Click for next page ( 11


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 10
CHAPIER 2 RESILIENT MODULUS TESTING OF ASPHALT CONCRETE INTRODUCTION This chapter evaluates resilient modulus testing methodology attest details for asphalt concrete. The measurement of He resilient properties of asphalt concrete has been the subject of considerable research. Different testing devices and techniques have been used in these studies. All of these efforts have led He American Society for Testing and Materials to standardize the resilient modulus testing method of asphalt concrete (ASTM D 4123-82~. However, as demonstrated in the "Workshop on Resilient Modulus Testing" held at Oregon State University in March 1989, there was strong consensus among pavement engineers that the ASTM D 4123 nroce`1ure is ~,nnen~nrilv fim~.-~nn~,mins' ~nr1 that the. tP..Ct results are difficult to reproduce. Recognizing the importance and existing problems of resilient modulus testing of asphalt concrete, the Strategic Highway Research Program (SHRP) has developed a resilient modulus test procedure for asphalt concrete (SHRP Protocol P07) as a part of He Long Term Pavement Performance Monitoring CUTUP) program. This testing procedure incorporates recent findings on resilient modulus testing into He existing ASTM D 4123-82. A comparison between ASTM D 4123 and the November 1992 version of SHRP Protocol P07 is summarued in Table I. An important overall objective of this study is to develop laboratory resilient modulus testing procedures suitable for use by a state transportation agency. To help achieve this goal, the emphasis of the study was placed on evaluating the effects on resilient modulus of laboratory testing details such as, for example, equipment calibration and testing conditions. Detailed laboratory studies were therefore carried out using different devices to evaluate the effect of laboratory test apparatus and testing details on resilient modulus test results. Based on the findings from the present study, a number of revisions are suggested to SHRP Protocol P07 (November, 1992~. METHODS FOR DETERMINATION OF MODULUS OF ASPHALT CONCRETE The resilient modulus of asphalt concrete has in the past been determined by two approaches: (~) predict the resilient modulus using physical and mechanical properties of He mixture using available correlations, and (2) measure the resilient modulus by laboratory testing. Empirical Predictive Methods The most well-known predictive methods are the Marshall stability-flow ratio, the Shell Nomograph, and the Asphalt Institute predictive model. Nijboer [~] suggested He use of the Marshall stability-flow ratio as follows: S60C 4 S4tC = ~ .6(stability/flow) where S is the modulus given in kilograms per square centimeter, stability in kilograms, and flow in 10

OCR for page 10
~o o - ~ ~- z ~ ~ ~ ~j ~ 2 ~ ~ e ~- ~ ~. 2 ~ , ~ . . ~ . . ~ . ~ I ~ ~ , L o ~ _ .c , E 11

OCR for page 10
- 5 o C) _' E 1 0 ;- 3 :, ~:r O ~OCL ~ ~ ~o ~ to C _ R ~ ~ ~ ~ ~ ~ - A (I.) :r ;- _ ~ C C ~; 3 o Id, ~ D ~ 0 ~.0 _ a o ~C .s ~ C ~ ~ ~ C ~ ~ U. U. ~ ~7 lo o - Hi o A - - et A: . Cal . U. . U. 8 o o V) o 2 ._ C~ - o O - c Ct c s ~ ~ U) ._ o C ~ o ,~ .E ~ ~ ~ o o ~ - o ~ :5 ~ ~ E * ~ * ~* 12 ~ U:) 04 0 o ._ Ct ~ _ O C~ C~ U. :: ~ ~ O C~ ~ e,a - _ C~ Ct ~ ~ ~ ~_ =, =0, 0 0 ~ ~n _ ~ ., ~ ~ = t_ - ;- 3 ~ C~ - - = ~ ~ - ~ - o c ~- o ~ ~ ~ ~ ~ - ~ s - ~o ~ e~ ce > o c ~ ~ oD ~ c ~o ~ s ~ ~ o ~ > o c: 3 ~ ~ U, _ U, _ C~ ~a S D - u, - s c~ 5 > 0 ~ ~ 3 _ ~ E . ~ ~ ~ o ~ ~ o ~o -~ C~ ~ ', ~ ~ ~ ~ _ o ~ ~ ~ ~ ~ ~ Z ~ s~ C~ ~ es =, . ~ o ~ ~- 1 ~ Q ,1 ~ ._ 8. ~ ~O O ~_ - ,~ t ~C~ O ~ o: ~E ~ U, ~o ._ E ~, _ . ._ _ Z 3 ~0 ~

OCR for page 10
be - - ~ in ~ ~ O - - c~ cat c ~ c) in - en - c~ ~ - ~ [L, 3 ~ ~ _ t- Cal t_ en C ~ ~ S - tell O ~ ~ a_ C- ~ Cal cat E ~ ~ . _ ~ ._ ~ en ~ ~ ~ C ~ ~ ._ C) D ._ us -3 E cat sE ~I'd ~' 8 3 ~ 4,, _ ~ I, ~ o ~ 8 ~ - ~. i~ ' ~ E j &= 5 C o ~1 = ~ ~ = ~ o , I,,, c: * * * * * G) Us O _ ._ By' ~~ AS 13

OCR for page 10
millimeters. This relationship was recommended for use in high temperature ranges by Heukelom and Klomp 191. McLeod [101 modified this equation using He English units: Modulus = 40(stability/flow) where modulus is given in pounds per square inch, stability in pounds, and flow in inches. Shell Nomograph. The Shell Nomograph was originally developed by Van der Poel [111. He defined tile stiffness as a modulus which is a function of temperature and loading time. Later Heukelom and Klomp [91 developed a relationship between He bitumen stiffness and He mixture stiffness based on volume concentration of aggregates. After McLeod [10] modified the nomograph by changing the entry temperature criterion, finally CIaessen et al. [121 produced a pair of nomographs used in the current Shell design manual. where E* = P = ac P - opt C - 1 - C2 = P - 200 - f = V = v n(106,70) = T = Asphalt Institute Method. The Asphalt Institute resilient modulus method was originally developed by Kallas and Shook [131 using cyclic biaxial test results. Their equation was refined by Witczak [141 from an expanded data base which relates dynamic modulus of asphalt concrete with percentage passing the No. 200 sieve, loading frequency, volume of voids, viscosity of asphalt cement at 70F, temperature, and percentage of asphalt cement by weight of mix. Since this data base was based on mixtures of crushed stone and gravel, Miller et al. [151 modified the equation for a broader range of material types. The final form of the equation by Miller et al. [15] is: logic ~ E ~ = C1 + C2 (pa pop + 4.0)0 s dynamic modulus (1Os psi) percentage of asphalt cement by weight of mix optimum asphalt content (2, 0.553833 + 0.028829(P2,,o/f l733) - 0.03476Vy + 0.070377~(106,70) + (0.93 1757/f 2774) 0.000005 T exp(1.3 + 0.498251ogl00 - 10.00189 T exp (1.3 +49825lglof~lfl ll percentage passing the No. 200 sieve loading frequency (Hz) volume of voids viscosity of asphalt cement at 70F (megapoises) temperature of pavement (F) 14

OCR for page 10
The Asphalt Institute Method Laboratory Test Methods i' s available in the form of an easy-to-use computer program. In He preceding section, the modulus of asphalt concrete was presented in several different forms including dynamic modulus, stiffness, and resilient modulus. In the following sections, typical laboratory testing methods are discussed for the determination of resilient modulus for asphalt concrete. Stress State. The stiffness (modulus) characteristics of asphalt-bound materials can be considered not to be significantly influenced by stress state at moderate to low temperatures. However, at temperatures above 25C the stress state, and therefore test configuration, have an influence on the stiffness characteristics of these materials. This influence becomes more pronounced as the binder becomes less stiff [161. Anisotropic Behavior. The determination of He resilient modulus of asphalt concrete involves using various types of repeated load tests. The most commonly used tests are as follows: Uniaxial tension test Uniaxial compression test Beam flexors (bending or rotating cantilever) test 4. Indirect diametral tension test S. Triaxial compression test A pavement layer has cross anisotropy in which radial properties are constant in all directions but are different from properties in He vertical direction. Wallace and Monismi~ [10] have claimed ~at, for an adequate description of the resilient characteristics of such a material, the following five parameters are required: 2. 3. 4. 5. Vertical strain due to an increment in vertical stress Radial strain due to an increment in vertical stress Radial strain due to an increment in radial stress Vertical strain due to an increase in radial stress Radial strain due to an increment in radial stress in a direction perpendicular to the strain They reported that the biaxial test measures the first and sometimes the second parameter whereas the diametral test measures a composite of the third and Fox parameters with roughly equal weight being given to each parameter. Due to anisotropy of asphalt concrete, the resultant discrepancy in resilient modulus between diametral testing and biaxial testing can be quite pronounced. Wallace and Monismith [171 carried out tests on an asphaltic concrete core taken from San Diego test road [id. They showed Hat as a result of placement and compaction efforts, the material was about twice as stiff in He radial direction as in the vertical direction. An asphalt layer of typical thickness is subjected to a bending action which is primarily resisted by He radial rather Man He vertical stiffness of the asphalt layer. Therefore for vertical cores taken from the pavement, the diametral test or flexural bending test should give a more relevant assessment of the 15

OCR for page 10
stiffness of the asphalt layer than tests performed in He vertical direction. Diametral test results are hence particularly attractive for evaluating radial tensile strain for a fatigue analysis. The diametral test has additional advantages since thin cores can be tested which permits more measurements over the depth of thick asphalt layers. FIexural Test. Early work to evaluate the resilient modulus of asphalt concrete was conducted by testing beam specimens under a third-point loading configuration. The flexural stiffness of beam specimens can be determined from the following equation: Pa `3L2 _4a2, where Es = 48 f ~3 Es = flexural stiffness (nisi) P = repetitive load applied on the specimen (Ib) a = ~h(L~) (in.) = reaction span length (in.) = moment of inertia of beam cross section (in.4) ~= measured deflection at the center of the beam specimen (in.) A number of different flexural test procedures have been developed to study the resilient and fatigue characteristics of asphalt concrete mixtures including: FIexure tests in which the loads are applied repeatedly or sinusoidally under center-point or third-point load Rotating cantilever beams subjected to sinusoidal loads Trapezoidal cantilever beams subjected to sinusoidal loads or deformations The advantages of the flexure test are [191: (~) it is well known, widespread in use, and readily understood; (2) The basic technique measures a fundamental property that can be used for both mixture evaluation and design; (3) Results of controlled-stress testing can be used for the design of thick asphalt pavements whereas results of controlled-strain testing can be used for the design of thin asphalt pavements. The method, however, is costly, time consuming, and requires specialized equipment [191. Also, the stress state within the pavement structure is biaxial, whereas the state of stress is essentially uniaxial in the flexure test. Triaxial Test. Numerous advantages are inherent in using the cyclic or repeated load biaxial test memos. The stress system that acts upon a specimen during the biaxial test approaches the system of stresses that are present in the upper portion of Be asphalt concrete layer of a pavement during loading. Furthermore, the strength of asphalt concrete can be determined when specimens are tested to failure under a single loading. The chief objections to the use of this method are its cost and the relative complexity of Me necessary testing equipment. In addition, the size of specimens required for testing of coarse aggregate mixtures and number of specimens needed for a test series discourage the adoption of the memos for routine testing. The analysis of biaxial data for bituminous mixtures is often complicated by a curved envelope of failure for which there is no well defined or proven application [201. One big advantage of 16

OCR for page 10
~ - biaxial testing is Mat stress levels and strains are generally much larger than for diametral testing so that greater testing accuracy can be achieved for stiff asphaltic materials. The influence in the biaxial test of secondary factors such as poor contact of deformation sensors, minor sample disturbance, etc. are less important than for the diametral test. Indirect Tensile Test. The indirect tensile test was developed simultaneously but independently in Brazil and in Japan [211. The test has been used to determine the tensile strength of Marshall-s~ze asphalt . . . . . concrete specimens. lne testing system includes indirect tensile loading apparatus, deformation measurement devices and data recording system. The indirect tensile loading apparatus consists of upper and lower loading plates and upper and lower 0.5 in. wide loading strips having the same curvature as the specimen. Load is vertically applied to the sides of the specimen and maximum tensile stress plane develops along the vertical diameter. The indirect tension test simulates the state of stress in the lower position of the asphalt layer which is a tension zone [221. Schmidt [23] proposed the use of a repeated load indirect tension test (which is called the diametral test) to determine the resilient moduli of asphalt concrete specimens. Figure 6 shows that the values of the resilient moduli obtained from this test compare favorably with those obtained from the direct tension, biaxial compression and beam flexure tests. Baladi and Harichandran 1241 conducted a comparative study of the following test methods: 1. 2. 3. 4. 5. Triaxial test (constant and repeated cyclic loads) Cyclic flexural test Marshall test Indirect tension test (constant and variable cyclic loads) Creep test The results of this study indicated that: 1. The repeatability of test results is poor. 2. . . 3. -line material properties obtained from the dltterent tests are substantially different. The results from the indirect tension test were Me most Promising although Rev were not consistent. - r ~ ~lo, _ _ _~ The advantages of the indirect tensile test are summarized as follows [17, 2l, 22, 251: 1. 2. 3. 4. 5. 6. The test is relatively simple and expedient to conduct. The type of specimen and the equipment can be used for other testing. Failure is not seriously affected by surface conditions. Failure is initiated in a region of relatively uniform tensile stress. The variation of test results is low compared to other test methods (refer to Figure 7~. A specimen can be tested across various diameters, and the results can be used to determine whether the sample is homogeneous and undisturbed. 7. The test can provide information on Me tensile strength, Poisson's ratio, fatigue characteristics, and permanent deformation characteristics of asphalt concrete. The main disadvantage of the test is its failure to completely simulate the stress conditions encountered in practice. As previously discussed, the diametral test does reasonably well simulate the tensile stress condition existing in the bottom of the asphalt concrete layer. The American Society of 17

OCR for page 10
7GO 6m 5m - ~ c 400 ~- ~ - - ~ c) . - - ~ c ~ 3a , 200 - 1W 1~0 :E ~Assumc V Q5 - c ~ _ ~ ~ ~ 4~30 ~ . 3m ~ ._ _ ~ c ~ ._ ~. 2m a: ~ - ~ Diam~ral [ensign _ x D'r~ Compress~on 0 Dir~ Tension `,_ ~Assumev~Q5 ~,~Assume2, 0.2 . , ~ ~ ~ A, ~ Notes: 1. Diametral tensile stress shown an the atcissa is the werage value from center to edge, i. e., 4Z ~ of max. ~x ' ~P.Q710. Oirect tensile or compressive stress aionq ar~s of 8" tall 4" diamMer specimen, strain measured w~th 2''long strain qa~es. I 1 I ,,,, I I ~ 0.2 0.3 0.4 0.5 I.0 2 3 S tres s, p s i 1 t.,, . I 4 510 (a) Direct tension, compression, and diametral methods Ffe~cural 8tams - t. &. ~,. Diametrai Cores - O. ~, c,0 ~V 0 35 - --~ ~ ~Flexural Values N50tes: 1Ca ~ Assume V ~ Q 2 . 1 ~ 1 2 1. Stresses on tIexural values are maximum fiber stross. Stresses 3n diametral values are maximum throwh center. _ . ~,, ., 1 1 1 1 ! , ,,, 1 1 3 ~510 20 30 40 50 100 150 Maximum Stress. psi (b) Flexural and diametral methods Figure 6. Companson of resilient modului of AC specimens using direct tension, compression flexural, ar~d diametral methods (After Ref (22) ) 18

OCR for page 10
o - x 100. _. us P. - LO i: u, ~4 ~4 - 10. us - ~1.0 . 0 c a cut _ _ O . 1 . . . , C = ( F-32)ll.8/ l()5psi = 689 spa/ 1 ~z,L, ~ / / ~ ~ A r ~- LEGEND: Tes c Temperature - 40F (23 Test Temperature ~ 70F Test Temperature - lOO.F 0.1 1.0 10. 100. Appeal t Concrete Modules (Compression Samples Without Confinement), psi x 105 Figure 7. Comparison of test results between the unconfined compression and indirect tension tests (After Ref (23) ) 19

OCR for page 10
Testing and Materials has adopted the repetitive indirect tensile test as a standardized method of measuring _ . , the resilient modulus of asphalt concrete (ASTM D 4123-82~. Control Mode. Two basic types of loading have been used in laboratory tests: controlled-strain and controlled-stress. Repetitive load is applied to produce a constant amplitude of repeated deformation or strain. Asphalt concrete in thin pavements (surface layer thickness of less than 3 in.) is considered to be in a controlled-strain condition. In controlled-stress tests, a constant amplitude of load is applied. The controlled-stress test simulates Me asphalt concrete in thick pavements (surface layer thickness greater than 6 ink. A comparative evaluation of controlled-stress and controlled-strain tests is presented in Table 2 [191. If Me testing procedure measures a real material property, the resilient moduli from the controlled- stress mode and controll~-strain mode must be the same because the sample does not know whether it is under the controlled-stress mode or controlled-strain mode. Bow Me controlled-stress and controlled-strain modes have been used in flexural beam and uniaxial tests. Only the controlled-stress mode has been applied to the indirect tensile test. The reason why the controlled-strain mode has not been used in the indirect tensile test is because the mechanism of forcing the deformation (either horizontal or vertical) back to the original position is not available. One can glue the upper loading strip to the specimen in order to control the vertical strain. However, this mechanism will develop a plane of maximum tensile stress along Be horizontal diameter when the loading head moves upward, which violates the theory behind the indirect tensile test. DIAMETRAL RESILIENT MODULUS TESTING DEVICES AND MEASUREMENT SYSTEMS USED IN EXPERIMENT Loading Devices Based upon He previous discussion of testing methods, the diametral test was selected for use in developing a standard test procedure for resilient modulus testing of asphalt concrete. The diametral test can be performed on small field core specimens and hence is practical for routine use. Also, the indirect tension to which the specimen is subjected during loading simulates reasonably well the tensile stress condition in He bottom of the asphalt concrete layer. Because of the popularity of the diametral test method, a number of test apparatuses have been developed which have important fundamental differences in equipment design concepts. Practically no research, however, has been previously performed to evaluate these testing systems. An experiment was therefore designed to identify the most reliable and accurate diametral test apparatus available and then to develop appropriate test procedures to allow the device to be used for routine testing. A representative group of the most promising diametral testing devices were carefully selected for evaluation In He testing program. The testing devices chosen are as follows: Retsina device, MTS device, Baladi's device and the SHRP Load Guide device. A detailed description of these testing systems is given in Appendix A. The simplest comparisons between He test devices can be made with respect to Heir loading configurations and diametrical deformation measurement systems. The Retsina device has fully independently aligned upper and lower loading strips and EVDTs that are clamped on the specimen to measure deformation. The MTS system has a guide rod that semi-rigidly aligns He upper and lower loading platens and extensometers Hat clamp on to ache specimen for deformation measurements. Both the Baladi and SHRP devices have heavy guide posts Hat rigidly align the upper and lower loading strips. 20

OCR for page 10
for the data collected for five consecutive load cycles, might indicate whether the deformations have become stable. At 41F, no significant trend was observed. The second preconditioning level (100 cycles) appeared to be reasonable with regard to the variation in Poisson's ratio and resilient moduli and was chosen as Me preconditioning level for 41F. As seen earlier, a significant trend was not evident in MR values so the choice of this level should not affect the resilient modulus value. At 77F, the coefficient of variation of the five cycle resilient modulus data decreased as the number of preconditioning cycles increased from 25 to 75 or 150 repetitions. Also, the resilient moduli decreased suggesting more damage. For 77F, the number of preconditioning cycles to use was selected to be between the 2nd and 3rd preconditioning levels. (i.e., between 75 and 100 preconditioning cycles). At 104F, We 2nd and 3rd levels showed improved performance of resilient modulus values, both being almost comparable. The MR values kept decreasing with increasing number of preconditioning cycles. Also, keeping the number of preconditioning cycles small means less damage to the specimen. Thus, at 104F a preconditioning level of 50 cycles was chosen for final testing. Calculation of Poisson's Ratio. In the first stage of testing, the values of resilient modulus and Poisson's ratio were computed in accordance with the SHRP P07(Nov. 1992) procedure as well as the elastic analysis. Poisson's ratios calculated from the SHRP P07 analysis (pr.sh.xv) were usually lower than those obtained using the elastic analysis (pr.el.xv). The difference was approximately 0.01 to 0.05 at 41F, 0.05 to 0.15 at 77F, and 0.05 to 0.20 at 104F. Also, the SHRP P07 analysis with an assumed value of Poisson's ratio always gave significantly higher values of resilient moduli than those calculated from the elastic analysis with an assumed Poisson's ratio (i.e., mr.sh.x.a were always significantly higher than mr.el.x.a). Discussion of Results of Stage 2 Testing The average values of resilient modulus and Poisson's ratio from five consecutive load cycles and the associated coefficients of variation have been tabulated in Appendix ~ (rabies I-! ~ to I-14~. The test ID consists of four letters which indicate He following: (~) specimen type: MI, M2, M3, CI, C2, or C3; (2) load amplitude level: I, 2, or 3; (3) preconditioning level: I, 2, or 3. The elastic analysis was used since He over me~ods do not consider different gage lengths for measurement of vertical and horizontal deformations. Poisson's Ratio. Better Poisson's ratio results were obtained at the higher temperatures. For the first trial at 77F (testing was done at 77F before testing at 104F), most of He Poisson's ratio values were between 0.2 and 0.5 with small five-cycle variances. The coefficient of variation for the Poisson's ratio values (pr.el.xm) improved (reduced) at higher load amplitudes. At the highest level of load, the coefficient of variation was less Can 5% in almost. all He cases. For the second trial at 77F (testing done at 77F after testing at 104F), all He same specimens showed an increase in Poisson's ratio values, indicating that damage does increase Poisson's ratio (Figure 42~. Hence, Poisson's ratio serves as an indicator to the damage occurring in the specimen. However, for high load amplitudes Poisson's ratio values were not, comparatively, as different as Hey were for low amplitudes. This behavior could be a result of the load-history dependence of asphalt concrete. The specimen goes through different thermodynamic states and some of this process is irreversible. The thermodynamic state of He specimen changes after the largest load has been applied during the first trial. Hence, although the magnitude of 86

OCR for page 10
\ L , ~ \' ~ 17S: X~ V ~ X X on . ~X\ ~ X~ 'X X At \ \ _= O _ _ O ~ ._ O o ~ - c Co O O ~ o Q CM _ \Lo lo T r T I I r ~cs~ ~ 000 ~COCal ~ 000 ~ LL 18 Ie!ll proofs '!~ s~uoss!od 87 i_ ._ 3 - V: `9 8 a CQ rat ' 1 o _ cat o be 0 x `0 0~ 0 . cat A) 1 . LO x ~ - ~ ~ - s~ c ~ - o ~ ~ o In c: o ~ c~ ~ ~ LO o ~ T~ ~ _ 4- . ~ Ct C~ ~ ~ C;, _ LL

OCR for page 10
loads for the smaller amplitude in the first and second trial is the same, the specimen behavior is different. However, when the largest load is applied again in the second trial, the specimen shows comparable behavior as it is very nearly at the same thermodynamic state as it was when the larger load was applied during the first trial. As before, the variation in Poisson's ratio for the second trial was small and it decreased with increasing load amplitude (Figure 43). At 77F it appears that higher load amplitudes result in more reasonable Poisson's ratios and smaller variance in five cycle data. At 104F, most Poisson's ratio values were between 0.4 and 0.6 with the coefficient of variation usually less than 5%. A trend of smaller coefficient of variation with increasing load amplitude was not seen. At load amplitude levels 2 and 3 the coefficients of variation were similar. For field cores, the use of a measured Poisson's ratio may lead to a higher estimation of resilient moduli values. For example, a constructed surface course has a resilient modulus of 500,000 psi and measured Poisson's ratio of 0.3. After 10 yes., the pavement needs maintenance and resilient moduli of field cores are evaluated. Because of an increase in Poisson's ratio due to damage, the measured resilient moduli after 10 yrs. might not show any significant decrease, thus resulting in an overestimation of the pavement condition. In such cases, the resilient moduli should be based on horizontal deformation and the initially measured value of Poisson's ratio of 0.3. Then, the calculated resilient moduli would represent the deterioration of the pavement. As more data becomes available, consideration should be given to the use of Poisson's ratio as a direct indicator of deterioration in an asphalt concrete pavement. Load Amplitude. Overall, at higher load amplitudes, the f~ve-cycle coefficient of variation for Poisson's ratio became lower with increased load amplitude at 41F and 77F. However, at 104F, the five-cycle variance was almost equivalent for load amplitude levels 2 and 3. At 41F there was an overall increase in resilient moduli (mr.el.x.a) with increasing load amplitude. The coefficient of variation in the resilient moduli decreases at higher load amplitudes. Also, the highest load level did not seem to cause any significant damage (based on measured vertical deformation as the tests progressed). Higher load amplitude is required to generate adequate values of deformation for measurement purposes. Thus, at 41F, testing at the SHRP P07 (Nov. 1992) level of 30% of the indirect tensile strength at 77F is recommended. The resilient moduli (mr.el.x.a) from the first trial at 77F exhibited a small decrease with increasing load amplitude. However, resilient moduli values from the EXSUM setup (mr.el.xm.c) usually increased with increasing load amplitude. At the higher loads the resilient moduli (mr.el.x.a and mr.el.xm.c) agree better than at other levels. As before, the coefficient of variation reduced with increasing load amplitude. The results from the second trial at 77F showed no trend in the mr.el.x.a values, but the coefficient of variation reduced with increases in load amplitude. A comparison of the resilient moduli values (mr.el.xm.c) at the highest load amplitude with the first trial at 77F, revealed a good similarity except for the 4 in. diameter and 2.5 in. thick specimens (types M! and CI). The reduction in resilient moduli resulting from damage to the specimens was counteracted by the increase in Poisson's ratio due to damage. Thus it seems that once we are able to measure Poisson's ratio with confidence, the resilient moduli values should be calculated based on that value. So, if slight damage is occurring to the specimen, the decrease in resilient modulus due to the damage can be compensated by an increase in Poisson's ratio values. A comparison of resilient modulus (mr.el.x.a ~ and Poisson's ratio (pr.el.xm) values between the two trials suggest that while there was a decrease in resilient modulus values (Figure 42) of approximately 5% to 20%, there was a corresponding 5% to 25% increase in pr.el.xm values (Figure 42~. 88

OCR for page 10
A: o c) cr. a: - ct ~ ~- o t - 1 //li! ~ l l l m ~{9 ~N O C] o O O O Hi o ~o o (ulx la Jd)= 89 -8 o . _ Ct Cat ~-= Cat Cat ._ o O m ~ o .= O tD -~ ~ o ~ .o s o Cat _ ~ ~ CQ c `_ Ct U: so As - Ct Ct o - - ._ 3 _ - U. ~ V) - to A ~ ~ o V ~ Ct so ._ - C~ so

OCR for page 10
At 104F, the overall trends indicate an increase in resilient moduli with increased amplitude for both mr.el.x.a data, from assumed Poisson's ratio and mr.el.xm.c data, from calculated Poisson's ratio. As with Poisson's ratio, the variance in the resilient moduli at the 2nd and 3rd load levels was reasonably similar although the scatter was relatively large. At 104F, the specimen undergoes significant damage, as can be seen from a rapid increase in the permanent vertical deformation as the test progresses. Thus, to Limit the damage to minimal values, it becomes important to keep the load levels as small as possible, but large enough to maintain adequate specimen deformations and load control. Significant deformations were obtained at 104F, even with small loads in the resilient modulus test. Hence, it is recommended that a smaller load amplitude of 3.5 to 4 % of He failure load should be used. This load should give essentially He same resilient moduli values (Appendix J. Table J-14). Recommended Seating Loads. SHRP recommended seating loads are suitable for testing at 41F and 77F, but at 104F, the seating load should be reduced. The 10% seating load recommended by the SHRP P07 protocol is not necessary. Instead seating loads of 5%, 4%, and 4% of the total load to be applied at each cycle for resilient modulus testing are recommended at 41, 77 and 104F, respectively. At 104F, a minimum seating load of 5 Ibs. must be maintained, and the seating load should not exceed 20 Ibs. Specimen Size and Type Specimen Diameter. The effect of specimen diameter can be studied by comparing the resilient moduli and Poisson's ratios obtained for specimen types M! and M3, and C! and C3. From a comparison of resilient moduli and Poisson's ratios and their coefficients of variation, an influence of specimen diameter on specimen response is not apparent at 41F. Tables I-~! to ]-14 (Appendix ]) show that the effect of specimen diameter on resilient modulus and Poisson's ratio seems to increase with increasing temperature. At higher temperatures the coefficient of variation reduced for the 6 in. specimen diameter. At 104F, there was a 24% (medium gradation specimens) to 50% (coarse gradation specimens) decrease in resilient modulus (mr.el.x.a) compared to the 4 in. specimens. An assumed Poisson's ratio was used for both the gradations for the 6 in. diameter specimen (Figure 44~. However, at 77F, only the coarse gradation specimen showed a decrease in resilient modulus with increase in diameter. Thus, it seems that at lower temperatures testing specimens of different sues is less likely to affect results than at higher temperatures. At higher temperatures specimens possess more non-homogeneity and this would cause a change in MR value with a change in diameter based on the aggregate to size ratio. Specimen Thickness. The effect of specimen thickness can be investigated by observing the difference in behavior between specimen types M! and M2, and C! and C2. Tables I-! ~ to I-14 (Appendix ]) show Hat no-significant trend was seen at any temperature in He five-cycle coefficients of variation. Also, there was no consistent trend in the resilient moduli values (mr.el.x.a). Since there is no difference in the five cycle variance, increasing the sample thickness will not help attain more repeatable results in the resilient modulus test. Specimen Gradation. The resilient moduli for coarse gradation specimens (C! & C2) were as much as 75 % higher than for the corresponding medium gradation specimens. This difference in MR increased with increasing temperature which is of significance for pavements constructed in regions having warm summer temperatures. For the 4 in. diameter specimens, the effect of gradation was smaller for the 4 in. thick specimens than for the 2.5 in. thick specimens (Figure 45~. Thus, as He specimen size, thickness, and diameter increases, He effect of gradation decreases. The coefficients of variation obtained for medium gradation specimens were less than for He coarse gradation specimens. The 90

OCR for page 10
4.0E+OS 3.SE+OS .= 3.0E+OS / /1 / 2SE+OS 20E+OS , , , 1.SE+OS 1.5E+OS 20E+OS 2SE+OS 3.0E+OS 3.SE+OS 4.0E+OS MR (psi), 4" diameter 3~ MG o CG Figure 44. Effect of specimen diameter on MR (mr.el.x.a - obtained from horiz. extensometer deformation using Elastic analysis and assumed By, for Stage 2 tests at 104F 91

OCR for page 10
-- 3~ - ~ 2- - zO~- ' 3~ 08 3~ 2 +~ . - - 2 . - 3.~ . - 1 .. - - ,~ 1 ~ 4.~ o ~ x~ ~ Isis 1 1 x+ - , Ate- l E+ - 1~ - o Cc3 (b) 77F qCE.gq~a art ~ 2~+a. I. - --,~ ~ ~ . x2 ~o ~ sol ~ USED .~ - 3 - - , z~- Z + - - SEAM +- Z~= Z.- 3~.- ~R - , USA l ' 4'.~2~ ~ sir ~ ~1= Figure 45. Effect of gradation on MR (mr.el.x.a - obtained from horiz. extensometer deformation using Elastic analysis and assumed it), at different temperatures and different specimen sizes for Stage 2 tests 92

OCR for page 10
differences In coefficient of variation for the two gradations were close to zero for the larger 6 in. diameter specimen that were 4.5 in. thick. Thus the medium gradation specimens can be tested using 4 in. diameter and 2;5 in. thick specimens, but the coarse gradation specimens should be tested using 6 in. diameter, 3.75 in. thick specimens to obtain good values of resilient moduli. Use of a 4 in. diameter specimen 2.5 in. thick is acceptable for testing medium gradation asphalt minces have a maximum aggregate size of about 3/4 in. However, for aggregate sizes greater than 3/4 in., a 6 in. diameter specimen 4.5 in. thick should be used to test these coarse gradation mixes. MULTI-LAB VALIDATION STUDY The details of the limited, multi-lab validation diametral resilient modulus test can be found in Appendix H. The general purpose of this study was to determine, with statistical analysis, the effects of multiple operators, retesting specimens, and assumed versus calculated Poisson's ratio on the resilient moduli determined for identical specimens. The specimens were tested at three different laboratories, using similar equipment but different equipment operators. The specific conclusions of the multi-lab validation study are: 1. The recommended testing protocol yields better estimates of Poisson's ratio, but the use of an assumed Poisson's ratio yields more consistent moduli. 2. The experience of the operator has a significant effect on the resilient modulus values during testing. The more experienced the operator, the less variation in the resilient moduli values. The difference in the resilient moduli for calculated versus assumed Poisson's ratio is also much smaller with an experienced operator. The coefficient of variation of the resilient moduli determined from calculated and assumed Poisson's ratios does, as indicated by the primary test program, increase with increasing temperature. 4. The number of times a specimen has been tested also has an effect on the resilient modulus. Structural damage to the retested specimens has a larger effect on resilient modulus than He lab-to- lab variation. The statistical analysis was performed on a very limited number of samples, making the interpretation of the statistical results somewhat less reliable. An extensive validation study is recommended to obtain better evaluations on sample-to-sample and lab-to-lab variations. COMPARISON OF LABORATORY AND BACKCALCULATED RESILIENT MODULI Resilient moduli for use in design are presently determined by both direct laboratory measurement and by backcalculation from falling weight deflectometer (FOOD) tests. Both laboratory tests and backcalculation procedures from field data have important limitations and advantages. In laboratory tests, fabrication of test specimens Hat duplicate field conditions and simulating in-situ stress states and environmental factors are difficult. In backcalculation procedures, the theory used assumes very ideal behavior of the materials (i.e., homogeneous, linear elastic, isotopic). Such materials are not found in pavements. The purpose of this study was to compare He resilient modulus determined using the proposed 93

OCR for page 10
laboratory procedure with FWD backcalculated values. A field test section on U.S. 421 in Norm Carolina was used to obtain field data and laboratory test specimens. The details of this study are given in Appendix T. however the specific conclusions of this study are: The backcalculated and laboratory AC resilient modulus values are similar if the field data is obtained on asphalt concrete layers less than 4 in. Hick. 2. There is a significant difference between the laboratory and field AC resilient moduli if the field data is obtained on sections with total asphalt concrete thicknesses of 9 in. or more. Since the data used In this study is rawer Emoted, it is not clear how much different the laboratory resilient modulus is from the FWD backcalculated modulus. Once LTPP field test data is fully collected and analyzed, a better assessment can be made on this issue. SUMMARY AND CONCLUSIONS Existing methods were reviewed for empirically predicting the resilient modulus for asphalt concrete. The Asphalt Institute Method, corrected for locally used materials and testing devices appears to offer a beginning point for developing a practical alternative to performing resilient modulus test on a routine basis. Different laboratory test methods, such as the repeated load biaxial test and the diametral test, apply different stress conditions to a specimen. As a result, the resilient moduli obtained from these different methods do not always agree. The repeated load diametral test was concluded to be the most practical, realistic method for evaluating the resilient modulus of asphalt concrete. An extensive resilient modulus testing program was, therefore, carried out using the diametral test. All tests were conducted using a 0. ~ sec., haversine shaped loading pulse using a closed loop, electro-hydraulic testing system. Experiments were performed to identify the most accurate and reliable diametral testing device. Loading equipment evaluated in the study were as follows: (~) Retina device, (2) MTS device, (3) Baladi's device and (4) SHRP Load Guide ~G) device. The following four deformation measuring devices were also studied: (~) stand along EVDTs, (2) an extensometer mounted on the specimen, (3) a gage-point-mounted (GPM) setup and (4) a special combined measurement system using a surface mounted EVDT to measure vertical deformation and externally mounted transducers to measure horizontal deformation. Diametral resilient modulus tests were performed on laboratory prepared asphalt concrete specimens having a coarse and medium gradation as well as on field cores and synthetic specimens. Temperatures of 41F, 77F and 104F were used in these tests. In performing these tests, equipment calibration, including the use of synthetic specimens, was found to be a critical aspect required to obtain reliable test results. Specimen rocking was also determined to be an important consideration in selecting an appropriate loading device. A square wave load pulse produces more damage and a significantly smaller resilient modulus compared to a haversine wave. Therefore the square wave load pulse should not be used for resilient modulus testing since it is not representative of He pulse developed in the field. 94

OCR for page 10
5. The loading pulse time significantly affects He resilient modulus. A loading time of 0.2 sec. reduces Me values. and produces more damage than for a 0. ~ sec. pulse. A shorter loading time ~ , . ~ , _ ~ rat ~- ~ ~ ~ ~ ~ ~ ~ ~ ~ . . . ~ ~ .~ . ~ a~ -, ~ .~ of ().()5 sec. is representative ot~ high vehicle speeds, but is not practical as the repeatability of the test is poor and accurate load control at higher temperatures is hard to achieve. A loading time of 0.] sec. is therefore proposed which is in agreement with the SHRP P07 (Nov. 1992) procedure. 6. 7 8. 9. The ratio of rest period to loading period of 4 and 24 used in this StU6Y 40 not have a significant _ , - , . . ~ . . . .. . . . . . . . . . #. .. . . enect on the resilient module values. Also past research has shown that a rest perlocl/loaulng time ratio greater Han ~ provides no extra benefit. A rest period/Ioading time ratio of 9, as presently used by SHRP, is a good choice. At 77 and 104F the resilient moduli decreased wig increasing number of preconditioning cycles. Based on a study of trends in the coefficient of variation for MR , the following preconditioning levels were Axed at the three deferent test temperatures to make resilient modulus test results more repeatable: 41 F: 77 F: 104 F: 100 cycles 100 cycles 50 cycles A significant difference between resilient moduli and Poisson's ratio is obtained using the SHRP P07 (Nov. 1992) analysis and the elastic analysis. The elastic analysis is essentially the same as the ASTM analysis except it allows the use of different deformation measurement gage distances while the ASTM analysis does not. The SHRP equations give resilient moduli values as much as 45% higher when an assumed Poisson's ratio is used. The elastic analysis with the appropriate coefficients for measurement geometry used in testing is the recommended approach. 10. ~- 12. Poisson's ratios obtained using the EXSUM system at higher temperatures are reasonable. The system can be modified with further use to make it simpler. The 4 in. diameter and 2.5 in. thick specimens may be acceptable for testing medium gradation mixes, but 6 in. diameter and 4.5 in. thick specimens should be used to test coarse gradation mixes. Grain sue distributions for the medium and coarse gradation mixes are given in Appendix B. Table Be-. SHRP recommended loads are suitable for testing at 41F and 77F, but at 104F, the load should be reduced. The 10% seating load recommended by the SHRP P07 protocol is not necessary. Instead seating loads of 5, 4, and 4% of the total load to be applied at each load cycle for resilient modulus testing are recommended at Ill, 77, and 104F, respectively. At 104F, a minimum load of 5 Ibs. must be maintained, and the seating load should not exceed 20 Ibs. An Improved diametral test was developed to evaluate the resilient modulus of asphalt concrete. The proposed test procedure is given in Appendix C. A closed loop, electro-hydraulic testing system and also a data acquisition system is used to apply a 0. ~ sec. haversine-shaped load pulse to a disk-shaped specimen. 13. Loading Device. The SHRP EG device minimizes rocking of the specimen. The good performance is apparently due to (~) use of two guide columns, (2) a counter-balance system, (3) 95

OCR for page 10
an innovative semi-rigid connection between the upper plate and the load actuator, and (4) its sturdiness. The disadvantages are its bulkiness, complication of use, possible inertia from the counter-balance system, friction in the guide columns, and limitation of the sue of the sample that can be used. 14. 15. 16. 17. Ad. 19. Mountable Extensometer. A mountable extensometer device, compared to the stand-alone EVDT measurement device, provides less variance and hence better repeatability within the five consecutive cycles used for resilient modulus determination. However, using the SHRP EG device EVDTs gave comparable performance to the mountable extensometer. Mountable deformation measurement devices are recommended for resilient modulus testing because of the smaller variability. Poisson's Ratio Importance. Poisson's ratio is one of the most important parameters influencing the resilient modulus. The variation In MR values due to the testing axis dependency and different lengths of rest periods are almost negligible compared to the magnitude of difference in the MR values from assumed and calculated Poisson's ratios. Poisson's ratio should be evaluated using the EXSUM deformation measurement system. EXSUM Deformation Measurement Device. The proposed EXSUM deformation measurement system provides a promising measurement method for determination of consistent and reasonable Poisson's ratios. At 41F, however, increase In variability occurs due to misalignment and rocking which become more Important for the small deformations occurring at low temperatures. Use of the SHRP EG device, or its modification, together with the EXSUM setup ensures obtaining reasonable values of Poisson's ratio even at low temperatures. The use of the EXSUM setup requires an increase in testing time compared to conventional measurement systems because of the significant time required for mounting the EVDT on the specimen. Preconditioning. Specimens were subjected to 3 different numbers of preconditioning load cycles at each temperature. No significant difference was observed in the variation of resilient moduli and Poisson's ratio for the last 5 load cycles for He largest two preconditioning cycles. However, MR values did decrease with increasing number of preconditioning cycles. For 41 F and 77F, 1 00 p recond itio ning cycl es are reco mmend ed wh il. e th e us e of 50 cy c I es i s r e co mm en d ed at 1 04 F . Calculation of MR. A significant difference exists between resilient moduli and Poisson's ratio values computed using the SHRP PO7 analysis and the elastic analysis used in this study which is similar to the ASTM analysis. The SHRP analysis gives higher values when an assumed Poisson's ratio is used as compared to the elastic analysis with an assumed Poisson's ratio. Load Amplitude. The load amplitudes recommended in SHRP protocol are suitable for testing at 41F and 77F, but at 104F a smaller load should be used. Load levels corresponding to 30, 15, and 4% of the indirect tensile strength at 77F are recommended for testing at 41F, 77F, and 104F, respectively. 96