Click for next page ( 25


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 24
24 Hollow Cylinder Test Laboratory MR Tests—Summary A hollow cylinder test simulates stress conditions close to A thorough review of several research reports and papers on the field traffic loading, including the principle stress rota- resilient modulus testing of subgrades and unbound bases in the tions taking place in the subgrade caused by wheel load laboratory conditions was next made. Several research reports movements (Barksdale et al. 1997). In this test, a hollow and articles have been published pertaining to this topic. These cylindrical soil specimen is enclosed by a membrane both papers and reports cover laboratory procedures and findings inside and outside the sample. Stresses are applied in axial as well as a few pavement design recommendations. or vertical, torsional, and radial directions. Repeated loads can be simulated in this setup and related moduli can be The investigator collected various state research reports determined. Because of the possible application of various that primarily describe DOT-funded research and how MR types of stresses, different stress path loadings simulating tests and their results are applied for the mechanistic pave- field loading conditions can be applied. Also, this setup can ment designs in the respective states. Additional information be used to perform permanent deformation tests. was also sought from the states during the surveys, which resulted in a few additional reports. Overall, resilient moduli A few other tests, including the Cubical Triaxial Test, are test results from the literature are presented in three phases: used in the literature. These tests are still under research evaluation, however, and they are yet to be used for practi- • First phase: reported before 1986 cal applications. Other test methods that provide parameters • Second phase: reported from 1986 to 1996 that are linked with moduli or stiffness properties of soils • Last phase: reported after 1996 include the CBR test, R value test, Texas triaxial value, and SSV test. These methods employ static loading of the soil In 1986, the AASHTO interim design guide recommended specimens until the failure, and the properties measured in the use of resilient modulus to characterize subgrades and the tests are reported as CBR, R, or SSV values. Because of bases for flexible pavement design. The starting year of 1996 the availability of large test databases of these properties, is randomly chosen for the last phase, because any reports various state agencies developed correlations between the presented after that year are considered to be recent litera- AASHTO-recommended moduli of soils with the CBR, R, ture (that is, literature presented in the last 10 to 11 years). or SSV values. Details on these correlations are presented in chapter five. MR Literature Before 1986 As noted previously, several test methods have been used In a 1989 workshop on the application of resilient modulus in to determine the resilient modulus of subsoils in the labora- pavement design, Vinson presented a comprehensive review tory conditions. However, the most prominent method for of the resilient modulus test work performed at the Univer- a laboratory modulus test has been the RLT test, primarily sity of California Berkley in the early 1960s. This paper pri- because AASHTO standardized this test and it features bet- marily includes the research performed by Seed et al. (1962), ter simulation of pavement subgrades under traffic loading. which focused on one of the earliest works on resilient The next section discusses several research studies that uti- modulus measurements of subgrade soils at various com- lized these tests and their findings. paction conditions and soil types. Other studies performed FIGURE 26 Stresses under traffic wheel load and stress reversal under wheel loading (Vinson 1989; Barksdale et al. 1997).

OCR for page 24
 25 by Barksdale (1971, 1972) and Terrel et al. (1974) addressed tute method of designing flexible pavements in which a bulk laboratory-related variables including the shape of the load- stress model was used to model the resilient behavior of ing pulse and duration of the laboratory test formulation. cohesionless soils. Uzan (1985) explained the limitations of the bulk stress model with two model constants and then Hicks and Monismith (1971) later described several introduced a two-parameter model consisting of both bulk factors that influence the resilient properties of granular and deviatoric stresses with three model constants to simu- materials including bases. Kalcheff and Hicks (1973), then late the resilient behavior of subsoils. documented the resilient moduli tests performed on granu- lar soils including unbound aggregates. This study described Hence, it can be summarized that the research to 1986 the potential stress sequence and stress ratios recommended mostly focused on (1) the development of test procedures for testing granular materials. Hull et al. (1980) reported MR and equipment modifications to test cohesive subgrades and test results on base materials following the test procedures granular base materials, (2) the development of appropri- recommended by Kalcheff and Hicks (1973). Allen and ate models to represent the resilient behavior, and (3) the Thompson (1974) addressed the response of lateral stresses introduction of a few correlations based on soil properties to on resilient properties of granular materials. predict resilient properties. In 1976, Thompson and Robnett studied resilient prop- MR Literature Between 1986 and 1996 erties of several Illinois subgrade soils at the University of Illinois, Urbana, and this study reported the test results and Following the recommendation by the 1986 AASHTO the development of correlations between resilient modulus interim design guide to use resilient modulus property for and subgrade soil properties. They proposed an arithmetic pavement material characterization, several research studies model to describe the resilient properties of fine-grained were initiated with the support of several DOTs. A few of soils. Shook et al. (1982) later discussed an Asphalt Insti- these studies and their findings are summarized in Table 5. Table 5 Findings from Research Reports and Papers on Resilient Modulus During 1986–1996 Description of References Research Summary and Findings Elliott et al. Resilient Repeated load triaxial (RLT) test setup and AASHTO T-274 procedure were used for performing (1988) Properties of laboratory resilient moduli tests. Elliott (1992) Arkansas Kneading and static compaction methods were followed for preparing compacted soil specimens. Subgrades Arkansas DOT Percent moisture content was varied between 70% (dry of optimum) to 130% (wet of optimum). Resilient moduli decreased with an increase in moisture content. Workshop on An overview A few of the following presentations were made at this workshop: Resilient Modulus of Resilient Vinson—Past work and current status Testing, Oregon Modulus Test State University, and its Ho—Florida DOT Experience Corvallis, Oregon Significance Thompson—Factors affecting MR of Subgrades and Granular soils (1989) Woolstrum (1990) Resilient RLT test setup and AASHTO T-274 procedure were used. Nebraska Properties of Conditioning was done with 15,000 cycles of a haversine loading. Department Nebraska Subgrades and Loading period of 0.1 s and a relaxation period of 0.4 s were applied. of Roads Aggregates Aggregates were tested as per granular material test procedures. (14 of them) Three compaction moisture conditions, optimum, wet, and dry of optimum were studied. Expected trends of decrease in resilient moduli with an increase in moisture content. Correlations between MR and Nebraska group index parameter were developed. Claros et al. (1990) Test RLT test setup and AASHTO T-274 (sample preparation) and P-46 procedure protocols (testing) Stokoe et al. Procedure were used on synthetic and subgrade and aggregate samples for performing laboratory resilient (1990) Development, moduli tests. Calibration Comparisons were made with torsional ring shear and resonant column test data, which showed Pezo et al. (1992) and Evalua- some variations. Pezo and Hudson tion of MR (1994) Equipment Synthetic samples provided a good methodology to calibrate the MR equipment. Texas Department and Modeling Pezo’s study on different subgrades of Texas with different PI values showed that an increase in of Transportation of Resilient moisture content resulted in the decrease of MR values. Properties A TxDOT model with seven correction factors was developed to predict resilient properties. continued

OCR for page 24
26 Table 5 (continued) Description of References Research Summary and Findings Wilson et al. Multiaxial Seven subgrade soils were taken from different sites. (1990) Testing of Specimens were compacted in cubical mold of 4 in. x 4 in. x 4 in. Ohio DOT Ohio Subgrade Cyclic tests were conducted after 200 cycles of conditioning. Resilient response was dependent on compaction moisture content. At low stress ranges, MR decreases with an increase in deviatoric stress. Ksaibati et al. Factors Nine subgrade soils were taken from different regions of Wyoming. (1994) Influencing RLT test and AASHTO TP-46 or T-294 were used. Ksaibati et al. MR Values of Subgrades FWD tests were conducted at the sampling sites and results were analyzed with backcalculation (2000) programs. Wyoming DOT Moisture content influenced resilient properties of A-4 and A-6 soils. EVERCALC program predictions appear to give accurate resilient moduli. Mohammad et al. Resilient RLT test method was used for MR tests. (1994, 1995, 1999) properties of T-292 and T-294 as well as internal and external displacement measurements were evaluated in Puppala et al. Louisiana providing realistic and repeatable resilient properties of silty clay and sandy soil. (1997, 1999) Subgrades Six different soils from different regions of Louisiana were sampled and tested in the RLT using Louisiana DOT T-294 procedure. Five compaction moisture conditions were studied. RLT test and AASHTO TP-46 or T-294 were used. Correlations between resilient moduli and soil properties were developed. Drumm et al. Resilient Eleven subgrades were taken from different regions of Tennessee. (1990) response of Compact densities as per standard Proctor compaction and specimen preparation as per kneading Drumm et al. Tennessee method. (1995) Subgrades and effects P-46 method and RLT tests were followed. Andrew et al. of post-­ Optimum moisture content condition and moisture content condition close to 100% saturation (1998) compaction were studied. Tennesse DOT saturation Saturation resulted in the reduction of resilient moduli. on resilient Saturation related reductions are highest in A-7-5 and A-7-6 soils than in A-4 and A-6 soils. response of subgrades A methodology to correct the resilient modulus due to increased degree of saturation was developed. Zaman et al. Assessment Resilient modulus testing using RLT test was performed. (1994) of Resilient Six different aggregate sources were taken from different locations from Oklahoma. Chen et al. (1995) Modulus Tests and T-292 and T-294 procedures were used for MR testing. Daleiden et al. Their Compactions were done at optimum and 95% of optimum dry density. (1994) Applications Resilient moduli of aggregates varied between 41 and 262 MPa and these results are lower than Oklahoma DOT for Pavement those reported in the literature at that time. Design Variability in MR due to test procedure appears to be higher than that of aggregate source. Santha (1994) Resilient Soils from 35 sites were studied. Georgia DOT Moduli of RLT equipment and T-274 procedure were used. Subgrade Soils from MR results varied considerably due to variations in soil types and compaction conditions. Georgia Soil type, compaction conditions, and Atterberg limits were used as independent variables for MR correlations.

OCR for page 24
 27 Elliott et al. (1988) reported resilient moduli test results obtained on different cohesive subgrades in Arkansas. The main intent of this research was to explain the effects of field moisture content on the resilient moduli of compacted sub- grades. This study also addressed the effects of compaction procedures and moisture content variation on the resilient moduli properties. Figure 27 presents the variation of resil- ient modulus of fat clay (CH) from Jackport, with respect to percent changes in optimum moisture content. A decrease in the MR value close to 1 ksi was reported for percent increase in the optimum moisture content. A low percent of optimum (below 100%) was termed as dry of optimum and a high percent of optimum (above 100%) was termed as wet of opti- mum conditions. Similar findings were reported on other cohesive subgrades in this study. In 1989, a workshop on resilient modulus was held at FIGURE 27  Resilient modulus of CH soil from Arkansas Oregon State University in Corvallis and several research- versus percent optimum moisture content (Elliott et al. 1988). ers, practitioners, and equipment manufacturers discussed various testing, design, and equipment-related aspects of resilient modulus testing. This workshop was instrumental Another study performed by Stokoe et al. (1990) in leading several DOTs to start research on the resilient addressed the use of other devices such as the RC test moduli response of their local soils. to measure moduli of calibrated synthetics specimens. Figure 29 presents a comparison of axial strains generated Woolstrum (1990) studied 14 different Nebraska soils from RC testing with those generated using RLT tests while and yielded similar conclusions to those of Elliott et al. calibrating the medium stiff and stiff urethane materials. (1988). This study also developed correlations by introduc- Both stiff and medium stiff materials (TU-960 and TU-900) ing a Nebraska group index parameter (G) as an indepen- extended over the majority of the MR range determined from dent variable. The AASHTO T-274 procedure was used to AASHTO T-274 RLT testing, whereas the soft material determine resilient properties of local soils. Several research (TU-700) extended over half of the range reported by the papers and reports were published in the early 1990s that RLT testing. This research also evaluated resilient or stiff- focused on the resilient modulus research performed in the ness properties of subgrades from different parts of Texas state of Texas (Claros et al. 1990; Pezo et al. 1992). These and also developed a universal model. studies show calibration of RLT test devices, use of other equip- ment including RC and torsional ring shear devices to measure the shear modulus, and estimating elastic or resilient modulus. Pezo et al. (1992) focused on a universal model development to predict resil- ient properties of subsoils. Research studies funded by Texas DOT (TxDOT) have evaluated the use of synthetic material samples for calibrat- ing various moduli measurement devices, including RLT, RC, and torsional ring shear devices (Cla- ros et al. 1990; Stokoe et al. 1990; Pezo et al. 1992; Pezo and Hudson, 1994). Figure 28 presents typical test results conducted on cohesive subgrade soil as per the TP-46 FIGURE 28  Moduli test results on synthetic sample from resonant column and RLT tests procedure. (Claros et al. 1990).

OCR for page 24
28 FIGURE 31  Resilient moduli results of granular subgrade from Ohio (Wilson et al. 1990). FIGURE 29  Comparisons of axial strains measured from compaction moisture content on the resilient moduli prop- RLT and resonant column tests (Stokoe et al. 1990). erties. These results indicate that an increase in deviatoric stress resulted in a decrease in resilient moduli properties. Also, an increase in compaction moisture content resulted in W ilson et al. (1990) presented research findings on Ohio a decrease of resilient moduli of granular soils. subgrades in which resilient properties of subgrades were measured using a multiaxial cubical triaxial test setup. Burczyk et al. (1995) reported resilient properties of Figure 30 presents a multiaxial testing device used in this Wyoming subgrades. An RLT test setup and AASHTO T-294 research. Researchers noted the uniqueness of this equip- procedure were used in this study, which also discussed vari- ment by applying stresses in all directions and concluded that ous fundamental soil properties that influence resilient prop- this equipment was capable of providing resilient properties erties of subgrades in Wyoming. This study also evaluated of subgrades. The test procedure used in this research was various backcalculation methods used to interpret the resil- unique, and it included application of one level of confining ient properties. pressure of 5 psi and different deviatoric stresses ranging from 1 to 8 psi. The deviatoric stress pulse was applied at a Mohammad et al. (1994a,b, 1995) described the resilient rate of 1 pulse/s. Resilient properties were determined based properties of Louisiana subgrades, including a complete on the measured strain response. evaluation of RLT setup and AASHTO test procedures T-292 and T-294, as well as measurement systems including linear Typical test data on a granular subgrade are presented in variable differential transformers (LVDTs) placed for the Figure 31, which shows the influence of deviatoric stress and middle third of the specimen (referred to as middle) and at the ends of the soil specimen (referred to as end), both being internal deformation measurement systems in yielding reli- able resilient properties. Figure 32 presents resilient moduli determined from both internal measurement systems of a sandy specimen tested in the research. Higher moduli values were measured from the middle LVDT system than the end LVDT system. The ratios between MR measured from the middle system to the end system varied from 1.15 to 1.22 at different compaction moisture content, indicating a 15% to 20% increase in moduli values obtained with the middle measurement system. Drumm et al. (1990, 1997) studied resilient moduli of different types of subgrades from Tennessee using a SHRP P-46 test procedure. Research was focused on the resilient FIGURE 30  Photograph of multiaxial testing device (Wilson et al. 1990). moduli data development for Tennessee subgrades, and also

OCR for page 24
 29 T-292 procedure, and this variation was attributed to stress conditioning and stiffening effects in the T-294 procedure. Also, this study reported that the MR variability with respect to test procedure was higher than the same with respect to aggregate source. Santha (1994) studied the resilient properties of subgrades from 35 test sites in Georgia, using T-274 procedure. These results show wide variations in MR with respect to soil type and compaction procedures used in the testing. The measured data were used and analyzed with two-parameter and three- parameter models. The three-parameter model captured the measured resilient properties better than the two-parameter model. Further discussion on these test results and their use in the development of multiple linear regression analyses of test results is presented in later sections. Overall, it can be summarized that the resilient modulus research performed between 1986 and 1996 focused on the FIGURE 32  Resilient moduli results measured from both end use of various laboratory and field equipment to determine and middle deformation measurement system (Mohammad et al. 1994a,b). the resilient properties of both subgrades and bases. Several test procedures (T-292, T-294, and P-46) were introduced and evaluated during this phase. Displacement measure- addressed the reductions in resilient moduli with respect to ment systems, one in the middle third of the soil specimen post-compaction saturation. Figure 33 presents the effects of and one at the ends of the soil specimen were researched post-compaction moisture content in terms of degree of sat- to provide realistic MR values. Also, a few of these studies uration on the resilient modulus of subgrades. As expected, investigated and tested various local subgrades and unbound an increase in saturation resulted in a decrease in MR values. bases for developing a database of resilient properties. The This research also led to the development of regression mod- measured data were later used to develop various models els for prediction of resilient properties of subgrades. These to predict resilient properties of subgrades and aggregates. details are presented in chapter four. Table 6 provides a summary of the literature findings from 1986 to 1996. In the early 1990s, Thompson and Smith (1990), Zaman et al. (1994), and Chen et al. (1994) reported the resilient MR Literature After 1996 moduli of unbound aggregate bases using AASHTO T-292 and T-294 procedures. Variations of moduli with respect to Modifications of AASHTO test procedures from T-294 to test procedures and aggregate types are explained in these T-307 with different testing-related specifications occurred studies. Figure 34 presents the effects of test procedures during this phase. This resulted in additional research being on the resilient properties of aggregates from Choctaw and Murray counties in Oklahoma. Test results indicate that AASHTO’s T-294 procedure yielded higher moduli than its FIGURE 33 Influence of post-compaction moisture content on FIGURE 34 Influence of AASHTO test procedure on resilient resilient moduli results (Drumm et al. 1990, 1997). moduli results of aggregate resources (Chen et al. 1994).

OCR for page 24
30 Table 6 Summary of Major Findings on Resilient Modulus from ‘1986 to 1996’ Topic Observations Equipment Used RLT Tests (Preferred Method) Resonant Column Test Torsional Ring Shear Test Cubical Triaxial Test Calibration Issues—Repeatability and Still a problem; One study reported the use of synthetic specimens with known stiffness Reliability Issues properties; Wide range of properties Test Procedure Used for RLTs T-274 was used in early studies followed by T-292 and T-294 methods; TP-46 was also used in the later studies; All studies used conditioning, followed by testing at various confining and deviatoric stresses; Stress sequence and stress ratios are different in each of these methods Material Types Most of the studies focused on natural subgrades; a few others studied aggregates and granular sandy soils Specimen Preparation Methods Most of the methods use impact compaction with standard Proctor energy efforts Others used static compaction and kneading compaction methods Compaction Conditions Close to optimum moisture contents; Several studies investigated variations in compaction moisture contents from dry to optimum to wet of optimum Models to Analyze MR Data Bulk stress (granular materials) and deviatoric stress (cohesive soils) models are frequently used since they are the recommended models by the AASHTO procedures performed by various DOTs to validate the previous research of these studies is impressive, more studies are still findings and new research to determine MR of the local soils. needed for a better understanding of resilient properties In this phase, all recent literature involving the use of resil- of several other states. Among the state DOTs, Florida, ient modulus testing from different state-funded research Minnesota, Texas, Mississippi, Ohio, and New Jersey studies using T-294 and T-307 methods are reported. lead considerable research efforts on laboratory testing and also implemented the use of resilient properties of This information is synthesized in Tables 7 and 8, with subgrades in their pavement design practice. Several Table 7 summarizing the test information, including test other states including Illinois, Louisiana, and others method followed, specimen preparation method adapted, and have started implementing or are planning to imple- types of soil and aggregate materials tested. Table 8 sum- ment the MR value in their pavement design practice. marizes salient findings and conclusions from the research • In the case of resilient modulus studies on aggregate performed by the reporting agencies. These references are bases, limited research studies have been performed. listed along with the state DOTs that provided funding for Studies in this area mainly came from the states of Ohio, the respective research. Texas, New Jersey, South Carolina, and Minnesota. Details on the aggregates research are presented in the Tables 7 and 8 present various available final research reports later sections. collected from the extensive literature review performed for this research. The research presented herein refers to the final The test equipment primarily used over the last 10-plus research reports, because they contain more detailed informa- years is RLT test apparatus. A few studies—including those tion on the tested materials and analyses of test results. by Nazarian et al. (1996), Davich et al. (2004b), and Gupta et al. (2007)—experimented with the use of bender element The following general summary provides information testing to determine the small-strain shear modulus, which on the resilient modulus of aggregates and subgrades that is was then correlated to the resilient modulus with an assumed relevant to both researchers and practitioners. Several DOTs Poisson’s ratio. Figure 35 presents a typical bender element funded research, which is summarized in the tables. Most of setup made of piezoelectric elements used by Davich et al. these studies addressed resilient properties of the subgrades (2004b). A wave pulse generated by emitter (bender ele- encountered in various states and a few addressed aggregate ment) from one end will travel through the soil specimen and granular bases used in the respective state. and reaches the receiver (another bender element) at the other end. Based on the travel time of a shear wave and mass • Several new research studies from different DOTs were density of the soil specimen, the small-strain shear modulus conducted during this phase. Though the total number value is determined.

OCR for page 24
 31 Table 7 Research Reports and Papers on Resilient Modulus After 1996—Test Information Reference & Description of Soil and Aggregate Supporting DOT Research Types Tested Test Procedure Followed Nazarian et al. Development of Four types of Gran- RLT test equipment was used; T-294 procedure was followed; bender element 1996 & Texas Test Method of ular bases including setups were also evaluated. DOT (TxDOT) Aggregate Bases Caliche, Limestone, Gravel and Iron Ore materials were used. Gandara et al. Effects of Fines Aggregate Bases Three confining pressures and three deviatoric stresses at each confining 2005 & TxDOT on Aggregate pressure (similar to T-307); Bases Specimens are grouted at the ends; Internal LVDTs were used to measure both axial and radial strains; Study focused on amount of fines on MR of Aggregates Berg et al. (1996), Resilient Modulus Two Clayey Specimens were compacted and then saturated and later subjected to freezing. Bigi and Berg Testing of Subgrades and Resilient modulus tests were conducted on frozen samples subjected to three (1996) & Untreated Two Bases with different temperatures below freezing temperature. Minnesota DOT Subgrades different amounts Some specimens in the triaxial cell were thawed and then subjected to resilient (MnDOT) of fines modulus tests. Loading pulse has 1 s of loading and 2 s of relaxation. Frozen samples were tested under 10 psi of confining and thawed samples were subjected at four different confining pressures. In all cases deviatoric stresses were varied. Davich et al. Small strain and Six types of bases LTPP P-46 method was followed for performing resilient modulus testing (2004b) & resilient modulus (granular types): whereas Bender element test was followed for performing small strain MnDOT testing of granular Fine to well graded modulus tests. soils sands with fines Gupta et al. Resilient Behavior Four subgrade soils Two compaction conditions and three matric suctions were applied at each (2007) of Unsaturated of different PI compaction condition and these suctions were 0, 22, and 50 psi; MR tests Minnesota DOT Subgrade Soils values and small strain bender element tests were conducted on all samples; Axis and Pavement translation technique was used to apply the suction to the soil samples; MR Design test procedure was based on NCHRP 1-28A George and Laboratory MR Several field sub- AASHTO TP 46 Protocol; Resilient modulus test setup; External LVDT Uddin (2000) Testing on Field grade samples at system; Laboratory MR at 14 kPa of confining pressure and 35 kPa of Mississippi DOT Cores and Corre- different depths and deviatoric stress was used as the subgrade layer property lations Between at different spacings MR and Dynamic (200 ft c/c); Sam- Cone Penetrome- pling depth up to 5 ter and FWD Data ft to a maximum of 10 ft; Soil samples from 12 sections George (2004) Prediction of Subgrade samples TP-46 was followed Mississippi DOT Resilient Modulus from 8 different RLT and impact compaction were used from Soil County/Roads Properties Ooi et al. (2006) Development of Subgrades from LTPP P-46 procedure (T-307) was followed for MR testing; Three confining Hawaii DOT Correlations for four locations in pressures were used as per the cohesive subgrade testing MR of Hawaii Hawaii Subgrades Lee et al. (1997) Resilient Proper- Soil specimens AASHTO T-274 method was used; Soil specimens were prepared using impact Indiana DOT ties of Indiana from three in- compaction method Cohesive Subsoils service subgrades Janoo et al. (1999) Resilient Modulus Five Subgrades TP-46 Procedure was used; Tests were conducted at several temperatures to New Hampshire of New Hamp- from New reflect freezing and thawing; Average MR from tested stresses were used as DOT shire Subgrades Hampshire effective moduli; Kneading compaction was used for testing; Tests were for Pavement conducted at optimum moisture content condition. Design Maher et al. Resilient Modulus Six Different TP-46 method was followed; Three compaction moisture content conditions (2000) Properties of New Subgrade Soils were studied New Jersey DOT Jersey Subgrade from New Jersey Soils continued

OCR for page 24
32 Table 7 (continued) Reference & Description of Soil and Aggregate Supporting DOT Research Types Tested Test Procedure Followed Bennert and Resilient Modulus Two different AASHTO TP-46 method was used Maher (2005) Properties of New aggregates of two New Jersey DOT Jersey Aggregates types from three regions with differ- ent gradation types and two types of blended aggregated with recycled concrete and asphalt aggregates Baus and Li Studies on South Seven types of Static and cyclic plate load tests were conducted on different aggregate beds. (2006) Carolina Aggre- aggregate bases South Carolina gate Bases DOT Ping et al. (2003) Resilient Moduli Two subgrade soils AASHTO T-292 and T-294 methods were followed for MR testing; LVDTs Florida DOT of Subgrades from from Florida were placed both in the middle and at the ends of the soil specimen; RLT Florida equipment was used for testing subgrades; Both vertical and horizontal LVDTs were used for monitoring vertical and lateral deformations Ping et al. (2007) Resilient Modulus 37 subgrade soils AASHTO T-307 method was followed for MR testing; Other procedures Florida DOT of Subgrades from from different dis- including T-292 and T-294 were used in the earlier MR testing; RLT equipment Florida tricts in Florida was used for testing subgrades; Both vertical and horizontal LVDTs were used for monitoring vertical and lateral deformations Hopkins et al. Resilient Modulus 128 tests on soil Specimens were molded at optimum moisture content and 95% of maximum (2001) of Kentucky samples collected dry density; TP-46 method was followed for the MR testing. Kentucky DOT Subgrade Soils from various loca- tions in Kentucky Masada et al. Resilient Moduli Several aggregate TP-46 method was followed (Bases and Subgrades) (2004) of Ohio Subgrade types were Ohio DOT Soils and Bases researched with dif- ferent types of gra- dations; Several subgrades from dif- ferent locations in Ohio were collected and tested Wolfe and Butalia Resilient Moduli Thirteen Subgrade Specimens were compacted at dry of optimum, optimum and wet of optimum; (2004) of Ohio Subgrade soil samples across T-294 method was followed; external LVDT system was used; Both Ohio DOT Soils the state of Ohio unsaturated (compacted) and saturated samples were tested. Malla and Joshi Resilient Moduli Connecticut—A-2 AASHTO T-307 tests were conducted (2006) of Subgrades from and A-4; Maine— Tests were conducted at various compaction moisture content conditions New England New England A-1, A-2, A-3, A-4, Transportation States; A-5, and A-6; Mas- Consortium sachusetts—A-1, (Connecticut, A-2, A-3, A-4, A-5, Maine, and A-6; New Massachusetts, Hampshire—A-1, New Hampshire, A-2, and A-4; Ver- Vermont) mont—A-1, A-2, A-4, A-6, and A-7. Titi et al. (2006) Determination of Seventeen soils AASHTO T-307 tests were conducted Wisconsin DOT Resilient Moduli from different Tests were conducted at different moisture content conditions Properties of regions in Wiscon- Typical Soils sin were tested. from Wisconsin A-1, A-3, A-4, Kim and Labuz Resilient Modulus Different aggre- NCHRP 1-28A test protocol was used and laboratory specimens were prepared (2007) and Strength of gates including using gyratory compaction method. Minnesota DOT Base Course with Recycled Asphalt Recycled Pavement and Bituminous natural aggregates Material

OCR for page 24
 33 Table 8 Research Reports and Papers on Resilient Modulus After 1996—Findings Reference Resilient Moduli Range Findings, Recommendations and Future Research Nazarian et al. Tested bases have MR values Conditioning was eliminated and grouting was recommended at both ends of the (1996 ) ranged between 70 MPa to 210 specimen, which resulted in better repeatability of test results. Nazarian et al. MPa with Limestone exhibiting Conditioning and testing sequence was modified for future testing in the state. (1998) close to 300 MPa at optimum moisture content Lateral deformations and Poisson’s ratio were directly measured. & TxDOT Gandara et al. Tested Aggregates have 25 to Effects of fines on aggregate resilient properties are minimal; however, it does (2005) 100 ksi influence the moisture susceptibility of the aggregate fine soil mixtures. TxDOT Type of fine material and its plasticity nature has major influence on permanent deformation. Tube suction tests on aggregates were used to address moisture susceptibility. The optimal fines in aggregates is between 5% and 10%. A field methodology is still needed to predict the performance of aggregates that can complete laboratory tests. Berg et al. (1996) Lower moduli (less than 1 ksi) Both unbound base and subgrades exhibited a 2 to 3 order increase in resilient MnDOT were reported, which was attrib- moduli due to freezing. uted to calibration error. With an MR decreased with an increase in saturation. inducement of freezing, the moduli of subgrades increased to 1000 ksi. MR of all soils are stress dependent. For unbound bases, the MR was Different regression expressions are presented. 10000 ksi at high freezing Test procedures are different from AASHTO recommended procedures. conditions. In few tests, calibration errors were reported. Davich et al. Resilient moduli of bases varied MR increased with a decrease in moisture content. (2004b) from 60 MPa to 800 MPa based Hyperbolic model using small strain modulus and shear strength properties are used MnDOT on confining and deviatoric to predict resilient modulus. stresses applied. Poor bonding and uneven specimen ends resulted in the variation of the three LVDT measurements. Field measurements of small strains can be used to estimate resilient moduli needed for the pavement design. More studies are needed to validate the procedures established in this research. Gupta et al. (2007) Resilient moduli based on external An increase in matric suction resulted in considerable increase in MR value. MnDOT displacement measurements varied A linear semi-logarithmic relationship between resilient modulus, MR and matric between 10 and 70 MPa and the suction was developed. same varied between 10 and 200 MPa when internal displacement Internal measurements always yielded higher MR values, which are 1.7 to 3 times system was used. that of MR values based on external measurements. Bender element data measured from unconfined compacted subgrades showed a correlation with the MR values measured from internal measurements. Soil water characteristic curves of all four soils were developed to determine the moisture content and suction relationships. Model formulated by Oloo and Fredlund (1998) was used to analyze the MR -Suction data. George and Uddin MR values of subgrades at 14 kPa Sample disturbance due to pushing of the Shelby tube resulted in higher MR values (2000) confining pressure ranged from for the top layer. Mississippi DOT 30 MPa to 270 MPa, with higher Laboratory MR varied considerably along the test section than along the depth of values being measured for upper the section. layer samples. Moisture content of the sample showed considerable influence on MR. George (2004) Resilient moduli (MR) varied Trends of MR values agree with those reported in the literature. MSDOT between 50 and 115 MPa. Ooi et al. (2006) MR values of subgrades ranged from An increase in deviatoric stress resulted in a decrease in moduli of subgrades. Hawaii DOT 10 ksi to 35 ksi for various silty sub- High PI silty material exhibited lower moduli than low PI silty soils. grades and this range is dependent on confining pressures and devia- Excessive drying of soil to meet the compaction moisture contents (soils sampled toric stressed applied in a test. from the field at high moisture levels) may compromise the resilient properties. Lee et al. (1997) Most MR results varied between MR is correlated with the stress at 1% axial strain of unconfined compression strength Indiana DOT 30 and 80 MPa test and this correlation did not show the dependency on compaction conditions. This relationship is valid for other cohesive soil types. continued

OCR for page 24
34 Table 8 (continued) Reference Resilient Moduli Range Findings, Recommendations and Future Research Janoo et al. (1999, MR results of subgrades varied Design moduli were determined by averaging moduli measured at different 2004) from 5 ksi to 2646 ksi based on confining pressures. New Hampshire the test temperature condition Poisson’s ratio values measured were different from those used for these types of DOT materials and this variation was attributed to conglomeration of particles. Corrections for effective moduli are based on subgrade temperature conditions, not based on air temperatures. All test results reported are based on those compacted at optimum moisture content condition and hence moisture and density related correction is needed when the moduli values at other compaction conditions are needed. Maher et al. (2000) Resilient moduli of soils varied Granular soils exhibit strain hardening as the MR values increased with deviatoric Maher et al. (1996) between 2 to 36 ksi and this load applications. variation is dependent on soil type Cohesive subgrades showed strain softening with a decrease in MR with deviatoric New Jersey DOT and moisture content condition. loading. Moisture content has a major effect on the moduli of both soils. Bennert and Maher MR values of natural aggregates Addition of recycled aggregates enhanced the resilient properties. (2005) varied from 13 to 23 ksi whereas Recycled asphalt aggregates blending has more profound effect on the MR values New Jersey DOT the same for blended aggregates than recycled concrete aggregates. varied from 20 to 40 ksi at stress conditions close to those expected As the gradation of aggregates become finer, the resilient properties decrease. under pavements Baus and Li (2006) MR values of aggregate bases from A structural coefficient of 0.18 is recommended for aggregate bases in the pave- South Carolina static plate load tests varied from ment design. DOT 14 to 77 ksi. Relaxation of coarse gradation did show a positive influence in enhancing the resilient modulus. Reevaluation of structural coefficients is needed by performing field tests. Ping et al. (2003) MR values of sandy soil varied Adjustment factors are introduced to convert MR from one test procedure to another Florida DOT from 90 to 340 MPa based on the test procedure and one measurement system (end system) to another (middle confining and deviatoric stresses system). applied during the testing and the same for cohesive soil varied from 60 to 170 MPa. Ping et al. (2007) The MR values of Florida subgrade Two databases of moduli were developed, one with comprehensive moduli data Florida DOT soils ranged from 7 ksi to 26 ksi, and the other with integration and analysis functions. with an average value of 14 ksi Physical properties such as density and moisture content and LBR influence the MR magnitudes of A-2-4 soils than A-3 soils. An increase in clay and fines content also influences the MR values. Gradation properties such as uniformity coefficient and coefficient of curvature have less influence on the MR. Several prediction models were developed for MR results. More soil information from previous MR tested soils will further enhances the model development. Hopkins et al. Compacted soil specimens Unsoaked soil specimens performed well under resilient modulus testing whereas (2001) exhibited MR values ranging soaked specimens experienced large deformations, which resulted in bulging in Kentucky DOT from 20 ksi to 28 ksi at a low some cases. confining pressure of 2 psi. Three models were developed for the analysis of the present results. Need for a resilient modulus test to perform on a nearly saturated soil specimen Acceptance criteria for MR testing should be developed. Masada et al. 2004 MR of bases varied from 3 ksi This research summarized various research studies conducted in the state Ohio and Ohio DOT to 65 ksi, and average values of several lab and field tests were conducted on both unbound bases and subgrades. the aggregate bases varied from Moduli results from these studies are summarized. 9 to 42 ksi. MR of subgrades varied from 12 to 24 ksi for A-4 soils; 3 to 23 ksi for A-6 soils; and 4.5 to 25 ksi for A-6 soils. continued

OCR for page 24
 35 Table 8 (continued) Reference Resilient Moduli Range Findings, Recommendations and Future Research Wolfe and Butalia MR values of subgrades varied An increase in moisture content resulted in the reduction of resilient properties. (2004) from 10 MPa to 120 MPa and In the case of A-4 soils, saturation resulted in the reduction of MR by 8% to 88% Ohio DOT lower values were measured when where as the same for A-6 and A-7-6 soils varied from 50% to 87% and 44% to tests were conducted on wet of 82%, respectively. optimum soil samples Field instrumentation with tensiometers and moisture probes are recommended for better understanding moisture flow patterns in subgrades, which in turn influences MR results. Malla and Joshi MR of tested soils varied from 30 to For coarser soils, MR increased with deviatoric axial stress and for cohesive soils, (2006) 160 MPa; MR decreased with deviatoric stress. New England Several prediction models were also developed for these soils. Transportation Consortium (Con- necticut, Maine, Massachusetts, New Hampshire, Vermont) Titi et al. (2006) MR of tested plastic coarse soils All tests showed good repeatability. Wisconsin DOT varied from 15 to 160 MPa; MR of Test database was used to develop correlations of Level 3 type. non-plastic coarse soils varied from 20 to 140 MPa; MR of tested sub- Trends with respect to confining and deviatoric stress are in agreement with those grade soils varied from 10 to 160 reported in the literature. MPa; Kim and Labuz Four percents of RAPs were mixed MR increased with an increase in confining pressure and deviatoric stress has no (2007) with natural aggregates and resil- effect on the moduli magnitudes. Minnesota DOT ient moduli tests on aggregates Increase in RAP amount resulted in an increase of moduli. showed that the MR varied between 125 and 700 MPa. The small-strain shear modulus (G max) is typically corre- lated with resilient modulus (MR or Emax) at low strains using the following relationship:  (4) where µ is Poisson’s ratio. Figure 36 presents a typical comparison of MR from RLT test results and Emax from bender element tests. Though the bender element method provided a similar trend as that of resilient properties from RLT, they are not the same. This is because of the variations FIGURE 35  Bender element setup used by Davich et al. 2004b. FIGURE 36  Comparisons of elastic moduli from RLT and bender element tests (Davich et al. 2004b).

OCR for page 24
36 in the strains at which these properties are measured. Other for stiffer materials, such as unbound aggregates, to limitations include the difficultly interpreting travel time achieve reliable measurements of MR values. Figure 39 periods and necessity for skilled personnel and persons with presents typical resilient properties of grouted granular considerable experience to interpret these results. specimens tested by Gandara et al. (2005). • Soil specimen conditioning before MR testing is not • Regarding the test procedures used, the majority of the unique and several research studies adapted different research studies used T-294, T-307, and TP-46 meth- approaches for the testing based on their experience ods. All these methods include conditioning and testing with the moisture information of the local subgrades. cycles with different sets of confining and deviatoric The majority of the studies reported that the testing is stresses for both cohesive and granular soil types. being performed on the soil specimens as compacted. Constant modification of test methods for MR measure- Most of these soil samples are in unsaturated condi- ments also contributed to a certain lack of interest in the tions. A few studies reported the use of soaked soil resilient modulus test procedures as indicated by one of specimens (Hopkins et al. 2004; Masada et al. 2004; the survey respondents. The newer test methodologies Wolfe and Butalia 2004) by saturating the compacted such as T-307 are considered as refined test procedures soil specimen. Figure 40 presents a typical resilient for determining the MR values. However, there is no modulus measurement of a clayey subgrade at differ- good one-to-one correlation between moduli measured ent moisture content and related saturation conditions. from one method to other methods unless a few correc- A decrease of close to 70 MPa was observed in the tion factors are applied, as suggested by Mohammad et moduli value when the clayey subgrade was subjected al. (1994) and Ping et al. (2003). to full saturation from the dry compaction state. Most • The majority of the research studies used either external of the studies reported over the last 10 years also noted or internal measurements, which showed certain varia- the importance of moisture content and moisture con- tions in the MR measurements (Ping et al. 2003; Ping and ditioning on the resilient moduli properties. Ling 2007). Figure 37 presents a typical comparison of MR measurements from end and middle measurements. These studies confirm the earlier findings reported by Mohammad et al. (1994). Though internal measure- ments tend to measure displacements that are free from system compliance errors, adaptation of this method in a routine test can take considerable time. Hence, exter- nal LVDT systems are recommended for the MR mea- surements in the recent AASHTO T-307 method. • Another important source of error is the need to grout the ends of the soil specimens, because irregular surfaces result in erroneous resilient deformation measurements. Gandara et al. (2005) recommended the use of grout- ended specimens for modulus testing. Figure 38 shows photographs of grouting being performed at the ends FIGURE 37  MR measured from LVDTs placed in the middle of specimens. Such procedure is highly recommended versus ends of A-3 Soil Specimen (Ping et al. 2003). FIGURE 38  (left) Grout used for specimen ends; (right) Leveling measurements on the top cap (Gandara et al. 2005).

OCR for page 24
 37 FIGURE 39 Typical resilient moduli of granular specimen grouted at the ends (Gandara et al. 2005). FIGURE 40  Resilient moduli of subgrade specimens at different saturation conditions (Wolfe and Butalia 2004).

OCR for page 24
38 Hopkins et al. (2004) noted that the saturation of the soil ing at a certain temperature. The frozen samples were specimen represents the subgrade conditions in the state of then subjected to MR testing, followed by a thawing pro- Kentucky. This study also noted that the current procedures, cess in the laboratory setup upon which it was subjected including T-307 and TP-46, do not specify a procedure for to additional resilient modulus testing. Figure 41 pres- performing tests on soaked specimens and recommends this ents a typical resilient modulus response of subgrades as an important future research need. subjected to different freezing conditions. • In the case of soils tested in cold regions, Berg et al. Janoo et al. (1999) also reported the freezing and thaw- (1996) described resilient properties of subgrades sub- ing effects on the resilient properties of New Hampshire jected to different freezing temperatures and thawing subgrades for “unfrozen to frozen” conditions followed by conditions in the laboratory environment. Specimens “frozen to thaw” conditions. Figure 42 presents typical test were compacted, saturated, and then subjected to freez- results for a silty sand subgrade. Both studies reported two to FIGURE 41  Resilient moduli of subgrade specimens subjected to different freezing temperatures (Berg et al. 1996). FIGURE 42  Resilient moduli of silty fine sand specimens subjected to different freezing temperatures (Janoo et al. 1999).

OCR for page 24
 39 three orders of decrease in the resilient properties as a result Wolfe and Butalia (2004) focused on the suction mea- of freezing and thawing effects. This signifies the need to surements in the field using tensiometers and then addressed determine the effective roadbed modulus in cold regions by the moisture content fluctuations in both base or subbase accounting for freezing and thawing effects. and subgrade environment. These measurements raised some concerns about the assumptions of using drainable On a different, but related topic, Ooi et al. (2006) rec- conditions in the bases. The Edil et al. (2006) and Gupta et ommended against drying of subgrades because drying may al. (2007) studies were conducted for the Minnesota DOT alter the resilient responses of the soil specimen as it exhibits (MnDOT), and they addressed the need to incorporate soil different moisture-related affinity than the one without dry- suction effects into the MR modeling. Edil et al. (2006) ing conditions. described an MR framework suggested by Oloo and Fred- lund (1998) that accounts for soil suction effects. Figure 43 • Recent studies performed by Wolfe and Butalia presents the effects of matric suction on the resilient moduli (2004); Edil et al. (2006); and Gupta et al. (2007) of subgrade soils tested at various suction conditions. An focused on the unsaturated soil principles and its increase in suction showed an increase in MR value of the soil implications to the resilient modulus property and because an increase in suction is always associated with dry the related mechanistic pavement design. The major- conditions in the soil specimens. ity of the subgrades are expected to be in unsaturated conditions for most of their design life and hence the This research is an important step in the better under- suction of subgrades in the unsaturated state plays an standing of resilient properties of unsaturated soils. More important role on both resilient properties and pave- such understanding will help in better mechanistic design ment design principles. of pavements built on subgrades in regions where full satu- FIGURE 43  Effect of matric suction on resilient modulus (Edil et al. 2006).