Elsevier

Construction and Building Materials

Volume 148, 1 September 2017, Pages 666-674
Construction and Building Materials

Rutting depth prediction of hot mix asphalts modified with forta fiber using artificial neural networks and genetic programming technique

https://doi.org/10.1016/j.conbuildmat.2017.05.088Get rights and content

Highlights

  • Adding fiber shows a higher rutting resistance due to lower thermal sensitivity.

  • Increasing the amount of forta fiber lead to reduce permanent deformation.

  • Better interlock between aggregates and forta fiber reduces the deformations.

  • Neural network model has shown good agreement with experimental results.

  • Genetic programming model has less error than the Burgers model.

Abstract

The most significant problems in the maintenance of highway networks are low strength against dynamic loads and short service life of pavements. In recent years using additive materials to improve the performance of asphalt mix under dynamic loading has been remarkably developed. Previous research show that adding appropriate polymer materials to hot mix asphalt improves the dynamic properties of these mixtures. A series of dynamic creep test were conducted under different temperatures and stress levels to evaluate rutting performance of asphalt samples. The proposed artificial neural networks (ANN) model for rutting depth has shown good agreement with experimental results. Beside, in this study a comparison is made between the Burgers model and genetic programming (GP) model in estimating the rutting depth of asphalt mix. Performance of the genetic programming model is quite satisfactory. The obtained results can be used to provide an appropriate approach to enhance the performance of asphalt pavements under dynamic loads.

Introduction

In the early twentieth century, the technology of building asphalt and concrete roads was transferred to out of the cities and to building the roads too [1]. Asphalt concrete is one of the materials widely used for pavements of the roads and airports. The researchers and engineers are constantly trying to improve the performance of asphalt pavements. Road pavement, as the surfaces frequently loaded by heavy axes, must have sufficient resistance to fatigue, cracking, creeping and sliding [2]. Life span of road pavement is an important topic in national economy. A good pavement must provide a smooth surface for driving, tolerate the high volume of traffic and transfer the tension to lower surface with the minimum loss [3]. The destructions occurred during useful life of the pavements primarily include permanent deformation in wheel track of the vehicles, rutting and thermal cracking. Since high costs must be spent for rehabilitation and reconstruction of these defects and errors, therefore early prevention is usually more affordable. To avoid these deconstructions, the pavement materials must be selected in a way to have sufficient strength and stability [4]. One of the problems of asphalt pavements is their considerable creep. Creep phenomenon is the gradual emergence of subsidence and sustainable displacements without cracks in pavements under the applied positive loads. Sustainable deformations that appear objectively as rutting of wheel tracks are considered as the primary criteria for an asphalt pavement project [5]. Excessive rutting usually referred to as the main cause of premature destruction and maintenance operations of road network, will lead to reduction in service life of the pavement [6], [7]. Excessive use of bitumen, increase of fine aggregates, and high content of river stone materials and rounded particles of stone materials are the usual reasons depending on material properties that influence the lasting transformation of the work [8]. Cracking is the most important destruction state of asphalt pavement. Constructing a pavement that does not crack within a certain period of time after exploitation is impossible. Cracking is an inevitable problem raised in over two million miles of paved roads of United States [9]. Since cracking is the major form of pavement destruction, it is usually considered as a decisive factor in determining the appropriate time and method for rehabilitation [3]. The horizontal shear force created under the wheels of vehicle is the cause of great stresses and strains on the pavement surface and it moves in a direction perpendicular to the path. These strains occur near edges of the tire and create longitudinal cracks. Considering the fact that width of the loading area is limited to vehicle’s track, these transverse cracks only develop in a limited length. These cracks are spread by flexural performance and shear performance of traffic load. Creation and spread of asphalt pavement cracks have various reasons, but the mechanisms involved in it can be classified in three forms: traffic, thermal and surface [10]. To enhance flexibility of pavements and also to increase their resistance to destructive factors such as fatigue, cracks caused by severe temperature changes and stable deformations, additives with the potential to improve the mechanical properties of asphalt pavements are applied in production of hot mix asphalt in recent years [11]. Anti-stripping additives and polymer modifiers are two common modifying methods for improving the fundamental properties of bitumen binders. Cohesion and adhesion are two important related factors of bitumen binders that can influence the performance of asphalt mixture. Results of the researches showed that the mixtures modified by polymer have a better performance compared to unmodified mixtures and the mixtures modified with anti-stripping additives. The studies have shown that the polymers improve the performance of adhesive bitumen against rutting and its adhesion and cohesion [12]. Yousofi and Ramzani (2005) studied the effect of modifying bitumen 60/70 of Isfahan refinery with two light polyethylene polymers (LDPE 200) and Styrene-Butadiene random co-polymer (SBR 1712), both of which are domestic products, on properties of the resulted asphalt mixes. It was observed that Marshal strength and fluidity of asphalt mixes was respectively increased and decreased in presence of polymer [13]. Zoorob et al. stated that plastic wastes can replace a part of stone materials or be used as bitumen modifiers. Dense Bitumen Macadam (DBM) along with plastic junks the major part of which is formed by light polyethylene (LDPE), decrease density of the mixture up to 16%. Also, they result in a 250% increase in Marshall Stability and an improvement in indirect tensile strength in plastic asphalt mixtures. Very detailed information about the applications of artificial neural networks in transportation engineering can be found in the relevant literature [14]. Shafabakhsh et al. evaluated the application of ANN in predicting permanent deformation of asphalt concrete mixtures modified by Nano-additives. A total number of 270 asphalt mixtures were constructed from two different aggregate sources (natural and steel slag) and were modified by micro silica and Nano TiO2/SiO2. All samples were tested at three different testing temperatures of 40, 50, and 60 °C and five stresses of 100–500 kPa. An ANN model developed using five input parameters including: aggregate source, additive type, additive content, temperature, and stress. The result indicates that the proposed model can be applied in predicting final strain of asphalt mixtures. The model is further applied to evaluate the effect of different percentages of Nano-additive on permanent asphalt deformation. Results show that an increase in percentage of Nano-additives is very effective in reducing the final strain of asphalt mixtures [15]. Xiao et al. studied the Prediction of Fatigue Life of Rubberized Asphalt Concrete Mixtures Containing Reclaimed Asphalt Pavement Using Artificial Neural Networks. In this study over 190 fatigue beams were made with two different rubber types (ambient and cryogenic), two different RAP sources, four rubber contents (0, 5, 10, and 15%), and tested at two different testing temperatures of 5 and 20 °C. The data were organized into nine or 10 independent variables covering the material engineering properties of the fatigue beams and one dependent variable, the ultimate fatigue life of the modified mixtures. The traditional statistical method was also used to predict the fatigue life of these mixtures. The results of this study showed that the ANN techniques are more effective in predicting the fatigue life of the modified mixtures tested in this study than the traditional statistical-based prediction models [16]. Another study in 2009, explores the utilization of the artificial neural network (ANN) in predicting the stiffness behavior of rubberized asphalt concrete mixtures with reclaimed asphalt pavement (RAP). A total of 296 asphalt mixture beams were constructed from two different rubber types (ambient and cryogenic), two different RAP sources, and four rubber contents (0, 5, 10, and 15%). All samples were tested at two different testing temperatures of 5 and 20 °C. The regression statistical method was used to predict the stiffness behavior of these mixtures via the 7 input variables covering the material engineering properties of the asphalt beams. In addition, the data were organized into 5 independent variables and one dependent variable (the stiffness values of the modified mixture beams) in ANN models. Results showed the ANN techniques to be more effective in predicting the fatigue life of the modified mixture than traditional regression-based prediction models [17]. Only a few studies have utilized genetic algorithms and genetic programming approaches in engineering applications such as pavement [18]. Tack & Chou developed genetic algorithm using a tool that optimizes the schedule for repair of multilayer pavements [19]. Alavi et al. employed a high precision model to estimate the flow index of high density asphalt mixtures by a new hybrid procedure that combines simulated annealing with programming [20]. The goal of this research is prediction of rutting depth of HMA modified with forta fibers by using artificial neural networks and genetic programming technique. To achieve this goal, a series of dynamic creep test were conducted under different temperatures and stress levels.

Section snippets

Used materials

The continual IV scale in ASHTOO standard [21] was used for grading the aggregates in this study. (Table 1) shows aggregates gradations used in this research.

Four sizes of stone materials (two coarse aggregates, one fine aggregate and a filler) were used in this research and these materials were graded according to ASTM D136 standard method and used in the ratios that the material mixture grading be within range of number 4 grading specification of AASHTOO standard with maximum size of 19 mm.

Creep test results

The effects of stress levels on deformation of the samples have been presented in Fig. 2. Considering the results obtained from the tests, it is observed that deformation of the samples is increased by increase of stress. By adding the fibers to the samples, we observe a considerable drop in deformation which occurs due to placing the fibers between aggregates of consumed materials and strengthening the interlock among the aggregates. Reduction process of deformation caused by 40, 50 and 60

Conclusion

This study presents an application of artificial neural networks (ANN) and genetic programming (GP) technique for the prediction of rutting depth of hot mix asphalt modified with forta fibers. According to this paper achievement:

  • The asphalt mixes containing 0.5% by weight of fiber have a higher true density compared to other samples, but their theoretical specific weight is less than other samples.

  • Performance of the samples containing 0.7% by weight of fiber was diagnosed to be weak, but they

References (21)

There are more references available in the full text version of this article.

Cited by (0)

View full text