During recent decades, growing interest in the developmentof so-called analytical or mechanistic road pavement designmethods has resulted in substantial research into themechanical behavior of the materials involved. In flexiblepavements, especially when unsurfaced or thinly surfaced,overall structural performance is largely dependent on thebehavior of the unbound granular base and subbase layers. Aproper understanding of the mechanical properties of granularmaterials is, therefore, a prerequisite for the success ofanalytical design procedures for flexible road pavements. Thepresent doctoral thesis is based on a five-year researchproject on the structural characterization of unbound granularmaterials and consists of six papers on different aspects ofthe research topic.
Papers I and II present the findings from an extensiveliterature survey and summarize the current state of knowledgein regard to resilient and permanent strain behavior ofgranular materials. The review of previous research shows thatthe structural response of these materials is affected, tovarying degrees, by several factors such as stress level,moisture content, stress history, density, fines content,aggregate type and number of load applications. Different viewson the significance and the extent of the impact of eachindividual factor are presented and discussed. A great numberof mathematical models have been developed over the years forprediction of stiffness properties and long-term performance ofgranular materials. The models found in the literature arelisted and their advantages and shortcomings are reviewed.
The focus of Papers III and IV is on the issue ofmathematical modeling of permanent strain development inunbound granular materials. The results of a series of repeatedload laboratory tests are used to assess the relationshipbetween permanent strain and both number of load repetitionsand stresses. The experimental data are used to compare thepredictive capabilities of some of the existing models found inthe literature, including the so-called Paute model which ispart of a newly proposed European standard for road materialcharacterization. Through further analysis of the test results,a new model is introduced expressing the accumulated permanentaxial strain, after any given number of load cycles, as afunction of the maximum shear to normal stress ratio and thelength of the stress path applied. Based on the application ofthis model, it is suggested that a shakedown approach, similarto that proposed by others for pavement analysis, would be auseful means for relating permanent strain development andlong-term performance of granular materials to stresscondition. Further, it is shown that permanent straindevelopment under repeated loading cannot be sensibly relatedto monotonic shear strength data, a common surrogateperformance measure referred to in the literature.
Paper V presents, in detail, the design work for alarge-scale triaxial testing apparatus developed as part ofthis research project. This triaxial apparatus uses a specimensize of 500x1000 mm and enables testing of granular materialswith up to 100 mm maximum particle size at cyclic deviatoricand confining stresses of up to about 1300 kPa and 600 kPa,respectively. The new triaxial equipment was employed in aseries of tests on different granular materials, the results ofwhich are also presented and discussed. The materials testedwere crushed limestone, granite and concrete, and natural sandand gravel. The structural response of crushed concrete provedto be the best, showing the highest stiffness and resistance topermanent deformation. The suitability of crushed limestone wasrather doubtful, as it exhibited high stiffness but lowresistance to permanent deformation. The performance of granitewas moderate, whereas sand and gravel behaved poorly.
Finally, Paper VI presents the findings of an investigationinto the effects of grading scale, or changes in the maximumparticle size of graded aggregates, on triaxial testingresults. Since most triaxial testing facilities available todate do not satisfy the specimen size requirements forcoarse-grained aggregates, these materials are commonly testedin so-called scaled-down gradings. This paper attempts toexplore the importance of grading scale on the basis of aseries of repeated load triaxial tests on crushed limestone andconcrete, and natural sand and gravel at gradings of 0/90,0/63, 0/32 and 0/16. The test results showed clearly that bothresilient and permanent strain responses are affected by thegrading scale of the triaxial specimen. The reduction ofgrading scale resulted in lower resilient moduli and higherPoisson's ratios, but the extent of the impact was dependent onaggregate type. The susceptibility to permanent strain was alsoinfluenced markedly, but the nature of the response proved tobe highly inconsistent. On the basis of the experimentalanalyses presented in this paper, it is found necessary fortriaxial tests on granular materials to be conducted at theirnatural gradings.
KEYWORDS:Granular materials, unbound aggregates,mechanical behavior, resilient response, permanent strain,mathematical modeling, shakedown, triaxial apparatus,large-scale triaxial testing, grading scale, maximum particlesize
Institutionen för infrastruktur och samhällsplanering , 1999. , 14 p.