In this project a bending strength test on thin beams of asphalt binders and mixtures was developed to evaluate low temperature properties of these paving materials. Work in Task 1 focused on preliminary development of a bending strength test method. The instrument manufacturer (Cannon Instruments) delivered a modified Bending Beam Rheometer (BBR) with a new proportional valve system and a heavier loading frame. A large number of beam replicates for asphalt binders and mixtures were prepared and tested, and a method detailing sample preparation and loading protocol was developed. Statistical analysis of the results showed significant differences between asphalt binder bending strength and direct tension strength. Asphalt mixtures were also tested under different conditions and it was found that strength was mainly affected by temperature, loading rate, and air voids. A review of fracture mechanics indicated that a coupled energetic statistical approach to strength and size effect was more practical in terms of keeping experimental tests simple. Work in Task 2 consisted of testing and comparing the results from the proposed testing method with the results from the standard strength test methods for asphalt binders and mixtures. The significant difference between BBR and direct tension test (DTT) asphalt binder strength results led to investigating the effect of cooling medium (ethanol and potassium acetate). It was found that ethanol significantly reduced the strength values, most likely due to environmental stress cracking. Weakest Link Model (WLM) confirmed that strength values obtained with DTT and BBR using potassium acetate were comparable. Simple statistical comparison between BBR and indirect tension test (IDT) strength results showed that the former test gave higher strength values. However, IDT test does not have a predictable size effect due to high sensitivity of strength results to test geometry and loading apparatus, which makes material strength extrapolation to larger structures difficult. Consequently, asphalt mixture size effect was investigated in Task 3 by testing beams of different sizes. Probabilistic theory and WLM were applied to determine the failure probability of the Representative Volume Element (RVE) of asphalt mixture and a set of closed form equations was derived to calculate the characteristics parameters of the mean strength size effect curve (energetic statistical size effect law).It was concluded that asphalt mixtures present a quasi-brittle failure behavior at low temperature. In Task 3 issues related to “tension” vs. three-point “bending” strength were also addressed, while, in Task 4, the effect of cooling medium (ethanol, potassium acetate and air) on BBR strength measurements of asphalt binder and mixture was further investigated. Testing in air was found to be the most representative of field conditions; however, cooling in air poses additional temperature control problems. A model to describe the substructure of asphalt mixture RVE was proposed, since the dimensions of the BBR beams are not large enough to fully describe statistically the strength properties of the material. This model was used to extrapolate the BBR mixture strength results to larger structure sizes. The final report documents the development of the low temperature bending strength test along with an assessment of the technology for implementation. A step-by-step method to analyze and interpret data from BBR strength testing of thin beams of asphalt binder and mixtures is also included.

The contractor's final report is available.