Determination of the Fatigue Behaviour of Asphalt Base Mixes using the Indirect Tensile and the 4 Point Bending Test

Transcription

1 Fakultät Bauingenieurwesen - Institut für Stadtbauwesen und Straßenbau - Professur für Straßenbau Determination of the Fatigue Behaviour of Asphalt Base Mixes using the Indirect Tensile and the 4 Point Bending Test Dr Christiane Weise, TU Dresden Dr Sabine Werkmeister, TU Dresden Dr Markus Oeser, UNSW Sydney Rolf Rabe, BASt

2 Background of the Research Determine input parameters of asphalt base mixes for the mechanistic pavement design procedure. The influence of the air void content, the aggregate variety, the bitumen content, the provenance and also the paving grade of the bitumen on the fatigue behaviour was investigated. Indirect Tensile Test (ITT) and Four Point Bending Test (4PBT) were carried out on three asphalt base mixes Tests displayed a difference of the E-Modulus values due to compression or tension particularly at temperatures above 5 C - the effects of the differences on the stress state within a specimen of the 4PBT were investigated using FE simulations. Slide 2

3 Aggregates and bitumen used for the tests Code Aggregate Type Bitumen Provenance G1B0 Greywacke 50/70 C (mixing plant) G1B1 Greywacke 50/70 A G1B7 Greywacke PmB 45A A Slide 3

4 4PBTs Test Conditions Parameter Wave Performed Test Conditions sinusoidal stress, with cycling through zero Frequency f [Hz] 30 Duration of Loading [s] 1/f Rest period [s] none Test Temperature [ C] stiffness: -10, 0, +10, +20 fatigue tests: +20 Stress amplitude triple variation Slide 4

5 Dynamic ITTs Test Conditions Parameter Wave Performed Test Conditions sinusoidal stress, without cycling through zero Frequency f [Hz] 10 Duration of Loading [s] 1/f Rest period [s] none Test Temperature [ C] -10, 0, +10, +20 Lower Stress Upper Stress contact stress / temperature-induced stresses triple variation Slide 5

6 Test Equipment Slide 6

7 Fatigue relations obtained by ITT at 10 Hz and 4PBT at 30 Hz Slide 7

8 Fatigue curves; influence of the test frequency on the number of endured load cycles (left); same results but plotted over the test duration until cracking (right) Slide 8

9 Fatigue relations obtained by ITT at 10 Hz and 4PBT at recalculated 10 Hz for the G1B0 mix Slide 9

10 Fatigue relations obtained by ITT at 10 Hz and 4PBT at recalculated 10 Hz for the G1B1 mix Slide 10

11 Fatigue relations obtained by ITT at 10 Hz and 4PBT at recalculated 10 Hz for the G1B7 mix Slide 11

12 FE mesh (quarter beam model) x 3 x 2 x 1 Slide 12

13 E-Modulus versus temperature for a SMA Slide 13

14 FE calculation results - stress σ 1 distribution in a 4PB quarter beam model induced by downward force beam dimensions: length: 260 mm width: 40 mm height: 40 mm Slide 14

15 Conclusions Both tests (ITT and 4PBT) are suitable to investigate the fatigue performance of asphalt base mixes Using appropriate test-analysing methods, similar test results (stiffness modulus and fatigue curves) can be determined If tests were conducted at different frequencies the test duration approach can be applied to recalculate the number of load cycles at different frequencies. The number of load cycles until fatigue failure was independently from the test frequency, but depends on the test duration and hence on the energy applied on the sample The deformation measurement equipment for the 4PBT needs to be improved - strain gauges should be glued directly underneath the test beam Slide 16

16 Conclusions cont. Due to the inhomogeneous sample structure caused by the grain assembly and large grain dimensions a disturbed stress distribution within the testing sample occurs. For analysing the test results according to the European Standard, a homogenous stress distribution is assumed. The test results might be inaccurate because of the disturbed stress state induced in the sample. For the ITT and 4PBT the crack-mechanism is similar. If a crack occurs, the stresses are redistributed and a stress concentration occurs at the peak of the crack. This stress concentration leads to an acceleration of the crack growths. However, due to the different specimen geometry the crack propagation mechanism and hence the crack growth rate differs for the 4PBT and the ITT tests. The crack mechanism occurring during the ITT and the 4PBT tests differ from the crack mechanism in a actual pavement. This must be always considered when using the results of these tests to predict the fatigue behaviour of pavements. Slide 17

17 Fakultät Bauingenieurwesen - Institut für Stadtbauwesen und Straßenbau - Professur für Straßenbau Thank you for your attention! Muito obrigado! Dr Christiane Weise, TU Dresden Dr Sabine Werkmeister, TU Dresden Dr Markus Oeser, UNSW Sydney Rolf Rabe, BASt