Frequently, load associated mode of failure, rutting and fatigue, are the main failure types found in some newly constructed roads within Baghdad, the capital of Iraq, and some suburban areas. The use of excessive amount of natural sand in asphalt concrete mixes which is attractive to local contractors could be one of the possible causes to the lack of strength properties of the mixes resulting in frustration in the pavement performance. In this study, the performance properties of asphalt concrete mixes with two natural sand types, desert and river sands, were evaluated. Moreover, five replacement rates of 0, 25, 50, 75, and 100% by weight of the fine aggregate finer than 4.75 were used. The performance properties including moisture susceptibility, resilient modulus, permanent deformation, and fatigue characteristics were evaluated using indirect tensile strength, uniaxial repeated loading and repeated flexural beam tests. Also, as a part of the research objective, the laboratory tests result were used to predict the performance using VESYS computer software. Results showed that mixes with high natural sand content (NSC) are more susceptible to moisture damage and rutting, lower resilient modulus and some improvement in fatigue resistance. Based on the obtained results, the necessity has rise to revise the current local specification for asphalt concrete which specifies the limits of natural sand content in the mixes of wearing and binder courses with 25% whereas for base course mixes no limit exist yet.
Arabic text classification is a challenging task because of the complex morphology of the language, the existence of different writing forms and a multitude of dialects, which can result in sparser common text representations. While transformer models such as AraBERT have obtained superior results on many Arabic NLP tasks, their high computational requirements make them difficult to deploy in environments with limited hardware resources. In some cases this can also make the model less practical for researchers working with basic computer systems. This study focuses on a more practical issue: how much accuracy a simple classifier may lose when the amount of required computation is reduced. We use a combined TF-IDF representation based on bo
... Show More