A Correlation Between Dynamic Cone Penetrometer Values and P

Geotechnical Testing Journal, Vol. 28, No. 1 Paper ID GTJ12312 Available Available onlin e at: www.astm.org www.astm.org

 Dar-Hao Chen,1  Deng-Fong  Deng-Fong Lin,2  Pen-Hwang  Pen-Hwang Liau,3 and John Bilyeu4

A Correlation Between Dynamic Cone Penetrometer Values and Pavement Layer Moduli

TheDynamic Cone Cone Penetr Penetrome ometer(DCP)is ter(DCP)is oneof the least least expens expensivetesti ivetesting ng devic devices es ableto charac character terizebase izebase and subgra subgrade de proper propertie ties. s. ABSTRACT: TheDynamic To fully use the DCP in pavement evaluation, an empirical relationship between DCP penetration rate and layer modulus is required. However, the literature on this correlation is limit ed. This study incorporates a total of 198 DCP and Falling Weight Deflectometer (FWD) tests done over 8 years on various types of highways (Interstate Highway, US, and Farm-to-Market). The computer program MODULUS was employed to backcalculate the layer moduli from the FWD results to build a correlation with DCP results. A comparison was made with the widely-used model by Powell et al. (1984). It is found that the difference between the two models decreases as the Penetration Rate (PR) increases. For a PR of less than 10 mm/blow, the difference between these two models was over 10 %. The difference is only about 1.7 % when the PR is 80 mm/blow. mm/blow. Without knowing the true moduli, it is impossi ble to tell which equati on is better. The correlation developed here provides another option and allows researchers to recognize the range of variability. KEYWORDS: dynamic cone penetrometer (DCP), falling weight deflectometer (FWD), correlation

Introduction

Although in pavement design the moduli are used to determine therequired therequired layer layer thickn thickness ess(es)of (es)of a pavem pavementstruct entstructure ure,, densit density y and moisture moisture content are the two parameters parameters for project project acceptance. acceptance. Density and modulus are very different material characteristics, even though density may be a good indicator of the strength of  granular materials (Chen et al. 1999B; Livneh and Goldberg 2001). There There is a clear clear gapbetweendesig gapbetweendesign n andconstruc andconstructio tion.Distr n.Districtdesi ictdesign gn and construction engineers often wonder what are the constructed base and subgrade moduli. The goal of lower cost and improved quali quality ty demand demandss more more precis precisee qualit quality y contro controll and qualit quality y assura assurance nce (QC/QA) programs that tie test results into design practice, making the in-place measurement of design parameters, like stiffness and modulus, necessary. Rapid in-situ stiffness/modulus testing would make it easier for transportation agencies to conduct reliable and defensible quality assurance programs. Currently there is no standard field test to determine the resilient modulus of base and subgrade soils for construction quality control/assurance purposes; though many transportation agencies are collecting stiffness/modulus stiffness/modulus along with density and moisture measurement. The Falling Weight Deflectometer (FWD), Geogauge, Dirt Seismic Pavement Analyzer (DSPA), and laboratory repetitive triaxial tests have been used to determine the pavement layer  modulus (Nazarian et al. 2002; Livneh and Goldberg 2001; Rahim ReceivedNovemb ReceivedNovember er 3, 2003; 2003; accepted accepted forpublicationJune forpublicationJune 7, 2004; 2004; published published January 2005. 1 APT System Manager, Construction Division, Texas Department of Transportation, 4203 Bull Creek #39, Austin, TX 78731, [email protected]. [email protected]. 2 Associate Professor, Department of Civil Engineering, I-Shou University, University, 1, Section 1, Hsueh-Cheng Rd. Ta-Hsu Hsiang, Kaohsiung County, 84008, Taiwan. dfl[email protected]. 3 AssociateProfessor AssociateProfessor,, Departmentof Departmentof Mathemati Mathematics, cs, National National Kaohsiun Kaohsiung g NorNormal University, Kaohsiung 802, Taiwan, R.O.C. [email protected]. 4 Transportation Transportation Engineer, Construction Construction Division, Division, Texas Texas Department of  Transportation, 4203 Bull Creek #39, Austin, TX 78731, [email protected]. tx.us.

42

and George 2002; Sawangsuriya et al. 2002). However, the limitations of each method are equally real. As many different sets of layer moduli would satisfy the same FWD deflection bowl, practicing pavement engineers struggle to identify the correct set. Also, the FWD often is unable to determine the extent of a weak base/subgrade layer due to a thick concrete layer that carries most of the load away. Laboratory repetitive triaxial tests are seldom used to determine the layer moduli for routine design or QC/QA tests in current DOT environments (Rahim and George 2002; Chen et al. 2001b). Seismic tests are quick and easy, but the seismically determined modulus is very high due to the high frequencies and miniscule loads used. The Geogauge shows some promise, but is highly sensitive to the surface preparation, and it only gives a composite stiffness that includes all layers to some uncertain depth. The Dynamic Cone Penetrometer (DCP) has become a costsaving alternative for characterizing the properties of pavement layers without digging test pits or collecting soil samples. The DCP serves as an excellent tool for construction inspection; it has the ability to verify both the level and uniformity of compaction (Burnh (Burnham am 1996; 1996; Siekme Siekmeier ier et al. 1999). 1999). In additi addition, on, the layer layer thickthickness can be determined from the changing slope of the depth versus accumulated blows profile. The DCP is an excellent tool to characterize base/subgrade properties under problematic jointed concrete pavement pavementss where there are voids voids or weak layers. layers. Also, the DCP is useful when the backcalculated moduli from the FWD data is in question, such as when the asphalt concrete (AC) thickness is less than 75 mm (3 in.), or when shallow bedrock is present. In addition, it is fairly easy to collect and analyze DCP data. The DCP has not been widely used in the pavement engineering community, community, partly due to the lack of a solid correlation between DCP results and modulu moduluss value values. s. Before Before the DCPcan evalu evaluate ate layer layer stiff stiffnes ness, s, an empirical correlation needs to be established. Many equations have been proposed to correlate DCP results to California Bearing Ratio (CBR) values. However, However, there are not many studies devoted to the correlation of CBR values to moduli or DCP results to moduli.

Copyright © 2005 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.

CHEN ET AL. ON PAVEMENT LAYER MODULI

43

TABLE 1 —Test site descriptions.

Location

FIG. 1—   DCP and FWD test sites.

The goal of this study is to develop an equation to estimate modulus through DCP testing. 198 DCP tests have been conducted on various highway classifications [Interstate Highway (IH), US, Farm-to-Market (FM)]. FWD tests were conducted before the DCP tests, at the same location, to determine the layer moduli through backcalculation. About 20 % of DCP tests were from the Mobile Load Simulator (MLS) project. The MLS is a full scale, accelerated pavement testing (APT) device. Under the MLS project, FWD, laboratory, Spectral Analysis of Surface Waves (SASW), and in-situ instrumentation results are available for comparison. This increases the confidence in the analysis results (Chen and Hugo 1998; Chen et al. 1999a; Chen and Hugo, 2001). The other 80 % of the tests were conducted for forensic or special study projects over an 8-year  period. Some of those results have been documented in literature (Chen et al. 2001b; Chen et al. 2002; Chen et al. 2003). Test Sections

Figure 1 and Table 1 depict the test sites and pavement structure [asphalt concrete (AC) surface and granular base thickness]. As indicated in Table 1, numerous tests were conducted at each test site at different locations with minimum spacing of 1.5 m. For  some test sites, the spacing was as much as 30 m between the test

IH35 on top of base US290 on top of base FM1644 US290/281 US82WB US82EB FM740 US281SB (FortWorth) US83 IH35 on top of 4 AC FM2818 FM2818 IH635FR US281 (SPS1) IH30 US290 US281SB (MLS) US281NB (MLS) Total

AC Thickness (mm)

Base Thickness (mm)

0 0 13 51 51 51 76 89 102 102 102 102 102 127 178 178 203 203

356 356 305 305 254 254 254 254 203 356 305 152 610 305 203 127 381 381

# Tests 3 5 8 15 24 2 5 2 5 7 31 12 21 13 11 5 24 5 198

locations. Some of the tests were conducted on top of the base layer  during construction (e.g., IH35 and US290). There were four test sites on US281 in different parts of the state. After one year of trafficking, severe alligator cracking and deep rutting were found in a foamed-asphalt project on US82, as shown in Fig. 2. Extensive field tests, including FWD, SPA, Ground Penetration Radar (GPR), and DCP, were conducted to investigate the cause of structural distress. Laboratory tests also were done to determine gradation, moisture content, capillary action, and indirect tensile strength. Four trenches were opened to test each layer directly and to obtain samples for laboratory testing. It was found that the subgrade modulus in the failed areas was similar to that in the intact areas. The foamed-asphalt base moduli on the intact areas were approximately 5–8 times higher than in the failed area. It was concluded that the rutting and alligator cracking were associated with failure of the foamed asphalt base. The foamed asphalt base exhibited a severe loss of strength when subjected to moisture (Chen et al. 2004a). A significant amount ( ∼20 %) of the DCP and FWD results used in this study were from the MLS project on US281, which had well-known pavement performance and extensive test results (Chen and Hugo 1998; Chen and Hugo 2001; Chen et al. 2001b). Layer  moduli from FWD, laboratory triaxial, SASW, and in-situ instrumentation results are available for comparison, which increases the

 DCP and FWD tests on US82 (a forensic study on a premature failure site). FIG. 2— 

44 GEOTECHNICAL TESTING JOURNAL

 DCP (left) and FWD tests (right) on FM1644 before the 1 million lb superheavy load move. FIG. 3— 

confidence of using those data. US281 is a two-lane highway (in each direction, a total of four lanes) in the Fort Worth District with an average of 3100 vehicles per day (1550 per direction) in 1994. Accelerated pavement testing was applied on southbound (SB) and northbound (NB) lanes of US281 to determine the effectiveness of  the rehabilitation treatments. Approximately 972 000 and 388 800 of 80 kN Equivalent Single Axle Loads (ESAL) were applied by the Texas MLS to the SB and NB lanes, respectively. After 972 000 and 388 800 ESALs, the average ruts were 4.31 mm and 10.11 mm, respectively. The 972 000 ESAL is equivalent to 8 years of traffic, which is the minimum life required by the FHWA for a rehabilitation work. The rehabilitation strategy used in the SB lanes was superior to the NB lanes for rutting resistance. Rutting was the primary distress mode. The DCP and FWD were applied to three job sites in the Dallas District (IH635FR, FM 2818, IH30) and one site in the Austin District (US290) (Chen et al. 2001a). These four projects were chemically treated with liquid stabilizers. TxDOT occasionally has to look for alternatives to lime stabilization because sulfates in the soil would otherwise cause heaving problems. In these projects, the effectiveness of chemical treatments was evaluated by stiffness measurements of the stabilized layers and adjacent non-stabilized layers. Two sections of US290 were testedwith theDCP and FWD. Both sections consist of 51 mm of AC over 305 mm of crushed limestone base. The only difference between these two sections is that the base of the first section was treated with EN1 liquid stabilizer. The pavement structure of the frontage road (FR) IH635FR consists of a 102 mm AC on top of 610 mm of liquid-stabilized (EMC squared/EMS) subgrade. The IH30 pavement consists of 178 mm of AC and 203 mm of EMC Squared/EMS treated subgrade. The FM2818 pavement structure also has 102 mm of AC. DCP and FWD tests were conducted on two sections of FM2818. The first pavement section included 102 mm of ACover 305mm of subgrade modified with EMC Squared/EMS. This section was constructed on approximately 4.57 m (15 ft) of fill material. Approximately 3.04 m (10ft) down withinthe fillmaterials there is another 305mm of subgrade modified with EMC Squared/EMS. The second section of FM2818 included 102 mm of HMAC over 152 mm of subgrade treated with EMC Squared/EMS. FM1644 is a load-zoned road with a Gross Vehicle Weight (GVW) restriction of 26 520 kg (58420 lb). Two requests for  identical superheavy loads slightly over 454 000 kg (1 million lb) were permitted on load-zoned road FM1644. This was to transport two generators for a power plant (Chen et al. 2004b). Before the superheavy load move, DCP, FWD, and other tests were per-

formed to improve the understanding of pavement behavior under  these conditions, as shown in Fig. 3. The maximum deflections of  FM1644 for a 40 kN load are in the range of 40–60 mils (1.02–  1.52 mm). Note that typical deflections for an interstate highway are in the range of 2–8 mils (0.05–0.20 mm). No visible damage was observed after the moves, even though the pavement structure was weak. The SPS1 pavement test sections on US281 (US281-SPS1 in Table 1) in south Texas is the largest Long Term Pavement Performance (LTPP) experimental site in the US. The project was opened to traffic in 1997, and pavement performance has been poor. Three of these test sections developed deep rutting within one year. Their  surfaces were milled to restore the ride quality. Three years after  construction, 14 of the sections had 10 mm or more rutting. A forensic study was initiated by TxDOT to identify the cause of  the problem (Chen et al. 2003). Nondestructive testing (FWD and GPR) and destructive field investigation was initiated, as shown in Fig. 4. The original plan was to cut nine trenches. However, after  four trenches were cut, the problematic layer was identified, and the trenching operation was terminated. DCP, stiffness gauge, seismic pavement analyzer, and nuclear density gauge tests were then conducted on top of the base and subgrade layers. Hot mix asphalt industries have been promoting a heavy duty and long lasting as phalt pavement called perpetual pavements. Perpetual pavements use multiple layers of durable asphalt to produce a smooth, long-lasting road. TxDOT is experimenting with perpetual pavement on IH35, one of the most heavily trafficked routes in Texas. DCP, FWD, and other tests were conducted on top of the 356 mmbase andafterplacement of 102mm of asphalt, as indicated in Table 1. It is important to note that when DCP tests were conductedon topof asphalt, 19 mm holes were drilled to enable testing. Existing Correlation Equations

Currently, there are no direct correlation equations between DCP penetration and modulus values. However, there are equations to correlate DCP penetration and CBR values. For example, the U.S. Army Corps of Engineers found a relationship for a wide range of  granular and cohesive materials, as given in Eq 1 (Webster et al. 1992). Equation 1 has been adopted by many researchers and practitioners (Livneh 1995; Webster et al. 1992, Siekmeier et al. 1999). logCBR = 2.465 − 1.12(logPR)

or

CBR = 292/PR1.12

(1)

where: CBR = California Bearing Ratio, and PR = the DCP’s penetration through the layer in units of mm/blow.

CHEN ET AL. ON PAVEMENT LAYER MODULI

45

 DCP, FWD, and other tests on US281 and US290 (forensic studies on premature failure sites). FIG. 4— 

Although the 1993 AASHTO Guide for Design of Pavement Structures adopted Eq 2 for calculating moduli (E) through CBR, the equation is only a rough estimate, as the moduli from which this correlation was developed ranged from 750–3000 × theCBR value. Also, the formula is limited to fine-grained soils with a soaked CBR of 10 or less. E(psi) = 1500 ∗ CBR or

E(MPa) = 10.34 ∗ CBR

(2)

The most widely accepted relationship between CBR and modulus (Eq 3) was proposed by Powell et al. (1984). E(psi) = 2550 ∗ CBR0.64

or

E(MPa) = 17.58 ∗ CBR0.64 (3)

Combining Eq 1 and Eq 3, a direct relationship between DCP penetration and modulus value can be established, as given in Eq 4: 0.7168

−

E(ksi) = 96.468 ∗ PR

or

0.7168

−

E(MPa) = 664.67 ∗ PR

(4) Correlation Equation Developed in This Study

 Data Preparation The MODULUS program developed by the Texas Transportation Institute (Uzan et al. 1988) was employed to backcalculate the layer  moduli from FWD results. This program normally does not yield reasonable estimates of layer moduli when the AC thickness is less than 75 mm (3 in.). In this case (or if there is no AC at all), the program BISAR was used, iteratively, to find a set of layer moduli where all deflections, measured on the surface by the FWD and at depth by multi-depth deflectometers, will match.

It is reported by Livneh et al. (1995) and Chen et al. (2001a) that the DCP results are affected by the test procedure; that is, there is at least a 10 % difference in penetration rate when tests are run through a hole drilled in the AC, as opposed to tests done directly on base without AC overburden pressure. Factors were proposed by those two studies (Livneh et al. 1995; Chen et al. 2001a) to correlate the results conducted directly on top of the base to those through a drilled hole. In the present study, the factors proposed by Chen et al. (2001a) were adopted with factors 1.19 and 1.12 for  base and subgrade layers, r espectively. DCPunc

=

1.19DCPcon

for base

(5)

DCPunc

=

1.12DCPcon

for s ubgrade

(6)

where DCPunc = test on top of base (mm/blow), and DCPcon = test on top of AC through a narrow drilled hole (mm/  blow). There were two sites (IH35 andUS290) where theDCP andFWD tests were conducted on top of the base. The data collections were made immediately after the base construction. Thus, the factors of  1.19 and 1.12 were applied to base and subgrade for these two sites. The combined IH35 and US290 data made up approximately 10 % of the overall data population. A slight reduction in data scatter  was found by applying the correction factors. The improvement was because only small portions of the DCP tests were conducted directly on top of the base. The authors believe that the correction factor impact will be greater when the population size increases.

46 GEOTECHNICAL TESTING JOURNAL

FIG. 5—   Determination of normal distribution.

 Regression Analysis Equation 4 was selected as the basic model for this study. Equation 4 is a linear equation that can be expressed in log form to facilitate a regression analysis: Log (EMPa) = a − b ∗ Log (PR)

(7)

where a and b are the regression coefficients. After applying Eq. 7, the “mean shift outlier model” by Sanford Weisberg (1985) was used for outlier detection. When the highest residual was determined to be an outlier, it was removed from the data set, and the remaining data were passed through the algorithm again. The process continues until no outlier can be found in the data set. After  this procedure was complete, 227 data points remained; 119 in base and 108 in subgrade. Therefore, only approximately 60 % of the data (119 out of 198) was s elected for model development because the outliers are expected for the DCP tests. Sometimes the DCP tip will encounter large aggregate or cobblestones that will skew the results. There are numerous instances where the FWD deflections are similar (e.g., less than 20 %) at the same test site, but the PR values vary by more than 400 %. It should be noted that not all of the remaining 119 bases have the subgrade PR results. There were 11 locations (119–108) without subgrade PR values because the DCP tests were only conducted on the base layer. This was because the base was too hard to penetrate, and tests were terminated before reaching the subgrade. The two basic assumptions for a linear regression analysis are: 1) the data are normally distributed, and 2) the data are random. Typically, with sufficient data (like this study), it is reasonable to assume a normal distribution. An effort was made to determine if  the data used in this study are really normal distribution. A straight line on the probability versus residual plot is a good indication of a normal distribution. As shown in Fig. 5, the line is nearly straight, and the p-value is 0.072. This means that under the null hypothesis and at a certain tolerance level (alpha = 0.05), the hypothesis was not rejected. Thus, the data used in this study are indeed a normal distribution.

In general, for a large data set, the normal distribution assumption is not a concern; rather, the randomness and constant variance assumption is more risky. Efforts were made to verify this assumption. As shown in Fig. 6, the residuals were spread almost equally above and below the zero line (x-axis), which is a good indication that the data set has the properties of randomness and constant variance. Note that fitted values and residuals in Fig. 6 are the results of  the regression on the log (E MPa) and log (PR) data. Thus, the 227 data set is not only normally distributed but also has the properties of randomness and constant variance. Table 2 presents the results of the ANOVA analyses performed on the 227 data set. The derived equation is presented in Eq 8. The R2 was found be 0.855, which could also be obtained by 197.05/230.47, as shown in Table 2. The corrected sum of squares (SS), which measures the total variability in the observations, was 230.47. And the amount of variability in the observation was divided into two components. One was accounted for by the regression line with variability of 197.05. The other one, residual variation left unexplained by the regression line, was 33.42. As shown in Table 2, the standard error of the coefficient of log (PR) was 0.01824, and the T (–36.42) was obtained by 0.66453/0.01824. The statistical test confirmed that the p -value of the variable log (PR) was less than 0.1 %. This indicates that this term was very significant and has a strong ability to explain the model variations. On the other  hand, the same result could be obtained by examining the F value (–36.422 = 1326.7 = 197.05/0.15). Furthermore, the standard error  of the observation was estimated by 0.3854 = (0.15)1/2 . 0.6645

−

E(ksi) = 78.05 ∗ PR

or

R2 = 0.855, MSE = 0.15

0.6645

−

E(MPa ) = 537.76 ∗ PR

,

(8)

where E is young’s modulus, and PR is the penetration rate of the DCP in mm/blow. Figure 7 shows the comparisons between the field test results (DCP mm/blow and FWD moduli) and the model from Eq 8. An effort was made to compare Eq 4 and Eq 8, and the results are presented in Fig. 8, which indicates that themoduli from these two

47

CHEN ET AL. ON PAVEMENT LAYER MODULI TABLE 2 —ANOVA analyses on the log (EMPa ) and log (PR) data.

Predictor

Coef

Constant log(PR)

4.3574 −0.66453

SE Coef 0.0405 0.01824

S = 0.3854 R-Sq = 85.5 % R-Sq (adj) = 85.4 %

T

Source

DF

SS

MS

F

107.51 † −36.42

Regression Residual Error

1 225

197.05 33.42

197.05 0.15

1326.7∗

Total

226b

230.47

∗

†

Significant at 0.1 %. Degree of Freedom (DF) is 226 for a 227 data set.

 Determination of randomness and constant variance. FIG. 6— 

FIG. 7— Correlation between field DCP (PR) and FWD moduli.

FIG. 8— Comparison of the models (Chen, Eq 4 & Powel, Eq 8).

48 GEOTECHNICAL TESTING JOURNAL equations increase as the PR decreases. For example, when the PR is 1 mm per blow, Eq 4 and Eq 8 yield modulus values of 665 and 538 MPa, respectively. This is a difference of 24 % of the Eq 8 result. Whenthe PRis 10mm per blow, the results fromEqs4 and 8 are 128 and 116 MPa, a difference of only 9.6 % of the Eq 8 result. This difference crosses through zero at about 57 mm/blow and is  –1.7 % when the PR increases to 80 mm per blow. Thus, it indicates that themain difference between this study and the model by Powell et al. (1984) is when the PR is small. Based on field DCP tests, it is also observed that the variability among DCP results increases as the PR decreases. Without ground-truth moduli, we are unable to differentiate which equation is more accurate. Equation 8 provides another optionand allows researchers to recognizethe range of variability. It is important to note that there is a limited number of equations to determine the modulus through CBR or PR, and the most widely accepted relationship between CBR and modulus (Eq 3) was proposed by Powell et al. in 1984. The FWD was not yet widely used in 1984.

Conclusion

It is well known that it is time-consuming and costly to repair  problems caused by an inferior base or subgrade. For the purposes of quality control/assurance and for tying test results into design practice, a technique that can provide the stiffness/modulus of the pavement layer is a rational choice. DCPs have been used by several agencies with much success. To fully use the DCP in pavement evaluation, an empirical relationship with layer moduli is required. Over an 8-year period, 198 DCP and FWD tests were conducted on different roadways in Texas. Some of the tests were conducted during construction, that is, before the AC placement. On the basis of the results obtained, the following conclusions were drawn:    r

   r

   r

   r

There are only a few equations to correlate DCP penetration rate and modulus. A correlation between DCP penetration rate and modulus was developed successfully in this study. The equation can be used for both base and subgrade s oils. An effort was made to compare the model developed in this study with that by Powell et al. The difference between these two models increases as the PR decreases. For a PR of less than 10 mm/blow, the difference between these two models was over 10 %. The difference is only about 1.7 % when the PR is 80 mm/blow. The correlation equation developed in this study provides an alternative for researchers and practitioners. It allows them to recognize the range of the variability.

 Acknowledgments This work could not have been completed without the assistance of Mr. Ralph Self, Mr. Carlos Peralez, Mr. Billy Pigg, Ms. Darlene Goehl, Mr. Cy Helms, Mr. Randy Beck, Mr. Norman Erickson, Mr. Magdy Mikhail, Mr. Bill Willeford, Dr. Andrew Wimsatt, and Dr. Mike Murphy of TxDOT.

References

AASHTO Designation: R-97, 1997, “Standard Recommended Practice for Acceptance Sampling Plans for Highway Construction,” American Association of State Highway and Transportation Officials, Washington, DC.

Livneh, M., 1989, “Validation of Correlations Between a Number  of Penetration Tests and In Situ California Bearing Ratio Tests,” TRR 1219. Burnham, T., 1996, “Application of the Dynamic Cone Penetrometer to Mn/DOT’s Pavement Assessment Procedures,” Revised Draft Report, Office of Minnesota Road Research. Chen, D.-H. and Hugo, F., 1998, “Test Results and Analyses of the Full-Scale Accelerated Pavement Testing of TxMLS,” Journal of  Transportation Engineering ASCE, Vol. 124, No. 5., Sept./Oct. 1998, pp. 479–490. Chen, D.-H., Bilyeu, J., and Hugo, F., 1999a, “Monitoring Pavement Response and Performance Using In-Situ Instrumentation Under Full-Scale Accelerated Loading,” Field Instrumentation  for Soil and Rock, ASTM STP 1358, G. N. Durham and W. A. Marr, Eds., ASTM International, West Conshohocken, PA, 1999 pp. 121–134. Chen, D.-H., Wu, W., He, R., Bilyeu, J., and Arrelano, M., 1999b, “Evaluation of In-Situ Resilient Modulus Testing Techniques,” Geotechnical Special Publication No. 89 ASCE, pp. 1–11. Chen, D.-H. and Hugo, H., 2001, “Comparison of the Effectiveness of Two Pavement Rehabilitation Strategies,” Journal of Transportation Engineering ASCE, Vol. 127, No. 1, Jan./Feb. 2001, pp. 47–58. Chen, D.-H., Wang, J. N. and Bilyeu, J., 2001a, “Application of the DCP in Evaluation of Base and Subgrade Layers,” Transportation Research Board, Transportation Research Record No. 1764, pp. 1–10. Chen, D.-H., Bilyeu, J., and Murphy, M., 2001b, “Stiffness Evaluation of Chemical Stabilizers used in the Dallas and Austin Districts,” Proceedings CD-ROM, Second International Symposium on Maintenance and Rehabilitation of Pavements and Technological Control, Auburn, AL, July 29–Aug. 1, 2001. Chen, D.-H., Scullion, T., Bilyeu, J., Yuan, D., and Nazarian, S., 2002, “Forensic Study of a Warranty Project On US82,” Journal of Performance of Constructed Facilities, ASCE , Vol. 16, No. 1, Feb. 2002, pp. 21–32. Chen, D.-H., Bilyeu, J., Scullion, T., Zhou, F., and Lin, D. F., 2003, “Forensic Evaluation of the Premature Failures of the Texas SPS-1 Sections,” Journal of Performance of Constructed Facilities, ASCE, Vol. 17, No. 2, May 2003, pp. 67–74. Chen, D.-H., Bilyeu, J., Scullion, T., Nazarian, S., and Chiu, C. T., 2004a, “Failure Investigation of A Foamed Asphalt Project,” Transportation Research Board Annual Meeting, 2004. Chen, D.-H., Bilyeu, J., and Chang, J. R., 2004b, “A Review of the Superheavy Load Permitting Program in Texas,” Transportation Research Board Annual Meeting, 2004. Livneh, M., Ishai, I., and Livneh, N., 1995, “Effect of Vertical Confinement on Dynamic Cone Penetrometer Strength Values in Pavement and Subgrade Evaluations,” TRR 1473, pp. 1–8. Livneh, M. and Goldberg, Y., 2001, “Use of Falling Weight Deflectomter and Light Drop-Weight for Quality Assessment During Road Formation and Foundation Construction,” 80th Annual Meeting Transportation Research Board, paper number  01-0106. Nazarian, S., Yuan, D., and Arellano, M., 2002, “Quality Management of Base and Subgrade Materials with Seismic Methods,” 81st Annual Meeting Transportation Research Board. Powell, W. D., Potter, J. F., Mayhew, H. C., and Nunn, M. E., 1984, “The Structural Design of Bituminous Roads,” TRRL Report LR 1132, 62 pp.

CHEN ET AL. ON PAVEMENT LAYER MODULI

Rahim, A. M. and George, K. P., 2002, “Automated Dynamic Cone Penetrometer for Subgrade Resilient Modulus Characterization” 81st Annual Meeting Transportation Research Board. Siekmeier, J. A., Young, D., and Beberg, D., 1999, “Comparison of the Dynamic Cone Penetrometer with Other Tests During Subgrade and Granular Base Characterization in Minnesota,” Nondestructive Testing of Pavements and Backcalculation of  Moduli: Third Volume. ASTM 1375, ASTM International, West Conshohocken, PA. Sawangsuriya, A., Edil, T., Bosscher, P., 2002, “Laboratory Evaluation of the Soil Stiffness Gauge (SSG),” 81st Annual Meeting of 

49

the Transportation Research Board, January 2002, Washington, DC. Uzan, J., Scullion, T., Michalak, C. H., Paredes, M., and Lytton, R. L., 1988, “A Microcomputer Based Procedure for BackCalculating Layer Moduli from FWD Data,” Research Report 1123-1, Texas Transportation Institute, Highway Materials Division, College Station, TX, 1988. Webster, S. L., Grau, R. H., and Williams, R. P., 1992, “Descrip tion and Application of Dual Mass Dynamic Cone Penetrometer,” U.S. Army Engineer Waterways Experiment Station, Instruction Report, No. GL-92-3.