How Soils Can Help Predict High Rebound of Prestressed Concrete Displacement Piles

Excessive pile rebound has been occurring for over a decade, when high displacement prestressed concrete piles are driven mostly with single acting diesel hammers in Florida’s low permeability very fine sand blends. This very complex engineering phenomenon, which occurs due to interactions between the soil, pile, and hammer; typically is observed at depths great than approximately 15 m (50 feet). Rebound greatly diminishes the end bearing capacities, causing various problems and significant cost increases. Findings from several research studies have produced some obvious trends based on grain size, Atterberg limits and cyclic triaxial testing. Excessive rebound was categorized as movements exceeding 12.5 mm (0.5-inches) while no-rebound was considered to be less than 6 mm (0.25-inches). Coarse grained rebound soils were very fine sands, passing the number 100 sieve, with silts and clays. According to the Unified Soils Classification System they classified as SM and have silt contents between about 20 and 40%. Cohesive rebound soils were highly plastic clays, classified as CH and also have silt contents between about 20 and 40%. Cyclic triaxial testing indicated that rebound soils are much more resilient than no-rebound soils, requiring many more cycles to produce strains of 2.5, 5, 10 and 15 percent. Historical Overview of Pile Rebound Historically, pile driving rebound has been referred to differently by various authors. Some call it pile “bounce” (Murrell et al., 2008), others call it “high pile rebound” (HPR) (Hussein, 2006; Cosentino et al., 2010, and 2016). Still others (Authier and Fellenius, 1980; Likins, 1983) use the terminology “high quake” (i.e., the limit or end of the elastic pile movement during a single hammer blow) as part of its description. Hussien (2006) notes that HPR increases as driving depths increase and that HPR is a function of the soils dynamic response, which is described terms of damping. During an evaluation of a 762 mm (30 -inch) instrumented PCP test pile along Florida’s State Road 528 in Brevard County, he documented a nearly 30 percent decrease in static capacity (5400 to 3900 kN) during 41 single acting air hammer blows and a corresponding 45 percent increase in rebound (9 to 13 mm). The instrumentation included strain gages and accelerometers mounted near the top of the pile. This decrease in capacity occurred even though the driving blows per 25 mm (1-inch) increased from 12 to 17 or 42 percent. The author produced excellent load versus deflection data from both static load testing and corresponding dynamic testing. Low displacement piles may rebound, but their rebound could mostly be the result of elastic compression. Elastic compression (δ) is defined as the load (P) times the pile length (L) divided by the pile area (A) and elastic modulus (E) or δ = PL/AE. When A is small, as is the case for low displacement piles, δ increases and may be a major component of the rebound (Nguyen et al., 2019). Murrell et al. (2008) found installing 500 mm (20-inch) square PCP displacement piles with single acting hammers, 1 The Edward H. Kalajian Professor, Florida Institute of Technology, Civil Engineering, 150 West University Boulevard, Melbourne, FL 32901-6975, USA. 2 State Geotechnical Materials Engineer, Florida Department of Transportation, State Materials Office, 5007 N.E. 39th Ave, Gainesville, FL 32609, USA. * Corresponding author, email: cosentin@fit.edu © 2020 Deep Foundations Institute, Print ISSN: 1937-5247 Online ISSN: 1937-5255 Published by Deep Foundations Institute Received 17 June 2019; received in revised form 2 July 2020; accepted 4 November 2020 https://doi.org/10.37308/DFIJnl.20190617.206 Introduction Florida’s complex alluvial sandy deposits have caused engineers, pile driving headaches for quite some time, no pun intended (Hussein, 2006). Typically, large diameter prestressed concrete piles (PCP’s) driven with open ended single acting diesel hammers, may rebound up to 75 mm (3-inches) once they reach depths greater than about 15.25 m (50 feet). The soils producing this rebound are thick deposits of very finesands, with silts and clays (Cosentino et al., 2010). Cone penetrometer testing with pore pressure (CPTu or Piezocone) measurements have produced both negative and positive pore water pressures ranging from – 1 to as high as 35 atmospheres’ (Dekhn et al., 2019, Cosentino et al., 2016). This phenomenon can produce high tensile stresses in PCP’s, may cause the concrete to fail in tension, lead to lengthy production delays and ultimately cause the engineers to question the pile capacity. vol14no2cosentino206.indd 1 27/01/21 4:57 PM © Deep Foundations Institute Cosentino, Horhota | How Soils Can Help Predict High Rebound of Prestressed Concrete Displacement Piles in saturated marine deposits with CPTu pore pressures above 20 atmospheres would produce bounce. They measured CPTu pore pressures as high as 37 atmospheres at the depths where the piles bounced. Significant soil quake, and consequently rebound, was observed by Regan and Higgins (2009) during driving with a single acting diesel hammer of square 355 mm (14-inch) PCP displacement piles that were designed to support the National Harbor Hotel in Washington, DC. Pile breakage and decreased hammer energy loss occurred during installation. The soils at this location were very dense sand interbedded with layers of highly variable plastic overconsolidated clays. The authors related pile rebound to a combination of the soil’s degree of overconsolidation and to the increase of excess pore water pressures during driving. Stevens (2012) evaluated 500 mm (20-inch) square PCP instrumented displacement piles driven with a Delmag D4623 single acting diesel hammer in Escambia Bay, Pensacola, Florida. These saturated soils consisted of green gray very soft to soft clay to a depth of 40 ft, underlain by medium dense to dense fine to medium sand to a depth of 78 ft, followed by dense to very dense silty fine sand. Pile driving was harder than anticipated, however, subsequent static load tests indicated lower capacities than expected. The sand grains were evaluated and found to have angular to subangular shapes. Large rebounds were observed during the installation of 11 indicator piles. CAPWAP analyses also indicated that the sands produced both large soil quakes and high damping values. The Florida Department of Transportation (FDOT) currently includes an acceptance specification which limits pile rebound to 6 mm (0.25-inches) and practical refusal as 20 blows per 25 mm (1-inch) (FDOT, 2019). Pile end bearing capacities require relatively large movements, or quakes, to be completely mobilized, therefore, any rebound jeopardizes this component of the piles’ total capacity. These movements can be 0.1 to 0.2 inches or 10 to 20 blows per 25 mm (1-inch), which are blow counts near or at refusal. When rebound occurs, these mobilization numbers increase by an order of magnitude, meaning that 25 to 50 mm (1 to 2 inches) may be required to fully mobilize end bearing. These movements are also dependent on numerous factors, including pile and soil type, soil engineering behavior, type of hammer used and hammer efficiency. Based on static testing, some engineers use Davisson’s (1975) relationship of D/120 to estimate this required mobilization movement, where D is the pile diameter. For example, using this criterion a 600 mm (24-inch) square PCP would need 5 mm (0.2-inches) of displacement to mobilize its end bearing capacity. This preceding information indicates that rebound is produced by a complex blend of hammer and pile type, occurring in saturated fine sandy soils with clays. During CPTu testing high pore water pressures occurred in the rebound zone. All piles were PCP displacement piles, driven with single acting diesel or air hammers in highly varied coastal type geologic deposits. Large quakes and large rebounds were observed, while driving became difficult yet pile capacities decreased. 2 | DF I JOURNAL Approach The key soil behavior results from a comprehensive research project funded by FDOT are being summarized (Cosentino et al., 2016). Undisturbed thin walled tube samples were obtained in accordance with ASTM D1587 (2015) from 26 locations at five sites throughout central Florida and the Florida panhandle. Each sampling location was within close proximity to a corresponding FDOT specified test pile, instrumented with pile driving analyzer (PDA) strain gauges and accelerometers. Rebound was determined by subtracting the maximum PDA displacement (i.e., DMX) from the set determined using the inspectors pile driving log, which was calculated using reciprocal of the blows per foot (i.e. set = 1/blows per foot). Pile rebound was separated into high rebound, if more than 12.5 mm (0.5 inches) occurred after each hammer blow, and no-rebound if less than 6 mm (0.25-inches) of rebound per blow occurred (Omar, 2015; Franqui, 2019). When 6 mm (0.25-inches) was used for high rebound, the resulting trends were not clear (Omar, 2015). All piles were 60 cm (24-inch) square prestressed concrete piles (PCP’s) and driven using single acting diesel hammers. Testing and Results To develop correlations and trends with rebound, basic geotechnical index properties were determined from the undisturbed samples, including grain size distribution, Atterberg limits, unit weight, and permeability. Correlations were then developed between rebound and these index properties. Correlations were also developed from a combination of consolidated drained triaxial tests (ASTM D7181, 2011) followed by cyclic triaxial (CT) stress level testing (Omar, 2016). Grain Size Results Table 1 is a summary of the sampling data for the coarsegrained soils tested at the 26 testing locations, where 15 locations had piles experience high rebound and 11 locations had piles experience no-rebound, according to the 6 mm (1⁄4-inch) criterion. The data were taken from 12 different lo