resis- tances the energy per blow will increase and the number of blows per minute decreases. Manufacturers can provide charts of energy versus rate of striking for diesel hammers, and when drivability predictions are being made the range of energies at which a particular hammer may operate should be considered. An example of this is shown in Fig. 1.17 in which back-analyses of static load tests have been used to generate blow count versus driving resistance curves for closed-ended tubularPART | DEEP FOUNDATIONS - PILES AND BARRETTES Pile: 914 mm 0.4. x 19 mm WT (lengths >16 m) Hammer: Delmag D82; Energy range 11 160-22 820 kg m © Test ple ‘A'~ Predicted resistance for energy and set at complotion of driving © Test ple ‘A'~ Failure load (from test load) approx. 600 t ‘4. Test ple Predicted resistance, based on test pile ‘A’ results, for energy and set al completion of driving (results of load test to 600 t Indicated resistance well above 600 1) 1000 3 Driving resistance: t 20 40 60 80 100 120 140 Blows per 300 mm Fig, 1.17 Pile drivability analysis for diesel hammer (after Tomlinson, 2001) steel piles being driven into calcareous sands and corals. In these ground conditions time-dependent effects are negligible, and so the driving resistance is considered to be the same as the static resistance. Although the original Smith idealization is incapable of modelling the interaction between driving resistance and energy generated for a diesel hammer, advances have been made. Goble and Rausche (1976) published details of a computer program which will model diesel hammer behaviour realistically; this program proceeds by iterations until compatibility is obtained between the pile-soil system and the energy-blows per minute performance of the hammer. Data required to undertake a pile drivability analysis include values for the stiffness and coefficient of restitution of the dolly/capblock and the quakes Q and damping constants J for the soil. Typical values of Q and J are given in Table 1.16, These may be used in pre-driving studies, but it must be remembered that they can be subject to significant variations and whenever possible they should be back-figured to fit load test results Since the late 1970s the exact form of damping of soil for Wave Equation analyses has been the subject of much investigation and Heerema (1979) and Litkouhi and Poskitt (1980) have published findings indicating that for end-bearing in clays and sands and for side resistance in clays, damping is dependent on the fifth root of velocity, that is, Equation 61SHORT COURSE IN SOIL-STRUCTURE ENGINEERING Table 1.16 Soil property values typically used in Wave Equation analyses (aiter Tomlinson, 2001) Soil type Damping constant Quake side ———— and end Skin friction End-bearing Q (mm) Jz (s/m) Jp (s/m) Clay 0.65 0.01-1.0 25 Sand 0.15 0.33-0.65 silt 0.33-0.5, 0.33-1.5 (1.15) is modified to R=R(1+JV°?) (1.16) Capacity evaluation in the field using Wave Equation The procedure of monitoring pile driving by dynamic measurements is well established. Early large-scale studies (e.g. Michigan State Highway Commission, 1965; Texas Highway Department, 1973; and Ohio Depart- ment of Transportation, 1975; and Goble ef al., 1970) led to the develop- ment of commercial systems which enable complete and relatively easy acquisition of dynamic measurements and analysis during driving. The PDA (Pile Driving Analyser, see Pile Dynamics Inc., 1990) which is the most commonly used device in the USA, utilizes a simplified pile-capacity evaluation method, known as the Case method (see Goble ef al., 1970) The formulation of the method is based on a simplification of the Wave Equation and employs the force and velocity measurements taken at the pile top in order to obtain the total resistance. The static resistance is then evaluated based on a dimensionless damping coefficient J. (Case damping) which was correlated to the soil type at the pile tip (Goble et al, 1967, 1975; Zhang et al. 2001a). The method encountered two fundamental difficulties: © the total resistance is time dependent and different variations of the method produce different results * the dimensionless damping coefficient was found to have question able correlation to soil type (e.g. Paikowsky, 1982; Thompson and Goble, 1988) and therefore had to be calibrated for the specific pile, soil and site conditions. Reliability of dynamic methods MeVay ef al. (2000) evaluated the parameters for load and resistance factor design (LRFD) of driven piles using dynamic methods. They consid- ered a database of 218 pile cases in Florida. Eight dynamic methods were [2 |PART | DEEP FOUNDATIONS - PILES AND BARRETTES studied: ENR, modified ENR, FDOT and Gates driving formulae, and the CAPWAP, Case method for PDA, Paikowsky’s energy method and Sakai's energy method. It was demonstrated that the modern methods based on wave mechanics, such as CAPWAP, PDA and Paikowsky's energy method, are toughly twice as cost-effective in reaching the target reliability indices of 2.0 to 2.5 (failure probability p; = 0.62 to 2.5%) as the ENR and modified ENR driving formulae. They concluded that the past designs with driving formulae reveal a large redundancy of pile groups against failure and correspondingly underestimate pile group effects. They therefore suggested that use of a relatively low reliability target index, 3; = 2.0 (p; = 2.5%), is sufficient for single-pile design. 63CHAPTER SEVEN Design of rock-socketed piles Research and practice in North America Ng et al. (2001a) report that Rosenberg and Joumeaux (1976) studied the results of a limited number of field load tests on small-diameter piles (200— 610mm). They proposed a tentative relationship between the unconfined compressive strength (UCS) of rock o, (MPaj and the ultimate side resistance q, (MPa) as follows: qs = 0.375(0,)°5!> (1.17a} Horvath and Kenney (1979) reviewed the results of load tests on rock- socketed piles and anchors from, mainly, Australia, Canada, the United Kingdom and the United States, including 49 tests on piles with diameters between 410 and 1220mm. The majority of the piles were socketed into sedimentary rock types, mostly shale or mudstone. They reported socket friction generally became fully mobilized at a displacement of around 6mm (0.5-1.5% of the pile diameter). A correlation was suggested between q, (MPa) and the compressive strength of the weaker socket, material (concrete or rock) c;- (MPa) for the piles given by Is = bVGee (1.17b) where b = 0.2-0.25 for piles >410mm in diameter. Horvath et al, (1983) later revised b to be equal to 0.2-0.3 for ‘large’-diameter piles, presumably >410mm, and proposed a method for estimating the effects of socket wall roughness on the capacity of side resistance using a ‘roughness factor’, which is dependent on asperity height, socket radius and socket length. However, in practice it is difficult to determine the roughness factor in the field accurately, particularly for piles constructed under water or ben- tonite Rowe and Anmitage (1987) recommended a correlation of q, = 0.45/70 for regular ‘clean’ sockets and q, = 0.6 /@;c for clean ‘rough’ sockets for an initial estimate of side resistance from the results of a large number of field load tests. ‘Rough’ sockets were defined as having grooves or undulations of depth and width >10mm at a spacing of 50-200mm. A more detailed design procedure taking into account socket dimensions was providedPART | DEEP FOUNDATIONS - PILES AND BARRETTES to refine the prediction. Rowe and Pells (1980) studied theoretical con- siderations of rock-socket behaviour. That paper is significant in that it predicted variations in side resistance capacity as a result of varying stress distributions, related to varying ratios of socket length to diameter and pile modulus to rock modulus. Reese and O'Neill (1988) recommended that the first of the following equations be used for side resistance for rock (or concrete) with a UCS up to 1.9 MPa and the second equation be used for higher strengths (MPa) qe = 0.1505 (1.18a) qs = 0.2 (1.18b) The first equation was taken from Carter and Kulhawy (1987) and the second equation from Horvath and Kenney (1979). The recommendations are adopted by the US Department of Transportation. Recently, Zhang and Einstein (1998) suggested relationships based on a review of the results of Rosenberg and Journeaux (1976), Meigh and Wolski (1979), Williams and Pells (1981), Horvath (1982) and Kulhawy and Phoon (1993). These are as follows for ‘smooth’ and ‘rough’ sockets (MPa), respectively: 4, = 04 VF (1.19) ds = 0.85 Vee (1.20) Research and practice in Australia In Australia, most experience in the design of rock-socket piles has been gained in mudstone, shale and sandstone. One example is the Melbourne mudstone, which Chiu and Johnston (1980) described as having charac- teristics very similar to over-consolidated clays. Williams et ai. (1980b) described the results of a number of field load tests in Melbourne mudstone, The socket friction-displacement relationships exhibited peak, followed by residual, friction behaviour, although the residual values were only around 5% less than the peak values for ‘rough’ sockets. ‘Rough’ sockets were defined by minimum statistical requirements for asperity height and angle and were achieved with excavation by con- ventional bucket augers. Williams et al. (1980a)} also described the results of tests in Melbourne mudstone of which at least one reached peak friction at a displacement of 6mm. Williams and Pells (1981) suggested a design method based on the results of Williams et al. (1980b), Pells et al. (1980) and Horvath (1978) but with emphasis placed on test results in Australia. For cases in which the socket wall is sufficiently rough to prevent brittle side shear behaviour, acurve for the ratio of q, to o,, termed the 'side resistance reduction factor’ a is used, The side resistance factor decreases with increasing oj. A 65SHORT COURSE IN SOIL-STRUCTURE ENGINEERING [ «| further factor J is applied to take into account a variation of side resistance with the rock mass factor (ratio of jointed rock modulus to intact rock modulus) giving Is = OB Ore (1.21) As mentioned above, Rowe and Pells (1980) predicted the relationship between socket side resistance capacity and the socket dimensions and modulus ratio. Design practice in Hong Kong In Hong Kong design practice, for large-diameter bored piles bearing on granitic and volcanic rocks, allowable values of toe resistance may be used as specified in PNAP 141 (BOO, 1990). The presumptive end-bearing values range from 3 to 7.5 MPa, depending on the rock ‘category’. The category is defined in terms of the rock decomposition grade, strength (UCS) and total core recovery. Generally, the presumed values may be used without a settlement check for structures which are not unduly sensitive to settlement. The details of the presumed end-bearing values from PNAP 141 are summarized in Table 1.17. Presumptive values of tock-socket side resistance are not defined. However, PNAP 141 allows Table 1.17 Presumed safe vertical bearing stress for foundations on horizontal ground in Hong Kong, simplitied from PNAP 141 (BOO, 1990; after Ng et al., 20014) Category Granitic and volcanic rock Presumed bearing stress (kPa) 1a) Fresh to slightly decomposed strong 1uck of 7500 material weathering grade II or better, with total core recovery of >95% of grade and minimum uniaxial compressive strength of rock material ¢, not less than 50 MPa (equivalent point load index. strength PLIsq" not less than 2 MPa) 1(b) Slightly to moderately decomposed moderately 5000 strong rock of material weathering graces Il or IIL or better, with total core recovery of >85% of grade and minimum uniaxial compressive strength of rock material ¢, not less than 25 MPa (PLIsg" not less than 1 MPa) uo) Moderately decomposed moderately stong to 3000 moderately weak rock of material weathering Grades Il] or lV or better, with total core recovery of >50% of grade * Point load index strength ( ccotes (ISRM, 1978), PLIay values) of rock quoted is equivalent value for 50mm diameterPART | DEEP FOUNDATIONS - PILES AND BARRETTES the use of a ‘rational design’ approach and rock-socket side resistance can generally be incorporated in a design, provided that the parameters and assumptions are verified by field testing, For caissons in granite, the Geotechnical Engineering Office (GEO, 1991) reported various findings from studies including those by Ladanyi (1977), Pells and Turner (1979) and Horvath et al. (1980) and acknowl- edges that socket side resistance can account for a significant portion of the load support capacity. For the preliminary design of side resistance in widely jointed rock, the GEO (1996) suggested the use of the empirical correlation proposed by Horvath et al. (1983) to estimate side resistance capacity. Horvath's correlation was developed using the results of load tests in mostly sedimentary rocks, however, whereas Hong Kong condi- tions are dominated by granitic, volcanic and metasedimentary rocks. ‘The GEO (1996) also suggested that the allowable working load should be estimated using a minimum mobilization factor of 1.5 on the ultimate side resistance. The mobilization factor is intended to take into account debris left at the base of the rock socket, which would cause the base resistance to be unreliable at working load. It is implicit in the GEO's recommendation that the contribution of toe resistance is only taken into account at the ultimate limit state. Therefore the contribution of toe resistance may potentially become irrelevant altogether if the working load requirement governs the design.CHAPTER EIGHT Case studies: rock-socketed piles [#] Overview Ng et al. (2001a) report that a number of empirical relationships have been published for estimating the side resistance of rock-socketed piles. All are based on studies of field load test results and laboratory tests and relate socket friction capacity to the UCS of rock or concrete, generally which- ever is the weakest. Most of the load tests considered were carried out in sedimentary rocks having lower strengths than typical rock types encountered in Hong Kong. Table 1.18 summarises details of the 13 tests carried out in Hong Kong. The tests were generally carried out in the preliminary stages of the respective projects. They were intended to investigate the feasibility of using rock-socket side resistance in design, rather than using the pre- sumed end-bearing stresses detailed in PNAP 141 (BOO, 1990), and to establish design rock-socket side resistance values. Also, some of the rock types tested were not volcanic or granitic rocks and therefore could not be designed using the presumed values in PNAP 141. The chosen methodology falls under the PNAP 141 clause permitting the use of a ‘rational’ approach to foundation design. Based on the loading test results, the use of rock-socket side resistance in design for each of the specific projects was approved by the relevant checking bodies and detailed design and construction proceeded accordingly. Of the buildings being supported by the piles, the settlements at this stage are, to the writers’ knowledge (Ng et al., 2001a), within acceptable structural limits. With the exception of the hand-dug caisson reported by Lam et al. (1991), excavation was by RCD (Rotary Core Drilling) under water or bentonite. The geological descriptions including the rock decomposition grades and the rock quality designations (RQDs| were assigned by geol- ogists. Rock types have also been classified generally as being granitic, volcanic or metasedimentary. The UCS o, values were either measured directly by laboratory tests or estimated from a correlation with the point load index test (PLIT) results. In the PLIT, rock specimens in the form of core, cut blocks, or irregular lumps are broken by application of a concen- trated load through a pair of spherically truncated conical platens with aPART | DEEP FOUNDATIONS - PILES AND BARRETTES suaytim £q poyordioiuy Pep ISO, [waoys 918 $9 9Por VENT $69] SA aVA_DUOD EM SoSH A|UO} SN a¥eIOUOD eeUNKOLde st sisaURTed UI BNYeA, ‘siaqua Aq payasdsayur yep pus yno pours S04, (¢¢‘d 995) xeput 1 jax0s-yoos UMUKPUL Una = "=D :ayON, wouraoetdsic =1q wowaorydsrp = 1b -yybua| 193005 s0n0 paBerane “oouEistsar: 530 1 aoueysisa1 apys yox208-3>4 v 969 S69 ga g(cdozr (onrues6) ayueIB TAT ape oss Sp-pueH gO ose (eer PP weT eT v= 098 oz (orueojos) eP2—1q VAL OPED — PEM sOPUNY—TT zoe ABMS SL ZI 2 orb ae (onmums6) oyues py] apeI sO} JOPIAGDUY OE oe ADDIS STL TT # 0001 000t alos) oz (onrues6) oyuess yyy] apex enuoquag apIN DY ST ze 2xpms sm. OF @ Off 096 68 297 (onrues6) oxues6 1 apr OTP apNGIY SZ cer Apmis SHE 6 D - $99% 06 — (os) szt (onyues6) oyues6 y ape25 ayuoTHAG EPUN COX GO cw SABRES ST ¥ 00s 019 69 or {onwes6) ayues5 TAT] apeID aMuoIEg JPR GY 9'E z09 SApms sm. @ OOLT o0Zt O01 los) s'z8 (onyues6) oyuei6 y] ape ayuoquaq spun cx ZT one “ApMS SUL 9 Vv ow om of (oq ue16) eyoypoues6 VAL ape HR |IPUN GDA 1% 1s2 Zxpms surg (Arewwounpas 2 - oo Ww 88 jou) oUO|SpUPSPIaUI I] ePID =A sepuNG]IY FT oozt = OF eApnis smL © - os 99 OF —forweDJ0) MI YsE osteOD TT PPE —_—sHJOMPUNGIY OZ —OZOT._—«-'EE. pms sm. 2 ort 9 lo (oruw9j0a) ra yse oug AR spe FEA LPN GIY 09 O90 SZE Apmis sy, Z D = 006% 76 — {08 GOT (OTUWDIOA) FM Ys oUY T ePEID —FBEMGEPUNGIY OZ — OOZES Sxpmis su 1 (aw) (a) (um) (ua) (ay) (ea) (%) som wSuat seyaurENp bua, oN Ia “bb qdy abeiony oda yoou WoReAex9 JO POMIA, YYHS JIS old —«BOUOIOF. of (01002 “Ie 12 BN 40110) 84901 pasoduiopap ojur pajayo0s sajrd paroq 1a}auTDIp-a610] Jo s]89) poo] Buoy Buoy Jo asvqning I'l P1GDL,SHORT COURSE IN SOIL-STRUCTURE ENGINEERING 70 tip radius of 5mm. Little or no specimen preparation is needed (Interna- tional Society for Rock Mechanics (ISRM), 1985). A factor of 24 was applied to the PLIT results, as suggested for weathered granite by the GEO (1991). The suggestions of the ISRM Commission on Testing Meth- ods were adopted for the PLIT tests (ISRM, 1985) and also the UCS tests (ISRM, 1978). The UCS values reported in Table 1.18 are the average values over the depth of the rock socket. Generally three or four tests in total were performed for each socket. It is suggested that, for purposes of comparison, the same or similar testing metheds should be adopted in the future by other researchers. An anomaly may be observed in the results in Table 1.18. Pile 12 was tested in rock, which was too decom- posed to allow a UCS test to be conducted, and the test may be considered to be not typical of the remaining tests in Table 1.18; the differences should be kept in mind in the analyses to follow. In the load tests, the axial load distribution in the rock socket was deter- mined using strain gauges installed within the pile shaft and by estimat- ing the pile shaft modulus. The average strain across the socket was taken to be equal to the average reading of the gauges at each level, For the piles with which the writers were directly involved, the average strain gauge readings were checked by calculating the total socket short- ening from the readings and comparing these with the shortening measured by rod extensometers. The difference in the results varied between 5 and 13%. The modulus was generally deduced using labora- tory test results on concrete core samples taken from the pile shaft and by considering the variation of modulus with strain over the duration of the loading test using gauges located near the top of each pile. With the axial load distribution known, the unit socket side resistance was calcu- lated accordingly, For most of the piles, the calculated toc resistance accounted for only up to 17% of the socket side resistance. The side resis- tance value Gnax Shown is the maximum recorded or extrapolated for each test and is the ‘mean’ value, averaged over the whole socket length. Table 1.19 summarizes key details of the 66 other tests from the litera- ture. Some contributed to the various empirical correlations for side resis- tance in the literature detailed earlier. The details and measurements shown in Table 1.19 are consistent with Table 1.18, although the level of detail is generally lower. The RQD is not shown and whether or not the excavation was ‘dry’ or ‘wet’ (constructed under bentonite or water) is unknown in many cases. For a number of tests in granitic rock in Singa- pore, the pile diameters are also unknown. Possibly some of these tests were on small-diameter rather than large-diameter piles. Note that, in a few cases, the sockets were artificially roughened, The UCS values shown were obtained using a range of methods, including empirical cor- relations with PLIT results and standard penetration test results.PART | DEEP FOUNDATIONS - PILES AND BARRETTES a 06% ——_ (ArequouNpes) auospnyy, wabne wna vet @ % ——_ (Areqwaupes) auorspnI sane 1oxong ozet urea 4 09:0 (ueyuaurpas) auoyspnyy, a oz SUIPRTEM 8 08:0 (Asejuaunpes) auoispnyy 2 099 (ogst) sumer, a os: (Are1uautpas) auois a oozt (6261) preuog pur ouoisuyor a ore (Arenuar a oz (6261) pleuod pue a v 0€% —— (Arequoumpas) ouorspary, raBine 19%: 099 (og61) 10 19 a ovo (daeyuounpas) aseqeia st9 (9261) aaem @ oct re,UOLNTP9S) HO a oat (reer) ste, 8 00 8) AIP 49g sOpUN parog osor (oc61) Suman a 06% —— (Arequoupos) ouorspMTY 0 (eu61) 1» 12 5 2 ore Lot (6261) Kouuay, pue ypearo} a ost or9 (0967) wosey @ ovtt oz (eg68) te pu Broqieso &) 00° Aroiwoumpas) a1eus ore (e610) weHoA ae a ozs (dzeiuounpes) 9724s sea (6261) Keuuey, pur yea TE a ovor (dapjuaumpas) ajeus oie (ea61) 1 12 wean 4 oss one (ea6t) 10 79 weAoH ¥ 09s ore (ea6t) 1m 70 wPArOHT v orn one (e961) (79 WeAIOH, v ors ove a oot spel 006 oez a 068 spel ajouussoct az a owe spel ajoyumog P57 4 oct Suyse> fq spor pur yaxong HamUCL + 2 00°F Asejwounpas) aseqeic, Bulseo ‘1q yo0r pur yaxanq Gun, (2661) eaan: £ 2 009 (dsejuompas) umsdis Buse ‘nq oor pur Yo¥onq Bum (i661) eqqmume) zz a 00'S (orUeDTOR} PONG arseqeIG, BuIseD nq yO PUR YexON HUTTE, (i661) Paqneue)z a 06: (Areyuaumpas) eypy Buise> H1q y90! pur Yxonq Burt oozt (660) Pane) 2 oe ojuaUTpas) au0} ‘a1oq Arp ‘jastyp pue reBne WOT ort a6 2 ose ‘a0q Arp ‘jastyp pue sobre WET ort er 3 006 ote ar D 006 ost a 2 oe x09 ose on > ooo (dreywammpas) auosIts 9109 Aup ‘Tosiy> pur 4 ore sk st a oer ontueB) qHUeID — o10q 19M ‘TosTYD PUD x 001 (9661) Buna] pt wav) (unua) (ws) eax) son ubuer seeuerp yi6u9y lab aberony ada 4004 UUoRPABOKO Jo POOH, y¥0S 10S Olle oua1a FR ‘ON atta (01002 “TP 1° BN 191]0) 84901 pasodurosap o1uT payayoos said paiog 1a}auNDIp-a610] Jo S189} pool Buoy Buop]-wou Jo asvqniog 6I'T B40 7SHORT COURSE IN SOIL-STRUCTURE ENGINEERING pauayfnox ArpeETRTe SPM 19x00, woneanand ut amnBy jo-po(e>s = 4 (ge 89s) xepu weoMto.eIdsIC = 1d ION 2 00°05 (onnes5) aque) aoe a a 6L 2 ‘00's (Garues6) onues aoe a a aL D #0007 a8 a a a D 1300'0% a8 a a 9 D 00°F 28 a a se D A00ET ra a a a 3 ‘00'S ra a a ee 2 ‘800'7 2 2 2 2 ze a OV a8 é é (e661) Buna] 1g 2 100" 8 é 2 (9661) Sunt og D 08% a a a (9661) Suna] 69, 2 OEE (onrues6) oer) ae a a (9660) Sune1 gg 2 00% (ones5) oynmessy 2 8 a 2 (9661) 5uney ag a 00% (ones6) aqme) a8 2 a 61) BunoT 99 2 00% fonrues6) oyuessy a8 a a jet) Bunoy 99 D OE a2 2 2 jet) BunoT Fa 2 OE a 8 a é (9661) Buna eg a OEE a 8 a ‘ (9661) Buna za D Ort aoa 2 2 (9661) Buna] Tg D AOVT a8 2 a (9661) Bunoy gg D one 2 Ot 006 OIE (oz61) buming 6s @ OFT (Areyuounpas) auorspnyy Su rsswispoH 0'T 006 ——«00'F (o26t) wosem =a a OT {Areyuourpas} auo Su rsseisy>0H 00'T 00600" (ozs) wosem es 2 Caag {dzejtounpas) 9724s seine qyang SEE S16 S'S (o96t} umqoy, 9g a 00'°9—_(Arewtourpas) ouoispues, reBny 960 OTL ¢ (0861) 18 stad gs 2 oorz (dzequaumpas) afeus seBne Bid Of 006 00% (og61) owour bs @ Lz ase ‘pOTUG 068 06 2 &s @ oso juaquaq Japun ase ‘PATUG O's 099k ws 4 ove ‘emuoiwog fopun yse> ‘pat OE 069 i 1s 2 aro obxy 009 O18 os @ oss s 2 00% ose er a 08% © (Aaeyuoumrpas) euorspnyy sobneyyond 00% —OseLt ay @ 0% ——_(Aterzounpes) auerspny veBne yong OZ —OFZL “& (an) (wa) (eu) a) son Guar swe qbue 1a ™b abereay oda 32031 woneseoxa Jo powna yIS WHS oT aouar9joy ON ofa (panunuos) (op002 “P19 BN 491]0) syoo1 pasodusosap our payayoos sald peiog 1a}aUIDIp-a610] Jo s}89} poo] Buoy 6UofJ-UOU Jo Bq} GI'] A1GDL, nPART | DEEP FOUNDATIONS - PILES AND BARRETTES Displacement Index (Dl) The database contains side resistance values recorded at varying displa- cements and degrees of mobilization. Therefore, a DI is assigned. This is an approximate measure of the pile ‘local’ displacement and degree of mobilization of side resistance for a given maximum side resistance value. The ‘local’ displacement is the pile displacement deduced at the vertical centre of the rock socket using the measured pile-head movement and shaft shortening at the rock socket midpoint. The DI values are assigned by observation of the field side resistance-displacement relation- ships. An index A is assigned to test results, which are observed to be close to the fully mobilized value, and is defined by the displacement reaching a minimum of 1% of the socket diameter and the gradient of the resistance- displacement relationship becoming <30kPa/mm at the end of the curve (at the completion of the test). An index B is assigned to test results not achieving an index of A but recorded at a displacement of at least 0.4% of the pile diameter and for which the gradient of the resistance-displace- ment relationship is <200 kPa/mm at the end of the curve. An index C is assigned to the remaining results. The various proposed displacement indices are illustrated graphically in Fig. 1.18. 00 os 10 18 20 25 8/0 (%4) Fig. 1.18 Proposed criteria for Displacement Index (DI) (after Ng et al., 20014) Capacity of side resistance and relationship with UCS and decomposition grade Relationship with unconfined compressive strength, UCS All of the various empirical relationships reviewed in the literature relate rock-socket side resistance to the UCS of rock or concrete, generally BSHORT COURSE IN SOIL-STRUCTURE ENGINEERING Gane KP Gaus KPa 6000 @ © Granite 4 Volcanic 5000 . ¢ Metasodimortary 1 Sedimentary Zhang and Einstein ‘ough socket (1998) ~~ Rowe and Armitage 4000 rough socket (1987) 7 Rowe and Armitage smooth socket (1987) 3000 - --7aoare are Eistoin = smoath socket (1958) _ 2000 awe Fiorath ot a (1969), * b=0 a5 o Figrai rat (1985) 1000 . ° 5000 4000 3000 2000 1000 ) Rowe ane Armitage rough socket (1987) Zhang and Eingein 7” rough socke/{1898) “Rowe and Armitage” Serene ae © Granite 4 Volearic 1 Sedimentary 10 20 30 40 50 60 8: MPa Fig. 1.19 Maximum achieved side resistance versus UCS: (a) all results, DI= A, B or C; (b) DI=A or B only (after Ng et al., 2001a) 4 whichever is weaker. In Fig, 1.19 the mean maximum socket side resistance wax has been plotted against the mean UCS ¢,, for all test results, Also shown are the suggested correlations of Horvath et al. (1983), Rowe and Armitage (1987) and Zhang and Einstein (1998). Overall, almost all the results lie below the correlation of Zhang and Einstein (1998) for ‘rough’PART | DEEP FOUNDATIONS - PILES AND BARRETTES Table 1.20 Classification of rock material decompos sockets but with a high degree of scatter. None of the correlations from the literature appear to be particularly suitable for predicting side resistance for all results. However, if the results in only sedimentary rock are examined, these are mostly higher in magnitude than results produced by the correla- tion of Horvath et al. (1983) for b = 0.2, Generally the results in granitic rock are lower than in sedimentary rock and tend to be lower than the empirical correlations. However, the UCS is also generally higher for these tests. The results in volcanic rock are few and scattered. The single result in metasedimentary rock indicates extremely high side resistance. Relationship with decomposition grade The Hong Kong tests were conducted in rocks varying from decomposi- tion grade IV/III to Il. From the descriptions of the rock decomposition grades in Table 1.20, it would be expected that rock strength and hence ‘ion grades of Hong Kong (simplified from GCO, 1988) Descriptive term Grade General characteristics for granite and volcanic rocks and other symbol rocks of equivalent strength in fresh state Residual soil Completely decomposed Highly decomposed Moderately decomposed Slightly decomposed Fresh VI Original rock texture completely destroyed Can be crumbled by hand and finger pressure into constituent grains V Original rock texture preserved Can be crumbled by hand and finger pressure into constituent grains Easily indented by point of geological pick Slakes when immersed in water Completely discoloured compared with fresh rock IV__ Can be broken by hand into smaller pieces Makes a dull sound when struck by geological hammer Not easily indented by point of geological pick Does not slake when immersed in water Completely discoloured compared with fresh rock I Cannot usually be broken by hand; easily broken by geological hammer Makes a dull or slight ringing sound when struck by geological hammer Completely stained throughout Tl Not broken easily by geological hammer Makes a ringing sound when struck by geological hammer Fresh rock colours generally retained but stained near joint surfaces I Not broken easily by geological hammer Makes a ringing sound when struck by geological hammer No visible signs of decomposition (i.e, no discoloration) 7sSHORT COURSE IN SOIL-STRUCTURE ENGINEERING 10 000 At 1% socket diameter . o an -— Ur 95% g Average Hong Kong result So Lo nt 1 _ ~3 Lower 95% ‘onfdonee limit 100 ' " Ww v Decomposition grade of rock Fig. 1.20 Maximum achieved side resistance versus decomposition Grade for Hong Kong granitic and volcanic rocks (after Ng et al., 20014) side resistance would decrease with the decomposition grade increasing from Il to IV/ILI. Figure 1.20 shows the relationship between decomposi- tion grade and side resistance. The side resistance results are as recorded at a displacement of 1% of the pile diameter in granitic and volcanic rocks. Each increment of decomposition grade is spaced evenly along the horizontal axis, and materials having two decomposition grades (identified by geologists) are shown halfway between the two increments, It is recognized that the evaluation of decomposition grade is subjective and is by no means accurate enough to provide rock strength, particularly between various rock types. Not surprisingly for some of the decomposl- tion grades, there is a range of results. This may be due to a variation in the UCS, rock type and construction method. However, a trend line can be drawn as shown in the figure. This could be used for preliminary prediction of side resistance capacity without having information such as UCS available. The 95% confidence limits of the trend line are also included. Side resistance capacity in rock of higher decomposition grades is particularly useful to designers if there is uncertainty over rock quality over the whole socket length and diameter.CHAPTER NINE Pile instrumentation: a case study of a well-instrumented barrette in Kowloon Bay, Hong Kong Overview: key questions As reported by Ng et al. (2000b), a well-instrumented barrette was tested in order to answer the following key questions about the behaviour of barrettes: * How are lateral stresses and pore water pressures distributed and how do they change during a load test? What is the importance of degree of mobilization of side resistance in ‘floating’ barrettes (i.e. that rely on side resistance for their load capacity)? What are realistic design values of side resistance and settlement? What is the influence of construction factors such as duration, use of bentonite and grouting? Introduction Limited space and high demand have made land in Hong Kong extremely expensive. Tall buildings are built to optimize the floor area to land area ratio, Many of the tall buildings located along the Victoria Harbour on Hong Kong Island and the Kowloon peninsula are commonly founded on reclaimed land. Thus deep foundations are required to resist both vertical and horizontal loads due to the weight of the building and ‘wind, respectively. The prevailing deep foundation types for tall buildings on these reclaimed lands are large bored and excavated piles, which are very long, normally in excess of 50m. These piles can be circular (bored piles/drilled shafts) or rectangular (barrette) in shape and must extend through the fill, underlying soft marine clay, sandy clay and alluvial sand deposit down into the deep weathered granite soil (saprolite), which is typically less weathered with an increase in depth. The thickness of the weathered granite can be up to 80m in some places, and its depth can extend to more than 100m from the ground surface. Over the past 15 years, barrette foundations have become increasingly popular in parts of Asia such as Hong Kong and Malaysia for many civilSHORT COURSE IN SOIL-STRUCTURE ENGINEERING 78 engineering structures and tall buildings. The construction method for barrettes is very similar to that adopted for diaphragm walls, where a rectangular trench is excavated under bentonite by heavy grabs or hydrofraise* and filled with tremie concrete. In Hong Kong, single barrettes up to a size of 1.5m wide by 6.6m long (on plan) have been constructed (Pratt and Sims, 1990). Due to their rectangular shape, barrette foundations are particularly suited to resisting large vertical and significant horizontal loads in a chosen direction. Site location and ground conditions The test site is located on the Kowloon peninsula of Hong Kong, to the east, of the runway of the old Kai Tak international airport, at the Kowloon Bay area, Figure 1,21 shows the geology and some relevant borehole informa- tion obtained at Kowloon Bay. The site is on marine reclaimed land and the ground level is at approximately 4.48 m above Principal Datum (PD). The groundwater level is about 3m below ground surface, The ground con- ditions consist of about 6 m fill material overlying a succession of approxi- mately 9.5m marine clay deposits, 1.5m of sandy clay (probably alluvial) 4.8m alluvial sand of Quatemary age, and about 12m of weathered granitic saprolites that overlie granitic rocks of the Upper Jurassic to Lower Cretaceous age (Strange, 1990). Detailed descriptions and measured N values (number of hammer blows per unit penetration) by Standard Penetration Tests for each type of material are given in Fig. 1.21 Details of construction The top 20m of the barrette consisted of a reduced-section sheathing zone (Fig. 1.21) built with the intention of minimizing the interface side re: tance developed between the barrette and the upper soil layers. This sheathing layer consisted of four layers: a 3mm steel plate welded onto the reinforcement cage, a coating of bitumen, a flexible and weak ‘voltex' layer (geotextile infilled with sodium bentonite), and a thin sheet of ply- wood. However, the final result was that the plywood was unfortunately attached to the steel plate with a dense matrix of high-strength screws, pre- cluding the possibility of shear between the intermediate ‘soft’ layers. As a consequence, the theoretical gap of about 80 mm between the plywood and the surrounding soil was not back-filled with gravel as planned, so that a ‘weak’ friction zone would hopefully exist. However, steel rods inserted into this suspected bentonite-filled gap about two weeks after concreting 4 A ‘hydrofraise’ is a drilling machine powered by three down-hole motors, with reverse mud circulation, A heavy metal frame serving as a guide is fitted at its base with two cutter drums carrying tungsten carbide tipped cutters which rotate in opposite directions and break up the soil. The hydraulic cutting device is designed to give the cutter drums a high torque at low speed of rotation,PART | DEEP FOUNDATIONS - PILES AND BARRETTES (qo002 “1 1 BN 1140) Buoy Buopy ‘hog uoopmoy 10 sanyoa Las pun sBo aoyo1og [2"t ‘Bry (Aen pues nim pews un ww € x09 188)8 PAL YB WW 90 ¥) “@8eq HOS (OFLet 021 14 001 06 08 o7 09 05 OF OE Oz OL 0 (yOu) nb Se ero els fy osm esre=e— a as (Gunvuo cauaHivan) Si o 8 . LINED pauresSunjpou m e+ Bit eocoauiaws fie ofc. | ert soot i "Buons 0} yeom Aiwasig < mo (avo Aas) FS s9fe| 2840, = vanou6 ou pur pues =] semen uum AYTO AS ‘ns OF (ssouyoryt wi 10) eames |_}—" ayaiouco pur syuoueg (sLisod3a 3NIEYN) “squewbey; yoys pue pues 0 $208. Win AYTO AIS HOS a ARE RH weer ‘oh 06181 01 001" 05" 08" Oz 08" 08 OF CE Oz OF 0 sores pms sities "L Sagemtaa "9SHORT COURSE IN SOIL-STRUCTURE ENGINEERING 10-725 EF, Note all dimensions are in mm 2 piezometer cable and hydraulic hose ° ® ® 24” PVC tube fr sain gauge cable Strain gauges @ Rod extensometer at middle cl CDG layer (36 m bal) {©} Horizontal embedment strain gauges S Inplace ineinometer % od extensometer at level 6.5 m below ground level (ba!) Rod extensometer at level 155 m gl Rod extensometer at top of C2G layer (30 m ba) Fd extensometer at bottom of CDG layar (41 m bal) 2 6” Castin flushing pipe Tremie pipe Fig. 1.22 Plan view layout of instrumentation of test barrette (after Ng et al., 2000b) [=| were unable to probe beyond a metre or two all round the barrette. Either concrete overflow had partially filled the gap, or construction activities failed the soil infilling the gap, or surface materials mixed with solidified bentonite. Thus, in the end, the sheathing zone was not expected to func- tion to effectively reduce side resistance over the top half of the barrette. At the bottom of the barrette, a ‘soft’ base was formed to minimize the effects of end-bearing for mobilizing full side resistance at the soil-wall interface. This was done by placing a 2.8 x 0.8 » 0.3m high steel box to the bottom of the trench, before the lowering of the main reinforcement cages. The box was made of 3mm thick steel plate, and it was initially filled with fine round sand, Seven days after concreting, the sand-filled steel box was drilled through and flushed with pressurized water via two cast-in flushing pipes and one concrete core hole in the middle of the barrette (Fig. 1.22). Great care was taken to ensure that most of the sand was flushed out to form a ‘soft base’ (i.e. void) underneath the barrette. Instrumentation To study the load transfer mechanism and load-settlement characteristics of the barrette constructed at Kowloon Bay, a substantial amount of instrumen- tation was installed. In addition, four sets of standard dial gauges together with surveying were used to monitor the vertical settlement of the top of the barrette and reference beams during testing. Strain gauges werePART | DEEP FOUNDATIONS - PILES AND BARRETTES Inplaco incinometer Strain gauge instrument Rod extensometer and Legend spacrg spacing Tocation” dummy gauge spacing 240000 4 Groundwater FI 4 a | g o e 8 3 o 3 4 ce: . 5 4 oa sift: -e L . ee i ot 8 we Hedy eral ai ies | yah! ol rele ae Jae a jee ES 3 at ys g| z i 2 Bet § | i] e wel S| r we qt (i]s wt ja 3 wel Overbreak (approx. 6:2 mx 2.4 era "Soft base @ Surface-mounted strain gauge & Embecment strain gauge “* The soft base is made by a box of 3 mm thick 11 Horizontal embed train gauge mild stee!, 300 mm in height ‘Dummy gauge 1 Earth pressure col and piezometer ® Rod extensometer anchor 4 inpace intinometer Fig. 1.23 Elevation layout of instrumentation of test barrette (aiter Ng et al., 2000b) placed at 27 levels on the reinforcement cages (Figs 1.22 and 1.23). Ten aluminium rod extensometers were sleeved individually in PVC tubes and installed to five depths at two different locations inside the barrette to moni- tor displacements between each depth and reference steel plate at the top of the barrette (Figs 1.22 and 1.23). A total of 38 biaxial servo accelerometer sensors were installed at 19 levels (most of them at 2m intervals) in two alSHORT COURSE IN SOIL-~STRUCTURE ENGINEERING cast-in pipes inside the barrette. The locations of the in-place inclinometers and the levels of sensors are indicated in Figs 1.22 and 1.23, respectively. The bottoms of the inclinometers were socketed in rock. A total of four vibrating wire total earth pressure cells, together with four vibrating wire piezometers, were installed at the barrette-soil interface at four levels within the layers of sandy clay, alluvial sand and weathered granite. The depths of the earth pressure cells and piezometers are shown in Fig, 1.23. Load and displacement behaviour of barrette ‘The test programme originally comprised four loading and unloading cycles (Fig. 1.24). However, after the applied load reached 7455 kN at the second cycle, substantial settlement was recorded and the applied load could not be held constant within the prescribed maximum settlement tolerance of Applied load: KN © 1000 2009 3000 4000 5000 6000 7000 8000 9000 1000011 000, ter 10 Lez 20 Settlement: xm 100 M0 120 190 140 Fig. 1.24 Load-settlement response of test barrette (after Ng et al., 2000b)PART | DEEP FOUNDATIONS - PILES AND BARRETTES Deduced axial load: KN. 0 45 ig. 1.25 Deduced axial load distribution at load test cycle 4 (after Ng et al., 2000b) 0.05mm/10 min. It was therefore decided to unload the barrette to 4555 kN and to hold it for 80h (about three days). After the holding period, the testing programme resumed and two more loading cycles were per- formed. The barrette-soil interface appeared to gain strength as a result of consolidation, which is discussed later. Due to the presence of the ‘soft base’ (i.e. void) underneath the barrette, the barrette ultimately settled about 100mm, enabling the side resistance to be fully mobilized along the shaft. Figure 1.25 shows the deduced axial load versus depth for the last loading cycle, LC4. Barrette side resistance From the gradient of the barrette normal stresses with depth, mobilized side resistance (interface shear stress, r) was calculated and plotted [=]SHORT COURSE IN SOIL-STRUCTURE ENGINEERING 60 50 40 30 Side resistance: kPa Local “1.98 to ~1.02 m PD (fi) “10.02 to 17.02 m PD (sandy © 22.52 to 34.02 m PD (weathe Fig. 1.26 Mobilization of side resistance | displacement: mm “* =1,02 to -10.02 mPD (marine deposi) clay) > ~17.02 to ~22.52 m PO (alluvial sand) ered granite) with local displacement (aiter Ng et al., 2000) Mobilzed side resistance: KPa oo ww a rr ra pti bd f 10 ; ' 15 ’ thi 20 ‘ 5 pee Sas ' bey 30 ' i 8 tou © 45 eee testes ean ER aE a TN Te OS Fig. 1.27 Distribution of mobilized side resistance (after Ng et al, 2000b)PART | DEEP FOUNDATIONS - PILES AND BARRETTES Plezometric level: m PD against deduced local displacement of the barrette in each soil stratum as shown in Fig. 1.26. Figure 1.27 shows the distribution of mobilized side resistance with depth. Pore pressure response at soil-barrette interface After installation of the piezometers, readings were taken continuously to compare them with the initial hydrostatic pore water pressures in the ground (the initial groundwater table was located at about 1.3m PD). These readings are shown in Fig. 1.28(a), (b) and (c). It was found that the measured pore water pressure at gauge P1 in the sandy clay layer (Fig. 1.28) was slightly higher than the corresponding hydrostatic value (1.3mPD) before the loading test. Piezometric level (head) is defined as the sum of the pore water pressure head and the elevation head at each location. The measured piezometric heads recorded at P3 and P4 were a little lower than the hydrostatic conditions in the weathered granite before the commencement of the load test. Labels LC1-LC4 denote the commencement of the first to the fourth load cycles, respectively. Simi- larly, labels UC1-UC4 represent the start of the first to the fourth unload- ing cycles. After the load test, dissipation of excess pore water pressures continued (Fig. 1.28(a)). As expected, the rate of dissipation was much slower in the sandy clay soil (P1) than in the weathered granite (P3 and P4). Dissipation at the alluvial sand interface was slow during loading, but generally the excess pore pressures generated returned to near their equilibrium value (around 2m of head) fairly quickly after slip. Nearly all the excess 2 F Second fallore ey (took place 10 P1 (-18 m PO, sandy clay) P2 (21 m PO, alluvial sand) 3 3 (-30 m PD, weathered granite) 4 (-82 m PD, weathered grrite) sf | Pt ‘ ee cher Pe Pas Ei _—_——_-— F4 o |B: Pa uca List te Fe ‘2eee7 128097 120187 118 1288 1s ae SMB TGR 7M Date: mmiddiyy Fig. 1.28 (a) Variations of piezometric level during load testing. 85SHORT COURSE IN SOIL-STRUCTURE ENGINEERING m Pi (18 m PD, sandy day) Uecachas P2 (-21 m PO, alluvial sand) io | 4555 kN Fa go m PO, weather grant) oe Pa (22 m PD, weathered Grant) | a 5570 KN| —_____| e { Fi E 5425 kN 26 2975 4N OF {Loading sip peak 2 sssoiat °F peak : oars Bo £ 3395 KN, i : a 2 hpi Pe 0 in 4 t t t nD uct ver deuce boting 2 2129197 8:00 12/20/07 12:00 12/29/97 16:00 12729197 20:00 12/80/97 0.00 12/90/97 4:00 12/9087 8:00 12/0087 12:00 Date: mmdtiyy time 2 Tew TEE RE camera @ NE eos mw PLB mE, sey ay) ei (atmo saucy 7 cers en | Bez sas my P55) PO: wet sae 6585 kN PORN | /6875 kN P4 (-32 m PD, weathered granite) 6295 an i 8905 kN sere hy out Piezometrc level: m PD Pa t t t =e : Loca ues Les uce ‘2i98 4:00 1/2198 8:00 1/2/98 12:00 1/2198 16:00 1/2/88 20:00 1/0/980:00 1/9/98 4:00 1/8198 8:00 Date: mmvadiyy time Fig. 1.28 (continued) (b) Variations of piezometric level during first two load test cycles. (¢) Variations of piezometric level during last two load test cycles (aiter Ng et al., 2000b) pore water pressure was dissipated in the weathered granite by 6 January 1998, that is, 80h after the load test. On the other hand, about 3m and 1m excess pore water head still remained in the sandy clay and alluvial sand layers, respectively, at this same time. What appeared to be tidal beha- viour was picked up at P2 and P4, but it is difficult to see since actual peak tide magnitudes dropped by 50% and the scan frequency of instru- ments was reduced substantially after the pile test.PART | DEEP FOUNDATIONS - PILES AND BARRETTES Lateral earth prossure: kPa Lateral earth prossure: kPa Changes of lateral stress at soil-barrette interface During the subsequent three-week curing period before the vertical load test, a gradual, continuous decrease of lateral earth pressure was measured at all lateral earth pressure cells (about 30kPa in the sandy clay and alluvial sand and about 60kPa in the weathered granite). The observed reduction in the contact earth pressure may in large part be due to soil consolidation as indicated by the dissipation of excess positive pore water pressures generated during the construction of the barrette, In Cd First failure took place ‘Second failure. PC1 (-18 m PD, sandy clay) Pos. ® took place PC2 (-21 m PD, alluvial sand) 500 Fos | PCS (-30 m PO, weathered granite) os POs (-32 m PD, weathered granite) 40 + Poa ; Consoicating a pos ae Pos seo Poe —— Poe ee 200 aa Por tt Lot Loe Les Los 100 (@) 0 ieees? 120087 120187 ine 108 1ORB Hes 1B 18GB 7B Date: mmiryy soo Pos PGI (18m PD, sandy cay) PCa, PC2 21 m PD, alluvial sand) oo ——— P63 (-20 m PD, weathered granite) "C3 4 PCé4 (-32 m PD, weathered granite) Poe 400 [toa sto so P02 = Pot TL Pai be + 4 t tt uct jek LC2 UC2 Holding 100 © ° . 12129197 8:00 12/2987 12:00 1229997 16.00 12297 2000 128097 0:00 120097 200 128097 800 129097 1200 Date: mnvdaiyy time Fig. 1.29 (a) Variations of lateral earth pressure during load testing. (b) Variations of lateral earth pressure during first two load test cycles (aiter Ng et al., 2000b) [ a7 |SHORT COURSE IN SOIL-STRUCTURE ENGINEERING 600 a 3 $s PC: (-18 m PD, sandy clay) a) oe G2 (-2 m PO, alluvial sand) 500 4} 2 ele 3 (20 m PO, weathered grant) : — a oH ala PCA (-32 m PD, weathered granite) 400 - PCa 5 §{5. g oo Poa $ 5 aol POS € Poe. Poe Z 200[ Por Baw 4 Tot PCT 3 Lc uca lca uce 100 (©) ° ‘¥2ie8 4:00 1/288 8:00 172/88 12:00 12/98 16:00 17298 20:00 110/98.0:00 1/0198 400 11988 8.00 Date: mmidaiyy time Fig. 1.29 (continued) (c) Variations of lateral earth pressure during last two load test cycles (after Ng et al,, 2000b) [= | addition, the reduction in earth pressure might be caused by a small shrinkage of the pressure ceil units as a result of a fall in temperature after hydration of cement in the concrete. Lings et al. (1994) also reported reductions in earth pressure at the soil-diaphragm wall interface of a heavily over-consolidated stiff clay. ‘The measured earth pressures during the four cycles of load testing are shown in Fig, 1.29. The earth pressure cells all remained virtually constant during the first load and unload cycles (LC1 and UC1) when the barrette displacement was small (Figs 1.24, 1,29(a) and 1.29(b)). With the begin- ning of load cycle 2, some small drops of lateral earth pressure were seen in the weathered granite soil layer until the onset of pile failure {interface slip).CHAPTER TEN Case studies: 15 full-scale loading tests on barrettes in Hong Kong Test barrettes reviewed Tables 1.21 and 1.22 review 15 case studies of barrettes in Hong Kong of which 13 were plain (non-grouted) barrettes (identified as B1-B13) and two were shaft-grouted barrettes (G1 and G2), founded upon granitic saprolites (weathered/decomposed granitic soils). All 15 barrettes were constructed between 1987 and 1998. The design plan dimensions and depths of the barrettes ranged from 2.2m x 0.6m to 2.8m x 1.0m and from 36 to 63m deep, respectively. The embedded depths of the barrettes into the granitic saprolites, d,, ranged from 11.6 to 37.6m. Site locations and ground conditions The typical ground conditions consist of fill material overlying a succession of marine deposits, alluvium and medium-to-coarse-grained granitic saprolites that overlie granitic rocks from the Jurassic to the Cretaceous age (see Fig. 1.30). Site investigations revealed that the typical thicknesses of the granitic saprolites range from about 10 to 40m, corresponding to mean SPT N values ranging from about 60 to 120. The water table is located at about 1.5m above the principal datum (PD) with a small fluctuation due to tidal action, as given in Table 1.23. Methods of interpretation of test data Interpretation of unit side resistance In order to investigate the side resistance behaviour at the interface between the barrettes and the granitic saprolites, the unit side resistances were interpreted from the test data using a consistent method. The unit side resistance is defined as the side shear stress acting on the barrette shaft. For a non-sleeved barrette segment in granitic saprolite, the unit side resistance was calculated by dividing the axial load difference between the two ends of the segment by the perimeter surface area of the segment. The axial load distribution along the depth of the barrette was obtained by multiplying the average vertical strain measured from readings of the strain gauges at each level by a derived pile modulus. 89SHORT COURSE IN SOIL-STRUCTURE ENGINEERING ‘sn9110q Jo wwoHog 04 doy sonyes Ay walapIp AnpuNsTD ssayuam om Aq paioadoyur Pep pI sro} oy) parade 1ur PD 188, ano poiues say ‘painoibisoq, Jou) Sows yuowo2x0}UIaE 1UaIOFFCL 1m saxtordes onruesb Jo S194) Ly ‘@harIeg Jo epbuo] paasays-~ *p ‘oaoudes outwP26 ul yeYs ojazNG poaaols-Low yo adap Pappaquid — °p :210N) p(PE00Z) 127 pue BN bro ozs oN: Let oT ore 0'9T =BOX8T LE zo p(P£00Z) @Tpue BN {Ur O'ZZ) FEO altt0°24) £6°0 (wo'zt) 98°09 ower ON «TSE OE EOP ETE BOXET IS 19 (P£00Z) eT pue BN 08'z oz/or ON Lie a o6e— ost gOxKas Ler eld a(2€002) eT pue BN ogz = Oz/OF ON OTE e Ofte 01d | 80x uc 0.Gr aa ql Tee) 99°0 ol'SE a(PE002) !9T pue BN (MET OE) ZF Oz/OF CN TOE 0% OBE «BOxXez ZED na a(P£00Z) 197 pue BN Q oz/oe ON FBT a 9% 9OXET OTF old (g661) 1 12 eats ep é oz/0e on S67 é Of 9OxXTZ sz 6a (8661) ‘1 12 eats ep é é oN, out é Stl 8Oxz~ ~—-S6E ea p(P£00Z) 197 pue BN aro oz/0e SeA 6'6T st 9IT 8Ox8t Fee id a(t o'ez) 6z°0 (4661) oT (eve ero — oz/0E oN 00 OL gis- ove OTx8z% oss og {v661) OHS HH 680 Q SOA 0% 0% e1s- OO BOxzz ESS sa (2€661) OHS “HH 6e0 == Oz/0e oN Ost st el~— TU BOXTT OLE va (ae661) OHS HH Lv0'0 — Oz/0E ON =e COZE: SEL 90xZ% BSE ea (6861) neId Q t SA OE 2 sis- 00% = BOXZZ ~— OBS za (e661) OH 'S HH FeO Oz/E SPA GETTER SGZSCOXTZ ~—OES 1@ (daw) (dq w) (%) ones aperb —aseq. (wi) age) nay (w) eouareyoy JUEUTBDIOFUTBY—-BJaIIUGD—— plo, ‘p PRM OL (wp (wy azgs deg = ANuepy (G02 ‘197 puD By 19410) Buoy Buoy wy sayaung 189) Jo ArDUNUNS 12°] a14DL, [=]PART | DEEP FOUNDATIONS - PILES AND BARRETTES Bunyjauanjatd 40} 100) UONPAPIN 5 ‘Bugazu0s inp —*) ‘GuRarUos e10jaq pur uoNPAvXe JOT pasdeyo aury DulpuPys = * *OTPAVIXe JO LORPINP = °) :210N a (.avz6 yworueqoony) abpapuey = — 8% 067 LIL ve Of SOK SPA 007 uy oor ‘astesjoxpAH wo (qv16 qeorreyoayy) abpanuey = 17 o6z 6 OF OZ = SPA SQA O'S oz — asteyorpAyT 1D (,a025 qeorueyooyy) sonianeg uorpeay 97 oo = ttt Se OW «PR MATL 9% asienorpaHH ELL (,a025 peorueyooys) sayanied uons°y OF 2 eB co cle A 88K OST 00 es eyoIpA ZIG (,ae16 reoraeyDay) s6ponuey ze a cor 09 et PA ASST 2 a aoreorpsHH 11 abpanuay 2 a t id took a a i ol sonoueg uonovey 2 vet 2 oy os oot qe jeormeyaw 6a aBpamuey 2 ‘ 2 book 00 oo qes6 warueyoW a abpapuey = 1 ole eOr Sr OFF SPA oN 079 ve elt qeah oymerpAH “a sonaieq uonseey Se 2 zel TP OW «A ON OTLB SL qpaB yeaa aSpantay, é a é é a $94 ON é a é qex6 opmespayy oa abpapuey = 67 - SIL €&€ SEs SOA oN, sor gid 06 qeab oynespAH ve seneneg donsey £7 a ou sccsec exes N cre Gir ost qes6 aqnespsty be aot 2 bob a 2 f i 2 qeab onneaps wa abpayuoy OF é Tit VE O96 S®A ON OORT OL Lv qua oynespay ta (sep) () sonoudes oun (3) ua) Suwop Burdens onnes6 — yeorqreno Buun:) omproduey ayy aseg apis (YY UL yeaIqIAAD — [PIO jooy, unoysAs wonDvo Banas Buuode query, woreavox Aanuepr (0g002 127 pud BN 49170) sajja110q 1891 Jo sTIDIaP UOHONIISUOD Za] AGLSHORT COURSE IN SOIL-STRUCTURE ENGINEERING [2] Te wore Fil Marine deposi 18 KN? Sleeve (04 m thick) Alluvium: 20 kNim® Barrette Granitic saprolite: = 20 kNien? | Toe level Grantic rock Fig. 1.30 Side view of typical test barrette (not to scale; after Ng and Lei, 2003a) The vertical strains were calculated with the readings just before the loading test being taken as the reference datum. The derived pile modulus is the composite axial rigidity of concrete and steel in the barrette. The Young's modulus of the barrette was obtained by calibration against the applied load at the barrette head and the average vertical strain at the top set of gauges. In deriving the unit side resistance, the actual as-built cross-sectional area of the barrettes from the concreting record was considered. To investigate the degree of mobilization of the harrette side resistance in granitic saprolites, the local displacements were also determined from the test data. The local displacement is the displacement deduced at the vertical centre of the non-sleeved barrette segment in the granitic saprolite. It was calculated by subtracting the measured barrette head movement from the shortening between the barrette head and the mid- height of the segment. The shortening was estimated using the measured vertical strains, Mobilization rating system Among many other factors, it is well recognized that the magnitude of the side resistance mobilized along the pile-soil interface is also a function of relative pile-soil movement, that is, the local displacement. A minimum. amount of local displacement is needed to mobilize the side resistance fully. In many circumstances, the full capacity of the side resistance may not be achieved in full-scale barrette loading tests. In order to comparePART | DEEP FOUNDATIONS - PILES AND BARRETTES ‘ononreq v Gor Le sb ve sez veo 009 42 2 (12) ev 4 uy got zee ay 6st zy ons oe 2 a2) o'4¢ v we 96 su or eve oer ooogt a v zee ey or z coy gis 1126 2 2 wes zat v7 a Loe os 066 ‘BL FE 88 ote é Q ua a eee or gor oc oue ou ousez 2 ora v oe. out ect oe oz or za ve 2 6a v ore ze an 1 69 808 00081 2 ea v see ve “ % v6 65 sosa (eo) 19% “al Le 998 99 82 '£9 9Aaa|S ON. ve vst SY 6Z a oa a orb 19 68 st ver ve oo00z 2 sa @ ad o8 vor of “9 vez 0081 2 ra a o6z zl ca oe eet Loe ooo8t a ea 2 29F 19 zst 2 2 ove 00st 2 za a ue 99 99 1S o ee ost oog6 (oz) e92 1@ wax) L(sdep)} Pays) Jo ssons (wary (unm) (ea) (shop) ego] iBuans [ona Ny aouvysisar quouresout yuoureaow = (N®) pour anyssarduroa aansoyo—LdS~—=SeSATUN (eas) GuOz peay and —prayand —_ props) sBunox-———epxeruin way ceoyy — ERUIXEYY panaafs UT TeNpISAY — UNUIIKEY uMMNXeY —@Beraay ——abeOAY Suge: aoumsisar uonezmqoyy awjoades apyuesb ur ypsex porordiom Pus Tun, ‘yep ase) BUIpeOT 189) a1@1DU0 Anwepl (oE002 ‘17 uD BN 121J0) sajjaizDq pajsay Wo syinsad Jo AipULUINS Ez" ATAPI,SHORT COURSE IN SOIL-STRUCTURE ENGINEERING [| the interpreted values of the side resistance among the test barrettes at similar degrees of mobilization, a simple three-grade mobilization rating (MR) system (Ng et al, 2001) was adopted to classify the degree of mobilization of the barrette side resistance. The mobilization rating A (i.e. an MR = A) is assigned to test results that fall within 10% of the ultimate unit side resistance determined using Brinch Hansen's (1963) 80% criterion. The ‘80% criterion’ defines failure as the stress (or the unit side resistance) at which the strain (or the local displace- ment) is equal to four times the strain at 20% less stress, The criterion assumes a hyperbolic relationship between stress and strain. The rating B is assigned to results that do not achieve a rating of A but that achieve or have been extrapolated to a local displacement of at least 1% of the equivalent diameter of the barrette, assessed on an equal area basis. A displacement of 1% of the pile diameter is used as a benchmark because of the availability of test results recorded at and extrapolated to this displacement in the literature. It is also defined by the GEO (1996) as corresponding to the ‘substantial’ mobilization of the capacity of the side resistance, although not necessarily the full mobilization of capacity. The rating C is assigned to results for which the degree of mobilization is unknown or the displacement was less than 1% of the equivalent diameter of the barrette The mobilization ratings assigned to the test barrettes are shown in the last column of Table 1.23, together with the loading test data and the inter- preted results. In addition, the concrete test results and the uncorrected mean SPT N values of granitic saprolites near the non-sleeved barrette shafts before barrette installation are given in Table 1.23. je resistance in granitic saprolites Figure 1.31 shows the mobilization of the unit side resistance with local displacement at the interface between the barrettes and the granitic saprolites. Figure 1.31b shows that, of the 13 plain barrettes, the lower portion of the shaft of barrette B11 (i.e, B11b) has the largest (yet seem- ingly far from fully) mobilized unit side resistance of 204kPa, whereas barrette B7 has the smallest ultimate unit side resistance of 37 kPa (Fig. 1.31a). This is most likely because their corresponding mean SPT N. values of granitic saprolites are, respectively, the largest and smallest (N = 187 for B11b and N = 24 for B7; see Table 1.23). Compared with most of the plain barrettes, the shaft-grouted barrettes Gl and G2 show a relatively larger unit side resistance response to a given local displace- ment and a higher ultimate unit side resistance, as expected, In this section, the correlations are studied between the maximum unit side resistance r, and the N’ value and the mean effective vertical stressPART | DEEP FOUNDATIONS - PILES AND BARRETTES 250 ‘Number n parenthesis represenis pan baete ‘the mean SPT Ni value eee “ae Shat-groutes parte 200 7 89 (19) Bis cei 3 B13 (96) BS (82) 3° 26 (66) so 212 49) aa ao 50 6070 BOO {Local displacement: mm @) 250 Bib (187) 1 (166) — Plain barrette MR te Shaft.grouted barrette 200 Number in parenthesis represents the mean SPT N value 3 (132) 3 ita (50) = 100 4 (80) B10 (109) 2 5 (61) 50 o ee Vt Oo oo Local displacement: mm, rc Fig. 1.31 Mobilization of unit shaft resistance with local displacement for barrettes with Mobility Rating (a) A and (b) B and C (after Ng and Lei, 2003a) ‘The traditional correlation between 7, and N is purely empirical, whereas the correlation between 7, and o, is based on the principle of effective stress with an assumption of zero adhesion at the barrette interface and the ratio 7,/o4, gives the well-known 8 value. Apart from these correla- tions, alternative empirical relationships are also studied, such as 9 versus N, z,/N versus N, 7,/N versus pile depth and 4 versus pile depth. Figure 1.32(a] shows the relationship between the maximum mobilized unit side resistance 7, and the mean SPT N value of granitic saprolite for barrettes with an MR = A, that is, more than 90% of the ultimate unit 9SSHORT COURSE IN SOIL-STRUCTURE ENGINEERING 300 w= 307 25/7 agi “Constructed by hydroraise with side soraping 3 8 $ os0 Nes i= 1.0 1 Soo *89 a & g i 100 150 200 Mean SPT value: @) 300 wli= 307 25 wit 18 “Constructed by hydrotraise with side scraping oor i= 40 \ \ : \ 0 —______] ° 100 150 200 Mean SPT value: A © Fig. 1.32 Relationship between measured unit shail resistance and mean Standard Penetration Test N value for barrettes with Mobility Rating (a) A and (b) B (after Ng and Lei, 2003a) side resistance has been mobilized. Tabulated values are presented in Table 1.23. Figure 1.32(b) shows the relationship between the maximum mobilized unit side resistance 7, and the mean SPT N value of granitic saprolites for barrettes with an MR = B, that is, the maximum mobilized unit side resistance is less than 90% of the ultimate unit side resistance. Tabulated values are presented in Table 1,23. Figure 1,33 shows the values of 7,/N plotied against N for plain A. Although the correlation between the values barrettes with an MI 96PART | DEEP FOUNDATIONS - PILES AND BARRETTES 20 Constructed by hydrotraise with side scraping “Trend line -tymean SPT fi value: KPa 05 ° 50 100 160 200 Mean SPT value: i Fig. 1.33 r,/N versus N for barrettes with Mobility Rating A (after Ng and Lei, 2003a) of 7,/N and N is fairly weak, it appears that 7,/N decreases slightly with increasing N values, as shown by the trend line (the regression line). This is similar, although the similarity is relatively weak, to the trend observed for barrettes in old alluvium in Singapore (Ho and Lim, 1998) Figure 1.34 shows the 7,/N values plotted against the embedded depths of plain barreties with an MR =A. It is evident that the 1,/N value decreases with increasing pile depth. This might be due to the increase in the degree of construction disturbance to the granitic saprolites as the “Constructed by hydrotraise with side scraping symean SPT W vale: KPa 20 30 40 50 60 70 80 Pile depth: m Fig. 1.34 7,/N versus pile depth for barrettes with Mobility Rating A (afler Ng and Lei, 2003a) 7SHORT COURSE IN SOIL-STRUCTURE ENGINEERING [=| 8 constructed by hydrofralse with side scraping © 280 B 200 3 3 & 180 3s E 100 2 50 2 ° ° 100 200 300 400 500 0 Mean ect verical stess:KPa (a) 300 “Constructed by hydrofraise uth side scraping 250 200 180 100 50 Measured unit shaft resistance 1,: KPa ° 100 200 300 “00 600 600 Mean effective vertical stress: «Pa >) Fig. 1.35 Relationship between measured unit shait resistance and mean effective vertical stress for barrettes with Mobility Rating (a) A and (b) B (after Ng and Lei, 2003a) depth increases. Relatively larger stress relief is likely to have been induced during the excavation of a deeper barrette trench. Figure 1.35(a) and (b) shows the relationship between the maximum unit side resistance 7, and the mean effective vertical stress o(, in granitic saprolites for MR =A and MR = B, respectively. For consistency, the mean effective vertical stresses ci, (see Table 1.23) were calculated at the mid-heights of the non-sleeved barrette shafts in granitic saprolites by assuming that the unit weights of the fill and marine deposits are 18.0kN/m’, the unit weights of the alluvium and granitic saprolites arePART | DEEP FOUNDATIONS - PILES AND BARRETTES o7 ~ - Upper 95% confidence ‘Average: f= 0.00370 iit: B = 0.0042N na limit: B = 0. / 05 j a o4 e —___Lower 95% confidence 03 limit: B = 0.0082 02 oF “Constructed by hydrofraise with side scraping 00 ° 50 100 150 200 Mean SPT value, Nv (a) o7 Uoper 85% contgence Average: B= 0.00368 ne limit: B = 0.00450 cr os , y : oa 7 Lower 95% confidence os lit = 0.00270 02 | on “Constructed by hydrotaise wit side soraping 09 ° 30 700 150 200 Mean SPT valu, fi (b) Fig. 1.36 Relationship between value and mean Standard Penetration Test N value for barreties with Mobility Rating (a) A and (b) B (after Ng and Lei, 2003a) 20.0 kN/m‘ (see Fig. 1.30), and the water table is at 1.5m PD if the values are not known. Figure 1.36(a) and (b) shows the correlation of J with N for barrettes with an MR = A and an MR = B, respectively. For plain barrettes, the 3 values lie between 0.0030N and 0.0046N, with an average value of 0.0037N, a 95% confidence range of 0.0032N-0.0042N, and a Coefficient of Variance (COV) of 0.17. Figure 1.37 shows the variation of 3 values with embedded depths of plain barrettes with an MR ~ A. Barrettes B8, B9 and B13 have N values greater than 80 (see Table 1.23), and barrettes B6, B7 [>|SHORT COURSE IN SOIL-STRUCTURE ENGINEERING os oa} = / ‘Trend line oee/ on *Constucted by hydrotraise with side scraping) 00. 20 30 40 50 60 70 0 Pile depth: m Fig. 1.37 Relationship between (3 value and depth for barrettes with Mobility Rating A (after Ng and Lei, 2003a) and B12 have N values less than 70. The 3 values for these two groups of barrettes lie between two distinctly different ranges. It is clear that the 3 values for the barrettes with N values greater than 80 are higher than those for the barrettes with N values less than 70. Although the results for these two groups of barrettes are very limited, a trend of an increase in 9 value with depth of the barrette can be observed, but a critical depth cannot be identified, as pointed out by Kulhawy (1984). The observed increase in the 4 value with increasing depth might be caused by an increase in the shear strength with depth at the interface. This is more likely to be the case if the interface contains both bentonite and in situ soil, as an increase in shear strength with depth would not be expected if failure occurs in the bentonite only, This observation (Wates and Knight, 1975; Littlechild and Plumbridge, 1998) suggests that if a bentonite filter cake does exist, it consists of some amount of in situ soil. Construction effects on side resistance Effects of grouting ‘The above-interpreted test results have shown that, compared with plain barrettes, the side resistance response to local displacement for shaft- grouted barrettes is relatively stiffer. The average 3 and 7./N values for shaft-grouted barrettes are about 1.3 and 1.45 times those for plain bartettes, respectively. The grouting effect in improving the side resis- tance for the barrettes in granitic saprolites is comparably as effective as that for a barrette in sand (Hamza and Ibrahim, 2000), which shows that the resistance of the grouted shaft is 1.5 times that of the non-groutedPART | DEEP FOUNDATIONS - PILES AND BARRETTES shaft, The stiffer side resistance response to local displacement and the higher maximum mobilized unit side resistance both result from the grouting which increases both the lateral pressure (i.e. the normal stress) on the barrette shaft and the shear strength of the barrette-granitic saprolite interface. Effects of side scraping and excavation tools After excavation and before base cleaning of a barrette trench, scraping the filter cake on the sidewalls of the barrette trench is sometimes carried out in practice, with the intention of minimizing the possible adverse effect of filter cake on the side resistance as reported in some references (O'Neill and Reese, 1970; Wates and Knight, 1975; Cernak, 1976; Baker et al., 1994; Cheng and Haberfield, 1998). However, it is postulated that the side scraping action would smooth the sidewalls of the barrette trench, leading to a lower frictional angle at the barrette-soil interface, although the filter cake may be fully or partially removed. In this study, side scraping was carried out on shaft-grouted barrettes G1 and G2 and plain barrettes B11-B13, but not on barrettes B1 and B3-B7 (see Table 1.22). Figures 1,32(a), 1.35(a) and 1.36 show the relationships between 7, and N, 7, and o,, and 8 and N, respectively, giving the 7,/N, 6 and 3/N values for barrettes with an MR = A. A superscript asterisk attached to the relevant symbols is used to identify the barrettes constructed with side scraping. Effects of duration of excavation and trench standing time Among many other factors, the maximum mobilized unit side resistance 7, is also a function of soil density and confining pressure at the barrette-soil interface, which can be characterized roughly by the oy and N values. In order to estimate the effects of duration of excavation and trench standing time on the maximum mobilized unit side resistance, the values of 9/N (,/a)/N)} for the plain barrettes were compared, for the sake of minimizing the influence of soil density and confining pressure on the side resistance. Therefore, the 3/N value is to some extent representative of a ‘reference’ capacity of the maximum unit side resistance mobilized under similar soil density and confining pressure conditions, The higher the 3/N value, the greater the mobilization capacity of the maximum unit side resistance. Ideally, the duration of excavation or the rate of excavation in granitic saprolites should be used for this purpose. However, due to the absence of these data, only the duration of excavation for the entire barrette, as presented in Table 1.22, was used. lol