Problematic Soils and Rocks and In Situ Characterization

This study uses lime and Class C fly ash, an industrial byproduct of electric power production produced from burning lignite and subbituminous coal, to study the plasticity reduction in highly expansive natural clays from Idabel, Oklahoma. This study is important, especially in Oklahoma, because most of the native soils are highly expansive and cause seasonal damage to roadways and structures. The addition of lime or fly ash helps to arrest the shrinkage and swelling behavior of soil. Four soil samples with the same AASHTO classification (A-7-6) were used in this study to show shrinkage variability within a soil group with the addition of lime and Class C fly ash. The plasticity reduction in this study was quantified using the linear shrinkage test. It was found that soils classified within the same AASHTO group had varying shrinkage characteristics. It was also found that both lime and fly ash reduced the linear shrinkage, however, the addition of lime reduced the linear shrinkage to a greater degree than the same percentage of Class C fly ash. Even though it takes much less lime than fly ash to reduce the plasticity of a highly expansive soil, it may be less expensive to utilize fly ash, which is a waste product of electric power production. Lime also has a lower unit weight than fly ash so weight percentage results may be misleading.

Numerous pavements constructed over expansive clay subgrades experienced heave distress although they were designed for swell movements which were predicted utilizing varying models. Such heaving results in high pavement roughness, which in turn requires constant remediation. A few of these models utilize Atterberg and soil suction parameters to predict the heave movements. It is possible that these models underpredict soil movements due to several reasons including soil variability, test procedures followed and lack of accounting for soil overburden depth. An on-going research study has focused on experimental investigations by performing 3-D free swell tests on two types of clays; high and low PI clays. Swell and suction tests were conducted on clay specimens compacted at three different compaction moisture contents and these results were used to compare with swell predictions from Potential Vertical Rise (PVR) and suction based models. These results explained the variations in swell predictions due to measurement procedures followed and assumptions used in the prediction models. Methods to enhance these predictions are discussed.

One-dimensional wetting-induced compression behavior of compacted soils is often studied in the laboratory using the single- or double-oedometer method. Samples prepared and tested in the laboratory may exhibit different compression behavior than soil in the field because of differences in soil fabric and differences in the size of the soil mass involved in each case. To examine the importance of these differences, this paper compares results of standard oedometer tests on compacted soil with those from samples tested in a very large oedometer. Large samples were prepared using a method to simulate compaction and fabric developed under field conditions.

The soil-water characteristic curve (SWCC) is a widely used experimental means for assessing fundamental properties of unsaturated soils for a wide range of soil suction values. However, a limited number of studies have focused on soil-water retention properties of natural and chemically treated expansive soils via the SWCC. The present work is motivated by the lack of experimental evidence of this type. The soil-water characteristic curves of an expansive clay from south Arlington, Texas, were measured under both natural and cement treated conditions using pressure plate (suction range of 0–690 kPa) and filter paper (suction range of 690–1,000,000 kPa) techniques. SWCC test results were analyzed using Fredlund and Xing's SWCC model equation. SWCC test and modeling results show a relatively minor, though systematic influence of cement dosage on the soil's entry values and SWCC model parameters.

This paper will present work over recent years, which has aimed to develop a reproducible collapsible loess sample in the laboratory. The idea was to create specimens that not only reproduce the degree of loess collapse found in the field, but did so repeatably, thereby allowing the various bond components and their relative contributions to be assessed. This paper will show how a global benchmark loess sample can be produced, how it behaves under loading and wetting. Inevitably there are differences that occur between a newly deposited sample and one that has existed for many thousands of years, and in passing how collapse of samples from across the world compare.

Internal erosion is a great concern in the construction of water retaining embankments. Erosion of soils can be related to their dispersion characteristics. However, although a number of tests can be used to classify soils in terms of their dispersion, they can give contradictory results. Therefore a modified pinhole test was developed to assess erosion of soils. This internal erosion test was used to evaluate the behaviour of Kaolin clay, where specimens were formed using compaction and consolidation techniques and these were compared with slaking tests. Results showed that was a significant difference between the erodibility of the two sample preparation methods of compaction and consolidation, in both the internal erosion test and slaking test. Furthermore, significant differences were also noted in both tests between specimens prepared using the same technique (compaction and consolidation) but different effort (light vs. heavy and 900 kPa vs. 1500 kPa, respectively).

Various field studies have revealed that sulfate induced heaving is caused by the formation of ettringite due to cement and lime treatments of sulfate rich clay. Variations in swell and shrink properties can lead to premature pavement failures, which can result in higher maintenance costs. This paper presents experimental results obtained from Kaolinite clay treated with different sulfate amounts and stabilized with cement and lime. Several laboratory experiments including swell, shrinkage, strength properties of these clays were conducted. It was observed that ettringite mineral formed in both lime and cement treated sulfate clays. Overall, lime treated Kaolinite clay experienced lesser swell strains and swell pressures and higher volumetric shrinkage than the cement treated Kaolinite clay due to variations in moisture and density conditions of the respective treated clays.

The material commonly known to engineers in Las Vegas, Nevada, as "caliche" results from cementation of sediments, both fine- and coarse-grained, calcium carbonate precipitated in soils where evaporation exceeds precipitation. For engineering purposes, caliche is defined as a rock-like material that occurs in soil deposits erratically in thickness, hardness, and lateral extent and it is therefore difficult to predict in terms of interference with below-grade construction. The greatest problem caliche creates is in its removal, which results in expense, construction delay, and potential vibration damage to adjacent improvements caused by some excavation methods. The presence of a substantial thickness of continuous caliche below a building site can provide an excellent bearing surface for conventional foundations which can support high loads where the alternative might be a more expensive foundation system. Caliche can cause damage to structures only in an indirect way such as differential settlement caused by foundations partially supported on caliche and partially on compressible soil. Seismic methods can be used to detect caliche nonintrusively. Delineation of thickness and lateral extent are more challenging.

Within the geotechnical and geological communities it is common knowledge that karstic limestone is highly variable however this variability is often not well defined or quantified. To assess the variability in karstic Ocala Limestone, an investigation was conducted at the shared University of North Florida and University of Florida karstic limestone geophysical/ground truthing test site in central Florida. The variability was assessed via delineating the limestone bedrock surface and through laboratory measurements made on core specimens taken from the site. The bedrock surface was identified using surface geological mapping, intrusive and non-intrusive (seismic refraction) techniques. The core specimens were characterized using compression wave (p-wave) velocity measurements. Analyses of the field data revealed a highly erratic limestone bedrock surface; the bedrock surface decreased in elevation from ground level to nine meters in depth within a horizontal distance of 24 meters. Additionally, seismic refraction tomography data revealed marked differences in the compression wave velocities at the top of the bedrock surface at various locations along one of the survey lines; wave velocities were highest within slots or valleys and lowest at the tops of blocks or pinnacles. Laboratory results also showed high variability; there were no clear velocity trends with depth within a single bore hole or at constant depths across the survey line. However, the geological history of the limestone provides explanations for the variability for both the limestone bedrock surface and compression wave velocity measurements.

The near surface bedrock geology of North Texas, USA consists of a number of weak shale formations or significant shale units within the various formations. These shales, though each with their own unique characteristics, present certain common challenges to the successful design and construction of engineered facilities. Lack of wide agreement within the local geotechnical engineering community pertaining to various engineering analyses for these materials is apparent in the number of failures and incidents of structure distress documented in the area, particularly slope failures and distress due to the expansive soils derived from these shales. This paper identifies the major North Texas formations of concern, with particular emphasis on the Eagle Ford Formation; explores some of the geologic and topographic aspects to be considered; and examines some of the geotechnical parameters of these materials and some of the controversy surrounding parameter value selection in various analyses.

Macroporous rock, such as lithophysal tuff and vesicular basalt, has long exasperated the engineering community due to its inability to be neatly characterized. Such rock can have extreme variability in the size, shape, distribution, and density of the voids which in turn causes great deviation in its engineering properties. Given the variability of the void size and shape, the proper specimen size can be difficult to distinguish. ASTM standards dictate the specimen diameter must be three times the nominal maximum size of voids or aggregate, calculating the nominal maximum size of macropores is not feasible. To simplify the characterization of the impact of specimen size, analog plaster models containing Styrofoam spheres of one-inch diameter were tested. Specimens are right cylinders with diameters of 5.08 cm, 7.62 cm, and 10.16 cm with 2:1 aspect ratios. All three specimen sizes were able to effectively capture trends in unconfined compressive strength and Young's modulus over a range of porosities, though the regression types differ. Only the 7.62 cm and 10.16 cm diameter specimens produced valid trends in the Young's modulus at various porosities, although as mentioned, the regressions are different.

Design of new transit stations in urban settings typically features underground caverns intersected by surface penetrations and utility tunnels. Moreover, system demands as well as related construction, operation, and maintenance costs often favor a shallow vertical alignment. Many factors, such as opening size, intersection geometry, excavation sequence, two-dimensional (2-D) loading of the individual openings, rock mass classification (Barton's reduced Q at intersections), and rock mass discontinuity orientation with respect to the intersection; impact the evaluation of the three-dimensional (3-D) problem of underground intersections under shallow rock cover. This paper will discuss the background and factors influencing the behavior of cavern-ancillary tunnel connections with low cover. A rational approach, which builds upon the typical 2-D analysis by considering these factors in order to evaluate the 3-D problem, will also be presented.

The Sowers dynamic cone penetrometer (DCP) has been used as a relatively low-cost near-surface soil investigation tool in the Southeastern United States for over 40 years. Typically, the Sowers DCP penetration resistance has been correlated to standard penetration test (SPT) N values for use in geotechnical analyses such as bearing capacity evaluation. Published correlations of DCP and SPT penetration resistances are limited to Piedmont residual soils in and around Atlanta, Georgia and soils in Idaho, Utah, and Wyoming. Correlations of Sowers DCP test results to piezocone penetration test (CPTu) and flat blade dilatometer test (DMT) measurements have not been examined. The following paper examines correlations between the Sowers DCP penetration resistance (n), corrected tip resistance (q_t) from piezocone penetration testing, and dilatometer modulus (E_D) from flat blade dilatometer testing in lower Atlantic Coastal Plain soils. Testing data were collected from eight sites in the Charleston, SC area. Correlations were conducted taking into account soil classification and groundwater location. Analysis of the data showed no clear correlations exist between these three test measurements with the exception of a linear relationship between n and q_t for silty sands below the groundwater table, where a relationship of q_t,ave(MPa)/n approx. 0.5 was observed regardless of classification method.

Since Ranzini suggested supplementing the SPT test with measurement of the torque required to turn the split spoon sampler after driving, many Brazilian engineers have been using this in the design of pile foundations. This paper presents a study of the rod length influence in the torque measurement. A theoretical study of material resistance considering torsion and bending in a thin wall tubular steel shaft was performed. It makes possible to conclude that the shearing tension caused by the proper weight represents less than 1% of the shearing tension caused by the turning moment. In addition, an experimental study was done with electric torquemeters fixed in a horizontal rod system. The tests were being carried out to analyze rods of one meter to twenty meters in length and the measurements were collected at the ends of each rod length verifying the efficiency data. As a result, it is possible to verify that the torque difference through rod length is lower than minimum scales of mechanical torquemeters that are used on practical engineering. Also a fact to be considered is a big torque loss for values under 20 N.m of applied torque. This way, the SPT-T is not adequate to low consistency soil.

The purpose of this paper is to evaluate the capability of existing interpretation methods and develop new methods to reasonably predict the overconsolidation ratio (OCR) of clayey soils utilizing the piezocone penetration test (PCPT) data. These existing methods include the Schmertmann method, Kulhawy and Mayne method, and the Chen and Mayne methods using pore pressure measurements either at the cone tip (u₁) or at the cone base (u₂). Seven sites in Louisiana were used for this study. At each site, in-situ PCPT tests were performed and soundings of cone tip resistance, q_c, sleeve friction, f_s, and pore pressures at different locations (u₁ and u₂) were recorded. High quality Shelby tube samples were collected close to the PCPT tests and used to conduct laboratory oedometer consolidation tests to evaluate the reference OCR values. The predicted OCR values from different methods were compared with the laboratory OCR values. The results showed that all the investigated PCPT methods overpredict the laboratory calculated OCR. The collected data were further used to develop two new PCPT methods that can estimate the OCR reasonably.

The linear curve fitting methods such as the Hvorslev method and the Bouwer and Rice method provide a rapid and simple means to analyze slug test data for estimating in-situ hydraulic conductivity (k) of geologic material. However, when analyzing a slug test in a relatively compressible aquifer, these methods have difficulties to fit a straight line to the semi-logarithmic plot of the test data that shows a concave-upward curvature because the linear curve fitting methods ignore the role of the compressibility or specific storage (S_s) of an aquifer. The comparison of the Hvorslev method and the Bouwer and Rice method is made for a partially-penetrating well geometry to show analytically that the Hvorslev method estimates higher hydraulic conductivity than the Bouwer and Rice method except that the well intake section locates very close to the bottom of the aquifer. The effect of fitting a straight line to the slug test data is evaluated along with the dimensionless compressibility parameter (α) ranging from 0.001 to 1.

This paper presents an innovative method that was developed for formation of a borehole that allowed for pressuremeter testing in otherwise unstable and variable material. The results of the testing are presented in the form of modulus parameters and compared to modulus values obtained from the literature and from geophysical methods used at the same project site. The testing and analysis indicate that the pressuremeter tests results were successful and the results appropriate for design.

A few traditional methods for determining water content in the field are either inaccurate or time consuming. As an alternative, the time domain reflectometry (TDR) technology has been used in the determination of the soil water content for geotechnical applications. This paper presents the preliminary results on the development of a new TDR probe for determining soil water content and dry density at different depths. This new probe is intended to be pushed into the ground using piezocone equipment. Different from the standard TDR probes with straight rods, the new probe consists of two parallel copper stripes coiled around a PVC-steel core. The probe diameter is the same as the standard 10 cm2 piezocone diameter. Through laboratory calibrations, it is possible to establish expressions relating the soil apparent dielectric constant and the bulk electrical conductivity with the gravimetric water content and the dry density.