Axial Response of Three Vibratory- and Three Impact-Driven H Piles in Sand

Transcription

1 136 TRANSPORTA TION RESEARCH RECORD 1277 Axial Response of Three Vibratory- and Three Impact-Driven H Piles in Sand }BAN-LOUIS BRIAUD, HARRY M. COYLE, AND LARRY M. TUCKER Three H piles were impact driven in a medim dense sand deposit, load tested in compression, and then extracted. The same three piles were then vibro-driven at a different location and load tested in compression. The top load-top movement crves show that the vibro-driven piles have, on the average, the same ltim ate capacity as the hammer-driven piles. These crves also show that at half the ltimate load, the movement of the vibro-driven piles is 2. times larger than the movement of the hammer-driven piles, on average. At half the ltimate load, however, the movement of the vibro-driven piles was only.2 in. Some of the piles were instrmented; this allowed researchers to obtain the load transfer crves. These crves showed that the vibro-driven piles carry mch more load in friction and mch less load in point resistance than the hammer-driven piles. In the spring and smmer of 1986, a series of vertical load tests was carried ot on instrmented piles driven in sand at Hnter's Point in San Francisco. The Federal Highway Administration (FHWA) sponsored two projects on impactdriven piles: one on the testing of five single piles, and one on the testing of a five-pile grop (J,2). The U.S. Army Engineer Lower Mississippi Valley Division (LMVD), Waterways Experiment Station (WES), and FHWA then sponsored a project on the comparison of impact-driven piles and vibratory-driven piles. Sbseqently, the Deep Fondation Institte drove a nmber of piles with varios vibratory hammers to compare the hammers' efficiencies. This article is a smmary analysis of the pile load tests sponsored by LMVD, WES, and FHWA comparing impactand vibratory-driven piles. The site is characteried, the load tests are described, the load tests reslts are analyed and discssed, and conclsions are made. The details of the work can be fond in Tcker and Briad (3). THE SOIL The soil has been described in detail by Ng, Briad, and Tcker (J). Below a 4 in. thick asphalt concrete pavement is a 4. ft thick layer of sandy gravel with particles p to 4 in. in sie. From ft t 4 ft 1.lt:plh is a hydralic fill made of clean sand (SP). Below 4 ft, layers of medim stiff to stiff silty clay (CH) are interbedded with the sand down to the bedrock. The fractred serpentine bedrock is fond between 4 ft to ft depth. The water table is 8 ft deep. Department of Civil Engineering, Texas A&M University, College Station, Tex Many tests have been performed at the site, inclding standard penetration tests (SPTs) with a dont hammer and a safety hammer, sampling with a Sprage-Henwood sampler, cone penetrometer tests (CPTs) with point, friction and pore pressre measrements, preboring and selfboring pressremeter tests, shear wave velocity tests, dilatometer tests, and stepped-blade tests. The CPT, SPT, and stepped-blade tests were performed before and after driving and testing of the FHW A test piles. Selected profiles are shown in Figres 1 throgh 6. The hydralic fill has the following average properties: friction angle 32 to 3 degrees, water content 22.6 percent, dry nit weight 1 pcf, D 6.8 mm, D 1.7 mm, SPT blow cont 1 bpf, CPT tip resistance 6 tsf, PMT net limit pressre 7 tsf, and shear modls (from shear wave velocity measrements) 4 tsf. THE PILES AND THE LOAD TESTS Three HP14x73 steel H piles were sed in this program. They were all embedded 3 ft below the grond srface; however, a 4. ft deep, 14 in. diameter hole was drilled prior to pile insertion for the impact driven piles, making the tre pile embedment eqal to 2. ft. For the vibratory-driven piles a hole was drilled throgh the 4 in. thick asphalt layer only, making the embedment eqal to 29. ft. Each pile had two angles (2. x 2. x 316 in.) welded t th1; sitles of the pile web as protection for the instrmentation. The first pile (Pile 1) was one of the single piles sed in the FHWA program. Pile 1 was instrmented with seven levels of strain gages and a telltale at the pile tip. This pile was calibrated before driving and the pile stiffness was measred for se in the data redction. The measred vale of pile stiffness (AE) was 614,98 kips. This vale was sed for all three piles. Pile driving analyer measrements were obtained dring the driving of Pile 1 with an impact hammer. Pile 1 was then load tested in compression 3 days after it was driven (FHWA program): this is load test ll. Pile 1 was then retesletl at 67 days after driving: this is load test UR. After Pile 1 was retested, the pile was restrck with an impact hammer and pile driving analyer measrements were obtained. Pile 1 was then extracted and vibro-driven abot 3 ft from the impact test. It was load tested at 33 days after driving (load test 1 V) and again at 63 days after driving (load test 1 VR). After load test lvr the pile was strck with an impact hammer and pile driving analyer measrements were obtained. The second pile (Pile 2) was instrmented with two levels of strain gages: one at the point and one at mid-length. Pile 2 was impact driven and tested at 6 days after driving: this is load test 2I. Pile 2 was then extracted and vibrodriven and

2 1 SPT N VAL UE CBLOWSFTl LOCAL CONE FRICTION Cksf) :i:: 2 I- UJ Cl 3 * Boring Bl + Boring B2 o Boring B3 x Boring BS x- - JC. Average N Vale SPT Tests Performed with a Dont Hammer. ".., 1 2 :i:: 2 I- UJ Cl 3 * CPT 2 X CPT 3 AVERAGE CPT Tests Performed with Standard Electric Cone FIGURE 1 Profiles of standard penetration test blow conts. FIGURE 3 Cone tip penetration test profiles for friction resistance. CONE TIP RESISTANCE Cksf) NET LIMIT PRESSURE Pl CkPol ,.. PBPMT P SBPMT P 2 1 PRESSUREMETER CPT 1 1 CPT 2 1 ".., "" CPT 3 2 AVERAGE, CPT Tests Per- I- Standard Elec- tric Cone. lj.j Cl ::r 2 formed with a , :i:: I- 3 lj.j Cl " ' ' ' ' ' ' ' FIGURE 2 Cone tip penetration test profiles for point resistance. '-'-..LLL'-'-..LLLLJ..-'-'1.LJ..LJcL.LJ..LJcL.1J..LJL.l-J FIGURE 4 Net limit pressre profile.

3 138 TRANSPORTATION RESEARCH RECORD 1277 FIRST LOAD MODULUS Ee MPo) RELOAD MODULUS Er MPc) I OD O n-rtt"'rrr,.,,..,,.'t't'",,,..,,..,..,,.,..-r-r--r-r--r-r-,..,.., ID 1 -PBPMT E PRESSUREMETER ID._.PBP!-1T E r... SBPT E SBP!-IT E r PRESSUREMETER ::c f- n. 3 ljj 3 ::c f- n. ljj so L_ L. FIGURE First load modls profile. FIGURE 6 Reload modls profile. tested at 64 days after driving. This is load test 2V. After load test 2V, the pile was strck with an impact hammer and pile driving analyer measrements were obtained. The third pile (Pile 3) was not instrmented. It was impact driven and tested at 12 days after driving: this is load test 31. Pile 3 was then extracted and vibrodriven. It was tested at 12 days after driving: this is load test 3V. Table 1 smmaries the eight load tests presented in this report. The impact-driven piles were installed sing a Delmag D22 diesel hammer at fll throttle. All restrike measrements were also obtained sing this hammer. The final blow cont for the impact driven piles averaged 11. blowsft (Table 2). The specifications for the Delmag D22 hammer are given in Table 3. The vibratory-driven piles were installed with an ICE-216 vibratory hammer (Table 4). The penetration rate over the last 2 ft of penetration varied between 12 and 2, averaging 16 ftmin (Table 2). The impact hammer and the vibratory hammer drove the piles in approximately the same time. The load test procedre consisted of applying the load of 1 percent of the estimated ltimate load. Each load step was held for at least 3 mintes, with the load being monitored continosly. All readings from electronically monitored instrments were recorded every mintes. RESIDUAL STRESSES Residal driving stresses were measred for Pile 11. The stress profile was not monitored between load test 11 and test lir. Therefore the same residal stress profile after driving was also sed for load test lir. Since additional residal stresses are sally indced de to a compression test, this profile and all sbseqent profiles for Pile lr are in error by an nknown amont. The residal stress profile is the first load profile shown in Figre 7. An attempt was made to record the residal driving stresses for Pile 21. However, the readings obtained were jdged nreliable. Therefore, the residal stress profile from Pile 11 was sed in redcing the load test data for Pile 21 (Figre 7). The vibratory-driven piles were assmed to have no residal driving stresses. No attempt was made to verify this. However, pblished data sbstantiate this assmption ( 4). LOAD TEST RESULTS The load movement crves for all the piles are shown on Figre 8. The point load-point movement crves backfigred from the measred data are presented on Figre 9. The load verss depth profiles for Piles 11 and 1 V are presented on Figres 7 and 1, respectively. The friction transfer crves were obtained after fitting the load verss depth profiles with a second order polynomial crve. The friction transfer crves are shown for Piles 11 and 1 V in Figres 11 and 12, respectively. PILE DRIVING ANALYZER RESULTS The pile driving analyer (PDA) consists of the dynamic monitoring of strain gages and accelerometers attached to the

4 Briad et al. 139 TABLE 1 LOAD TEST PROGRAM SUMMARY Test Piles 11 - FHWA Pile (Impact) 3 Day Test UR - FHWA Pile (Impact) 67DayTest lv - FHWA Pile (Vibratory) 33 Day Test lvr - FHWA Pile (Vibratory) 63 Day Test 2I - New Pile (Impact) 6 Day Test 2V - New Pile (Vibratory) 64 Day Test 3I - New Pile (Impact) 12 Day Test 3V - New Pile (Vibratory) 13 Day Test A 11, UR, 1 V, and 1 VR - Seven strain gage levels B. 21 and 2V - Two strain gage levels C. 31 and 3V - No strain gages TABLE 2 DRIVING RECORDS Pile 1 Pile2 Pile3 Impact Driven - last foot 12 blft NA 11 blft - total 96 blows NA 131 blows Vibro-Driven - last foot 2 ftmin 1 ftmin 12 ftmin - total 8 sec. 96 sec. 8 sec. TABLE 3 SPECIFICATIONS FOR DELMAG D22 IMPACT HAMMER Rated energy Ramweiht Blowsnnte Maximm explosive pressre on pile Working weight Drive Cap weight 39,7 ft-lbs 4,8 lbs ,7 lbs 11,27 lbs 1, lbs TABLE 4 SPECIFICATIONS FOR ICE-216 VIBRATORY HAMMER Eccentric moment Freqency Amplitde Power Pile clamping force Line pll for extraction Sspended weight with clamp Length Width Throat width Height with clamp Height withot clamp 1 in-lbs 4-16 vpm inches 11 HP tons 3 tons 482 lbs 47 inches 16 inches 12 inches 78 inches 68 inches pile close to the top. The strain measrements are sed to obtain force measrements as a fnction of time, and the accelerometer measrements are integrated to obtain velocity verss time. From these two measrements, varios parameters may be calclated, inclding the maximm energy delivered to the pile and the static resistance of the pile by the CASE method. This static resistance represents the PDA capacity prediction. PDA measrements were made on Pile 11 (driven with the impact hammer) for both the initial driving and the restrike seqence after the load tests were performed, and on Piles lv and 2V (driven with the vibratory hammer) for restrike after the load tests were performed. The reslts are presented in Table. DISCUSSION OF RES UL TS Top Load-Movement Crves The top load-movement crves for the eight load tests are shown together in Figre 8. Two main observations can be made. First, in all cases the vibratory-driven piles have a lower initial stiffness than the impact-driven piles. Second, the impact-

5 2 LOl\D CKIPS> so 7 JOO " 1 J: 1 a w c 2 2 PILE II INCLUDING RESlDUl\L LOl\OS 3'-'-'--'-._._.._..._..._._...._._..,_.._._._ FIGURE 7 Corrected load verss depth profiles for Pile 11. TOP LOl\O Cl(J PS>, so JOO ISO 2 'i; \. 2 in, ' ' \ \.s -- JMPl\CT \ VJBRl\TORY '\ - i..., \ JVR w 2: w..j UI "' L 2V I I I I. S D1 + PLAE FIGURE 8 Comparison of load-settlement crves.

6 1 " Ill "" Ill -Ill 12 JOO 7 "" ct: - so _. - ::J. 2.,,, I PILE IV oc.._...' ',..---'...., s 1. MOVEMENT JN.> FIGURE 9 Comparison of point load transfer crves. L\ CKIPS> 2S 7S JOO 12 lso 17S 2 " 1 J: IS w Q 2 2S 3L-L-'-.LJLJ..-'-.._._.,_._.._._.._,._._._.._..._. _,_._...._.._.,_.._._._.._, FIGURE 1 Raw load verss depth profiles for Pile lv. PILE II DEPTH 2 FT Ill 2 FT -.s If I- JS FT ct: JO FT I- FT ::J -. '-...,....., i..'--..s l. 2.S HOVEHENT CIN. > FIGURE 11 Friction verss movement crves for Pile 11.

7 142 TRANSPORTATION RESEARCH RECORD PILE IV DEPTH 2 FT 2 FT 1 FT ID FT G: UI -I -: i.. I- - ::J. FT FIGURE 12 o ,... ;..... L J MOVEMENT l N. > Friction verss movement crves for Pile lv. TABLE PILE DRIVING ANALYZER RESULTS (7) Maximm Maximm Estimated Force Energy Capacity" Pile Blowcont (kips) (kip-ft) (kips) lib 1212 in c 616 in vc 716 in vc 1412 in "Case method capacity assmes a damping constant J =.2. Initial driving. crestrike. driven piles show more consistent response between piles than the vibratory-driven piles. Third, on the average, the ltimate load is the same for the vibro-driven and the impact-driven piles, bt the ltimate load is more erratic for the vibro-driven piles. In an effort to qantify these observations, for measrements have been made from the load-movement crves. First, the ltimate load has been defined as the load corresponding to a movement of one-tenth of the eqivalent pile diameter pls the elastic compression of the pile nder that load as if it acted as a free-standing colmn. This line has been drawn on Figre 8. Second, the load at a movement of.2 in. has been obtained from the load-movement crves. Third, the movement at one-half the defined ltimate load has been obtained. Forth, the piles' stiffness response has been calclated as one-half the defined ltimate load divided by the movement occrring at that load. These for items are tablated in Table 6 for the eight load tests. Table 6 shows that, on the average, there is only 1 percent difference in the average ltimate load between the impactdriven piles and the vibratory-driven piles. However, the coefficient of variation of the ltimate loads for the vibratorydriven piles is 4.3 times higher than that of the impact-driven piles. The factor of safety wold therefore have to be higher for the vibro-driven piles in order to obtain the same risk level as for the hammer-driven piles. The second colmn of Table 6 shows that at a movement of.2 in. the impact-driven piles carry 33 percent more load than the vibratory-driven piles. The coefficient of variation of this load for the vibratory-driven piles is 4.2 times higher than that of the impact-driven piles. The movement at one-half the ltimate load for the vibratory-driven piles is over two times larger than that of the impact-driven piles, and the coefficient of variation is. times larger. The movements, however, are very small even for the vibro-driven piles. The initial stiffness response of the pile, defined as onehalf the ltimate load divided by the movement at that load, is 1.91 times higher for the impact-driven piles than for the vibratory-driven piles. The coefficient of variation of this stiffness for the vibratory-driven piles is 3.6 times that of the impact-driven piles. Load Distribtion The load distribtion of the vibratory-driven piles differs greatly from that of the impact-driven piles. At the maximm load, the impact-driven piles carried approximately 1 percent of the load in point resistance, whereas the vibratory-driven piles carried only 13 percent of the load in point resistance (see for example Figres 9, 11, and 12). In the reload test on the vibratory-driven pile, the point resistance had increased to 29 percent of the total load. This indicates that the difference in the driving process cases a different soil reaction, bt the difference becomes less pon repeated loading. Indeed, one compression load test can be considered as a slow blow. The nit friction profiles at maximm loading are shown in Figre 13 for the five instrmented pile tests. Again it can be seen that there is a definite difference in soil reaction between the impact-driven and vibratory-driven piles, and that the difference becomes less prononced pon repeated loading. Load Transfer The H pile presents a problem when compting nit point resistance and nit friction vales. What is the failre srface?

8 TABLE 6 ANALYSIS OF PILE TEST RESULTS Pile 11 Load at Load at Movement Initial oco+.2 in. at O2 Stiffness P AE (kips) (in. (kipsin.) (kips) llr Average Standard Deviation Coefficient of Variation IV lvr 2V 3V Average Standard Deviation Coefficient of Variation Impact Vibratorv UNIT FRICTION AT HAXIHUM LOAO. s KSF> J. :c w c PILE ''P:L7 JV JVR ' ' ' 2 ' ' ' ' ' FIGURE 13 Friction verss depth profiles.

9 o _ , 144 TRANSPORTATION RESEARCH RECORD 1277 One possible assmption is that the pile fails along a rectangle which encloses the H section, with a soil plg forming between the flanges. Another possibility is that the pile fails along the soil-pile interface, with no soil plg forming. Previos research has shown that a better assmption may be that the failre srface is in between the previos two assmptions, with a soil plg filling half the area between the flanges (1). For the HPl 4x73 piles sed in this stdy, this assmption gives the following properties: perimeter, 7.47 in.; tip area, 19.8 in. 2 This assmption is sed for all frther analyses. Figre 9 shows the nit tip resistance verss tip movement crves for the five instrmented piles. Figre 14 shows the same crves with the tip resistance normalied by dividing by the maximm tip resistance. These two figres show a fndamentally different reaction between the vibratory-driven piles and the impact-driven piles: the tip resistance of the vibratory-driven piles is mch lower than that of the impactdriven piles, and the initial slope of the crve is different. However, the difference in shape almost vanishes pon reloading. The shape of the tip resistance crve of the initial loading of the vibratory-driven pile sggests that the sand immediately nder the pile point is initially loose bt densifies as the pile is loaded. Indeed, at higher loads the tip resistance begins to increase at a faster rate, rather than reaching ii limiting vale as the other tests show. By comparing the crves for tests 11 and UR, it can be seen that the impad-riven piles do not ndergo sch a change in behavior between initial loading and reloading. Figres 1 and 16 show the normalied friction movement crves for the impact-driven piles and vibratory-driven piles, respectively. Figre 1 shows again that the impact-driven IJI -IJI a: c N -..J x a: x I I r i; { - r - _,. NORMALJZEO Cl-W CURVES JMPACT ORJVEN VJBRATORY DRIVEN. I. MOVEMENT CJN.) FIGURE 14 Normalied point load transfer crves :: c.., N..J x :: c. 7.,i I. 2 I I ) I " " - NORMALJZEO F-W CURVES FOR JMPACT DRIVEN PJLES -- PJLE JJ PJLE JJR. PJLE 21 "-"-'"'-L--'--'---L-L-L--&---''--'--'--.._ ,...._...,. J MOVEMENT CJN. > FIGURE 1 Normalied friction transfer crves for impact-driven piles.

10 Briad et al. 14 a It: i.. NORMALIZED F-W CURVES.., FOR VIBRATORY DRIVEN PILES. I -- PILE JV x PILE JVR It: a. 2 o _ ,' ,. s I. 2. MOVEMENT JN.) FIGURE 16 Normalied friction transfer crves for vibratory-driven piles. piles are very consistent in their responses and that no change in behavior occrs between initial loading and reloading. However, Figre 16 shows that the vibratory-driven pile exhibits a great change in behavior between initial loading and reloading. The five crves for the initial loading have a mch softer initial response than the reloading crves. The initial loading crves also become softer as the depth increases, whereas the reload crves become stiffer as the depth increases. The reloading crves exhibit a pattern that matches the impactdriven piles very well. Effect of Time Piles 11 and 1 V were tested abot 31 days after driving and then retested abot 6 days after driving. The plots ofltimate load verss time for two criteria are shown in Figres 17 and 18. The trend given by those two figres does not allow s to conclde that there is an increase in capacity verss time. Indeed, Figre 17 shows an increase in stiffness, bt Figre 18 shows a decrease in capacity for the vibratory-driven piles; for the impact-driven piles, Figre 17 shows an increase while Figre 18 shows a slight increase. At Lock and Dam 26, where impact-driven H piles were also load tested (6), the capacity obtained in the load test was 67 percent higher than the capacity predicted by the wave eqation method, on average. This cold have been de to a 67 percent average gain in capacity of the piles between the time of driving and the time of the load test (1 week). Note that the sand at Lock and Dam 26 had an average of 7. percent passing the No. 2 sieve and a blow cont averaging 3 bpf, while the sand at Hnter's JV 21 llr II - I VR - IV. x 2V?. I > FIGURE 17.. I TIME DAYS> Pile capacity at.2 in. settlement verss time. I I.. 1

11 146 TRANSPORTATION RESEARCH RECORD Ill " 21 IIR IL. x IV 3V x UI 11 x. IVR 2V..J 1 - IL. Q. c i UI..J IL. -. FIGURE 18 1'--. -'----.,., ljj 1 Pile ltimate capacity verss time. TIME CDAYS> Point had percent passing the No. 2 sieve and a blow cont averaging 1 bpf. In the case of Hnter's Point, several factors inflence the change of capacity, one of which is the effect of time. The others inclde the inflence of soil heterogeneity and the inflence of the first load test on the second load test. In order to isolate the effect of time, one soltion wold be to se lowstiain testing to obtain the variation of stiffness verss time. Another soltion wold be to se a series of load tests more closely spaced in time, performed on the same pile over a longer period of time (e.g., 1 day, 4 days, 1 days, 4 days, 1 days). CONCLUSIONS AND RECOMMENDATIONS The main conclsions to be drawn from this stdy are the following: 1. Compared to impact driving, vibratory driving of piles leads to approximately the same maximm load at large movements, to a large scatter in this maximm load from one pile to the next, and to a larger movement at working loads. This larger movement was only.2 in. and may not reqire a redction in allowable load; however, the scatter in the maximm load may reqire a higher factor of safety for the same risk level and therefore a lower allowable load. 2. The load distribtion is inflenced greatly by the driving process. For the piles in this stdy, impact driving led to a point resistance that was 1 percent of the total load at maximm loading, compared to 13 percent for vibratory driving. The load transfer crves are also affected by the driving process. Vibratory driving led to load transfer crves that reqired mch larger movements to reach maximm loading. 3. The effens of vibratory driving listed above are lessened pon reloading of the piie. Upon reloading, a vibratory-driven pile carries a larger percentage of the load in point resistance, and the load transfer crves take on the same shape as the impact-driven piles. It wold therefore seem desirable to have a vibratory hammer that can switch to an impact hammer and impart a few seating blows at the end of the vibro-driving seqence. 4. The data show that, in this relatively loose clean niform sand, there is a trend for impact-driven piles towards a slight increase in capacity verss time. For vibratory-driven piles however, the data does not show any clear trend.. The piles encontered very easy driving, and the time reqired to driven the piles was the same for the impact and vibratory hammer. ACKNOWLEDGMENTS This project was sponsored by the U.S. Army Engineer Division, Lower Mississippi Valley, and monitored by the U.S. Army Engineer Waterways Experiment Station. Frank Weaver and Rich Jackson of LMVD and Britt Mitchell of WES are thanked for their valable comments and spport. REFERENCES 1. E. S. Ng, J.-L. Briad, and L. M. Tcker. Pile Fondations: Tiie Behavior of Piles in Cohesionless Soils, Report FHWA-RD FHWA U.S. Oepartmcnl of Transportation E. $. Ng, J.-L.Briad, and L. M. Tcker. Field Sr11dy of Pile Grop Action ;,, Sa11d. Report FHWA-RD-88-SJ. FHWA, U.S. Department of Transporta1ion, L. M. Tcker and J.-L. Briad. A:i:ia Response of Three Vibratory

12 Briad et al. and Three Impact Driven H Piles in Sand. Misccllanco. Paper GL U.S. Army Engineer Waterwy Experiment tation, Vicksbrg, Miss., A. H. Hnter and M. T. Davisson. Measrements of Pile Load Transfer. In Performance of Deep Fondations, Report STP 444, American Society for Testing and Materials, New York, 1969, pp F. Rasche G. G. Goble, and G. E. Likin. Dynamic Determination of Pile Capacity. Jomal of Geoteclmical E11gi11eeri11g, AS E, Vol. 3, No. 3, 198, pp L. M. Tcker and J.-L. Briad. Analysis of the Pile Load Test Program at the Lock & Dam 26 Replacement Project. Miscellaneos Paper GL U.S. Army Engineer Waterways Experiment Station, Vicksbrg, Miss., D. M. Holloway. Dynamic Monitoring Program Reslts, FHWA Vibmtoiy Hammer-Pile Testing, Hmer's Point, California. Report to GeoResorce on hant, Pblication of this paper sponsored by Committee on Fondations of Bridges and Other Stmc111rc. 147