An Electro-Vibrocone for Evaluation of Soil Liquefaction Potential

by Alec McGillivray1, Thomas Casey1, Paul W. Mayne2, and James A. Schneider3

ABSTRACT

    An electric downhole vibrocone has been designed for the site-specific evaluation of soil liquefaction susceptibility in high seismicity regions. The device induces localized cyclic pore pressures while concurrently measuring dynamic tip and friction resistances. The latest design couples a dual-element penetrometer with a piezo-actuator shaker. This vibrocone offers a number of advantages over its predecessors including downhole vertical oscillation, adjustable dynamic force and frequency of excitation, and the simultaneous measurement of porewater pressures at multiple positions

INTRODUCTION

    The evaluation of soil deposits in high seismic regions includes laboratory testing and field investigations to determine the likelihood of liquefaction, as well as the subsequent post-cyclic strength behavior. Current laboratory and field methods are frought with empirical corrections and modifiers that result in high uncertainty for many sites, particularly the recently-recognized seismic zones of New Madrid/MO and Charleston/SC. In the laboratory approach, cyclic triaxial and cyclic simple shear tests on cohesionless materials rely on obtaining high-quality samples. Preservation of the soil structure by freezing is very costly and not readily available for use on routine projects. Moreover, the utilization of cyclic laboratory tests relies on a number of empirical modifiers that attempt to account for initial stress state (Ko), overburden stress level (Ks), depth effect, number of cycles to failure (Nf), and accumulated strains. Effects of aging, sampling disturbance, inherent fabric, and re-establishment of the in-situ stress state are difficult to take into account.

    Accordingly, for most projects, the greatest weight in the assessment of liquefaction potential is given to the results of field in-situ tests. To date, the standard penetration test (SPT) has been most widely used for this purpose. Unfortunately, the number of corrections that must be made to the raw data from even the most carefully conducted tests leads to similar difficulties as noted for lab methods. The SPT-N value must be modified to include corrections for energy efficiency, overburden stress level, fines content, borehole diameter, barrel liner, rod lengths, aging, and other factors (Skempton 1986; Kulhawy & Mayne 1990; Robertson & Wride 1997, 1998).

    For liquefaction analysis, Seed et al. (1975) established a relationship between the adjusted (N1)60 and cyclic stress ratio (CSR = cyclic shear stress normalized to vertical effective stress). The CSR relates earthquake acceleration to the dynamic shear stress of the soil and can be obtained through a simplified procedure by Seed and Idriss (1971):

                image002.gif (691 bytes)   (1)

where tcyc = equivalent uniform cyclic shear stress, tave = average cyclic shear stress, svo˘ = effective vertical stress, sv = total vertical stress, amax = peak ground acceleration, g = gravitational constant, and rd = stress reduction coefficient accounting for the stiffness of the soil column. The energy-corrected SPT value, normalized to a stress level of one bar and designated (N1)60, is used in the well-known curves for a binary assessment of either "liquefaction" or "no liquefaction" (e.g. Glaser & Chung, 1995; Robertson & Wride, 1998; Olson & Stark, 1998).

    More recently, interest has focused on development of CSR curves from the cone penetration test (CPT) for the in-situ evaluation of liquefiable soils (e.g., Suzuki, et al., 1995; Stark and Olson, 1995). The CPT offers several advantages over the SPT including better standardization, a continuous record with depth, and the ability to measure several parameters, including: tip stress qt, sleeve friction fs, porewater pressure ub, and vertical inclination i. However, modification factors for overburden stress level, aging, and fines content are still necessary.

    Other in-situ tests have been correlated for use in liquefaction assessment, including the flat plate dilatometer test (Reyna and Chameau, 1991) and shear wave velocity (Andrus and Stokoe, 1996). The seismic cone (Campanella, 1994) produces four measurements, including tip resistance and shear wave, thus allowing two independent evaluation of liquefaction potential. However, since each of the above approaches provide only an empirical indirect assessment, a more rational and direct method was sought using a vibrocone.

VIBROCONE CONCEPT

    The basic concept of a vibrocone consists of a cone penetrometer with trailing vibrator unit, as shown by Figure 1. Prior versions of a vibrocone device were developed in Japan, Italy, and Canada, and a detailed review is given by Wise, et al. (1999). The original model of Japanese vibrocone was built by the Public Works Research Institute (PWRI) that applied a downole horizontal centrifugal force of 32 kgf and operated at a set frequency of 200 Hz (Sasaki and Koga, 1982). Each test required two sister soundings, one static and one dynamic. This was followed by later designs with dynamic forces of 80 and 160 kgf at 200 Hz (Teparaksa, 1987). A similar model developed at ISMES in Italy also used horizontal vibration (Piccoli, 1993).

image004.gif (3294 bytes)

Figure 1. Schematic of the Original Vibratory Piezocone (Sasaki and Koga, 1982)

    Previous field studies used simple side-by-side comparisons of static tip resistance (qcs) with the dynamic tip resistance (qcd). Figure 2 illustrates the recorded profiles of qcs and qcd from vibrocone tests at two Japanese sites. For site 1 which historically shows no evidence of sand boils or settlements following earthquakes, both the static and dynamic tip resistances are relatively similar (except for a small localized zone about 3 m deep), inferring no major liquefaction problems. In contrast, the results of vibrocone tests at site 2, which is known to have liquefied repeatedly, illustrate that the dynamic resistance is considerably reduced at depths between 2 to 5 meters, reflecting the contractive nature of the sand deposit and high likelihood of liquefaction.

image006.gif (3622 bytes)

Figure 2. Vibrocone tests (a) at site 1 which shows no apparent damage during seismic events and (b) at site 2 with historical liquefaction evidence following seismic events (modified after Sasaki et al., 1984).

 

Table 1.  Contributions to the Development of the Vibrocone 

Country

Author(s)

Details

Results

Japan

Sasaki & Koga, 1982
Sasaki et al., 1984 (PWRI)

·         Down-hole vibration at 200 Hz

·         32 kgf horizontal centrifugal force

Reduction in qc reflected possible liquefiable zones

Japan

Teparaksa, 1987 (PWRI)

·         Down-hole vibration at 200 Hz

·         80 kgf horizontal centrifugal force

Compared sister sets of static and dynamic soundings

Canada

Moore, 1987 (UBC)

·         Vibration applied at top of rods

·         Vertical force at 75 Hz frequency

Shoulder pore pressures did not identify liquefiable layers

Italy

Mitchell, 1988

Piccoli, 1993 (ISMES)

·         Down-hole horizontal vibration

·         200 Hz

Qualitative interpretation of tip resistances

USA

Wise, et al. (1999)

Schneider, et al. (1999)

·         Downhole vertical-impulses (5 Hz)

·         Midface porewater pressures

Increased porewater pressures in liquefiable layers

 

    A UBC vibrocone (Moore, 1987) was constructed using an uphole vertical vibrator consisting of an oscillating pair of eccentric counter-weights to the actuator assembly above-hole in the cone rig. The electric power for the vibrator was coupled with the rig power, causing fluctuations in the operating frequency. Tests in a silt showed a reduction in qc, but a shoulder element did not show evidence of excess pore pressures.

VIBROCONE DESIGN

    Under funding from the USGS and NSF, the development and calibration of a piezovibrocone penetrometer has been initiated in a joint research program at Georgia Tech and Virginia Tech (Wise, et al. 1999; Schneider, et al. 1999). The new vibrocone is an improvement over prior vibrocones for the following reasons:

    Seismic strong motion records from most earthquakes show a spike in their power spectra at frequencies of 1 to 5 Hz, yet previous vibrocones operate at frequencies of 75 to 200 Hz, which are considerably higher. Therefore, from an operational standpoint, a lower frequency vibro-unit was desired having adjustable force excitation.

    For the initial design, a downhole pneumatic system was built to offer simplicity in operation and economy in construction. Impulse-type loading was used for the initial trials. Two units were needed, one for field use by Georgia Tech and one for CPT calibration chamber testing at Virginia Tech. For chamber calibration tests, a 15-cm2 triple-element piezocone was donated by Fugro Geosciences. This allows for the simultaneous measurement of cum-ulative cyclic pore pressures at three locations: midface (u1), behind the tip (u2 or ub), and behind the sleeve (u3). For initial field trials, a 10-cm2 Davey-type piezocone with a single midface (u1) porous element was used.

    The fabricated prototype attached to the triple-element piezocone is illustrated in figure 3. Key resistance parameters are measured, including tip resistance (qc), sleeve friction (fs), multiple porewater pressures (u), and frequency content of excitation. Both an HP oscilloscope and spectrum analyzer have been used for the latter.

    The vibratory module consists of a solenoid valve, air cylinder, an impact mass, velocity geophone, a housing assembly, and nitrogen gas. An electronic timer provides electrical pulses to a solenoid valve which opens and closes at the rate dictated by the timer setting. The solenoid pressurizes the air piston which in turn drives the excitation. Two modes of excitation are available to increase the versatility of the unit and range of applicability of the results. One mode is driven by a spring-mass oscillation and another uses impulse with an impacting mass. To measure the applied force and frequency, a small OYO Model 14-L9 geophone is installed in the unit. In future designs, an inexpensive micromechanical system (MEMS) will be used to measure dynamic accelerations. In lieu of an oscilloscope or analyzer, the MEMS can be read directly using a notebook computer.

    Without reconfiguring the vibrator, the dynamic force can be adjusted at the control panel by increasing or decreasing the pressure from the tank or compressor. The air cylinder in the prototype has a bore size of 2.7 cm corresponding to a force of 100 kgf at 1.7 MPa pressure. The frequency of excitation can be varied by simply changing the timer setting which is capable of duty cycles from 0.001 seconds to 9999 hours and varied by increments of 0.001 seconds.

image008.gif (6083 bytes)

Figure 3. Prototype Pneumatic Impulse Generator and Triple Element Piezocone.

    The solenoid is rated to perform adequately at duty cycles down to 60 per second. Note that calling the timer setting a ‘frequency’ is not quite accurate. The timer setting is actually a duty cycle for which the frequency content of the system can be determined by performing a Fourier analysis on the voltage versus time data. A spectrum analyzers has also been used to determine the frequency content in real-time. Figure 4 illustrates the various piezovibrocone components.

image010.jpg (13025 bytes)

Figure 4. Vibrocone Model with Downhole Pneumatically-Driven Impulse Module.

 

FIELD PERFORMANCE

    The impulse vibrocone system has been tested to check its ability in generating excess porewater pressures. The device was advance in historically-liquefiable sands in a series of soundings near Charleston, South Carolina, USA. This seismic region was selected for trial soundings with the piezovibrocone due to the noted abundance of mapped paleoliquefaction features resulting from the Charleston Earthquake of 1886 with M = 7.5 to 7.7 (Martin & Clough, 1994).

    The specific test site locations were based upon the prior documentation of accessible sites where prior SPT, CPT, and grain size information was reported (Martin & Clough, 1990). Gregg In-Situ, Inc. of Aiken, South Carolina provided a 25-tonne cone truck and an electronic 10-cm2 seismic piezocone with a porous element at the u2 (shoulder) position. These data provided a baseline reference for comparison with both static and dynamic soundings for the VCPT.

    The Charleston site selected was the Thompson Industrial Services in the Atlantic coastal plain region of South Carolina. The water table was within 1.5 m of the ground surface. Based on published first-hand accounts and field reconnaissance, the extent of liquefaction due to the 1886 Charleston EQ had been characterized as extensive and moderate at the Ten Mile Hill and Eleven Mile Post sites, respectively (Martin and Clough, 1994).

    Static reference soundings were performed with both with an electronic ConeTec (type 2) piezocone and the electric Davey (type 1) piezovibrocone (see Figure 5). The tip stresses and sleeve resistances indicate favorable comparisons between the two soundings. The midface u1 readings mirrored the shoulder u2 measurements, yet midface pressures were generally above hydrostatic while shoulder pressures were slightly below hydrostatic conditions.

image012.gif (11300 bytes)

    Adjacent to these two static tests, a dynamic sounding was also performed using the vibro-unit at 5 Hz, thus completing a VCPT. At depths of between 3.5 to 5.1 meters, Fig. 5 shows a significant spike in the porewater pressure response of the dynamic sounding. It is believed that this indicates a zone of liquefiable soil. Moreover, it can be seen that tip resistances for all three soundings were relatively low in this zone, but the dynamic tip resistance (VCPT) does not noticeably fall below either of the tip resistances from the two PCPTs. Noteably, however, a complete and rational interpretation for the VCPT will require a complex consideration of all possible responses for the loading conditions, including those due to static liquefaction, steady-state, quasi-liquefaction, cyclic mobility, as well as dilatant behavior of soils, as planned in future studies (Schneider et al. 1998). The measured qt and u readings, in fact, may reflect totally different portions of the induced effective stress paths.

image014.jpg (13589 bytes)

Figure 6 . Components of the next generation design: ElectroVibrocone System Using Dual-Element Penetrometer, Frequency Generator, Amplifier, and Piezo-Stack for Dynamic Excitation.

SUMMARY

    A specialzed in-situ tool, termed the piezo-vibrocone, has been constructed for the direct evaluation of soil liquefaction potential on site-specific projects. The device utilizes a multi-element piezocone coupled with a downhole impulse vibro-unit operating vertically at low frequencies. Preliminary trials in historically-liquefied sands of the Charleston seismic region show generation of cyclically-induced porewater pressures. Additional improvements are ongoing in producing more uniform and sinusoidal forces, dynamic stress measurements, and controlled soil conditions involving large calibration chamber testing.

ACKNOWLEDGMENTS

    The authors thank Dr. Cliff Astill of NSF, Dr. John Unger of USGS, and Dr. Dan Abrams of the Mid-America Earthquake Center for funding this project. Additional appreciation is given to Recep Yilmaz and Rick Klopp at Fugro Geosciences, Brad Pemberton at Gregg In-Situ, Craig Wise at Black & Veatch, as well as James K. Mitchell, Tom Brandon, and John Bonita at Virginia Tech.

REFERENCES