Estimating the delayed strains in concrete, namely creep and shrinkage is very important to asses the condition of the structure. Time dependent deformations in concrete, both creep and shrinkage, play a critical role in prestressed concrete structures, such as bridge girders, nuclear containment vessels, etc. These strains result in lossess, through release of prestress, and thereby influence the safety of these structures. Recognizing the role of free and bound moisture movement is the primary ingredient responsible for the development of both creep and shrinkage stains as well as the degradation of concrete under high temperature, the present study has also examined the effects of high temperature on concrete degradation, experimentally and also analytically in the same modelling framework. Fire in concretes deteriorates mechanical properties of the material and lead to col-lapse under loads. Two types of spalling occur in concrete when exposed to high temperature, namely explosive and thermal spalling. Explosive spalling occurs once the hydrostatic stress (developed due to pore pressure) exceeds the tensile strength of the concrete. Where as thermal spalling of concrete happens due to degradation of material properties (elastic modulus, compressive and tensile strength) when exposed to high temperature due to decomposition of chemical bonds that release the bound water. The present study comprises of an experimental and analytical program to assess the levels of creep and shrinkage in different concrete under various loads and environmental conditions. Deformations due to high temperature in di erent concretes forms another component of the present study. Total six concrete mixes has been studied to investigate and asses the extent of creep and shrinkage taking place in the concretes under different environmental conditions, load level and age at loading. In total six mixes, three that are self compacted concrete mixes (35MPa, 55MPa and SCC70MPa), a high volume y ash concrete mix ( 45 MPa) and two normal concrete mixes (35 MPa and 45 MPa) have been considered in this study. To study the high temperature effects, the same mixes considered in the creep and shrinkage study and in addition a heavy density concrete mix (25 MPa) is used. A normal concrete having a 28 day uniaxial compressive strength of 45 MPa after proper curing, referred to as M45 concrete, was one of the six mixes. Likewise a heavy density concrete designated as H25, corresponding to a 28 day uniaxial compressive strength of 25 MPa was another mix that was studied and was made using iron ore aggregate and iron ore sand. A concrete having high volume y ash replacing cement designated as F45 offered a 28 day strength of 45MPa. Three self-compacting concretes with uniaxial compressive strengths of 35, 55 and 70 MPa were designated as SCC35 SCC55 and SCC70, respectively is studied for creep, shrinkage and high temperature effects. F45 concrete shows lower creep strain when compared to normal M45 concrete, under similar casting, curing and testing condtions. This is due to increase in stiffness of y ash based concretes with time. Where as in shrinkage it is observed that a little higher strain takes place in F45 at initial ages than in M45 concrete mix for the same conditions. But in the later age, F45 concrete shows a decreasing rate of shrinkage strain. This is because, water to cement ratio of y ash concrete is higher than the M45 concrete. The SCC35 concrete shows higher creep and shrinkage than M35 concrete even though both the concretes have the same water cement ratio. This difference comes from the aggregate cement ratio (a/c). The lower the aggregate cement ratio, the higher the creep and shrinkage. M35 concrete has a higher aggregate cement ratio than the SCC35. Concretes exposed to higher temperature and lower humidity shows higher creep and shrinkage due to its higher rate of drying. An analytical model has been developed to simulate the drying phenomena in concrete based on poromechanics. The hydration effects of blended cements is considered while developing the model. This models prediction of degree of hydration, temperature and relative humidity is used to model creep and shrinkage in concrete. To model creep and shrinkage, micro prestress solidi cation theory is implemented and validated with the present experimental results. The model is able to predict the drying phenomena of concrete realistically. Further, a benchmark problem reported in the literature is solved through this model and validated through a comparison with the experimental results (beam detection due to creep and shrinkage). Under high temperature tests, H25 concrete shows better resistance for all the ranges of temperatures. This may be because of the hematite aggregate having a high melting point and strong interfacial transition zone (ITZ) properties between aggregate and cement mortar. The SCC70 shows poor performance against explosive spalling at both the ages (28 and 365 days) due to its lower permeability when exposed to high temperature. The intensity of explosive spalling is higher in SCC70 concrete tested at 28 days than at 365 days of age. This is because of variation in moisture content. SCC70 concrete failed due to explosive spalling at temperature of 398oC when tested at 28 days and failed at 575oC when tested at 365 days. This indicates the amount of moisture content in the concrete plays an important role while causing explosive spalling. F45 concrete shows a poor resistance against temperature beyond 500oC in its residual properties. SCC55 contains cement and y ash and shows higher residual properties when compared to normal vibrated M45 mix under similar high temperature conditions. Two geopolymers pastes prepared with y ash and metakaolin as a complete cement replacement were studied for passive re protection capability. The study shows MF70 mix (containing 70% y ash and 30% metakaolin) gives better resistance against heating than MF50 mix (50% each of metakaolin and y ash). Hence y ash geopolmer is a choice of material for passive re protection. An analytical model has been developed based on poromechanics to simulate high temperature e ects in concrete. Two type of spalling is considered while modelling the high temperature e ects of concrete, namely explosive and thermal spalling. Explosive spalling is estimated based on the hydro static stress (Biotech efficient times the pore pressure). If the hydrostatic stress increases beyond the tensile strength of concrete then explosive spalling occurs. Where as the thermal spalling is estimated based on the stresses developed due to applied mechanical and thermal loading. To validate this model, two benchmark problems from the literature have been solved and validated with the reported results. This model is able to predict pore pressure and temperatures gradients accurately, and this in turn helps to predict explosive and thermal spalling realistically in concrete under elevated temperature conditions.

In respect of autogenous and drying shrinkage and the effects of relaxation creep during the hydration cycle, roller compacted concrete in dams has to date been universally assumed to behave in the same manner as conventional mass concrete, despite notional evidence to the contrary on prototype dam structures, particularly in respect of high-paste RCC. While the results of laboratory materials testing and associated early behaviour analyses for RCC have been published, no conclusive example exists in the public domain whereby predicted behaviour is confirmed through measured behaviour on a comprehensively-instrumented prototype dam structure. In his PhD thesis, Quentin Shaw presents evidence to indicate that the early behaviour of RCC, and particularly high quality, high-paste RCC in dams, is quite different to that of CVC. Referring to instrumentation records from Wolwedans and Knellpoort dams in South Africa, Çine Dam in Turkey, Wadi Dayqah Dam in Oman and Changuinola 1 Dam in Panama, indications of less than expected shrinkage and stress relaxation creep during the hydration cycle in the constituent RCC are documented. Taking the comprehensively-instrumented and monitored Wolwedans Dam, the actual materials behaviour of the constituent RCC is evaluated through the replication of the prototype behaviour on a finite element model. Through this analysis, it is clearly demonstrated that the level of shrinkage and stress relaxation creep that would be traditionally assumed in RCC simply did not occur. In fact, the analyses suggested that no shrinkage, or creep was apparent. The reasons for the different behaviour of high-paste RCC compared to CVC are subsequently explored. With Wadi Dayqah Dam as the only example evaluated where some drying shrinkage and/or stress relaxation creep was obviously apparent, the evident susceptibility of this lean RCC mix, with a high w/c ratio, a high content of non-cementitious fines, natural gravel aggregates, a high aggregate water absorption and placement in a very dry environment, is noted. However, it is considered to be the combination of a strong aggregate skeletal structure developed through roller compaction and a low w/c ratio that results in a particularly resilience in high-paste RCC to early shrinkage and creep. It is also recognised that temperature and gravity effects in an arch dam structure will tend to limit, or even eliminate containment stresses in the critical load-carrying upper section and that this will reduce the risk and impact of stress relaxation creep. Consequently, a new understanding of the early behaviour of RCC in large dams is presented, suggesting that a high quality RCC mix in an arch dam can be designed for a cumulative shrinkage and stress relaxation creep under the hydration cycle of approximately 20 microstrain, compared with a more traditionally accepted value of between 125 and 200 microstrain. The implications of these findings on the design of large RCC dams are demonstrated to be significant, particularly in respect of RCC arch dams. In addition, suggestions are made for the requirements in respect of RCC mix design for negligible shrinkage and creep, while an approach to combine the use of field measurement with structural modelling to predict and demonstrate actual RCC behaviour is briefly discussed. / Thesis (PhD)--University of Pretoria, 2011. / Civil Engineering / unrestricted

Alkali activated materials, especially when activated with waterglass, are subjected to substantial drying shrinkage that hinders their broader industrial application. The effect of shrinkage reducing admixtures (SRA), based on poly(propylene glycol), on drying shrinkage of alkali activated blast furnace slag (BFS) mortars was studied. The aim of this thesis is to determine the efficiency of SRA as well as the influence of different type of alkali activator with varying silicate modulus on drying shrinkage characteristics. It was observed that the high amount of alkalis positively influenced the effect of SRA. The higher the amount of alkalis, the lower the drying shrinkage was. It also caused more water to be incorporated in the alkali activated structure and prevent it from evaporation. Thus, the weight change of such mortars was the lowest. Reduced effect of SRA on mechanical strength properties was observed in mortars with low amount of alkalis which was connected with delayed hydration resulting in insufficient amount of created C-S-H gel. The presented thesis further discussed and underlines the role of different kinds of alkali ions as well as their amount on the properties of alkali activated BFS systems.

While many major cities in the United States are once again gaining population, growing their economies, and attracting talent, many small and mid-sized cities are in decline. The reasons for this growing disparity are multi-faceted. A growing body of research has been exploring planning challenges in declining cities and towns. This body of research—often called “shrinking cities” and “urban shrinkage” research—is premised on the belief that many declining places will continue to shed population, jobs, and industries, and planning smartly for this decline is the only sensible path forward. So far, research in the U.S. has focused primarily on Northeast and Midwest cities where population and industrial decline has been the most severe. Less scholarship has studied places that have declined more slowly and more recently. This thesis examines the current trends impacting the decline of mid-sized cities in the Midwestern United States, focusing on four cities in the State of Illinois. It also explores whether these cities are ready to consider the possibility that population decline is not temporary and change their planning strategies accordingly. Finally, this thesis will introduce an emerging paradigm in contemporary urban planning practice that fuses growth and decline strategies, to prepare mid-sized cities for an uncertain demographic and economic future.

In this paper, a rheumatoid arthritis (RA) medicine clinical dataset with an ordinal response is selected to study this new medicine. In the dataset, there are four features, sex, age,treatment, and preliminary. Sex is a binary categorical variable with 1 indicates male, and 0 indicates female. Age is the numerical age of the patients. And treatment is a binary categorical variable with 1 indicates has RA, and 0 indicates does not have RA. And preliminary is a five class categorical variable indicates the patient’s RA severity status before taking the medication. The response Y is 5 class ordinal variable shows the severity of patient’s RA severity after taking the medication. The primary aim of this study is to determine what factors play a significant role in determine the response after taking the medicine. First, cumulative logistic regression is applied to the dataset to examine the effect of various factors on ordinal response. Secondly, the ordinal response is categorized into two classes. Then logistic regression is conducted to the RA dataset to see if the variable selection would be different. Moreover, the shrinkage methods, elastic net and lasso are used to make a variable selection on the RA dataset of two-class response for the purpose of adding penalization to increase the model’s robustness.The four model results were compared at the end of the paper. From the comparison result, logistic regression has a better performance on variable selection than the other three approaches based on P-value.

Worldwide diarrheal disease is a leading cause of morbidity and mortality in children less than five years of age. Incidence and disease severity remain the highest in sub-Saharan Africa. Kenya has an estimated 400,000 severe diarrhea episodes and 9,500 diarrhea-related deaths per year in children. Current statistical methods for estimating etiological and exposure risk factors for moderate-to-severe diarrhea (MSD) in children are constrained by the inability to assess a large number of parameters due to limitations of sample size, complex relationships, correlated predictors, and model assumptions of linearity. This dissertation examines machine learning statistical methods to address weaknesses associated with using traditional logistic regression models. The studies presented here investigate data from a 4-year, prospective, matched case-control study of MSD among children less than five years of age in rural Kenya from the Global Enteric Multicenter Study. The three machine learning approaches were used to examine associations with MSD and include: least absolute shrinkage and selection operator, classification trees, and random forest. A principal finding in all three studies was that machine learning methodological approaches are useful and feasible to implement in epidemiological studies. All provided additional information and understanding of the data beyond using only logistic regression models. The results from all three machine learning approaches were supported by comparable logistic regression results indicating their usefulness as epidemiological tools. This dissertation offers an exploration of methodological alternatives that should be considered more frequently in diarrheal disease epidemiology, and in public health in general.

Thesis (MScEng (Civil Engineering))--University of Stellenbosch, 2011. / ENGLISH ABSTRACT: Plastic shrinkage cracking (PSC) is the cracking caused by the early age shrinkage of concrete within the first few hours after the concrete has been cast. It results in unsightly surface cracks that serve as pathways whereby corroding agents can penetrate the concrete which shortens the expected service life of a structure. PSC is primarily a problem at large exposed concrete surfaces for example bridge decks and slabs placed in environmental conditions with high evaporation rates. Most precautionary measures for PSC are externally applied and aimed to reduce the water loss through evaporation. The addition of a low dosage of polymeric fibres to conventional concrete is an internal preventative measure which has been shown to reduce PSC. The mechanisms involved with PSC in conventional and low volume fibre reinforced concrete (LV-FRC) are however not clearly understood. This lack of knowledge and guidance leads to neglect and ineffective use of preventative measures. The objective of this study is to provide the fundamental understanding of the phenomena of PSC. To achieve the objective, an in depth background study and experiments were conducted on fresh conventional concrete and LV-FRC. The three essential mechanisms required for PSC are: 1→ Capillary pressure build-up between the particles of the concrete is the source of shrinkage. 2→ Air entry into a concrete initiates cracking. 3→ Restraint of the concrete is required for crack forming. The experiments showed the following significant findings for conventional and LV-FRC: PSC is only possible once all the bleeding water at the surface has evaporated and once air entry has occurred. The critical period where the majority of the PSC occurs is between the initial and final set of concrete. Any preventative measure for PSC is most effective during this period. The bleeding characteristics of a mix have a significant influence on PSC. Adding a low volume of polymeric fibres to concrete reduces PSC due to the added resistance that fibres give to crack widening, which increases significantly from the start of the critical period. The fundamental knowledge gained from this study can be utilized to develop a practical model for the design and prevention of PSC in conventional concrete and LV-FRC. / AFRIKAANSE OPSOMMING: Plastiese krimp krake (PSK) is die krake wat gevorm word a.g.v. die vroeë krimping van beton binne die eerste paar ure nadat die beton gegiet is. Dit veroorsaak onooglike oppervlak krake wat dien as kanale waardeur korrosie agente die beton kan binnedring om so die dienstydperk van die struktuur te verkort. Dit is hoofsaaklik ʼn probleem by groot blootgestelde beton oppervlaktes soos brug dekke en blaaie wat gegiet is in klimaat kondisies met hoë verdamping tempo’s. Meeste voorsorgmaatreëls vir PSK word ekstern aangewend en beperk die water verlies as gevolg van verdamping. Die byvoeging van ʼn lae volume polimeriese vesels is ʼn interne voorsorgmaatreël wat bekend is om PSK te verminder. Die meganismes betrokke ten opsigte van PSK in gewone beton en lae volume vesel versterkte beton (LV-VVB) is vaag. Die vaagheid en tekort aan riglyne lei tot nalatigheid en oneffektiewe aanwending van voorsorgmaatreëls. Die doel van die studie is om die fundamentele kennis oor die fenomeen van PSK te gee. Om die doel te bereik is ʼn indiepte agtergrond studie en eksperimente uitgevoer op gewone beton en LV-VVB. Die drie meganismes benodig vir PSK is: 1→ Kapillêre druk tussen die deeltjies van die beton is die hoof bron van krimping. 2→ Lugindringing in die beton wat krake inisieer. 3→ Inklemming van die beton is noodsaaklik vir kraakvorming. Die eksperimente het die volgende noemenswaardige bevindinge opgelewer: PSK is slegs moontlik indien al die bloeiwater van die beton oppervlakte verdamp het en indien lug die beton ingedring het. Die kritiese periode waar die meerderheid van die PSK plaasvind is tussen die aanvanklike en finale set van die beton. Enige voorsorgmaatreël vir PSK is mees effektief gedurende die periode. Die bloei eienskappe van ʼn meng het ʼn noemenswaardige effek op die PSK. Die byvoeging van ʼn lae volume polimeriese vesels tot beton verminder die PSK deur die addisionele weerstand wat die vesels bied teen die toename in kraakwydte. Die weerstand vergroot noemenswaardig vanaf die begin van die kritiese periode. Die fundamentele kennis wat in die studie opgedoen is, kan gebruik word vir die ontwikkeling van ʼn praktiese model vir die ontwerp en verhoed van PSK in gewone beton en LV-VVB.

Thesis (MSc)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: The increasing awareness of climate change causes a growing interest in pavement rehabilitation. Pavement rehabilitation by in-situ stabilisation with bitumen reduces the extraction of natural aggregate resources while enhancing flexibility and durability properties, which lowers maintenance costs over the design-life of the pavement structure. Incorporating Bitumen Stabilised Materials (BSMs) into a pavement structure can therefore have economic and environmental benefits, but more research is needed to fully understand the behaviour and potential of these materials. Stabilising materials with bitumen provides useful properties to pavement layers. The “TG2 2nd Edition, Bitumen Stabilised Materials” was published by the Academy of South Africa in May 2009, which provides a good understanding of the usage of Bitumen Stabilised Materials (BSMs). However, the shrinkage and flexible behaviour of these materials are still not fully understood and therefore more research on these materials is needed. The aim of this project is thus to determine the shrinkage and flexible behaviour of BSMs to incorporate these behavioural types in the revised design method for flexible pavements in the SAPDM. In addition, the influence of several additives on the shrinkage and flexible behaviour of BSMs have been evaluated to get an improved understanding of these properties. The additives included cement (1% and 2% content) and both bitumen emulsion (0.9% and 2.4% content) and foamed bitumen (only 2.4% content). Two linear shrinkage testing methods have been designed to test the shrinkage potential of BSMs, including a beam testing method and a cylindrical testing method. Based on the usage of the shrinkage measurements the applicable method can be used to determine the shrinkage potential of a BSM. The flexibility is a more complex property and was tested using a simple monotonic beam test. The strain-at-break parameter obtained from this test provided an indication of the material flexibility. Slight differences in the trends were observed between beam and cylindrical shrinkage due to specimen geometry, exposed surface area and shrinkage testing direction. Beam specimens initially show swelling when bitumen is added to the mixture and shrinks thereafter. Cylindrical specimens on the other hand show initial shrinkage followed by a slight length increase, where after shrinkage continues. The additives had the same influence on the shrinkage for both the beam and cylindrical specimens. Although all shrinkage measurements were small, an increase in bitumen reduced shrinkage and an increase in cement increased shrinkage. Stabilisation with foamed bitumen rather than bitumen emulsion proved to show less shrinkage, but only in combination with 2% cement. The strain-at-break, dissipated energy and material stiffness calculated from the monotonic beam tests provided a good indication of the flexibility behaviour of BSMs. Higher bitumen content increased the flexibility potential and an increase in cement decreased the flexibility potential of BSMs. This project has provided good insight on both the shrinkage and flexibility behaviour of BSMs, which can be used in the revised copy of the SAPDM. Increased bitumen contents decreases the shrinkage potential and increases the flexibility of a BSM. Increased cement contents on the other hand, increases shrinkage and decreases flexibility of BSMs. The correct combination of cement and bitumen in a BSM can thus provide a material with the wanted flexibility while keeping the shrinkage to a minimum. / AFRIKAANSE OPSOMMING: Die toenemende bewustheid van klimaatsverandering veroorsaak toenemende belangstelling in die rehabilitasie van plaveisels. Plaveisel rehabilitasie deur in-veld stabilisasie met bitumen verminder die ontginning van natuurlike hulpbronne, terwyl die verbetering van buigsaamheid en duursaamheid eienskappe die onderhoudskoste verlaag oor die ontwerp-lewe van die plaveiselstruktuur. Die inkorporasie van Bitumen Gestabiliseerde Materiale (BGM) in 'n plaveisel struktuur kan dus omgewings en ekonomiese voordele inhou. Meer navorsing word wel benodig om die gedrag van hierdie materiale beter te verstaan. Die stabilisering van materiale met bitumen verskaf nuttige eienskappe aan 'n plaveisellaag. Die "TG2 2de Uitgawe, Bitumen Gestabiliseerde Materiale" is gepubliseer deur die Akademie van Suid-Afrika in Mei 2009 en verskaf 'n goeie begrip van Bitumen Gestabiliseerde Materiale (BGM). Die krimpings en buigsaamheid gedrag van die materiaal word wel nog nie ten volle verstaan nie en daarom word meer navorsing oor hierdie materiaal benodig. Die doel van hierdie projek is dus om die krimpings gedrag sowel as die buigsaamheid gedrag van 'n BGM te bepaal en sodoende die kennis te gebruik in die hersiende ontwerp metode vir buigsame plaveisels in die SAPDM. Die invloed van verskeie bymiddels op die krimpings en buigsaamheid gedrag van 'n BGM is ook geëvalueer om 'n beter begrip van hierdie eienskappe te verkry. Die bymiddels sluit sement in (1% en 2% inhoud) asook beide emulsie bitumen (0,9% en 2,4% inhoud) en skuim bitumen (slegs 2.4% inhoud). Twee lineêre krimpings toets metodes was ontwerp om die krimping potensiaal van BGM's te bepaal, wat 'n balk toets metode en 'n silindriese toets metode insluit. Die metode wat gebruik sal word om die krimping van 'n BGM te bepaal moet baseer word op die toepassing waarvoor die krimpings resultate gebruik gaan word. Die buigsaamheid is 'n meer komplekse eienskap en was getoets met behulp van 'n eenvoudige monotoniese balk toets. Die spanning-by-breekpunt waardes wat verkry was vanuit die balktoetse, het 'n goeie aanduiding van die buigsaamheid van die materiaal verskaf. Klein verskille in krimpingstendense tussen balk en silindriese proefstukke is opgemerk tydens die projek en is veroorsaak deur die geometrie van die proefstuk, die blotgestelde oppervlakte asook die rigting van kimp toetsing. Balk proefstukke toon aanvanklike swelling wanneer bitumen bygevoeg is, gevolg deur krimping. Silindriese proefstukke aan die ander kant toon aanvanklike krimping gevolg deur 'n effense toename in lengte, waarna krimping weer plaasvind. Die bymiddles het dieselfde invloed op die krimping van beide die balk en silindriese proefstukke. Alhoewel al die krimpingswaardes baie klein was, het 'n toename in bitumen 'n vermindering in krimping voortgebring en 'n toename in sement het 'n toename in krimping voortgebring. Stabilisasie met skuim bitumen in plaas van emulsie bitumen toon verlaagde krimping, maar slegs in kombinasie met 2% sement. Die spanning-by-breekpunt, verkose energie en materiaal styfheid wat bereken is vanaf die monotoniese balk toets resultate, het 'n goeie aanduiding van die buigsaamheid gedrag van BGM's verskaf. 'n Hoër bitumen inhoud verhoog die buigsaamheid potensiaal van BGM‟s terwyl 'n toename in sement die buigsaamheid potensiaal van BGM's verlaag. Hierdie projek bied goeie insigte vir beide die krimpings en buigsaamheid gedrag van BGM's, wat in die hersiende ontwerp metode van die SAPDM gebruik kan word. Verhoogde bitumen inhoud verminder die krimping potensiaal en verhoog die buigsaamheid van 'n BGM. Verhoogde sement inhoud aan die ander kant, verhoog krimping en verminder buigsaamheid van BGM's. Die korrekte kombinasie van sement en bitumen in 'n BGM kan dus 'n materiaal produseer met die gewenste buigsaamheidseienskappe en terselfde tyd die krimping tot 'n minimum beperk.

This paper considers Bayesian estimation of the threshold vector error correction (TVECM) model in moderate to large dimensions. Using the lagged cointegrating error as a threshold variable gives rise to additional difficulties that are typically solved by relying on large sample approximations. Relying on Markov chain Monte Carlo methods we circumvent these issues by avoiding computationally prohibitive estimation strategies like the grid search. Due to the proliferation of parameters we use novel global-local shrinkage priors in the spirit of Griffin and Brown (2010). We illustrate the merits of our approach in an application to five exchange rates vis-á-vis the US dollar and assess whether a given currency is over or undervalued. Moreover, we perform a forecasting comparison to investigate whether it pays off to adopt a non-linear modeling approach relative to a set of simpler benchmark models. / Series: Department of Economics Working Paper Series

Le processus de retrait-gonflement des sols argileux est à l'origine des dommages au bâti durant les périodes de sécheresse. Dans cette thèse, les processus de retrait-gonflement de deux formations de sol argileux en France qui sont les Argiles Plastiques du Sparnacien et l'argile d'Héricourt du Lias sont étudiés à différentes échelles et différents états (intact, remanié, compacté et traité à la chaux) en considérant spécialement le rôle de la microstructure. Pour ce faire, une partie de la recherche est réalisée sur la caractérisation des sols étudiés pour analyser leurs propriétés minéralogiques, microstructurales, géotechniques et hydriques par rapport à leur propriété de retrait-gonflement. La deuxième partie de la recherche est consacrée à l'analyse des processus de gonflement, de retrait, de fissuration des sols au cours d'un cycle simple ou de cycles répétés d'humidification-séchage à l'échelle de l'éprouvette. Les processus de retrait-gonflement de sol à une échelle plus grande sont étudiés dans la troisième partie en réalisant les essais d'infiltration et d'évaporation dans une chambre environnementale. Les résultats obtenus mettent en évidence les différences de comportement au retrait-gonflement pour différents états des sols étudiés. Les minéraux argileux ainsi que le quartz, les carbonates et des hydrates, mais aussi la microstructure des sols contrôlent ces phénomènes de retrait-gonflement lors des échanges hydriques. Tous les changements de volume de sol argileux au cours de gonflement-retrait sont gouvernés principalement par la famille de pores inter-agrégats naturels. Ces résultats sont apportés par les analyses microstructurales porosimétriques et au MEB. / The shrink-swell process of clayey soils is causing damage to the structures during periods of drought. In this thesis, the shrink-swell process of two clayey soils in France that are the Plastic Clays of Sparnacian age and Héricourt clay of Lias is studied at different scales and different conditions (intact, remoulded, compacted and lime treated) by insisting on the role of microstructure. To do this, a part of the study is carried out on the characterisation of soils studied to analyse their mineralogical, microstructural, geotechnical and hydraulic properties in relation with their shrink-swell property. The second part of the study focuses on the analysis of swelling, shrinkage, cracking processes of soil during a single cycle or repeated wetting-drying cycles at specimen scale. The shrink-swell process of soil at a larger scale is studied in the third part by performing infiltration and evaporation tests in the environmental chamber. The results highlight the difference in the shrink-swell behavior at different conditions of soils studied and emphasize the importance of the presence and nature of clay minerals also quartz, hydrates and microstructure of soil to the shrinkage and swelling during water exchange. All changes in the volume of clayey soil during shrinkage and swelling are mainly governed by natural inter-aggregate pores, which are showed by microstructural analyses of MIP and SEM tests.