Evaluating Properties of Weak Shales in Western Missouri

Evaluation of the geomechanical properties of shales, especially weak ones is always problematic. The Missouri Department of Transportation (MODOT) recently undertook a major research initiative to achieve significant and recurring cost savings for MODOT by developing improved, technically sound design specifications. Test drilling in shale was conducted; Boreholes were typically drilled in pairs, side-by-side, with one boring being used for core sampling, and the other being used for in situ penetration testing. Coring methods were modified to provide better quality samples. Testing was conducted on or as near the site immediately after recovery. On site point load testing was introduced and used along with slake durability testing to rank the shale in the Franklin Shale Rating System. Samples of shale too weak for point load testing were tested for plasticity index, which is also part of the Franklin Shale rating system. In the penetration boreholes, alternating split-barrel sampler penetration and Texas cone penetration tests were conducted at 2.5 foot intervals using a standard automatic safety hammer. Between tests, the borehole was cleaned and drilled to the next testing level using a tri-cone roller bit. Once in the core box the core should be examined and logged and samples selected immediately. If RQD (Rock Quality Designation) is measured, it needs to done quickly as in some cases the shale core will spontaneously break into smaller pieces as a result of stress release. Samples need to be tested as soon as possible, and protected from deterioration due to desiccation by sealing them with wax, cellophane, and/or aluminum foil. 2.2. Testing of Shale Various lab and field tests can be used or have been specifically designed for testing the geomechanical properties of shales. These can be divided into strength, strength index, and durability tests. There are several examples of durability tests including the slake test, jar slake test, free swell test [1], and slake durability test [2] (ASTM D4644-04). The slake durability test is probably the most common and useful test that takes 10 lumps (approx. 500 g) of material and measures the % loss of material (by dry weight) after two cycles of being mechanically agitated in a partially submerged wire mesh drum (Figure 2), and then dried. Strength and strength index tests include both insitu penetration tests and lab strength tests. Penetration tests are performed by driving split spoons or steel cones (Figure 3) into the shale and counting the number of blow required to penetrate a given distance. Typically, when a split spoon is hammered into shale, it is the blow count that is of interest; there is typically very little if any sample. For the split spoon or Texas cone [3] (TexDOT Designation TEX-132E) there is often very little penetration, and results are recorded not as blows per foot but rather as penetration per 100 blows [4]. An expendable tip cone can be used as well, but can possibly only work in very soft shale because it needs to be continuously driven, not incrementally as with the Texas cone or split spoon. Lab tests include of uniaxial or triaxial compression tests as well as point load testing [5] (ASTM D5731-07). Point load testing (Figure 4) is quick and easy and can readily be done in the field. Point load index testing can be correlated to uniaxial or unconfined compressive strength (UCS) test results using a straight line best fit. Rasnak and Mark [6] report two different studies in shale of which both result in a conversion of UCS=12.6 * pointload strength. Figure 2: Shale durability testing apparatus. Figure 3: Driven tools. Right: split spoon. Center: Texas Cone. Left: Expendable tip cone. Figure 4: Point load testing apparatus. Figure 5: Franklin’s shale rating system [1]. Additional testing to be considered for very weak shale is Atterberg limits (ASTM D 4318-05). 2.3. Classification and Empirical Design Classification and empirical design methods abound in rock engineering. Santi [7] describes methods for field characterization of weak rock. Bienwaski’s Rock Mass Rating (RMR) system has long been used for design of underground openings [8]. Barton’s Q-system is used to design support in underground openings [9]. Numerous other classification systems include empirically derived design guidelines based on the specially designed classifications [10]. For shales, Franklin suggested a similar classification system called the Shale Rating (R) system [11, 12]. The system can be used for design purposes when both strength and durability are issues, and is comprised of three parameters (Figure 5). The horizontal axis is slake durability index (Id2), while the vertical axis it point load strength (Is50) (for Id2 > 80%) or plasticity index (for Id2 < 80%) Franklin [11] proposed various design criteria based on the shale rating system, including lift thickness for embankments (Figure 6), embankment slope angles and heights (Figure 7), and cut slope angles in shales (Figure 8). Figure 6: Franklin’s design lift thickness and compacted field density as a function of shale rating [1]. Figure 7: Franklin’s design chart for embankment height and slope angle as a function of shale rating [1]. Figure 8: Franklin’s design chart for cut slope angles as a function of shale rating [11]. 2.4. Shale Foundations When considering allowable bearing pressures on shale, especially for deep foundations, durability is typically not considered, and designs are based primarily on measured strengths. It is not that weathering of the shale has not occurred at depth (that will be reflected by lower strengths in more highly weathered sections) but rather by the assumption that no additional deterioration of the shale will be expected during the engineering lifespan of the structure being supported. 3. MISSOURI SHALE INVESTIGATIONS 3.1. Major Missouri DOT Initiative The Missouri Department of Transportation (MODOT) in 2009 undertook a major research initiative along with Missouri University of Science and Technology (MS&T) and University of Missouri-Columbia (MU) to “achieve significant and recurring cost savings for MODOT by developing improved, technically sound design specifications”. Part of the research effort is intended to evaluate common site characterization practices to quantify the variability in parameters used for Load and Resistance Factor Design (LRFD). The expectation is that, by quantifying variability, the benefits of improved practices will become apparent. MODOT has had issues with reliability and confidence in applying shale testing results to designs of deep foundations and retaining walls. The problems in general were poor or damaged core recovery and highly variable unconfined compressive tests. 3.2. Shales in Western MO The shale formations investigated in western Missouri are Pennsylvanian in age. These are part of predominantly clastic sediments, with some limestone and coal beds [13]. An example of a stratigraphic sequence very similar to the one in encountered in the Grandview Site is shown in Figure 9 [14]. Shales are in general gray, silty, and slightly commonly calcareous and fissile [14]. In some places thin coal beds are encountered. The shales are variably weathered. In some places the shales could be more aptly characterized as clays. The highly weathered shales are not only seen near the ground surface or top of the succession, but rather are distributed throughout the succession. Figure 9: Stratigraphic section representative of the Kansas City location. 3.3. Testing Sites and Geology During the phase of the MODOT program that related to shale investigations, MODOT conducted drilling at five different sites; results from four of which are reported here. At all the shale sites, field load testing (Osterberg Cell) has been or will be completed on full-scale drilled shaft foundations. In all, twelve borings were drilled for the purpose of this. Boreholes were typically drilled in side-byside pairs, with one boring being used for core sampling, and the other hole being used for penetration testing. Test site locations are shown in Figure 10. Figure 10. Drilling test site locations in western Missouri. Clockwise from top left: Kansas City, Lexington, Warrensburg, and Grandview. 3.4. Drilling and Testing Several new investigative approaches were used. Coring methods were modified to provide better quality samples. Core runs were carefully extruded, logged and photographed (Figure 11). Shear strength testing was conducted on or near the site via Unconsolidated-Undrained (UU) and Unconfined Compression (UC) procedures according to ASTM D2850 and D2166, respectively. Specimens for strength testing were cut to length from individual core pieces that were at least 150 mm (~6”) long. Samples were sealed with plastic wrap and aluminum foil in the field (Figure 10), transported to an on-site laboratory, and trimmed to specimen lengths averaging approximately 100 mm (4”) using a rock saw. Specimens were not trimmed along the diameter, which averaged approximately 50 mm (~2”). Unconsolidated-Undrained (UU) triaxial compression tests were conducted by encasing the specimens in a latex membrane and applying isotropic confining pressure without allowing drainage. Isotropic confining pressure (σ3) was applied with a magnitude approximately equal to the in-situ confining stress which was assumed to be 0.75z (in psf), where z was the sampling depth in units of feet. All specimens were loaded to failure under strain-controlled axial loading using an axial strain rate of 1%/min. The peak deviator stress (σ1 – σ3) was used to calculate undrained shear strength [su (UU)] and compressive strength [UCS (UU)]. Additional specimens were tested under unconfined compression (UC) to determine undrained shear strength and compressive strength [UCS (UC)] [su (UC) = qu/2] [15]. These results were used for comparison with the UU test results and to assess any variably and bias between this testing protocol and conventional MoDOT practice. All UU and UC testing was conducted where possible the same day (and generally within 5 hours of sampling) to minimize stress release and other deteriorating effects by bringing the testing apparatus to a nearby MODOT field office (Figure 12). Figure 11. Shale core samples for on-site triaxial strength testing. Samples were wrapped in plastic wrap and foil and transported to an on-site laboratory within five hours of sampling. (Photo: Dory Colbert) Figure 12: Triaxial testing in MODOT field office. (Photo: Dory Colbert) On site pointload testing was conducted (Figure 13) and correlated with unconfined compressive strength, and used along with slake durability testing to rank the shale in the Franklin Shale Rating System. Because diametral testing of the horizontally bedded shale makes no sense axial testing was performed using approx 25 mm lengths of core cut with a tile saw (Figure 14). Samples of shale too weak for point load testing were tested for plasticity index, which is also part of the Franklin Shale rating system. Weak shales were also tested with a specially adapted pocket penetrometer with an indentation cross sectional area that was one half of the standard size (Figure 12), a tool and method currently used by MODOT. In penetration sampling boreholes, alternating, splitbarrel sampler, and Texas cone penetration test were conducted at 2.5 foot intervals a standard automatic safety hammer (Figure 3). Between tests the borehole hole was cleaned and drilled to the next testing level using a tri-cone roller bit. Figure 13: Point load test machine in the field showing axial testing. Figure 14: Point load testing in the field. Because axial testing was deemed necessary, a tile saw was used in the field to prepare samples. Figure 15: Specially modified pocket penetrometer. With indentation cross section one half of the standard size.