Research Activities

Full-Scale Pile Group Lateral Load Testing in Soft Clay

by K. Rollins, K. Peterson and T. Weaver

This article presents research resulting from NCEER's Highway project. This project was jointly funded by NCEER, the Utah Dept. of Transportation, the Federal Aviation Administration, and the Federal Highway Administration. More information about this study is available in a forthcoming technical report which will be distributed by NCEER and the Utah Dept. of Transportation early next year. Comments and questions should be directed to Kyle Rollins, Brigham Young University, at (801) 378-6334; or e-mail:

The lateral load capacity of pile groups is critically important in the design of structures subjected to earthquakes. Although fairly reliable methods have been developed for predicting the lateral capacity of a single pile, there is very little information to guide engineers in the design of closely spaced pile groups (spacings less than six pile diameters). In addition, very little field data is available regarding the behavior of pile groups under short period dynamic loads which would be produced by an earthquake.

To improve our understanding of pile group behavior, a series of full-scale lateral load tests has been carried out on a 9-pile group (3 piles in 3 rows) driven in soft clay at the Salt Lake International Airport. The objectives of the test program were to:

Pile Group Configuration

The test piles are 32.4 cm O.D. steel pipe piles (wall thickness = 9.5 mm) with a concrete in-filling which are spaced at about 90 cm on center. The piles were driven to a depth of 9.1 m which was sufficient to provide fixity. Full-scale instrumented lateral load tests have been performed for four different conditions as outlined below:

  1. Static, free-head loading condition on a single pile with lateral load applied with a 4.4 kN hydraulic jack.
  2. Static, free-head loading condition on a nine-pile group with lateral load applied with a 1.33 MN hydraulic jack.
  3. Dynamic, free-head loading condition on a nine-pile group with load applied by the 14.4 MN Statnamic device at 180?to the static loading.
  4. Dynamic, fixed-head loading condition with load applied to a concrete pile cap on nine-pile group with the 14.4 MN Statnamic device at 90?to the static loading. This testing was performed with and without a compacted gravel backfill behind the pile cap.

Because of space constraints, only results from the free-head load testing will be presented in this article. Other results and recommendations will be included in a forthcoming technical report.

Geotechnical Conditions at the Test Site

Due to the complex pile-soil-pile interaction anticipated in this series of tests, a comprehensive geotechnical investigation was carried out to accurately characterize the soil properties at the site. This investigation included in-situ testing as well as conventional sampling and laboratory testing. Laboratory testing included hydrometer and mechanical analysis tests for grain-size determination, Atterberg limit tests, natural moisture content tests, U-U triaxial shear tests and consolidation tests. In-situ testing included standard penetration (SPT) testing, cone penetration (CPT) testing, dilatometer (DMT) testing, pressuremeter (PMT) testing, and vane shear testing (VST). Since lateral loads on piles only stress a zone 10 to 15 pile diameters deep, the soil characteristics in the upper 5 m of the soil profile are of most importance. The soil profile in this depth range consists of a soft to medium cohesive layer 3 m thick overlying a medium dense to dense sand. The cohesive surface zone consists of layers of low-plasticity silt, clay and sandy silt which are typical of near surface deposits in the Salt Lake Valley.

Static Free-Head Tests

For the static free-head tests, load was applied to a frame and each pile was connected to the frame using a tie-rod with a pinned connection. The frame ensured that each pile underwent about the same displacement, however the load carried by each pile could be different. Load on each pile was measured using strain gauges on each tie rod. In addition, bending moment and displacement were measured as a function of depth using strain-gauges and inclinometer probe measurements. Load-deflection curves for the single pile test and the pile group test are shown in figure 1. In order to make comparisons possible, the total load on the pile group has been divided by the number of piles to obtain an average pile load. Figure 1 highlights the fact that for the same average pile load, the displacement of the pile group may be 2 to 2.5 times higher than that of the single pile. Figure 2 shows the average pile load versus deflection curves for each row of piles in comparison with the single pile load-deflection curve. These results indicate that the load distribution in the pile group is not uniform but is a function of the row position. For a given displacement, front row piles carry the greatest load while middle row piles carry the lowest load. Back row piles carry loads somewhat higher than the middle row piles but significantly less than the front row piles.

The ratio between average load carried by a pile in each row and the load carried by a single pile is shown for the three rows in figure 3. While the ratios are somewhat dependent on deflections, they are relatively constant for deflections greater than about 10 mm. In relation to the single pile load, the front, middle and rear piles typically only carry about 80%, 50% and 60%, respectively, of the single pile load.

Analysis of Static Free-Head Pile Test Results

Lateral pile response is typically analyzed using finite-difference models of the pile along with non-linear springs to represent the resistance provided by the soil. The load-displacement curves for the soil are known as p-y curves, where p is the horizontal soil resistance (force per length) and y is the horizontal displacement. Generic p-y curves have been developed for soft clays, stiff clays, and sands and have been widely incorporated in computer models. If the generic soil-type dependent p-y curves are not used, the designer must provide site-specific p-y curves using techniques based on in-situ tests such as the pressuremeter or the dilatometer.

P-Multipliers for Group Effects

Research has generally found that when piles are further than six pile diameters apart, group effects are not significant. For closer spacings, group effects become significant and this group behavior has been modeled using p-multipliers (Brown et al., 1988) to reduce the soil resistance depending on the position of the pile in the group (i.e., leading row vs. trailing row). With this approach, it is possible to reduce the computed load carrying capacity of the piles in a group relative to the single pile capacity as has been observed in field and model tests. The back-calculated p-multipliers from the Salt Lake load tests are 0.7, 0.4 and 0.5 for the front, middle and back row piles, respectively. These p-multipliers are relatively consistent with those obtained from the three other full-scale tests where load distribution has been measured (Meimon et al., 1986; Brown et al., 1987; and Brown et al., 1988) even though the soil properties at these sites vary considerably.

Dynamic Free-Head Test Results

Dynamic load tests were performed using the Statnamic method in which a solid-fuel rocket is used to produce a pulse load. The load was typically applied over a 120 to 150 millisecond time period and produced maximum accelerations between 1 and 2 gs. Three firings were conducted and the average load versus displacement curves are shown along with the static single pile and group curves in figure 4. It may be observed that the dynamic resistance is greater than the static resistance and approaches that produced by the single pile. The increase in the slope of the Statnamic load-displacement curve in comparison with the static group curve is apparently due to soil damping resistance. It is interesting to note, however, that for a maximum average load of 69.3 kN, the maximum displacement was about 25 mm which is almost identical to the displacement obtained under the static loading of the same magnitude. Furthermore, the maximum bending moments for the Statnamic loading case were generally within about 15% of the values measured when the same load was applied statically. Finally, the load distribution among the piles for the Statnamic loading was very similar to that for the static case except that the trailing row load was somewhat higher. These results suggest that the damping resistance under dynamic loading conditions may tend to compensate for any reduction in soil stiffness so that much of the response of the pile group is similar to that for the static loading. Additional analyses are currently underway to separate out the relative contribution of damping to the measured lateral resistance and to evaluate the ability of existing computer models to predict the measured response.


Brown, D.A., Morrison, C., and Reese, L.C., (1988), "Lateral Load Behavior of a Pile Group in Sand," Journal of Geotechnical Engineering, ASCE, Vol. 114, No. GT11, pp. 1261-1276.

Brown, D.A., Reese, L.C., and O'Neill, M.W., (1987), "Behavior of a Large Scale Pile Group Subjected to Cyclic Lateral Loading," Journal of Geotechnical Engineering, ASCE, Vol. 113, No. GT11, pp. 1326-1343.

Meimon, Y., Baguelin, F., and Jezequel, J.F., (1986), "Pile Group Behaviour Under Long Time Lateral Monotonic and Cyclic Loading," Proceedings, Third International Conference on Numerical Methods in Offshore Piling, Inst. Francais Du Petrole, Nantes, pp. 286-302.

Click here to return to the Table of Contents

| Home | NCEER | Information Assistance | Publications | Databases, Software |
| Events | New! | Links | Data Resources |