INTRODUCTION

The Cone Penetrometer Test (CPT) yields in situ measurements of soil characteristics allowing detailed assessment of site-wide liquefaction potential. CPT measurements are converted to equivalent Standard Penetration Test (SPT) blowcounts and then used in the blowcount method of liquefaction potential assessment (16). The CPT measurements are treated as equivalent SPT blowcounts for several reasons: the SPT method was developed based on observed cases of liquefaction (14); the method is in routine use, making the expansion of the method's data base likely; and most importantly, the SPT and CPT are both similarly influenced by most soil compositional and environmental variables (8). Reasons for use of the CPT-SPT-liquefaction potential method instead of the SPT method are that the CPT is faster and less expensive (2), is not subject to the number of equipment and procedural error sources as is the SPT (3,9), and provides better resolution of overall site characteristics.

The use of the CPT method requires that the CPT field measurements yield identification of material types and that ground water elevations are known or determined through use of a CPT tip equipped with a pore-fluid-pressure sensing element. In addition, CPT data must be correlateable to SPT data either before or after measurements are normalized for the effects of overburden pressure. The applicability of CPT data to these determinations and, therefore, to assessment of liquefaction potential using the simplified blowcount method is analyzed in this paper through use of a set of field measurements performed using both the CPT and SPT at liquefaction susceptible sites.

FIELD INVESTIGATIONS

Field investigations included electric CPT measurements together with SPT measurements where both a standard rope-around-the-cathead  "donut” hammer and a mechanical "free-fall” trip donut hammer were used. Eight CPT soundings and four standard hammer and four trip hammer borings were performed at each of two sites. Soundings were performed prior to SPT borings which were placed at horizontal separations of less than ten feet.

The standard electric friction cone penetrometer described in ASTM D3441-75T (7) was used for all measurements. Data recordings were taken on a two-channel strip-chart recorder having the chart drive locked to the advance of the sounding rods. All penetrations were at 2 cm/sec and instrument inclination was continuously monitored.

The standard donut hammer and trip donut hammer were both used on the same drill rig with the same crew. The standard hammer has a total weight of 225 pounds including anvil, guide, and coupling, while the trip hammer has a total weight of 210 pounds. Drop height was controlled by the driller with the standard hammer but was pre-set for the trip hammer. Rotary wash drilling was used with a baffled fish-tail bit for all borings. Blows were applied at a rate of between 30 to 40 per minute for both hammers. All procedures and equipment followed ASTM D-1586 (1) except that liner samplers were used without liners as is common practice (8).

DATA ANALYSES

Data  obtained  during  field  investigations  and  from subsequent laboratory soil classifications (ASTM D421-58; D422-63) were used to evaluate the adequacy of CPT measurements for determination of soil type and prediction of SPT measurements. Although correlations between CPT and SPT were made for soils ranging from sands to clays, it  is assumed that the SPT method (16) of liquefaction potential assessment is valid only for cohesionless soils with low  fines  content;  no  fines –  content  adjustments (l'7, 19) were used.

FIGURE 1: CPT BEHAVIOR TYPE CLASSIFICATION CHART

The assessment of soil types using electric CPT measurements has been fully described elsewhere (2). A summary of the assessment method is given in Figure 1, where a chart defining boundaries between soil-behavior types determined from CPT data is shown. During initial computer processing of the CPT logs the data is checked against those soil-behavior type boundaries and then tabulated. Likewise, prior to calculation of liquefaction potential, expressed  as Factors of Safety, the data is scanned for material susceptibility, which is assumed as low-fines  content  cohesionless  sands  represented  by materials falling  in Zone 1 of Figure 1. Higher-fines content sands and cohesionless silts are represented by Zone 2, and cohesive soils by Zones 3 and 4.

Several methods were employed to investigate the effect of increasing overburden stress upon CPT data and CPT-SPT correlations (5). Although  trends  were  found  relating the CPT overburden-pressure normalization factor, Cp, and the SPT factor, C_N, the lack of CPT data showing the effect of overburden alone, and the difference in C_N and Cp  relations proposed by various authors (4, 6, 10, 13, l5) prevents assignment of a high degree of certainty to any one relation. Therefore, CPT-SPT correlations are made prior to  overburden stress  adjustments. The predicted SPT values  are  then  adjusted  using  the  relation suggested by Seed (15):

where

C_N = 1-1.25 log r_v'  ……….. (1)

where r_v' is given in tsf.

The conversion of electric CPT measurements into equivalent SPT  measurements  has  been  investigated  by  several researchers (5, 7, 8). The method described herein builds upon the results of those investigations from which it appears that the most rational approach is the detailed modeling of SPT measurements in terms of components of resistance; these components can be estimated from CPT measurements. Schmertmann (11, 12) provides detailed information regarding the entire SPT process, from energy contained in the falling hammer through to the energy used in static or dynamic advance of the SPT sampler. The CPT-SPT conversions presented herein draw heavily upon that information.

 

FIGURE 2: ELEMENTS OF ENERGY DISSIPATION MODEL

The geometries of the SPT sampler and the electric friction-cone penetrometer are shown in Figure 2. The stresses q_ci and f_sj developed on the cone penetrometer tip and sleeve at each depth increment d_i and d_j during continuous penetration are imposed upon the SPT sampler end and outside and inside areas at the same depth increments, yielding quasi-static penetration forces. The sampler is then considered held stationary while the soil moves past it, and the forces acting at each location on the sampler are converted, by multiplication of distance moved, to energy expenditures required for that quasistatic penetration of the sampler.

Adjustments need be made to those CPT-based energies to account for the difference between static and dynamic soil resisting stresses, the influence of a borehole and initial six inches of SPT sampler embedment, the difference between static and dynamic energy expenditures during penetration, and the difference between total energy available for sampling and energy provided by the theoretical thirty-inch free fall of a 140 pound SPT hammer. Finally, the potential energy lost by the SPT hammer-rod system during the sampling interval needs be considered. The concept is summarized as forcing equivalence between available energy, E_i, and dissipated energy, E_d.

where E is energy; W_h and W_hr are buoyant weights of hammer and hammer-rod system; q is the fraction of theoretical hammer energy actually delivered to the sampler; d_h, d_i, d_i and d_j, are distances of hammer fall, (30 inches)  sampler total  penetration  (12  inches),  sampler incremental tip penetration, and incremental penetration of each area of outside sleeve; q_ci, f_si, and f_sj are incremental stresses measured by cone tip and sleeve and applied to SPT sampler tip, outside sleeve, and inside of cutting shoe; A_e, A_sj, and A_s are areas of SPT sampler projected end area, outside sleeve element, and inside of cutting shoe; C₁, C₂, C₃ are conversions between cone stresses and SPT sampler stresses on tip, outside sleeve, and inside cutting shoe, and C_i is an adjustment for the difference between static- and dynamic-penetration energy losses resulting from rate, damping, and energy transfer differences between the CPT and SPT.

The adjustment factors, C₁, C_2, C₃, C_i, have been selected based upon limited research data. Schmertmann (11) suggests that C₁, and C₂, the constants relating quasi-static CPT and SPT end bearing and local side friction stresses, are equal to 1.0. For the studies presented herein, C₁ is weighted by decreasing percentages of end bearing measured beyond the current elevation of the tip to account for the dependence of end bearing on soil resistances ahead of the tip (18). The present use of C₂ is also somewhat modified to account for the observed rapid increase in CPT sleeve friction stresses immediately beneath a soil surface. Thus the sleeve stress applied to the SPT sampler side is decreased beneath the value measured with the CPT in the zone immediately beneath the boring. This term also accounts for the indeterminate loss of side friction at borehole bottom due to disturbance during the SPT boring. C₃ is an adjustment of the measured CPT side friction to account for the difference between soil-sampler friction stresses on the inside of the cutting shoe as compared to the outside. Without any positive information available defining C_3, it has been assumed at 1.0. In addition, the inside friction force has been limited to the end-bearing force of a soil-plugged sampler, although no adjustment of end bearing stress is made to account for the increased area of the plugged sampler. C_i is an adjustment to account for the difference between static- and dynamic-penetration energy dissipation, and has been directly linked, based upon the static to dynamic penetration ratios summarized by Schmertmann (11, 12), to the ratio of friction to end-bearing resistance contributions during equivalent quasi-static SPT sampler penetration. Finally, assuming a constant g for each hammer, the ratio of calculated energy expenditures to theoretically available energy should equal the energy transfer efficiency, g.

Plots showing measured N value against calculated loss of energy (including the potential energy of the hammer-rod system) and theoretical maximum available energy during the SPT sampler penetration were prepared for each boring-sounding comparison. Typical plots are shown in Figures 3 and 4 for two sets of standard and trip hammer borings where one set is from each of two different sites.

FIGURE 3: CPT ENERGY DISSIPATION (FT-TON) VS SPT N VALUES

Examination of these plots reveals nonlinearity between calculated maximum available energy and calculated dissipated energy, even for the essentially constant material type represented in Figure 3. As g is assumed independent of depth, material type, and blowcount (12), and because the transfer ratio between hammer kinetic energy and incident rod energy is constant for the trip hammer, then Figures 3 and 4 show that with increasing blowcount more energy is being lost than is accounted for. A possible cause of this behavior, in addition to the falseness of any of the previous assumptions is that as the site penetration resistance increases, the ratio of average dynamic to average static penetration forces decreases, resulting in a higher  "quake” threshold to be overcome by each pulse of the SPT rod energy train. Thus fewer pulses produce "set" of the sampler, resulting in the decrease in time to a given percentage of total set as blowcount increases (12). This reduction in number of pulses containing sufficient force to induce sampler penetration results in the apparent loss of calculated dissipated energy with increasing blowcount. It also appears that the deficit in energy dissipated during sampling increases with the decreasing ratio of energy transmitted into rods divided by site penetration resistance.

FIGURE 4: CPT ENERGY DISSIPATION (FT-TON) VS SPT N VALUES

The estimation of some energy efficiency factor becomes complicated by this phenomenon. However, the data in Figures 3 and 4 reveal very similar behavior for the trip hammer measurements at the two sites. Thus comparing at equal depths the slopes of the best-fit relations for the different hammers should yield a ratio reflecting the ratio of the two hammer efficiencies. Once the controlled trip hammer energy is actually measured, then the energy dissipation relation for any particular hammer can be scaled to any desired ratio of the calibrated energy efficiency. For example, a comparison of the standard and trip hammer energy ratios at one location reveals a ratio of hammer efficiencies of 0.48 at one site and 0.72 at the other. If it was known that the trip hammer efficiency was 70 percent at that location, then the standard hammer efficiencies would be calculated as 33.6 and 50.4 percent, respectively. Considering the number of variables that can affect energy delivered by the standard hammer (3), such a variance is not surprising, and is supported by the average blowcount ratios determined for the two sites (5).

Regardless of actual hammer energy, the CPT method may be used to calculate blowcounts representative of any particular SPT hammer through hammer – specific calibration. For example, a comparison of a predicted continuous versus measured N value profile is given for one boring-sounding combination in Figure 5. Although the predicted SPT values are shown continuous, the value at any depth rep-resents the equivalent SPT blowcount for a depth interval starting 12 inches above that depth. In Figures 6 and 7, the range of predicted and measured blowcounts for each hammer at each site are shown. Several features are of interest in these figures. First, the range of predicted blowcounts agrees well with the range of measured blowcounts at any depth. Next, the difference in blow-counts predicted, or measured, for the same site but using different hammer energies is very significant. For corrected blowcounts less than about 35, the Factor of Safety against liquefaction calculated using the simplified method (16) changes in direct proportion to a change in measured blowcount. Thus the routine investigation using blowcounts taken at five foot intervals can provide very misleading results, depending primarily upon variability of the profile. In addition, examination of the ratio of friction to end bearing components of the SPT blowcount reveals that in all but clean sands, the friction component dominates the total resistance. As has been shown (3), borehole disturbance can cause significant reduction of effective stresses at the borehole bottom, thus greatly reducing the measured blowcount. It appears then, that any but the most careful of field operations may produce a wide range in measured blowcounts.

The calculation of liquefaction potentials from CPT data is easily accomplished once the CPT data are converted to equivalent SPT blowcounts. The equivalent SPT blowcounts are first normalized for overburden effects using equation (1). The normalized equivalent blowcounts, N_1' are then converted to cyclic strengths and induced stresses are calculated following the procedures of the simplified method (16). Dividing the cyclic strength ratio by the induced stress ratio yields a factor of safety against liquefaction. This procedure is followed for each CPT sounding and the vertical profiles of factors of safety versus depth are compiled into horizontal cross sections of the site. Examination of the cross sections reveals if there is continuity of susceptible materials with low factors of safety. Statistical processing of any continuous vertical and horizontal zones of weak and susceptible materials can be performed to allow calculation of average factors of safety. In addition, zones of questionable susceptibility are readily identified from the cross sections. These zones can then be sampled for subsequent laboratory testing.

In summary, as the range comparisons of Figures 7 and 8 show, although a very extensive and carefully controlled sampling program can define site blowcount variability, the continuous CPT profiles give a more complete description. The increased description of site variability coupled with the greater degree of measurement repeatability afforded by the CPT method as compared with the SPT method allows increased confidence in the results of some subsequent deterministic or probabilistic evaluation of site liquefaction potential.

CONCLUSIONS

The following conclusions are drawn based upon the investigation results:

1.      The blowcounts predicted using CPT measurements agree well with measured blowcounts, indicating that both test methods are similarly influenced by the same soil characteristics and that CPT measurement may therefore be directly used in SPT type liquefaction potential assessments.

2.      The blowcount profile predicted from CPT measurements provides  much  clearer  resolution  of  stratigraphy  in terms of both material type and material penetration resistance than is normally obtained with SPT results;

3.      The blowcounts  determined  from  CPT  measurements  may be  more  reliable  than those  actually  measured  using uncalibrated equipment or poorly controlled procedures during the SPT;

4.      Differences  between  any  one  CPT-predicted  and  SPT-measured set  of  blowcounts  are less  than potential differences in blowcounts themselves due to  changes of  equipment,  procedure,  and,  in  certain  cases, sampling interval;

5.      Blowcounts predicted from CPT measurements can be adjusted, once the CPT is calibrated against actual blowcount hammer energy measurements, to account for the dependence of any particular SPT-based design procedure upon a specific SPT-sampler energy.

ACKNOWLEDGEMENTS

Most of the field investigations referenced in this paper were performed under USGS Contract No. 14-08-0001-17790. In addition, data analyses were supported by Ertec Research and Development grants. Finally, the interpretation of the volumes of field data was facilitated by the conscientious efforts of Messrs. George Edmonds, Mike Moore, Mike Leue, and Richard Miller.

APPENDIX I. – REFERENCES

APPENDIX II – NOTATION

The following symbols are used in this paper.

A_c = cross-section area of CPT tip;

A_e = cross-section area of SPT sampler;

A_s = incremental surface areas of SPT sampler;

C₁ = factor relating end bearing stresses from CPT to SPT;

C_2,3 = factors relating friction stresses from CPT to SPT;

C_i = ratio of dynamic to static penetration energy dissipation;

C_N = SPT overburden correction factor;

d_h = distance of SPT hammer fall;

d_I,j = depth of SPT sampler penetration;

d_p = total SPT sampler penetration depth;

E_D = energy dissipated during SPT sampler penetration;

E_i = energy available for SPT sampler penetration;

f_s = CPT and SPT quasi-static end bearing stresses;

g = SPT hammer energy transfer efficiency;

r_v' = effective vertical overburden stress