Discussion of “Load tests on full-scale bored pile groups”

Abstract

The authors are correct in stating for static loading tests that “there are few full-scale ... pile group load tests reported in the literature.” I agree. However, a few more references are available than those listed by the authors, e.g., O'Neill et al. (1982a, 1982b), Phung (1993), and O'Neill and Reese (1999). Figure D1 shows the load–movement measured by Phung (1993) in comparing the response of a single pile to a group of five piles driven at a centerto-center spacing (c/c) of 5.7 pile diameters in a fine sand. The center pile (pile #1) was installed and tested as a single pile before the other piles were installed and connected by a rigid cap. The load–movement response of the five piles was different, but the difference was limited to the development during the initial loading. Beyond the first about 4 mm of movement, the load– movement curves were essentially parallel. Measurements of load distribution showed that the difference was mostly due to the difference in shaft resistance — the toe resistances were essentially equal for the five piles — and no correlation to location within the group could be discerned. The differences are considered caused by unsystematic compaction of the sand with no apparent effect of the driving sequence or other driving effect. A main observation was that the response in the loading of the center pile as a part of the group in effect was a reloading of the pile. The response at first loading of the pile as a single pile was considerably less stiff. Similar to the authors' work, the references mentioned above involve pile groups that at most consist of nine piles. A group of just a few piles — and nine piles is a very small number, where pile groups are concerned — will show minimal interaction and variation between the piles in supporting a structure. For a small pile group, the difference in load response between the piles in a piled foundation subjected to working load from the supported structure will be more affected by load variations, such as load center, load inclinations, and lateral loads, as opposed to when subjected to a static loading test on the group. Very few well-documented case histories are available in the literature with regard to full-scale studies of the performance of large pile groups under working load. However, a few are; for example, Golder and Osler (1968), Badellas et al. (1988), Goossens and Van Impe (1991), and Savvaidis (2003). The case histories show beyond doubt that the capacity and the load distribution of an individual pile in a large group of piles is of little relevance to the response of the piled foundation. Instead, the response of a piled foundation made up of a good-size pile group constitutes a settlement problem, and the capacity and load distribution of either an individual pile or the group is not the governing issue for a design. Despite that the authors (as do so many others) imply that the static loading test measures pile settlement, what is measured in a loading test is amovement response to a series of applied loads, not settlement. The authors' paper presents the movement response to applied load for a single pile and a few very small pile groups, not the settlement. Of course, settlement assessment relies verymuch on the results of a static loading test; in particular, on the response of the pile toe. However, the actual settlement of a piled foundation due to a working load, whether composed of a single pile, a few piles, or a large group of piles, is a very different issue. The authors present the load–movement response of two single piles and state the capacity criterion that the capacity of the piles is based on the “traditional 10% relative settlement criterion”. Although the criterion is used less often these days, it does keep appearing in the literature. A couple of years ago, I searched an assortment of successively older papers, textbooks, and standards that essentially stated the same criterion — sometimes with a slight modification away from the 10% value. Many did not give reference to the source, but some did, and I found the original source. The criterion has its origin in a mistaken quotation of a now 70 year old statement by Terzaghi (1942). Terzaghiwrote: “the failure load is not reached unless the penetration of the pile is at least equal to 10% of the diameter at the tip (toe) of the pile.” (For full quotation and context, see Likins et al. 2011). Note, Terzaghi did not define the capacity as the load generating a movement equal to 10% of the pile diameter, he emphatically stated that whatever definition of capacity or ultimate resistance used, it must not be applied until the pile toe has moved at least a distance corresponding to 10% of the pile toe diameter. (The pile head will then have moved an additional distance equal to the pile shortening.) Most certainly, Terzaghi did not suggest that a fixed movement value, however determined, could serve as a definition of capacity. Figure D2 shows the load–movement plot of the authors' static loading test on pile DZ1L. The usually very conservative definition called “offset limit”, or Davisson limit, indicates a lower-bound value of 1400 kN. It is here offered for reference. I do not suggest that the offset limit would be the pile capacity, but it does show the load for which the ultimate shaft resistance would have been reached. The Hansen 80-percent method results in an interpreted capacity of 1700 kN, coincidentally the maximum load applied in the test — the pile seems to be plunging. The Chin–Kondner and Decourt extrapolationmethods indicate 1850 kN. Thus, coincidentally, the 1540 kN capacity per the authors' “traditional” definition happens to be a reasonable value to choose from the load– movement curve. For full definitions and description of the methods for determining the capacity from the pile-head load– movement response, see Fellenius (1975, 2012). Figure D3 shows the authors' load distributions as evaluated from the strain-gage measurements in pile DZ1L up to a load of 1440 kN. The authors did not include the distribution for the maximum applied load (1700 kN). I have supplemented the figure with the authors' qc diagram from the sounding pushed nearest the test pile, soil descriptions, and layer boundaries. I have also added a distribution determined bymeans of both total stress and