![Sample our Engineering & Technology journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5933/TOC-Subject-Sample-2021-Engineering-Tech.jpg"})
Abstract
Cracking is one of the most prevalent types of distresses in asphalt pavements. There are different cracking index parameters that are determined from tests conducted on binders and mixtures to assess cracking potential. The objective of this study is to compare binder and mixture parameters and evaluate the similarities and differences between the rankings and values obtained. This study includes binder and mixture testing on 14 plant-produced mixtures including 3 different binder grades, 3 binder sources, 3 aggregate gradations, and mixtures containing a range of reclaimed asphalt pavement and/or recycled asphalt shingles contents. Testing included PG grading and 4 mm dynamic shear rheometer testing on the extracted and recovered binders that were long-term aged. Mixture testing included complex modulus, simplified viscoelastic continuum damage (SVECD) fatigue, and disc-shaped compact tension testing on short-term-aged mixtures. Parameters evaluated included high and low PG temperatures, ΔTcr, Glover–Rowe parameter (binder and mix-based), R value, dynamic modulus, phase angle, number of cycles to failure from SVECD and layered viscoelastic critical distresses analysis, and fracture energy. The results show that generally the binder parameters correlate well with each other but the mixture parameters do not. Good correlation was observed between binder and mixture stiffness-based parameters, but there was generally low correlation observed between binder and mixture cracking parameters for the mixtures evaluated in this study, possibly a result of differences in ageing level. Recommended future work includes non-linear statistical analysis, incorporation of field performance, and testing on long-term-aged mixtures.
Acknowledgements
The authors would like to acknowledge Beran Black and Matthew Courser at NHDOT for their efforts in conducting binder testing, and Mary Westcott, David Duncan, and Peter Moore at Pike Industries Inc. for fabricating specimens and providing production information.
Disclosure statement
No potential conflict of interest was reported by the authors.
7. Discussion
LUDO ZANZOTTO: This nice paper of yours is talking about the correlation of the properties of binders and the properties of paving mixes. That includes also the correlation of the low temperature cracking of mixtures and binders. What I would like to note is that there is one method for the determination of the cracking temperature of paving mixes that was not mentioned in your presentation and which is working for us very well for a very long time. The method was proposed by Dr. Rowe and we are using it for more than ten years. The method is using the indirect tensile test. We perform the indirect tensile test so that the lowest testing temperature is below the expected critical cracking temperature – we can go as low as minus 40°C. That strongly improves the accuracy of the results. For the determination of the cracking temperatures of binders we are using the direct tension test with modified sample preparation procedure. We have been constantly finding that the critical cracking temperatures of binders and the critical cracking temperatures of paving mixes prepared with the same binders are within 2°C. In the determination of the critical cracking temperatures of the paving mixes we are using software developed by Dr. Rowe.
REYHANEH RAHBAR-RASTEGAR: Thank you. Thank you for your comment.
AMIR GOLALIPOUR: Thanks so much, Reyhaneh, for a great presentation. This is a great study and a lot of good data that you put together for both binder and mixture. Those binder tests were conducted on only as-recovered binders or did you age them too?
REYHANEH RAHBAR-RASTEGAR: Those are recovered binder and then after 20 hours RTFO.
AMIR GOLALIPOUR: So, you did RTFO aging and then 20 hours PAV on them?
REYHANEH RAHBAR-RASTEGAR: Yes. Sorry, I meant 20 hours PAV.
AMIR GOLALIPOUR: Have you done any 40 hours PAV on those materials?
REYHANEH RAHBAR-RASTEGAR: We did 40 hours on some of the mixtures because we didn’t have enough material to do 40 hours on all of them. That’s why we didn’t include 40-hour PAV results in this paper. Just 20-hour PAV.
AMIR GOLALIPOUR: Did you see any type of correlation for those materials at 40-hour PAV?
REYHANEH RAHBAR-RASTEGAR: Honestly, that’s something I need to take a look at.
AMIR GOLALIPOUR: Also, I have one comment. All the binder parameters that you looked at – the G*, phase angle, ΔTC, and Glover-Rowe parameter – they're all captured in linear viscoelastic region of binder behavior. But when you are looking at your mixture testing, you are running some tests like AASHTO TP 107 (AMPT cyclic fatigue) that you call it, [S-VECD] and also DCT. Those are cracking tests or damage tests where the material is in a very different behavior state. So, I would suggest that if you want to look at apple to apple comparison for these types of properties, maybe it is a good idea to look at some cracking or damage test for binder as well. Dr. Don Christensen looked at DENT test and LAS. I know that in NCHRP 9-59 project, they are looking at some of those tests. Maybe you will find a better correlation between binder and mixture properties using those test results.
REYHANEH RAHBAR-RASTEGAR: Exactly. Thanks for your comment. That was something that I wanted to mention but I forgot. That’s another thing that can make a difference between mixture and binder cracking parameters. Thank you.
AMIR GOLALIPOUR: Great job. Thanks.
REYHANEH RAHBAR-RASTEGAR: Thanks.
JO DANIEL: I just wanted to clarify that one of the objectives of this New Hampshire DOT project was to see if the specification binder parameters are enough to do the prediction. So, part of the overall goal of the project was specifically to do that comparison.
IMAD AL-QADI: Thank you. That was a very interesting presentation. So, I’m going to comment on Jo Daniel’s comment. Logically there should be a good correlation between binder and mixture fundamental properties. This is due to the fact that during testing most of the strain is happening in the asphalt and almost nothing in the aggregate; such as cracks occur in asphalt binder. When adding RAP, it will take some time for its binder to diffuse, not all of it may be diffused, however. On the other hand, in the case of RAS, I think maybe a very small amount of RAS binder is diffused, and possibly none. During extraction, the binders are forced to mix. So, when running the test on the extracted binder, the behavior would be different than the actual mix binder. Logically you would see a correlation if you separate the two binders.
REYHANEH RAHBAR-RASTEGAR: You’re right, but sorry I think I didn’t get your question.
IMAD AL-QADI: What I’m saying is in this case here, when you extract a binder that has RAP binder or RAS, you are forcing all binders to mix. So, they are different than the material tested as part of the mixture itself. Hence, some problems in the presented correlation.
REYHANEH RAHBAR-RASTEGAR: You're right. Thanks for your comment. But I’m thinking that if you want to predict the performance of asphalt mixture based on asphalt binder data, testing on extracted and recovered binder is what we usually do. So finally, we might get to that point that it’s not the right way and maybe we want to consider the results of mixture side, but is there any other way that we can do instead of testing extracted and recovered binder?
IMAD AL-QADI: No, what you are doing is correct. I didn’t criticize what you are doing.
REYHANEH RAHBAR-RASTEGAR: No, thank you. I was just curious about that.
IMAD AL-QADI: I’m just saying you may need to look at the data from a different point of view. Thank you.
REYHANEH RAHBAR-RASTEGAR: Yes. Thank you.
GEOFFREY ROWE: Reyhaneh, very nice presentation. I’ve just got a point to make on the correlation concept when you're looking at master curves, and maybe a suggestion to consider. You have got in your presentation, a table (slide 14) looking at the correlation between some of the parameters you used. I think you had three slides in your presentation where you looked at different parameters on master curves versus the correlation factors (based upon your Table 6 and 7 in the paper). What you’ve got there is you’ve got really two types of parameters. I would suggest you’ve got parameters which are more descriptive of the shape of the master curve, and you’ve got parameters which are really point parameters on that master curve. I think of a ΔTC – it is more descriptive of the shape of a master curve. A low number of ΔTc is more akin with a gel binder or the very high curvature on a Black space, whereas with a high negative ΔTc, you're getting a more oxidized and a flatter line on a black space. So that’s really capturing something akin to a rheological index or those sorts or types of parameters, capturing the shape of the master curve. I wouldn’t necessarily expect G*, which is a single point parameter, to correlate well with a shape parameter. I think when you're looking at some of these correlation studies, it might be better to split the parameters out into point parameters and shape parameters. I know you didn’t have time to allude to it in the slides, but you also looked at the shape of the mixture master curve and the inflection point. You had some discussion on that in your paper. I think there are some things that we can do with the shapes of these curves that would be nice to further investigate, to try to look at, say, shape parameter correlations and then point parameter correlations and then see how they relate to performance! Nice paper, thank you very much.
REYHANEH RAHBAR-RASTEGAR: All right. Thank you for your comment. And actually, we considered R value and shape parameters in the paper, but because of the time, I didn’t include the results here. There was a good correlation between R value and G-R parameter, for example. But you’re right, I think it’s a good idea to split shape and point parameters.
GEOFFREY ROWE: My point really is that your poor correlations in slide 14 between the parameters would be expected. I wouldn’t have expected them to correlate anyway because you're basically trying to correlate a point parameter with a shape parameter.
REYHANEH RAHBAR-RASTEGAR: Yes.
GEOFFREY ROWE: I wouldn’t expect them to correlate, if you see what I’m saying.
REYHANEH RAHBAR-RASTEGAR: Yes.
DON CHRISTENSEN: Very good presentation, very interesting. I have a comment that I think is related in some ways to what Geoff just said, and that is when you look at these rheological parameters like ΔTc, G* and this kind of thing, in a lot of cases they’ve been correlated to field performance and you have to think about why that is or what the physical reason is. And in a lot of cases, I think it’s complicated. It might have to do with things like thermal fatigue or healing or lack of healing or loss of adhesion during aging. So, these are phenomena that you're not necessarily going to be able to correlate directly to a mixture fracture property. So as Geoff said, it’s not surprising to me that you're not really seeing good correlations, but that doesn’t mean that these binder tests aren’t going to correlate to field performance. I don’t know that you necessarily need to respond to that. It’s just an observation that you need to, I think, be careful sometimes when you do these sorts of correlations.
FRANK FEE: In terms of relating the binder and the mixture, what you might try is just adding another half percent of binder to the mix and then test it.
GERALD REINKE: I wasn’t going to say anything until Frank mentioned the half percent AC. We’ve done the testing and it doesn’t make any difference whether you add the extra half percent AC or not. At some point, we may show that here, but it doesn’t make any difference at all in terms of the aging of the material.
REYHANEH RAHBAR-RASTEGAR: Thanks for your comment.