1triaxial

TRANSPORTATION RESEARCH RECORD 1577 Paper No. 970314 27

Triaxial Characterization of MinnesotaRoad Research Project Granular Materials

NAVNEET GARG AND MARSHALL R. THOMPSON

Department of Civil Engineering, University of Illinois at Urbana-Champaign,205 North Mathews Avenue, Urbana, Ill. 61801.

Six granular materials were used as base and subbase materials in the flex-ible pavement test sections for the Minnesota Road Research (Mn/ROAD)project. Crushed/fractured particles are not allowed in aggregate classesCL-1Fsp, CL-1Csp, CL-3sp, and CL-4sp. Ten to 15 percent crushed/frac-tured particles are required for CL-5sp. One hundred percent crushed/fractured particles are required for CL-6sp. A comprehensive laboratorytesting program was established to determine pertinent engineering prop-erties of the granular materials. Rapid shear tests and repeated-load testswere conducted to determine the shear strength parameters (friction angleand cohesion), resilient modulus, rutting potential, stress history effects onshear strength, and moisture susceptibility. The results from the rapidshear tests and permanent deformation tests show that the rutting poten-tial of a granular material can be characterized from rapid shear test at aconfining pressure of 15 psi (103.35 kPa). The rutting parameter A was afunction of the shear strength of the granular materials. The shear strengthresults obtained from rapid shear tests performed at a confining pressureof 15 psi reflect the rutting trends observed in the low-volume road testsections at the Mn/ROAD project. Results from repeated-load tests wereused to develop the parameters for K-θ, UT-Austin, and Uzan’s modelsfor evaluating the resilient modulus of granular materials. The axial strainvalues calculated from the resilient modulus models appear to be in goodagreement with the measured axial strain values, except for the very lowshear strength material CL-1Csp.

Unbound aggregate materials, such as crushed stone and gravel, areused as surface layers, bases, and subbases. The load-deformationresponse of unbound aggregates is an important pavement designconsideration. Both permanent and resilient deformation character-istics are important. The shear strength of unbound materials is alsoimportant relative to the behavior and performance of the material asa pavement layer. Since unbound granular materials have little or notensile strength, shearing resistance of the material is used to developa load-distributing quality that greatly reduces the stresses transmit-ted to the underlying layers. Some important factors influencing theshear strength of untreated granular materials are gradation, mois-ture and density, maximum particle size, amount and plasticity offines, particle geometric properties, and confining pressure. Uponapplication of vertical load to a granular layer, deformation occurs.The deformation includes two components: resilient (or recoverable)deformation, and permanent (or nonrecoverable) deformation.

Over the years, the University of Illinois (U of I) has developedand successfully used a triaxial testing procedure for characterizingand evaluating granular materials (1). In the first phase, a triaxialshear test is performed at a rapid shearing rate to determine the shearstrength of material. In the second phase, a specimen is subjected to1,000 load repetitions at 45-psi (310-kPa) deviator stress and 15-psi(103.35-kPa) confining pressure (referred to as the conditioningstage). The second stage consists of subjecting the “conditioned”specimen to 100 load repetitions at different stress states (described

later) for measuring resilient modulus. In the third stage, followingresilient modulus testing, the specimen is subjected to rapid sheartest at 15-psi (103.35-kPa) confining pressure to establish the stresshistory effects on shear strength.

The resilient modulus of granular materials is an important inputvariable for the design of pavement structures. Statistically devel-oped models (from laboratory test results) are used to characterizethe resilient behavior of granular materials. The K-θ model has beenthe most popular model used in granular material characterization.Even though it is a popular material, the K-θ model neglects theimportant effect of shear stress and shear strain on the resilient mod-ulus (2,3). The UT-Austin model presented by Pezo (4) predicts theaxial strain and includes in the prediction parameters the confiningpressure and the deviator stress instead of resilient modulus. Sinceaxial strain is the response variable in the model, the model is sta-tistically sound because the prediction variables are independentfrom the response variables.

This paper summarizes the results obtained from comprehensivelaboratory testing of the granular materials used as base and subbasein the Minnesota Road Research (Mn/ROAD) project test sections.Mn/ROAD is the largest and most technologically advanced researchfacility in the world. The test facility is located parallel to Interstate94 in Otsego, Minnesota. The pavement sections are designed so thatdifferent combinations of materials, layer thicknesses, and designdetails can be evaluated Rapid shear tests and repeated-load testswere conducted to determine the shear strength parameters (frictionangle φ, and cohesion c), resilient modulus (ER), rutting potential,stress history effects on shear strength, and moisture susceptibility.

MATERIALS TESTED

Six granular materials were used as base and subbase materials inthe flexible pavement test sections of the Mn/ROAD project. Thematerial specifications are given in Table 1. Crushed/fractured par-ticles were not permitted in aggregate classes CL-1Fsp, CL-1Csp,CL-3sp, and CL-4sp. Ten to 15 percent crushed/fractured particleswere required for CL-5sp; 100 percent crushed/fractured particleswere required for CL-6sp. The specifications for plasticity index andliquid limit are given in Table 1. Laboratory testing showed that allthe materials are nonplastic.

TESTING PROGRAM

Specimen Preparation

Rapid shear tests and repeated-load tests were conducted on cylindri-cal specimens 6 in. (152.4 mm) in diameter and 12 in. (304.8 mm)

28 Paper No. 970314 TRANSPORTATION RESEARCH RECORD 1577

TABLE 1 Mn/ROAD Base and Surfacing Aggregate (SP 8680-123)

high. A split aluminum mold was used for preparing the specimens.A 31-mil (0.787-mm-thick) neoprene membrane was placed insidethe mold. The material (aggregate mixed with the required amount ofwater) was compacted in the mold in five lifts. Target moisture con-tents and densities were selected on the basis of AASHTO T99 testresults and the field-measured values. A pneumatic vibratory com-pactor was used for compaction. Specimen density was monitored bymeasuring the compacted thickness of each lift. The final height anddensity of specimen were noted after compaction. The split aluminummold was then removed, and a 25-mil (0.635-mm-thick) latex mem-brane was placed on the specimen. The second membrane (latexmembrane) was required because the neoprene membrane was gen-erally punctured while compacting the specimen. The drainage portwas left open and the test performed under drained conditions.

Testing Equipment

A Material Testing System (MTS) closed-loop electrohydraulic sys-tem, Model 407, was used for testing. The main part of the systemconsists of controller, loading frame, and hydraulic power supply.The system is fitted with a ram capable of applying a load up to 10kips. The ram is fitted with an internal linear variable differentialtransformer (LVDT). The MTS-407 controller provides the elec-tronics for closed-loop control and controls the system operation.An IBM Personal Computer AT, fitted with an eight-channel DataTranslation 2801-A analog:digital (A/D) board, triggered the MTSand recorded the data. Two external LVDTs, mounted on the topplate of the triaxial chamber, were used to measure the displace-ments during the resilient modulus tests. A T-bar was attached to theloading piston that actuated the external LVDTs.

Rapid Shear Testing

A triaxial shear test performed at a rapid shearing rate is more rep-resentative of highway loading conditions than the conventionalslow triaxial shear test (strain rate of 1 to 3 percent per minute).Rapid shear tests were performed at confining pressures of 5, 10, 15,20, and 30 psi (34.45, 68.9, 103.35, 137.8, and 206.7 kPa) to deter-mine the friction angle (φ) and cohesion (c) used to define the Mohr-Coulomb failure envelope. Deviator stress was applied at a constantdisplacement rate of 1.5 in./sec (strain rate of 12.5 percent /sec, 5 percent strain in 400 msec) for a 12-in. (304.8-mm) specimen.

The Mohr-Coulomb envelope was determined by regression tech-nique using the failure deviator stress and confining pressure. Linearregression was performed to obtain a best fit equation of the form

σ σ1 3= +a b p

where σ1 is major principal stress and σ3 is minor principal stress.Cohesion (c) and angle of internal friction (φ) were evaluated asfollows:

Conditioned rapid shear tests evaluate the effect of stress history onthe shear strength of the material. After the completion of therepeated-load resilient modulus sequences, the sample was subjectedto a rapid shear test at a confining pressure of 15 psi (103.35 kPa).The peak shear strength of the unconditioned sample and that of theconditioned sample were compared.

Repeated-Load Testing

Conditioning and Permanent Deformation Testing

Conditioning cycle data indicate rutting potential. The specimenswere conditioned for 1,000 load repetitions at a deviator stress of45 psi (310 kPa) and a confining pressure of 15 psi (103.35 kPa)(stress state referred to as 45/15). Some materials did not surviveconditioning at 45/15. They were then conditioned at a 30-psi(206.7-kPa) deviator stress and 15-psi (103.35-kPa) confiningpressure (30/15). CL-1Csp was conditioned at 15/10. Permanentdeformation, resilient deformation, and applied deviator stressmeasurements were made at 1, 10, 50, 100, 500, and 1,000 loadapplications. The following model was used to characterize therutting potential:

where N is the number of load repetitions and A is the antilog of a in

where b represents the percentage of strain accumulated per logcycle.

Resilient Modulus Testing

Specimens were subjected to various repeated triaxial stress statesless than failure. A haversine load waveform was applied [pulseduration of 0.1 sec (10 Hz)/rest period of 0.9 sec]. After condition-ing, modulus testing was conducted at various stress states given inTable 2. The sample was subjected to 100 load repetitions at each

log % loge p a b N= +

e pbA N% =

c a bb b

= [ ]= −( ) +( )[ ]−

/sin /

21 11

p

φ

Deviator Stress, σd [psi (kPa)] Confining Pressure, σ3 [psi (kPa)] No. of Load Repetitions

45a (310) 15a (103.35) 1,000b

10 (68.9) 5 (34.45) 10015 (103.35) 5 (34.45) 10020 (137.8) 10 (68.9) 10030 (206.7) 10 (68.9) 10030 (206.7) 15 (103.35) 10045 (310) 15 (103.35) 10045 (310) 30 (206.7) 10060 (413.4) 30 (206.7) 100

a Standard stress state for conditioning. Some materials conditioned at 30/15 and 15/10.b Sample conditioning.

Garg and Thompson Paper No. 970314 29

sequence (stress state). If the difference between the modulus val-ues at the 50th and 100th load repetition was more than 5 percent,the sequence was repeated.

The resilient modulus for each stress state was evaluated as follows:

where

Er = resilient modulus (psi),σd = applied deviator stress (psi), ander = resilient (recoverable) strain.

LABORATORY TEST RESULTS

Rapid Shear Testing

A summary of the rapid shear test results is given in Table 3. Figure1 shows the peak deviator stress achieved at a confining pressure of15 psi (103.35 kPa) for different materials. Materials CL-1Csp andCL-1Fsp were not tested at higher moisture contents because of thedifficulty in preparing specimens. Considering the results from rapidshear tests at confining pressure of 15 psi, materials can be placedinto three groups:

• Group 3: peak deviator stress less than 60 psi (413.4 kPa); CL-1Csp.

Er d r= σ /e

• Group 2: peak deviator stress between 60 psi and 120 (413.4and 826.8 kPa); CL-1Fsp, CL-3sp, CL-4sp, CL-5sp.

• Group 1: peak deviator stress higher than 120 psi (826.8 kPa);CL-6sp.

Variable moisture sensitivity was observed Materials CL-1Cspand CL-1Fsp were most sensitive. CL-1Csp and CL-1Fsp couldnot be tested at higher moisture contents because of difficultiesencountered during specimen preparation. For CL-3sp, a changein moisture content from 8 to 6.8 percent resulted in no change inthe friction angle. However, a reduction in peak deviator stressfrom 115 to 94 psi (792.35 to 647.66 kPa) was observed. For CL-4sp, a reduction in moisture content from 9.4 to 7.9 percentresulted in a 45 percent increase in the friction angle. The increasein peak deviator stress at 15-psi (103.35-kPa) confining pressurewas 17 percent. For CL-5sp, when the moisture was reduced from7.7 to 6.8 percent, the friction angle increased from 39 to 43 degrees (10 percent increase) and peak deviator stress increasedby 42 percent. For CL-6sp, an increase in moisture content from5.3 to 6.3 percent resulted in a decrease in the friction angle from51 to 47 degrees. No change was observed in the peak deviatorstress at 15 psi confining pressure.

Permanent Deformation Testing

Specimens were subjected to a repeated-load deviator stress of 45 psi (310 kPa) and a confining pressure of 15 psi (103.35 kPa),

TABLE 2 Stress States for Resilient Modulus Testing

TABLE 3 Results from Rapid Shear Testing on Mn/ROAD Granular Materials


FIGURE 1 Rapid shear test results for confining pressure of 15 psi.

for 1,000 load repetitions. This stress state is referred to as 45/15.Some specimens did not survive conditioning at 45/15 and wereconditioned at lower stress state. Material CL-1F was conditionedat 30/15. Material CL-1C showed the highest rutting potential andwas conditioned at 15/10. Table 4 gives the A and b values (in themodel ep % =A Nb) for different materials at different moisture anddensity levels. Higher A and b values represent increased ruttingpotential.

Resilient Modulus Testing

Resilient modulus test data were used to develop K and n param-eters for the K-θ model. Table 5 gives the K and n values for materials tested. The following relationship (5) was establishedbetween K and n:

The K-n relationships obtained by Rada and Witczak (5) and U of Iare shown in Figure 2. The results of the Rada and Witczak studycame from a broad data base and various testing procedures.

log . . .K n R= − =3 996 0 0893 0 812p

DISCUSSION OF LABORATORY TEST RESULTS

Rapid Shear and Permanent Deformation Tests

The materials that developed at least 90-psi (620.1-kPa) deviatorstress at 2 percent axial strain, survived conditioning at 45-psi (310-kPa) deviator stress and 15-psi (103.35-kPa) confining pres-sure. Parameter A is a function of deviator stress at 1 percent axialstrain obtained from the rapid shear tests conducted at the confiningpressure of 15 psi (103.35 kPa). The relation is

where σd1% is the deviator stress in pounds per square inch at 1 per-cent axial strain. Figure 3 shows parameter A as a function of devi-ator stress at 1 percent axial strain. The b parameter generally varieswithin fairly narrow limits (0.1–0.2)(6–8). For specimens condi-tioned at 45/15, the b values ranged from 0.08 to 0.24 (Table 4). Forlow-shear-strength materials (conditioned at 30/15 and 15/10), theb values were in the 0.31–0.45 range.

For materials conditioned at 45/15, an attempt was made todevelop a correlation between b and the shear strength of material.The correlation is

A Rd= − =1 10386 0 007911 0 9712. . .%p σ


where σd2%, σd3%, and σd4% are deviator stresses in pounds persquare inch at 2, 3, and 4 percent axial strains. Even though the R2

value for the correlation is very high, the 3.0172 * log(σd3%) termdefies the engineering logic that rutting potential decreases withincrease in shear strength. Several other combinations of more thanone “stress term” were tried. There was always some “term” in theregression equation that indicated “increased rutting” with increasein shear strength.

If only Materials CL-3sp and CL-4sp (no crushed/fractured particles) are considered, the following regression equation isobtained:

where σd1% is the deviator stress at 1 percent axial strain.A similar relationship could not be developed for materials with

crushed/fractured particles (CL-5sp, CL-6sp), as only CL-6sp (100 percent crushed/fractured particles) survived conditioning at45/15. Material CL-5sp (15 percent crushed/fractured particles)developed a peak shear strength of 115 psi (792.35 kPa) at 15-psi

b Rd= − =0 5987 0 00616 0 9912. . .%p σ

bR

d d

d

= − ( ) + ( )− ( ) =0 6205 2 6916 3 0172

0 5905 0 9962 3

42

. . log . log. log .

% %

%

p p

p

σ σσ

(103.35-kPa) confining pressure; still the specimen failed duringconditioning at 45/15 because the deviator stress developed at 2 percent axial strain was less than 90 psi (620.1 kPa).

It is apparent that there are no clear, overall, and comprehensiverelations for estimating the b term. A is the dominant term in therelationship ep % = A Nb and can be estimated accurately from therapid shear test results performed at 15-psi (103.35-kPa) confiningpressure.

Resilient Modulus Tests

The resilient modulus test data were used to develop parameters forthree (K-θ, UT-Austin, Uzan) granular material resilient modulusmodels. These models are used to estimate the resilient modulus ofgranular material as a function of stress state

K-θ Model

Linear regression was performed to obtain a best fit equation of theform

E a nR = + ( )log θ

TABLE 4 Results from Permanent Deformation Testing on Mn/ROAD Granular Materials

TABLE 5 Results from Resilient Modulus Testing


FIGURE 2 Relationship between resilient modulus parameters K and n.

where

θ = bulk stress (σ1 + 2 * σ3),σ1 = major principal stress, andσ3 = minor principal stress.

The response was transformed into the following model:

where K is the antilog of a. The stress sensitivity is depicted by n. TheK and nparameters for different materials tested are given in Table 5.

Uzan’s Model

Linear regression was performed to obtain a best-fit equation ofthe form

E KRn= p θ

where θ is the bulk stress and σd is the deviator stress. The responsewas transformed into the following model:

where K3 is the antilog of a. The values of K3, K4, and K5 for dif-ferent materials tested are given in Table 5.

UT-Austin Model

Linear regression was performed to obtain a best-fit equation ofthe form

E KRK

dK= 3 4 5p plog θ σ

E a K KR d= + ( ) + ( )4 5p plog logθ σ

FIGURE 3 Rutting model parameter A as function of deviator stress.


where

ei = measured resilient axial strain,σd = deviator stress, andσ3 = minor principal stress.

The response was transformed into the following model:

where

N6 = 102a,N7 = 1 – K7, andN8 = –K8.

The values of N6, N7, and N8 for different materials are given inTable 5.

The R2 values for the UT-Austin model were comparativelyhigher than for the K-θ model and Uzan’s model. The axial strainswere calculated from the estimated modulus values from the threemodels and were compared with the measured axial strains andresilient modulus values. The axial strain values calculated from theresilient modulus models were in good agreement with the mea-sured axial strain values, except for the very low shear strengthmaterial CL-1Csp. Figures 4 and 5 show the resilient modulus–axialstrain relationship for CL-1Csp and CL-6sp, respectively.

E NR dN N= [ ][ ]6 7

38σ σ

log log logea da K K= + ( ) + ( )7 8 3p pσ σ LABORATORY SHEAR STRENGTH RESULTS ANDOBSERVED RUTTING IN FIELD

The low-volume road test sections at Mn/ROAD are loaded by afive-axle tractor-trailer. In one lane, the tractor-trailer is loaded to80,000-lb (16-kip drive axle, 32-kip dual tandems on the trailer,100-psi tire pressure). In the other lane, the tractor-trailer travels inthe opposite direction loaded to 102,500 lb (13.2-kip drive axle, 41.5-kip front dual tandems on the trailer, 47.8-kip rear dual tandemon the trailer, 100-psi tire pressure). The number of equivalent single-axle loads (ESALs) corresponding to each pass of the tractor-trailer is presented in Table 6 (per The Mn/ROAD data base). Table6 gives the Mn/ROAD low-volume road test sections and the gran-ular materials used. The unsurfaced test sections Cell 33 (LVR-A1)and Cell 35 (LVR-A3) experienced severe rutting in the first fewweeks of trafficking. Severe washboarding and corrugationsoccurred in Cell 35. The results from rapid shear tests showed CL-1Csp to be material of very low shear strength (50 psi, Table 3),compared with 100-psi (689-kPa) tire pressure, and to have veryhigh moisture susceptibility. Even though CL-1Fsp achieved a peakdeviator stress of 87 psi (599.43 kPa) at 15-psi (103.35-kPa) con-fining pressure (Table 3), the material did not perform well in thefield because of high moisture susceptibility (discussed earlier). Thepoor performance of test sections LVR-A1 and LVR-A3 was pre-dicted after rapid shear and permanent deformation tests were con-ducted in the laboratory. Test sections LVR-A1 and LVR-A4(CL-1Csp granular material) experienced higher rutting than test

FIGURE 4 Axial strain–resilient modulus relationship for CL-1Csp.


FIGURE 5 Axial strain–resilient modulus relationship for CL-6sp.

sections LVR-A3 and LVR-A2 (CL-1Fsp granular material),respectively, as predicted by the laboratory tests.

Figure 6 shows the rut depth measurements on the lane loadedby an 80,000 lb tractor-trailer. The 80,000-lb lane had experienced16,553 passes (38,700 ESALs) of the tractor-trailer. The resultsfrom rapid shear tests reflect the rutting trends observed in thefield, except for test section LVR-F11 (Cell 27), which experi-enced higher rutting than expected. Similar trends were observed(Figure 7) on the lane trafficked by the tractor-trailer loaded to102,500 lb. The 102,500-lb lane had experienced 5,812 passes(40,000 ESALs) of tractor-trailer.

The following model was used to characterize rutting in the testsections:

where RD is rut depth in inches and N is the number of tractor-trailerpasses. (In Table 7, when the parameters A′ and Bare evaluated usingnumber of passes, N in the preceding equation is the number ofpasses of the tractor-trailer, and when the parameters A′ and B areevaluated using number of ESALs, N in the equation is the numberof ESALs.) The A′ and B values for test sections are summarized in

RD = ′A N B

TABLE 6 Low-Volume Road Test Sections at Mn/ROAD Project


Table 7. Table 7 gives the rut depths measured on July 22, 1996.This is the last set of useful rut depth data on the LVR test sectionsbecause the rut depth profiles were disturbed by the trucks carryingmaterial for the rehabilitation of aggregate test sections. The lanewith 80-kip loading experienced higher rutting than the 102.5-kiplane. No particular trends were observed in the A′ and B values forthe 80 and 102.5-kip lanes.

Except for Cell 31, the A′ values were lower for the 102.5-kiplane. For Cell 31, the A′ value was higher and the B value was lowerfor the 102.5-kip lane.

Rutting can occur in the asphalt concrete surface, granular base,and subgrade. Khedr (9) showed that the rutting parameter Aa (in themodel ep/N = Aa * N–m) is a function of the resilient modulus and theapplied stress. A study conducted at University of Minnesota (10)on the Mn/ROAD asphalt concrete mixes showed that all the threemixes (35, 50, and 75-blow) had similar modulus values. Therefore,the asphalt concrete mixes should show similar rutting trends.

Analysis of falling weight deflectometer (FWD) tests showed thatthe backcalculated subgrade “breakpoint” modulus (ERi) valueswere similar for all the test sections. The trends observed for theparameter A from laboratory tests on granular bases reflect thetrends for parameter A from the field results. This suggests that therutting is probably occurring in the granular base layer. RecentMn/ROAD staff trenching studies in the aggregate test sectionsshowed that the rutting was primarily in the granular base and not inthe subgrade.

CONCLUSIONS

Laboratory test results on granular materials used in the Mn/ROADlow-volume road test sections have been presented. The resultsfrom the rapid shear tests and permanent deformation tests showthat the rutting potential of a granular material can be characterized

FIGURE 6 Rut depth measurements on LVR Mn/ROAD test sections on 80,000-lb lane.

FIGURE 7 Rut depth measurements on LVR Mn/ROAD test sections on 102,500-lb lane.


from rapid shear test at a confining pressure of 15 psi (103.35 kPa).The stress-strain curve from the rapid shear test is used to predictthe rutting potential. Granular material that achieves at least 90-psi(620.1-kPa) deviator stress at a confining pressure of 15 psi, dis-plays a low rutting potential The shear strength results obtainedfrom rapid shear tests performed at a confining pressure of 15 psiappear to reflect the rutting trends observed in the low-volume roadtest sections at the Mn/ROAD project.

Results from repeated-load testing were used to develop the pa-rameters for K-θ, UT-Austin, and Uzan models for characterizingthe resilient modulus. The estimated axial strain and resilient modulus values from the three models are in good agreement withthe measured values. Less agreement between the measured and estimated axial strain and resilient modulus values was noted forCL-1Csp (poor-quality, very low shear strength material).

The aggregate layer in pavement must possess enough shearstrength and rutting resistance (for a given asphalt concrete thick-ness) to minimize rutting within the layer. Adequate asphalt con-crete and granular layer thicknesses must be provided to protect thesubgrade. This paper demonstrates that rapid shear and repeated-load triaxial testing can be used to predict and rank the permanentdeformation behavior of granular materials.

ACKNOWLEDGMENTS

This paper was sponsored by the Illinois Department of Trans-portation (IDOT) Division of Highways as part of a cooperativeeffort between the Minnesota Department of Transportation(Mn/DOT), IDOT, and the Department of Civil Engineering, Uni-versity of Illinois at Urbana-Champaign, to provide broad utiliza-tion of the data obtained from the Mn/ROAD project. Thecooperation is facilitated through a memorandum of understandingbetween Mn/DOT and IDOT. The generous cooperation and help ofthe MnROAD staff are acknowledged, particularly the efforts ofThomas Burnham, David Palmquist, and David Van Deusen.

REFERENCES

1. Thompson, M. R., and K. L., Smith. Repeated Triaxial Characterizationof Granular Bases. In Transportation Research Record 1278, TRB,National Research Council, Washington, D.C., 1990, pp. 7–17.

2. May, R. W., and M. W. Witczak. Effective Granular Modulus to ModelPavement Responses. In Transportation Research Record 810, TRB,National Research Council, Washington, D.C., 1981, pp. 1–90.

3. Uzan, J. Characterization of Granular Material. In TransportationResearch Record 1022, TRB, National Research Council, Washington,D.C., 1985, pp. 52–59.

4. Pezo, R. F. A General Method of Reporting Resilient Modulus Tests ofSoils—A Pavement Engineer’s Point of View. Presented at the 72ndAnnual Meeting of the Transportation Research Board, Washington,D.C., 1993.

5. Rada, G., and M. W. Witczak. Comprehensive Evaluation of Labora-tory Resilient Moduli Results for Granular Material. In TransportationResearch Record 810, TRB, National Research Council, Washington,D.C., 1981, 23–33.

6. Calibrated Mechanistic Structural Analysis Procedures for Pavements,Vol. 2. Appendices. National Cooperative Highway Research ProgramProject 1-26. TRB, National Research Council, Washington, D.C.,March 1990.

7. Garg, N., and J Budiman. Effect of Large Size Aggregates on the Per-formance of Base Materials for Flexible Pavements. Presented at the73rd Annual Meeting of the Transportation Research Board, Washing-ton, D.C., 1994.

8. Thompson, M. R., and D. Nauman. Rutting Rate Analyses of theAASHO Road Test Flexible Pavements. In Transportation ResearchRecord 1384, TRB, National Research Council, 1993, pp. 36–48.

9. Khedr, S. A. Deformation Mechanism in Asphaltic Concrete. Journalof Transportation Engineering, ASCE, Vol. 112, No. 1, Jan. 1986.

10. Stroup-Gardiner, M., and D. E. Newcomb. Investigation of Hot MixAsphalt Mixtures at Mn/ROAD—Final Report. Minnesota Departmentof Transportation, July 1996.

The contents of this paper reflect the views of the authors, who are respon-sible for the facts and accuracy of the data presented. The contents do notnecessarily reflect the official views or policies of IDOT or Mn/DOT. Thispaper does not constitute a standard, specification, or regulation.

Publication of this paper sponsored by Committee on Mineral Aggregates.

TABLE 7 Results from Rut Depth Measurements at Mn/ROAD Test Sections

Documents

1triaxial