CREEP TESTING, A SIMPLE TOOL TO JUDGE ASPHALT MIX STABILITY

Research is reported which consisted of 3 stages: (a) the development of theoretical models; (2) the correlation of these models with unconfined constant-load creep compression tests; (3) and the correlation of the creep test with laboratory rutting experiments. The first stage (theoretical model) was based on relative sliding displacements of pairs of particles of mineral aggregates in a visco-elastic matrix. The results of the first two stages show that the presentation of creep data in the form of the stiffness modulus of the mix as a function of the bitumen is very valuable, because the materials are then described in terms of characteristics that are independent of any test variables and which can be applied in design calculations. The third stage, the results of which are presented in this paper, consisted of unconfined creep tests under continuous and repeated loading. An investiation was made of the two main aspects in which the two test methods (laboratory test track and creep test) differ, viz the contrasts: unconfined versus confined and statis versus dynamic. To single out the influence of the contrast static versus dynamic, parking tests with a static wheel were carried out on the test track pavement, under geometrical conditions equal to those in the rutting test. Details are given of the parking and repeated loading tests. The problems connected with a ranking procedure of asphalt mixes are discussed. The sand asphalt and asphaltic concrete mixes investigated in the creep tests had stiffness modulus which was independent of the stress applied. A comparison of the Marshall results in terms of the Marshall stability and Marshall quotient stability/flow, the rutting deformations measured in the open-air test track at the end of the experiment and the rut depth calculated from the creep testing, revealed poor correlation both with the Marshall stability and Marshall quotient. A mix design procedure quideline is proposed in which the rut depth can be calculated at a chosen value of the design life. These values of rut depth and design life are related with the position of a fixed point on the creep curve. The slope of the creep curve gives an indication of the sensitivity of this method.