Relationship of pavement layers’ bearing capacity between laboratory DCP tests and field performance

Date
2023-03
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH SUMMARY: This research study forms part of the SANRAL Project 3.5, focused on the improvement of performance-based test methods utilised in pavement engineering in the South African context, specifically aimed at unbound granular material. As most existing test methods in South Africa are based on empirical formulations, a lack of correlation exists between the test results and in-situ material conditions, and therefore the prediction of the material performance has often been inaccurate. The performance-based test methods are aimed at characterising the material strength as well as a prediction of the performance of the material under the influence of repeated traffic loading. This study considers the influence of moisture content, density, specimen geometry, maximum particle size and aggregate packing on the performance on unbound granular materials in South Africa. To investigate the influence of moisture content and density, three different moisture contents were incorporated in the test procedure, combined with two different compaction energies. To investigate the influence of specimen geometry, two specimen sizes are considered in the experimental methodology. The influence of maximum particle size and aggregate packing is considered by implementing two material gradings, with different maximum particle sizes implemented by scalping the material at the required sieve size. The smaller specimen grading, scalped at the 20 mm sieve, was labelled as “S20” material. the larger specimen grading, scalped at the 28 mm sieve, was labelled as “S28” material. The investigation of the material performance was conducted by means of laboratory dynamic cone penetrometer (DCP) testing, as well as laboratory California bearing ratio (CBR) testing on a graded crushed G2 Hornfel material. The DCP testing was conducted on both specimen sizes, with one test conducted per small specimen and three tests conducted per large specimen. The CBR testing was only performed on the smaller specimens due to the availability of test equipment. Preliminary testing was performed on the material to classify the material characteristics. A sieve analysis was conducted on the material to understand the material proportions. A moisture-density analysis was performed on the material to establish optimal moisture contents (OMC) corresponding to the optimal density (MDD) achievable in the material. The test specimens were compacted by means of vibratory energy, using the vibratory hammer to ensure an accurate simulation of in-situ conditions. From the DCP testing and determination of the DCP-DN values, it was evident that specimens prepared at lower moisture contents and higher densities yielded lower DN values, higher DCP-CBR values and higher estimated elastic stiffness values. It was subsequently concluded that the G2 material is sensitive to moisture and density changes, and that these factors will significantly impact the material performance. The effects of soil suction were considered, and it was determined that soil suction has a negligible impact on the G2 material as the degree of saturation in the material was well above 20%. The larger specimens, with larger maximum particle size also displayed lower DN values, indicating that these specimens had improved shear strength when compared to the smaller specimens with smaller maximum particle size. This conclusion was based on the improvement in aggregate packing due to the presence of larger aggregates. This was confirmed by considering the dominant aggregate size range (DASR), which improved the understanding of the dominant aggregate skeleton governing the behaviour of the G2 material. It was therefore concluded that larger laboratory specimens, with larger aggregates, yield improved material performance and are more representative of in-situ conditions. The correlation between DCP-CBR and laboratory CBR was investigated from the tests conducted in this study, as well as DCP-CBR correlations from previous studies. It was evident that various DCP-CBR formulations displayed large variance in the results when compared to each other with a standard deviation of 23% for the 100% MDD specimens and 26% for the 98% MDD specimens, suggesting that a lack of confidence is associated with the DCP-CBR formulations. When comparing the DCP-CBR and laboratory CBR test results, poor correlation was observed. It was therefore concluded that DCP-CBR formulations do not correlate to laboratory CBR values with a minimum required level of accuracy. It was also concluded that CBR values are not suited for performance-based material classification due to the variability observed in these values. In summary, it was concluded that the performance of granular material is influenced by to the moisture content and density at which the material is prepared. It was evident that other factors such as grading and maximum particle size also have a significant impact on the performance of the material. From this study, it was evident that certain current laboratory techniques do not accurately simulate in-situ conditions due to the adoption of a scaled-down approach. It was also concluded that the use of CBR is not suited for the accurate classification of UGM due to the large variability in CBR results and poor correlation to in-situ material performance. The DCP test, using the DCP-DN values, is recommended and is an effective performance-based test procedure to establish estimations for the behaviour and strength of unbound granular material. The DCP has proven to be cost-effective and easy to use and may therefore be a valuable tool in pavement engineering, especially in developing countries.
AFRIKAANSE OPSOMMING: Hierdie navorsing studie maak deel uit van die SANRAL Projek 3.5, wat gefokus is op die verbetering van uitkoms-gebasseerde toetsmetodes wat toegepas word in plaveisel ingenieurswese in die Suid-Afrikaanse konteks, spesifiek gemik op ongebonde granulere materiaal. Aangesien die meeste van huidige toetsmetodes in Suid-Afrika gebasseer is op empiriese formulerings, bestaan daar ‘n leemte in die korrelasie van toetsresultate en in-situ materiaal omstandighede en daarom is die voorspelling van die materiaal gedrag dikwels onakkuraat. Die uitkoms-gebasseerde toetsmetodes is gemik op die identifisering van die skuifsterkte van materiaal, sowel as die voorspelling van die gedrag van die material onder die invloed van herhaaldelike verkeerlading. Hierdie studie oorweeg die invloed van voginhoud, digtheid, proefstukgeometrie, maksimum aggregaat grootte en aggregaat verpakking op die gedrag van ongebonde granulere materiaal in Suid-Afrika. Om die invloed van voginhoud en digtheid te ondersoek, was drie verskillende voginhoude ingesluit in die toets prosedure, gekombineerd met twee verskillende kompaksie vlakke. Om die invloed van proefstukgeometrie te bepaal is twee verskillende groottes gebruik in die eksperimentele metodologie. Die invloed van maksimum aggregaat grootte en aggregaat verpaking is ondersoek deur die gebruik van twee materiaal graderings met verskillende maksimum aggregaat groottes bekom deur die material te “scalp” teen ‘n vereiste sif grootte. Die kleiner monster gradering was gemerk as “S20” materiaal. Die groter monster gradering was gemerk as “S28” materiaal. Die ondersoek van die materiaal gedrag en eienskappe was uitgevoer deur middle van laboratorium DKP toetse, sowel as die laboratorium KDV toetse op G2 Hornfel materiaal. Die DKP toetse was uitgevoer op beide monster groottes, met een toets uitgevoer op die kleiner monsters en drie toetse uitgevoer op die groter monsters. Die KDV toetse was alleenlik toegepas op die kleiner monsters as gevolg van die beskikbaarheid van toets toerusting. Voorlopige gedragstoetse was uitgevoer om die materiaal te klassifiseer volgens die materiaaleienskappe. ‘n Sifanalise was uitgevoer op die G2 klip om die material verhoudings van die materiaal te verstaan. ‘n Vog-digtheidsanalise was uitgevoer op die materiaal om vas te stel watter optimum voginhoed (OVI) korreleer met die maksimum digtheid (MDD) bereikbaar in die materiaal. Vir die kleiner S20 monsters, was ‘n maksimum droe digtheid van 2350kg/m3 bevind, by ‘n voginhoud van 5.3%. Vir die groter S28 monsters was die maksimum droe digtheid bevind teen 2350kg/m3 by ‘n voginhoud van 4.9%. Die toets monsters was gekompakteer deur middel van vibrasie energie, met die gebruik van ‘n vibrasie hammer om akkurate simulasie te verseker met in-situ omstandighede. Met behulp van die DKP toetse en die bepaling van die DKP-DN waardes, was dit duidelik dat monsters voorberei teen laer vogvlakke en hoer digtheid, laer DN waardes, hoer DKP-KDV waardes, hoer geskatte styfheid waardes gelewer het. Dit was vervolgens afgelei dat die G2 materiaal se sensitiwiteit vir vog en digtheid verander en dat hierdie faktore ‘n betekenisvolle impak het op die materiaal se gedrag. Die groter monsters, met groter maksimum aggregaat grootte het ook laer DN waardes getoon, aanduidend daarvan dat hierdie monsters beter skuifsterkte het in vergelyking met die kleiner monsters met kleiner aggregaat groottes. Hierdie gevolgtrekking was gebasseer op die verbetering van aggregaat verpakking as gevolg van die teenwoordigheid van groter aggregaat. Die dominante aggregaat grootte (DASR) was oorweeg, wat dit moontlik gemaak het om die dominante aggregaat geraamte wat hoofsaaklik bydrae to die material sterkte, beter te verstaan. Daar was tot die gevolgtrekking gekom dat groter laboratorium monsters, met groter aggregaate verbeterde materiaal gedrag en eienskappe gelewer het en dit was meer verteenwoordigend van in-situ omstandighede. Die korrelasie tussen DKP-KDV en laboratorium KDV was ondersoek deur die toetse uitgevoer in hierdie studie, sowel as DKP-KDV korrelasies vanuit vorige studies. Dit was duidelik dat verskeie DKP-KDV formulerings groot variasies getoon het in die uitslae wanneer dit in vergelyking tot mekaar gestel was met ‘n standaard deviasie vir 23% vir die 100% MDD monsters en 26% vir die 98% MDD monsters, aanduidend van ‘n gebrek aan vertroue, geassosieerd met die DKP-KDV formulerings. Wanneer die DKP-KDV en laboratorium KDV toets resultate vergelyk word, was swak korrelasie waargeneem. Dit was dus aanvaar dat DKP-KDV formulerings nie korreleer met laboratorium KDV waardes met ‘n minimum vereiste vlak van akkuraatheid nie. Dit was ook aanvaar dat KDV waardes nie gepas is vir uitkoms-gebasseerde materiaal klassifikasie is nie as gevolg van die veranderlikheid waargeneem in hierdie waardes. Ter opsomming, was dit bevind dat die gedrag van granulere materiaal onderhewig is aan die voginhoud en digtheid waaronder die materiaal voorberei is. Dit was duidelik dat ander faktore, soos gradering en maksimum aggregaat grootte ‘n betekenisvolle impak het op die gedrag van die materiaal. Vanuit hierdie studie is dit duidelik dat sekere huidige laboratorium tegnieke nie akkuraat in-situ kondisies simuleer nie as gevolg van die aanneming van ‘n afgeskaalde benadering. Dit word ook bevind dat die gebruik van KDV nie geskik is vir die akkurate klassifikasie van ongebonde granulere material is nie as gevolg van die groot veranderlikhede in KDV resultate en swak korrelasie met in-situ materiaal gedrag. Die DKP toets, met die gebruik van DKP-DN waardes word aanbeveel en is ‘n effektiewe uitkomsgebasseerde toetsprosedure om waarnemings van die gedrag en sterkte van ongebonde granulere materiaal te bepaal. Die DKP was bevind om koste-effektief en maklik te wees om te gebruik en mag daarvolgens ‘n waardevolle apparaat in plaveisel Inginieurswese wees, veral in ontwikkelende lande.
Description
Thesis (MEng)--Stellenbosch University, 2023.
Keywords
Granular Material
Citation