Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
1 CHAPTER 1- INTRODUCTION AND RESEARCH APPROACH 1.1 Background Moisture induced damage in asphalt concrete has been widely acknowledged as a serious cause for diminishing the long-term performance of asphalt friction courses. For this reason, determining moisture susceptibility of asphalt mixtures has attracted serious attention of highway agencies and the pavement industry nationwide. An extensive effort has been made to improve moisture susceptibility laboratory experiments so that they properly characterize and predict the behavior of asphaltic mixes in the field. Currently, the majority of the transportation agencies try to control moisture induced damage failures in the field by specifying such laboratory tests. The most common moisture susceptibility test is AASHTO T283 (1), in which asphalt/aggregate mixtures are subjected to mechanical loading after they have been exposed to moisture. Because of the frequent use of the test, it is important that the precision estimates that include the information on allowable difference between test results that are measured in one laboratory and the allowable difference between test results measured in different laboratories to be available. In this regard, the AASHTO Materials Reference Laboratory (AMRL) as part of NCHRP 9-26 looked into the variability of the test by conducting an interlaboratory study, in which Tensile Strength Ratio (TSR) data on two mixtures with expected different levels of moisture susceptibility were collected from over 40 different laboratories. Based on the interlaboratory results, repeatability and reproducibility statistics of the TSR results were determined. Even though this statistical evidence will give a clear indication of the variability of the test, it will not give any insight into the reasons of the possible discrepancies of the test, nor gives any direction towards its improvement. Therefore, to develop a more fundamental understanding of the results of the precision estimates and possible solutions towards an improvement, finite element analyses were made with the Computer Aided Pavement Analyses finite element system, CAPA-3D (2), developed at Delft University of Technology. In the finite element analyses, various micro-scale finite element meshes were made to represent the investigated mixtures. For the finite element meshes, X-Ray tomography scans were made of the representative samples of the mixtures. To simulate the moisture infiltration into the mix components, the finite element meshes were exposed to the same moisture conditioning and temperature cycling as in the laboratory test. As a first step in the project, a short background is given on the T283 test procedure. The concerns about the test resulted from previous experimental studies and the challenges that have so far been encountered with the test are summarized.
2 1.2 Background of the test protocol The AASHTO T283 test method (1) is the result of several alterations to the original Lottman test in an attempt to improve its reliability (3, 4). The basic concept of the test is to compare the indirect tensile strength of dry samples and samples exposed to saturation, freezing, and thawing. The method is used for testing samples prepared as part of the mixture design process, plant control process, and for cores taken from the pavement. The indirect tensile strength test is conducted on the dry and conditioned specimens according to ASTM D 6931 (5). In addition to visual observation for stripping, the ratio of average tensile strength of the conditioned and dry specimens is reported as the tensile strength ratio (TSR): 2 1 STSR S = (1) Where S1 is the average dry and S2 is the average conditioned tensile strength of the sample. For the laboratory mixed-laboratory compacted specimens a minimum TSR of 0.80 is recommended for correlation with field performance (6, 7). Although AASHTO T283 is still the most widely used method for determining HMA moisture susceptibility, highway agencies have reported several shortcomings of the method. One of the major complaints about the test is that the test does not always correctly predict moisture sensitivity of the mixtures as it has been observed in the field. Mixtures that performed well in the field have exhibited unexpectedly low TSR values and poor performing mixtures have indicated unexpectedly high TSR values (8). The research by Epps et al. (9) which included five different mixtures from various states indicated that the sensitivity of the mixtures to moisture damage, as described by the state highway agencies, did not satisfactorily match the observed T283 behavior of a number of mixtures in the study. Another frequently made complaint with regard to the test is the disagreement of the test results between 100 mm (4â) and 150 mm (6â) in diameter specimens. In a survey of 89 agencies compiled by AMRL, a number of state DOTs reported that 100 mm (4â) Marshall specimens indicate better agreement with the field performance than 150 mm (6â) gyratory specimens. However, Epps et al. (9) have shown that 150 mm (6â) gyratory specimens provide less variable results than 100 mm (4â) Marshall Specimens. The other complaint about the AASHTO T283 test method is regarding the conditioning of the test. It has been stated that the duration and severity of saturation and moisture conditioning does not always promote the stripping of the mastic. Choubane et al. (10) have suggested saturation levels above 90 % and multiple freeze-thaw cycles in order to promote stripping. They found that degrees of saturation of 55 % versus 80 % would result in significantly different tensile strength of the mixtures. In addition, Kandhal and Rickards (11) showed that in four different case studies of stripping in asphalt pavements, the asphalt pavement was nearly 100 % saturated with water, which is much higher than the saturation level that is recommended in AASHTO T283.
3 An additional reported complaint about the T283 test is the mode of mechanical testing of the specimens. Kandhal and Rickards (11) have argued that a cyclic load which can simulate the pumping action of traffic load is a better test than loading the samples with a constant rate. Finally, a last complaint about the test that is often reported by state DOT engineers is that the test is very time-consuming. Several state highway agencies follow a shortened version of AASHTO T283 test method, which might provide different findings than if all steps of the test are followed (12). In this study the variability of the test is being quantified by a means of an interlaboratory study involving two asphalt mixtures with different levels of moisture susceptibility tested by more than 40 laboratories. As a result, precision estimates of AASHTO T283 based on the TSR results from the laboratories will be developed. In addition, a theoretical and computational analysis of the moisture infiltration in the chosen mixtures and a discussion regarding the structural nature of the test is given, which address both some of the complaints regarding repeatability and comparisons between laboratory results and the field. Toward the end of the report some additional comments are made with regard to the general reported complaints about the test, as summarized in the above. 1.3 Problem Statement The accurate and precise characterization of moisture resistance of asphalt mixtures is an important aspect of selecting appropriate mixtures for various paving projects. AASHTO T283 has been the most commonly used test method for detecting moisture susceptibility. There are reports on high variability of the test results; however there is no information on the precision estimates for the test method. In addition, the causes of the variability in the test results are not clearly defined. 1.4 Research Objectives The overall goal of this study is to determine the precision estimates of AASHTO T283 test methods. The following objectives follow from this goal: 1. To evaluate causes of variability of the test results. X-ray tomography images and finite element modeling will be used to examine the effect of specimen structure in moisture conditioning. 2. To recommend modifications for improvement of asphalt mixture moisture damage test. 1.5 Scope of Study The scope of the project involved the following major activities: I. Conduct Preliminary laboratory test according to AASHTO T283: a. Select materials and mixture design.
4 b. Prepare the test specimens. c. Scan the specimens using X-ray computed tomography to obtain insight information of specimensâ structure. d. Condition the specimens according to AASHTO T283. e. Conduct strength test on the dry and conditioned specimens to evaluate moisture sensitivity of the selected mixtures. f. Analyze the results of the preliminary study. II. Design and conduct an interlaboratory study (ILS): a. Prepare instructions for preparation, conditioning, and testing of the ILS specimens. b. Identify the laboratories participating in the ILS. c. Send the materials (aggregate, asphalt) and instructions to the participating laboratories. d. Analyze results of the ILS to evaluate accuracy and precision of the AASHTO T283 test method in determining moisture susceptibility of the selected mixtures. e. Prepare a precision statement for AASHTO T283. III. Examine the causes of variability of the AASHTO T283 test results: a. Analyze the X-ray images for 3-D computation of size and distribution of air voids in the compacted specimens. b. Use the CAPA-3D finite element program to simulate the moisture infiltration into the specimens during moisture conditioning. c. Discuss the possible causes of variability in laboratory moisture damage test. IV. Make conclusions and recommendations based on the findings of the study.
