The autogenous shrinkage due to self-desiccation of high- and ultra-high performance concretes with very low water-cement ratio in case of restraint leads to considerable stresses starting from very early age. The resultant risk of cracking presently cannot be adequately investigated. Parameters that are particularly difficult to capture experimentally are the concrete temperature and the viscoelasticity. The primary objective of this work was to assess as precise as possible the autogenous shrinkage cracking propensity of representative concretes at strong restraint and constant room temperature. Test methods needed to be chosen and enhanced in a way that preferably allowed for the efficient and precise investigation of all relevant factors in the future. Ideally, a method suitable for a complete empirical modeling was provided. First the methodological requirements and the advantages and disadvantages of existing test methods were discussed. Based on this, optimized test methods were proposed. Their suitability was verified using the example of ultra-high strength concrete. The choice of concrete compositions considered the essential measures for reducing shrinkage (internal curing, shrinkage-reducing admixtures, reduction of the fraction of Portland cement in the binder). The autogenous shrinkage was measured with the shrinkage cone method. This new test method was validated by investigations of the repeatability and reproducibility and proved efficient and precise. It allows for measurements under non-isothermal conditions; no established test method exists for that purpose to date. The autogenous shrinkage of the ultra-high strength concretes at the age of 24 h, investigated under quasi-isothermal conditions (20 °C), was between 0,25 mm/m and 0,70 mm/m. It was particularly low when a shrinkage-reducing admixture was added and when superabsorbent polymers were used. The stresses due to restraint were determined with the restrained ring test. A large part of the stresses to be expected according to Hooke’s Law were eliminated by creep and relaxation. The relaxation capacity being very pronounced at very early age was the main reason that no visible cracking occurred, not even with the concretes with high autogenous shrinkage. The development of the autogenous shrinkage cracking propensity was described as ratio of restraint stress and splitting tensile strength. By means of modified ring tests, used to determine the maximum tensile stress, it could be shown that the ratio of stress to strength is an appropriate failure criterion. However, the cracking propensity can be calculated correctly only if the strongly age-dependent ratio of uniaxial to splitting tensile strength is accounted for. Besides, it needs to be considered that at very early age a plastic stress redistribution may occur in restrained ring tests. The reference concrete showed a high cracking propensity of up to 0.68. The fact that shrinkage-reducing measures led to significantly lower values reveals their relevance for the safe application of ultra-high strength concrete. However, the investigations carried out here at 20 °C do not allow for a final assessment of the cracking propensity under typical on-site conditions. To empirically model the autogenous shrinkage cracking propensity as a function of temperature and stress level in the future, an analytical stress solution for non-isothermal restrained ring tests and a new approach for investigating the residual stress and relaxation capacity by means of non-passive restrained ring tests was suggested. / Das durch Selbstaustrocknung verursachte autogene Schwinden von besonders leistungsfähigen Betonen mit sehr niedrigem Wasserzementwert führt bei Dehnungsbehinderung bereits in sehr frühem Alter zu erheblichen Zwangsspannungen. Die Gefahr der Rissbildung, die sich daraus ergibt, lässt sich bislang nur unzureichend untersuchen. Experimentell besonders schwer zu erfassende Faktoren sind die Betontemperatur und die Viskoelastizität. Das vorrangige Ziel der Arbeit war die möglichst genaue Ermittlung der autogenen Schwindrissneigung repräsentativer Betone bei starker Dehnungsbehinderung und konstanter Raumtemperatur. Dabei waren die Prüfverfahren möglichst so zu wählen und weiterzuentwickeln, dass sich zukünftig alle relevanten Faktoren effizient und genau untersuchen lassen. Im Idealfall sollte eine Methode entstehen, die eine vollständige empirische Modellierung erlaubt. Zunächst wurden die methodischen Anforderungen und die Vor- und Nachteile existierender Prüfverfahren diskutiert. Darauf aufbauend wurden optimierte Verfahren vorgeschlagen. Ihre Eignung wurde an ultrahochfestem Beton überprüft. Bei der Auswahl der Betone wurden die wesentlichen Maßnahmen zur Schwindreduzierung berücksichtigt (innere Nachbehandlung, schwindreduzierende Zusatzmittel, Verringerung des Portlandzementanteils am Bindemittel). Das autogene Schwinden wurde mit dem Schwindkegelverfahren gemessen. Das neue Verfahren wurde durch Untersuchungen zur Wiederhol- und Vergleichsgenauigkeit validiert und erwies sich als effizient und genau. Es ermöglicht Messungen unter nicht-isothermen Bedingungen; hierfür existiert bisher kein etabliertes Verfahren. Das autogene Schwinden der untersuchten ultrahochfesten Betone unter quasi-isothermen Bedingungen (20 °C) betrug im Alter von 24 h zwischen 0,25 mm/m und 0,70 mm/m. Besonders gering war es bei Zugabe eines schwindreduzierenden Zusatzmittels bzw. Verwendung superabsorbierender Polymere. Mit dem Ring-Test wurden die bei Dehnungsbehinderung entstehenden Spannungen ermittelt. Ein großer Teil der gemäß Hooke’schem Gesetz zu erwartenden Spannungen wurde durch Kriechen und Relaxation abgebaut. Die im sehr frühen Alter stark ausgeprägte Relaxationsfähigkeit war der wesentliche Grund dafür, dass es selbst bei Betonen mit hohem autogenen Schwinden zu keiner erkennbaren Rissbildung kam. Die Entwicklung der autogenen Schwindrissneigung wurde als Verhältnis von Zwangsspannung und Spaltzugfestigkeit beschrieben. Durch modifizierte Ring-Tests, mit deren Hilfe die maximale Zugspannung ermittelt wurde, konnte gezeigt werden, dass das Verhältnis von Spannung und Festigkeit als Versagenskriterium geeignet ist. Die Rissneigung lässt sich aber nur dann korrekt berechnen, wenn das stark altersabhängige Verhältnis von einaxialer Zugfestigkeit und Spaltzugfestigkeit berücksichtigt wird. Außerdem ist zu beachten, dass es im sehr frühen Alter zu einer plastischen Spannungsumlagerung in Ring-Tests kommen kann. Der Referenzbeton wies eine hohe Rissneigung von bis zu 0,68 auf. Dass die schwindreduzierenden Maßnahmen zu deutlich geringeren Werten führten, zeigt deren Bedeutung für den sicheren Einsatz von ultrahochfestem Beton. Die hier bei 20 °C durchgeführten Untersuchungen erlauben allerdings keine abschließende Bewertung der Rissneigung unter baustellentypischen Bedingungen. Um die autogene Schwindrissneigung zukünftig als Funktion der Temperatur und des Lastniveaus empirisch modellieren zu können, wurden eine analytische Spannungslösung für nicht-isotherme Ring-Tests und ein neuer Ansatz zur Untersuchung der Resttrag- und Relaxationsfähigkeit mit Hilfe nicht-passiver Ring-Tests vorgeschlagen.

The autogenous shrinkage due to self-desiccation of high- and ultra-high performance concretes with very low water-cement ratio in case of restraint leads to considerable stresses starting from very early age. The resultant risk of cracking presently cannot be adequately investigated. Parameters that are particularly difficult to capture experimentally are the concrete temperature and the viscoelasticity. The primary objective of this work was to assess as precise as possible the autogenous shrinkage cracking propensity of representative concretes at strong restraint and constant room temperature. Test methods needed to be chosen and enhanced in a way that preferably allowed for the efficient and precise investigation of all relevant factors in the future. Ideally, a method suitable for a complete empirical modeling was provided. First the methodological requirements and the advantages and disadvantages of existing test methods were discussed. Based on this, optimized test methods were proposed. Their suitability was verified using the example of ultra-high strength concrete. The choice of concrete compositions considered the essential measures for reducing shrinkage (internal curing, shrinkage-reducing admixtures, reduction of the fraction of Portland cement in the binder). The autogenous shrinkage was measured with the shrinkage cone method. This new test method was validated by investigations of the repeatability and reproducibility and proved efficient and precise. It allows for measurements under non-isothermal conditions; no established test method exists for that purpose to date. The autogenous shrinkage of the ultra-high strength concretes at the age of 24 h, investigated under quasi-isothermal conditions (20 °C), was between 0,25 mm/m and 0,70 mm/m. It was particularly low when a shrinkage-reducing admixture was added and when superabsorbent polymers were used. The stresses due to restraint were determined with the restrained ring test. A large part of the stresses to be expected according to Hooke’s Law were eliminated by creep and relaxation. The relaxation capacity being very pronounced at very early age was the main reason that no visible cracking occurred, not even with the concretes with high autogenous shrinkage. The development of the autogenous shrinkage cracking propensity was described as ratio of restraint stress and splitting tensile strength. By means of modified ring tests, used to determine the maximum tensile stress, it could be shown that the ratio of stress to strength is an appropriate failure criterion. However, the cracking propensity can be calculated correctly only if the strongly age-dependent ratio of uniaxial to splitting tensile strength is accounted for. Besides, it needs to be considered that at very early age a plastic stress redistribution may occur in restrained ring tests. The reference concrete showed a high cracking propensity of up to 0.68. The fact that shrinkage-reducing measures led to significantly lower values reveals their relevance for the safe application of ultra-high strength concrete. However, the investigations carried out here at 20 °C do not allow for a final assessment of the cracking propensity under typical on-site conditions. To empirically model the autogenous shrinkage cracking propensity as a function of temperature and stress level in the future, an analytical stress solution for non-isothermal restrained ring tests and a new approach for investigating the residual stress and relaxation capacity by means of non-passive restrained ring tests was suggested.:1 Introduction 2 Autogenous shrinkage 5 2.1 Shrinkage and hydration 5 2.2 Definitions and research approaches 10 2.3 Metrological issues 14 2.3.1 Multitude of test methods 14 2.3.2 Time-zero 16 2.3.3 Other metrological issues 18 2.4 Corrugated tube method 19 2.5 Influencing parameters 21 2.5.1 Concrete composition 21 2.5.2 Temperature 23 2.5.3 Specific countermeasures 25 2.6 Summary and conclusions with respect to the own work 25 3 Concretes used in the own investigations 27 3.1 Preliminary remarks 27 3.2 Concrete compositions 27 3.3 Constituents 28 3.3.1 Cement 28 3.3.2 Ground-granulated blast furnace slag 28 3.3.3 Silica fume 28 3.3.4 Admixtures 29 3.3.5 Aggregates 29 3.4 Mixing 29 3.5 Basic properties 30 3.5.1 Compressive strength 30 3.5.2 Splitting tensile strength 31 3.5.3 Modulus of elasticity 33 3.5.4 Analysis of mechanical properties 35 3.5.5 Coefficient of thermal expansion 38 3.5.6 Isothermal calorimetry 39 3.6 Summary 39 4 Shrinkage cone method for measuring autogenous shrinkage 41 4.1 Introduction 41 4.2 Setup and measurement procedure 41 4.3 Temperature control 44 4.4 Precision under quasi-isothermal conditions 47 4.4.1 Repeatability 47 4.4.2 Reproducibility 49 4.4.3 Shrinkage cone method vs. corrugated tube method 49 4.5 Autogenous shrinkage of the investigated concretes at 20 °C 54 4.6 Tests under non-isothermal conditions 55 4.7 Summary 56 5 Stress and cracks due to restrained autogenous shrinkage 58 5.1 Introduction 58 5.2 Degree of restraint 58 5.3 Formation of cracks 60 5.4 Very early age and importance of stress relaxation 63 5.5 Creep and cracking - further methodological aspects 65 5.6 Autogenous shrinkage cracking propensity 69 5.7 Role of temperature history 70 5.8 Further state of knowledge 72 5.8.1 Preliminary remarks on test methods 72 5.8.2 Quantitative investigations under restraint conditions 73 5.8.3 A full-scale model for assessing the cracking risk at very early age 77 5.9 Summary 78 6 Investigation of the autogenous shrinkage cracking propensity 80 6.1 Introduction 80 6.2 Suitability of temperature-stress testing machines 80 6.2.1 Development, setup and use 80 6.2.2 Results of round robin tests 83 6.3 Restrained ring test - methodological foundations 86 6.3.1 Setup and use 86 6.3.2 Evaluation of restrained ring tests 90 6.3.3 Use of temperature changes for the investigation of creep and relaxation 96 6.4 Own investigations with the restrained ring test 97 6.4.1 Setup 97 6.4.2 Compensation of disturbing temperature effects 99 6.4.3 Repeatability 100 6.4.4 Measured steel ring strains 101 6.4.5 Simple stress analysis 102 6.4.6 Autogenous shrinkage cracking propensity - further analysis 106 6.4.7 Thermal stress component 116 6.4.8 Period of maximum cracking propensity 118 6.4.9 Restraint stress versus autogenous shrinkage 119 6.4.10 Cracking propensity versus autogenous shrinkage 120 6.4.11 Further considerations on creep 121 6.5 Summary 126 7 Summary, conclusions and outlook 128 7.1 Summary and conclusions 128 7.2 Outlook 130 8 Literature 131 9 Annex 159 / Das durch Selbstaustrocknung verursachte autogene Schwinden von besonders leistungsfähigen Betonen mit sehr niedrigem Wasserzementwert führt bei Dehnungsbehinderung bereits in sehr frühem Alter zu erheblichen Zwangsspannungen. Die Gefahr der Rissbildung, die sich daraus ergibt, lässt sich bislang nur unzureichend untersuchen. Experimentell besonders schwer zu erfassende Faktoren sind die Betontemperatur und die Viskoelastizität. Das vorrangige Ziel der Arbeit war die möglichst genaue Ermittlung der autogenen Schwindrissneigung repräsentativer Betone bei starker Dehnungsbehinderung und konstanter Raumtemperatur. Dabei waren die Prüfverfahren möglichst so zu wählen und weiterzuentwickeln, dass sich zukünftig alle relevanten Faktoren effizient und genau untersuchen lassen. Im Idealfall sollte eine Methode entstehen, die eine vollständige empirische Modellierung erlaubt. Zunächst wurden die methodischen Anforderungen und die Vor- und Nachteile existierender Prüfverfahren diskutiert. Darauf aufbauend wurden optimierte Verfahren vorgeschlagen. Ihre Eignung wurde an ultrahochfestem Beton überprüft. Bei der Auswahl der Betone wurden die wesentlichen Maßnahmen zur Schwindreduzierung berücksichtigt (innere Nachbehandlung, schwindreduzierende Zusatzmittel, Verringerung des Portlandzementanteils am Bindemittel). Das autogene Schwinden wurde mit dem Schwindkegelverfahren gemessen. Das neue Verfahren wurde durch Untersuchungen zur Wiederhol- und Vergleichsgenauigkeit validiert und erwies sich als effizient und genau. Es ermöglicht Messungen unter nicht-isothermen Bedingungen; hierfür existiert bisher kein etabliertes Verfahren. Das autogene Schwinden der untersuchten ultrahochfesten Betone unter quasi-isothermen Bedingungen (20 °C) betrug im Alter von 24 h zwischen 0,25 mm/m und 0,70 mm/m. Besonders gering war es bei Zugabe eines schwindreduzierenden Zusatzmittels bzw. Verwendung superabsorbierender Polymere. Mit dem Ring-Test wurden die bei Dehnungsbehinderung entstehenden Spannungen ermittelt. Ein großer Teil der gemäß Hooke’schem Gesetz zu erwartenden Spannungen wurde durch Kriechen und Relaxation abgebaut. Die im sehr frühen Alter stark ausgeprägte Relaxationsfähigkeit war der wesentliche Grund dafür, dass es selbst bei Betonen mit hohem autogenen Schwinden zu keiner erkennbaren Rissbildung kam. Die Entwicklung der autogenen Schwindrissneigung wurde als Verhältnis von Zwangsspannung und Spaltzugfestigkeit beschrieben. Durch modifizierte Ring-Tests, mit deren Hilfe die maximale Zugspannung ermittelt wurde, konnte gezeigt werden, dass das Verhältnis von Spannung und Festigkeit als Versagenskriterium geeignet ist. Die Rissneigung lässt sich aber nur dann korrekt berechnen, wenn das stark altersabhängige Verhältnis von einaxialer Zugfestigkeit und Spaltzugfestigkeit berücksichtigt wird. Außerdem ist zu beachten, dass es im sehr frühen Alter zu einer plastischen Spannungsumlagerung in Ring-Tests kommen kann. Der Referenzbeton wies eine hohe Rissneigung von bis zu 0,68 auf. Dass die schwindreduzierenden Maßnahmen zu deutlich geringeren Werten führten, zeigt deren Bedeutung für den sicheren Einsatz von ultrahochfestem Beton. Die hier bei 20 °C durchgeführten Untersuchungen erlauben allerdings keine abschließende Bewertung der Rissneigung unter baustellentypischen Bedingungen. Um die autogene Schwindrissneigung zukünftig als Funktion der Temperatur und des Lastniveaus empirisch modellieren zu können, wurden eine analytische Spannungslösung für nicht-isotherme Ring-Tests und ein neuer Ansatz zur Untersuchung der Resttrag- und Relaxationsfähigkeit mit Hilfe nicht-passiver Ring-Tests vorgeschlagen.:1 Introduction 2 Autogenous shrinkage 5 2.1 Shrinkage and hydration 5 2.2 Definitions and research approaches 10 2.3 Metrological issues 14 2.3.1 Multitude of test methods 14 2.3.2 Time-zero 16 2.3.3 Other metrological issues 18 2.4 Corrugated tube method 19 2.5 Influencing parameters 21 2.5.1 Concrete composition 21 2.5.2 Temperature 23 2.5.3 Specific countermeasures 25 2.6 Summary and conclusions with respect to the own work 25 3 Concretes used in the own investigations 27 3.1 Preliminary remarks 27 3.2 Concrete compositions 27 3.3 Constituents 28 3.3.1 Cement 28 3.3.2 Ground-granulated blast furnace slag 28 3.3.3 Silica fume 28 3.3.4 Admixtures 29 3.3.5 Aggregates 29 3.4 Mixing 29 3.5 Basic properties 30 3.5.1 Compressive strength 30 3.5.2 Splitting tensile strength 31 3.5.3 Modulus of elasticity 33 3.5.4 Analysis of mechanical properties 35 3.5.5 Coefficient of thermal expansion 38 3.5.6 Isothermal calorimetry 39 3.6 Summary 39 4 Shrinkage cone method for measuring autogenous shrinkage 41 4.1 Introduction 41 4.2 Setup and measurement procedure 41 4.3 Temperature control 44 4.4 Precision under quasi-isothermal conditions 47 4.4.1 Repeatability 47 4.4.2 Reproducibility 49 4.4.3 Shrinkage cone method vs. corrugated tube method 49 4.5 Autogenous shrinkage of the investigated concretes at 20 °C 54 4.6 Tests under non-isothermal conditions 55 4.7 Summary 56 5 Stress and cracks due to restrained autogenous shrinkage 58 5.1 Introduction 58 5.2 Degree of restraint 58 5.3 Formation of cracks 60 5.4 Very early age and importance of stress relaxation 63 5.5 Creep and cracking - further methodological aspects 65 5.6 Autogenous shrinkage cracking propensity 69 5.7 Role of temperature history 70 5.8 Further state of knowledge 72 5.8.1 Preliminary remarks on test methods 72 5.8.2 Quantitative investigations under restraint conditions 73 5.8.3 A full-scale model for assessing the cracking risk at very early age 77 5.9 Summary 78 6 Investigation of the autogenous shrinkage cracking propensity 80 6.1 Introduction 80 6.2 Suitability of temperature-stress testing machines 80 6.2.1 Development, setup and use 80 6.2.2 Results of round robin tests 83 6.3 Restrained ring test - methodological foundations 86 6.3.1 Setup and use 86 6.3.2 Evaluation of restrained ring tests 90 6.3.3 Use of temperature changes for the investigation of creep and relaxation 96 6.4 Own investigations with the restrained ring test 97 6.4.1 Setup 97 6.4.2 Compensation of disturbing temperature effects 99 6.4.3 Repeatability 100 6.4.4 Measured steel ring strains 101 6.4.5 Simple stress analysis 102 6.4.6 Autogenous shrinkage cracking propensity - further analysis 106 6.4.7 Thermal stress component 116 6.4.8 Period of maximum cracking propensity 118 6.4.9 Restraint stress versus autogenous shrinkage 119 6.4.10 Cracking propensity versus autogenous shrinkage 120 6.4.11 Further considerations on creep 121 6.5 Summary 126 7 Summary, conclusions and outlook 128 7.1 Summary and conclusions 128 7.2 Outlook 130 8 Literature 131 9 Annex 159

Purpose – The purpose of this study is to analyze a supply chain, where manufacturing is outsourced, to identify risk zones where any quantity difference could occur, the cause of quantity difference and the measures that could be taken to reduce the quantity difference.   Method – In order to achieve the purpose a literature study and a case study has been carried out. The case study consisted of interviews, observations and document studies and have generated empirical data. In the analysis of the collected data a fishbone diagram has been established and pattern matching has been applied. Findings – The case company's supply chain was mapped to identify risk zones, risk factors, and measures for the risk factors. The biggest risk zones identified in the cas company was picking and the store's inventory. The risk factors that were considered to be contributing most to the quantity difference was the lack of communication, mistakes, lack of standards, lack of security and barcode scanners. The measures that have been presented are improving communication, introduction of new standards, upgrading of bar code readers, enhancing safety and eliminating the possibility of making mistakes. Implications – With the help of the proposals that the authors have developed, the identified risk factors could be counteracted. Before proposed measures are implemented, all personnel should first and foremost take part of the mapping of the supply chain to understand how things work together. Method discussion – To achieve reliable results triangulation has been used when interviews, observation and document studies have been conducted. The internal validity is considered to have been strengthened by pattern matching when empirical data has been compared with theory. Since the respondents' answers have been similar the internal validity has been further strengthened and therefore credible. / Syfte – Studiens syfte är att analysera en försörjningskedja, där tillverkningen är outsourcad, för att upptäcka riskzoner där en eventuell kvantitetsdifferens skulle kunna uppstå, orsak till kvantitetsdifferens samt vilka åtgärder som skulle kunna vidtas för att minska kvantitetsdifferens. Metod – För att uppnå syftet har litteraturstudier samt en fallstudie genomförts. Fallstudien har bestått av intervjuer, observationer samt dokumentstudier och har genererat empirisk data. Vid analys av insamlad data har ett fiskbensdiagram upprättats samt mönsterpassning tillämpats. Resultat – Fallföretagets försörjningskedja kartlades för att identifiera riskzoner, riskfaktorer samt åtgärder till riskfaktorerna. De största riskzonerna som identifierades på fallföretaget var plockning samt butikslager. De riskfaktorer som ansågs vara mest bidragande till kvantitetsdifferens var kommunikationsbrist, misstag, bristfälliga standarder, brist på bevakning samt streckkodsläsare. Åtgärderna som har presenterats är förbättring av kommunikation, införande av nya standarder, uppgradering av streckkodsläsare, förhöjning av säkerhet samt eliminering av möjlighet att begå misstag. Implikationer – Med hjälp av de åtgärdsförslag som författarna har tagit fram kan de identifierade riskfaktorerna motverkas. Innan studiens åtgärdsförslag implementeras bör all personal först och främst ta del av kartläggningen av försörjningskedjan för att förstå hur aktiviteterna samverkar. Metoddiskussion – För att uppnå ett tillförlitligt resultat har triangulering använts då intervjuer, observationer och dokumentstudier har genomförts. Den interna validiteten anses ha stärkts genom mönsterpassning då empiri har jämförts med teori. Eftersom att respondenternas svar har varit likartade anses den interna validiteten stärkts ytterligare och är därmed trovärdig.

This thesis work deals with the volume changes during the solidification of cast irons and peritectic steels. The volume changes in casting metals are related to the expansion and/or contraction of the molten metal during solidification. Often, different types of shrinkage, namely macro- and micro-shrinkage, affect the casting quality. In addition to that, exposure of the metal casting to higher contraction or expansion during the solidification might also be related to internal strain development in samples, which eventually leads to surface crack propagation in some types of steel alloys during continuous casting. In consequence, a deep understanding of the mechanisms and control of the solidification will improve casting quality and production. All of the experiments during the entire work were carried out on laboratory scale samples. Displacement changes during solidification were measured with the help of a Linear Variable Displacement Transformer (LVDT). All of the LVDT experiments were performed on samples inside a sand mould. Simultaneously, the cooling curves of the respective samples during solidification were recorded with a thermocouple. By combining the displacement and cooling curves, the volume changes was evaluated and later used to explain the influence of inoculants, carbon and cooling rates on volume shrinkages of the casting. Hypoeutectic grey cast iron (GCI) and nodular cast iron (NCI) with hypo-, hyper- and eutectic carbon compositions were considered in the experiments from cast iron group. High nickel alloy steel (Sandvik Sanbar 64) was also used from peritectic steel type. These materials were melted inside an induction furnace and treated with different types of inoculants before and during pouring in order to modify the composition. Samples that were taken from the LVDT experiments were investigated using a number of different  methods in order to support the observations from the displacement measurements:  Differential Thermal Analysis (DTA), to evaluate the different phase present; Dilatometry, to see the effect of cooling rates on contraction for the various types of alloys; metallographic studies with optical microscopy; Backscattered electrons (BSE) analysis on SEM S-3700N, to investigate the different types of oxide and sulphide nuclei; and bulk density measurements  by applying Archimedes' principle. Furthermore, the experimental volume expansion during solidification was compared with the theoretically calculated values for GCI and NCI. It was found that the casting shows hardly any shrinkage during early solidification in GCI, but in the eutectic region the casting expands until the end of solidification. The measured and the calculated volume changes are close to one another, but the former shows more expansion. The addition of MBZCAS (Si, Ca, Zr, Ba, Mn and Al) promotes more flake graphite, and ASSC (Si, Ca, Sr and Al) does not increase the number of eutectic cells by much. In addition to that, it lowers the primary austenite fraction, promotes more eutectic growth and decreases undercooled graphite and secondary dendritic arm spacing (SDAS). As a result, the volume expansion changes in the eutectic region. The expansion during the eutectic growth increase with an increase in the inoculant weight percentage. At the same time, the eutectic cells become smaller and increase in number. The effect of the inoculant and the superheat temperature shows a variation in the degree of expansion/contraction and the cooling rates for the experiments. Effective inoculation tends to homogenize the eutectic structure, reducing the undercooled and interdendritic graphite throughout the structure. In NCI experiments, it was found that the samples showed no expansion in the transversal direction due to higher micro-shrinkages in the centre, whereas in the longitudinal direction the samples shows expansion until solidification was complete.   The theoretical and measured volume changes agreed with each other. The austenite fraction and number of micro-shrinkage pores decreased with increase in carbon content. The nodule count and distribution changes with carbon content. The thermal contraction of NCI is not influenced by the variation in carbon content at lower cooling rates. The structural analysis and solidification simulation results for NCI show that the nodule size and count distribution along the cross-sections at various locations are different due to the variation in cooling rates and carbon concentration. Finer nodule graphite appears in the thinner sections and close to the mold walls. A coarser structure is distributed mostly in the last solidified location. The simulation result indicates that finer nodules are associated with higher cooling rate and a lower degree of microsegregation, whereas the coarser nodules are related to lower cooling rate and a higher degree of microsegregation. As a result, this structural variation influences the micro-shrinkage in different parts. The displacement change measurements show that the peritectic steel expands and/or contracts during the solidification. The primary austenite precipitation during the solidification in the metastable region is accompanied by gradual expansion on the casting sides. Primary δ-ferrite precipitation under stable phase diagram is complemented by a severe contraction during solidification. The microstructural analysis reveals that the only difference between the samples is grain refinement with Ti addition. Moreover, the severe contraction in solidification region might be the source for the crack formation due to strain development, and further theoretical analysis is required in the future to verify this observation. / <p>QC 20170228</p>

La variabilité des propriétés du bois ainsi que son hygroscopicité pourraient être un frein à son utilisation dans la construction, même s’il peut être considéré comme un matériau de choix dans le contexte environnemental et économique actuel. Il est donc primordial de mieux comprendre les origines physiques du comportement du bois pour être capable d’améliorer la prédiction de ses propriétés, et pouvoir ainsi le rendre plus compétitif par rapport aux autres matériaux de construction. Le comportement hygromécanique du bois, caractérisé par des variations dimensionnelles en présence de variations d’hygrométrie, est particulièrement difficile à prédire, du fait de sa microstructure multi-échelle et de ses interactions complexes avec l’eau.Dans ce contexte, la thèse vise à comprendre et enrichir les relations entre la microstructure du bois, ses propriétés de sorption et son comportement hygromécanique, en étudiant l’influence de l’hétérogénéité de l’accroissement annuel (cerne), constitué de bois initial et de bois final dont la structure et les propriétés présentent de nombreuses différences. Cette étude est menée sur du Douglas (Pseudotsuga menziesii (Mirb.) Franco), actuellement référencé comme un matériau de structure intéressant. Des outils de caractérisation avancés sont utilisés : la Résonance Magnétique Nucléaire (RMN) du proton pour caractériser les mécanismes de sorption ; la corrélation d’images volumiques (DVC) pour mesurer les champs de déformations à partir d’images 3D de microtomographie aux rayons X (µTRX), donnant aussi accès à la densité locale du bois.Après une introduction sur le matériau bois et un état de l’art sur son comportement hygromécanique, une caractérisation préliminaire de la microstructure (angle des microfibrilles, largeur de cerne, densité) et du comportement hygromécanique d’échantillons de bois initial et de bois final prélevés dans différents cernes est menée. Une forte anisotropie du bois initial est mise en évidence en opposition au comportement isotrope transverse du bois final. Les déformations suivant la direction des fibres présentent aussi de fortes non-linéarités peu discutées dans la littérature. Une discussion sur la variabilité des propriétés est par ailleurs engagée, ainsi que sur les relations structure-propriétés à l’échelle macroscopique.L’origine des différences de comportement hygromécanique entre le bois initial et le bois final est tout d’abord recherchée au niveau des mécanismes de sorption, au travers une étude de relaxométrie RMN du proton en 2D (cartes T1-T2). Deux types d’eau liée situés dans des environnements distincts sont mis en évidence et leur isotherme de sorption diffère dans les deux types de bois. Une hypothèse sur leur localisation dans la paroi cellulaire est proposée, puis une modélisation simplifiée 2D est effectuée pour évaluer leur impact respectif sur le comportement hygromécanique du bois initial et du bois final, en particulier dans la direction des fibres.Enfin, les champs de déformations locaux et globaux sont étudiés en analysant par DVC des images de µTRX de bois initial et de bois final soumis à différentes sollicitations hydriques. Le couplage entre ces deux matériaux est aussi étudié pour évaluer leurs interactions et comprendre le comportement du bois à l’échelle du cerne. Un protocole de DVC adapté aux images de bois est proposé. Les comportements hygromécaniques du bois initial, du bois final et du cerne sont comparés. A l’échelle locale, des hétérogénéités du champ de déformations sont mises en évidence et corrélées à la densité locale. Leur effet sur le comportement du cerne et sur la courbure des échantillons induite par le chargement hydrique est analysé. Une modélisation 3D par éléments finis, tenant compte des gradients locaux de propriétés, vient enfin compléter cette étude pour améliorer la compréhension des interactions mécaniques entre le bois initial et le bois final / Wood has highly variable properties and is also hygroscopic. These characteristics may restrict its use in construction even if it can be considered as a material of choice with the current environmental and economical concerns. Therefore, it is essential to better understand the physical origins of the behavior of wood in order to improve the prediction of its properties, and making it competitive with respect to other building materials. Dimensional changes of wood appear when it is subjected to relative humidity variations. This hygromechanical behavior is particularly difficult to predict because of the multiscale structure of wood and its complex interactions with water.In this context, the present work aims to understand and enrich relationships between microstructure, sorption properties and hygromechanical behavior of wood. More specifically, it is focused on the influence of the growth-ring heterogeneity, constituted of earlywood and latewood which have different structures and properties. The study is performed on Douglas fir (Pseudotsuga menziesii (Mirb.) Franco), which is a species of significant interest for structural applications. Advanced characterization tools are used: proton Nuclear Magnetic Resonance (NMR) to characterize sorption mechanisms; digital volume correlation (DVC) to measure deformation fields from X-Ray microtomography 3D images (XRµT), also providing local density of wood.At first wood properties and its hygromechanical behavior are described through a literature overview. Preliminary microstructural (microfibril angle, growth-ring width, density) and hygromechanical behavior characterizations of earlywood and latewood samples with different cambium age are performed. Earlywood reveals a strong anisotropic behavior compared to latewood which is isotropic in the transversal plane. Moreover, strains along the fiber direction nonlinearly evolve with moisture content. This phenomenon has been hardly reported and studied in the literature. Discussions on variability of properties and on relationships between structure and properties are also initiated.Sorption mechanisms are then studied by 2D NMR relaxometry (T1-T2 correlation spectra) in order to investigate differences between earlywood and latewood hygromechanical behaviors. Two types of bound water located in distinct environments are highlighted and their sorption isotherms are shown to be different in the two types of wood. A hypothesis on their location in the cell-wall is proposed and a simple 2D model is developed to evaluate their respective effect on the hygromechanical behavior of earlywood and latewood, especially in the fiber direction.Furthermore, local and global strains fields are studied using DVC from XRµT images of earlywood and latewood subjected to relative humidity variations. The coupling of these two materials is also investigated in order to evaluate their mechanical interactions and to understand the behavior at the growth-ring scale. A specific DVC procedure is developed for images of wood. The hygromechanical behaviors of earlywood, latewood and a growth-ring are compared. At the local scale, strains fields heterogeneities are highlighted and correlated to the local density. Their effect on the growth-ring behavior and the samples curvature is analyzed. A 3D finite elements model which takes into account local gradients of properties is finally developed to better understand earlywood-latewood mechanical interactions

För att kunna prediktera fel under processen precisionsgjutning, används simulering som ett verktyg att bestämma hur detaljer skall gjutas. Det finns flertalet parametrar som är möjliga att manipulera eller ändra. För att komma så nära ett verkligt scenario som möjligt behöver dessa parametrar optimeras mot verkligheten. Syftet med detta arbete är att analysera faktorer som påverkar värmeöverföring för att sedan presentera optimeringsförslag åt TPC Components AB mot de gjorda mätningarna. Detta görs för att kunna förutsäga till exempel hur snabbt ett föremål kyls i processen och hur det påverkar det slutgiltiga resultatet. Resultat från de optimerade simuleringarna jämförs mot ett praktiskt försök och även TPC Components tidigare standard utvärderas mot arbetets optimerade simuleringar. En viktig fråga är att undersöka varför simuleringsprogrammet inte blir identiskt med verkligheten och hur detta undersöks. Känslighetsanalyser utgör därmed en stor del av arbetet och det är av vikt att undersöka enskilda parametrar för att få en uppfattning om vad som påverkar resultatet och vad som går att bortse från. Resultatet av arbetet visar att ändrade parametrar för indata vid precisionsgjutning i TPC Components AB:s vakuumprocess har resulterat i en mer pålitlig simulering. Simuleringen bygger nu på vetenskapliga bevis och faktiska uppmätningar under processen för att kunna simulera så likt verkligheten som möjligt. Arbetet resulterar i att färre gjutförsök innan godkänt tillvägagångssätt för nya detaljer behövs, då simulering kommer att visa resultat närmare verkligheten. Arbetet har förbättrat TPC Components simuleringsarbete vilket kommer påverka både arbetstid, kostnader och miljö på ett positivt sätt. / Simulation of investment casting play a big role for prediction of problems in the process. TPC Components AB need their simulations to be optimised against the practical process in order to get results as close as possible to the results from real practice. This thesis is mapping the parameters sensitivity against TPC’s earlier standards to put forward what parameters the result is depending on. The optimization of the parameters is towards measurement of heat and time and practical trial in order to get a scientific ground. The heat transfer for complicated structure needs to be simulated in order to get a reasonable value to work with. Results shows that TPC’s earlier model of simulation have lack of measurement and important parameters are based on assumptions. The work done gave TPC a model closer to the real practice which can both mean faster decisions of how to cast new products and less costly practical trials.

Made available in DSpace on 2016-12-23T14:05:55Z (GMT). No. of bitstreams: 1 1- Capa.pdf: 18716 bytes, checksum: 1dddf42a26f2caa98fb2c3c97cb02a55 (MD5) Previous issue date: 2004-04-06 / O estudo realizado nesta dissertação é motivado principalmente pelo interesse em se avaliar a resposta térmica e tensional em estruturas de concreto massa nas primeiras idades considerando a viscoeleasticidade linear do material. Neste trabalho desenvolvem-se modelos numérico-computacionais e implementam-se os softwares PFEM_2DT e PFEM_2DAT, de análise térmica e de tensões no domínio bidimensional, respectivamente. Inicialmente, foi realizada uma revisão bibliográfica do problema termomecânico considerando os mecanismos e modelos físicos para a representação do mesmo, utilizando-se de ferramentas matemáticas e artifícios numéricos para a resolução do problema. Foram pesquisados ainda os parâmetros influentes na análise de temperaturas e de tensões no concreto massa, destacando-se as propriedades viscoelásticas de fluência e retração onde de forma resumida, porém esclarecedora, foram descritos seus principais conceitos, causas e efeitos bem como alguns dos principais modelos de representação dessas deformações diferidas. Por meio desses estudos teóricos e das conseqüentes implementações computacionais, foram analisadas diferentes estruturas de concreto através dos programas computacionais desenvolvidos e, posteriormente, realizadas comparações com resultados experimentais e numéricos obtidos por outros pesquisadores objetivando validar os aplicativos. Na elaboração deste software fez-se uso do novo paradigma de programação Programação Orientada a Objetos, associado à resolução de sistemas de equações lineares pelo Método dos Gradientes Conjugados. Analisou-se o comportamento das temperaturas de diferentes estruturas sujeitas ao calor interno gerado no concreto sob a influência das condições ambientais, que permitem que esse calor seja dissipado, gerando como conseqüência, distribuições de temperatura não-lineares no interior da estrutura. Nos estudos tensionais das estruturas de concreto massa analisadas nesta dissertação pôde-se avaliar a evolução das tensões de origem térmica considerando o material viscoelástico linear, comparando-as com a resistência à tração do material. Por fim, efetuaram-se estudos comparativos de tensões em um bloco de concreto massa utilizando os modelos elástico e viscoelástico linear, avaliando o nível de importância e o grau de magnitude da consideração dos efeitos diferidos associados quando da análise do comportamento termomecânico de estruturas maciças de concreto em fase de construção. / The study accomplished in this dissertation is motivated mainly by the interest in evaluating the temperature and stress field in mass concrete structures at first ages considering material s linear viscoeleasticity. The development and implementation of numeric-computational models are also scope of this work by means the softwares PFEM_2DT and PFEM_2DAT, of thermal stress analysis in the two-dimensional domain. Initially at first step, the bibliographic revision of the thermomechanical problem was accomplished considering physical mechanisms and models for its representation, using mathematical tools and numeric artifices for the resolution of the problem. The influential parameters in the analysis of temperatures and stresses in the mass concrete were also researched, standing out to the creep and shrinkage rheological properties, where in a summarized and elucidated way, its main concepts, causes and effects were described, as well as some of the main representation models of those differed deformations. Considering theoretical studies and the consequent computer implementations, different concrete structures were analyzed through the developed softwares, and compared with experimental and numeric results obtained by other researchers with the objective of validating the applications. In the elaboration of this software was made use of the new programming paradigm Object-Oriented Programming, associated to the resolution of linear systems equations for the Method of the Conjugated Gradients. The temperature field s behavior of different structures subject to the internal heat into the concrete was analyzed, under the influence of the environmental conditions, that allows the heat to be dissipated, generating as consequence, non- linear temperature distributions inside the structure. The termomechanic problems studied in mass concrete structures were analyzed in this dissertation, where the thermal stresses behaviour could be evaluated considering the linear viscoelasticity properties of the material, comparing them with the tensile strength of the material. Finally, comparative studies of the stresses in a mass concrete block were executed using elastic models and linear viscoelasticity, evaluating the importance and magnitude levels of the associated differed effects in the termomechanic behavior analysis of solid concrete structures in a construction stage.

Plasma treatment of polypropylene fibers presents a new progressive method, how to increase the utility properties of these fibers. The thesis is focused on verifying the effect of surface treatment of polypropylene fibers in concrete with low-temperature plasma discharge in the normal atmosphere. The paper describes the procedure for treatment of polypropylene fibers with low-temperature plasma and evaluate the impact of this adjustment on the volume changes of cement composites. It should also be emphasized that the thesis is focused on the initial volume changes, ie volume changes in the early stages of solidification and only for your own mixture, which is not considered an external load. Subsequently, the experimental verification of the effect of the addition made commercially available fibers and fibers treated plasma volume changes to a selected physico-mechanical properties of the test compounds.

<p>U disertaciji su prikazani rezultati sopstvenog eksperimentalnog istraživanja mogućnosti primjene biopepela, nastalog sagorjevanjem žetvenih ostataka, kao mineralnog dodatka u cementnim kompozitima: malterima i betonima. Istraživanje je zasnovano na komparativnoj analizi 27 vrsta maltera u kojima je varirana vrsta biopepela i vrsta sitnozrnog agregata i 12 vrsta betona koje se razlikuju u vrsti veziva i u krupnom agregatu. Na očvrslom malteru su ispitani: konzistencija, čvrstoća pri pritisku i kapilarno upijanje vode, a na očvrslom betonu ispitivani su kapilarno upijanje vode, skupljanje pri sušenju, čvrstoća pri pritisku, dinamički modul elastičnosti, otpornost na habanje, vodonepropustljivost i čvrstoća na savijanje. Istaknuto je da se primjenom biopepela kao zamjene dijela cementa mogu dobiti konstrukcijski malteri i betoni, odnosno betoni sa zadovoljavajućim fizičkim i mehaničkim karakteristikama.</p> / <p>The dissertation presents the results of an own experimental research on the possibility of using biomass ash as a mineral additive in cement composites: mortar and concrete. The research is based on a comparative analysis of 27 types of mortars in which the type of biomass ash and the type of fine-grained aggregate were varied, as well as 12 types of concrete that differ in the type of binder and coarse aggregate. Following properties were tested on hardened mortars: consistency, compressive strength and capillary water absorption, and on hardened concrete: capillary water absorption, drying shrinkage, compressive strength, dinamic modulus of elasticity, wear resistance, waterpermeability and flexural strength. It was pointed out that biomass ash can be used as a replacement of a part of cement, thereby obtaining satisfactory physical and mechanical characteristics.</p>

Le changement climatique et les événements climatiques sévères tels que les périodes de sécheresse/humidification prolongées sont à l'origine du phénomène de retrait-gonflement dans les sols argileux. Ce phénomène est affecté par les interactions sol-végétation-atmosphère (SVA) et peut causer d’importants dommages structurels aux constructions légères telles que les bâtiments résidentiels. L’objectif de ce travail de recherche est de modéliser le comportement in situ du retrait-gonflement des sols gonflants dans un contexte SVA en se basent sur des outils numériques. Une méthode d'interaction sol-atmosphère est initialement présentée accompagnée d’un modèle couplé hydro-thermique du sol. Cette approche a été principalement mise en place afin de déterminer les conditions aux limites temporelles à la surface du sol en se basent sur la notion du bilan de masse et d'énergie pour déterminer a posteriori, les modifications spatio-temporelles de la succion du sol, de la teneur en eau et de la température. Cette approche a été validée à l'aide des observations in situ des sites instrumentés. Par la suite, l’influence de l’absorption d’eau par les végétations a été intégrée dans le terme source de l’écoulement de l’eau dans un milieu non saturé, à l’aide d’un modèle d’absorption d’eau de racine existant.Les variations temporelles de succion ont été postérieurement reliées au comportement volumique du sol en appliquant une approche simple développée à partir des résultats expérimentaux des essais de séchage/humidification réalisés dans la littérature. Les indices associés dans le plan indice des vides-log succion, ainsi que les paramètres complémentaires du modèle linéaire ont été corrélés aux paramètres géotechniques de base. L'approche proposée a été ultérieurement validée avec des données in situ fournies par la surveillance d’un site expérimental. Le site expérimental de Roaillan a été instrumenté afin de surveiller les modifications physiques du sol ainsi que le comportement structurel du bâtiment. Les comparaisons entre les résultats de la modélisation et les observations in situ de la succion du sol, la teneur en eau, la température et les mouvements du sol dans le temps ont montré une performance acceptable du modèle. L’approche a ensuite été appliquée pour étudier l’influence des projections climatiques futures (2050) sur les variables physiques et les mouvements du sol sur ce site. Trois scénarios RCP relatifs aux changements climatiques ont été examinés dans cette étude, qui ont révélé des différents comportements possibles à court terme et à long terme. Finalement, l'approche développée a été appliquée au territoire français en le divisant en six régions climatiques. Différents paramètres de sol ont été attribués à chacune de ces régions climatiques afin de définir les conditions de référence. En conséquence, l’influence de différents facteurs externes sur les mouvements du sol a été analysée sur une période donnée. Enfin, l’étude suggère les mesures adéquates à prendre pour minimiser l’amplitude du phénomène de retrait et de gonflement dans un contexte SVA. / Climate change and severe climatic events such as long drought/rehydration periods are at the origin of the shrinkage and swelling phenomenon in expansive soils. This phenomenon is affected by Soil-Vegetation-Atmosphere (SVA) interactions and can cause severe structural damage to lightly loaded constructions such as residential buildings. The objective of this re-search work is to simulate the in-situ behavior of the shrinkage-swelling in expansive soils in a SVA context using numerical tools. A soil-atmosphere interaction method is primarily presented along with a coupled hydro-thermal soil model. This approach was established in order to determine primarily, the natural time variable boundary conditions at the considered soil surface based on the mass and energy balance concept, and secondly to determine the spatial-temporal changes of the soil suction, water content and temperature. This approach was validated using in situ observations of monitored sites. Thereafter, the influence of the water uptake by vegetation was incorporated in the source term of the unsaturated water flow theory, using an existing root water uptake model. Subsequently, the temporal variations of the soil suction were related to the volume change behavior using a simple approach developed based on the experimental results of drying/wetting tests performed in the literature. The associated volumetric indices in the void ratio-log suction plan, along with the complementary parameters of the linear model were correlated with basic geotechnical parameters. The proposed approach was validated with in situ data provided from an experimental site. The Roaillan experimental site was instrumented in order to monitor the soil’s physical changes along with the structural behavior of the building. Comparisons between the simulated and observed soil suction, soil water content, temperature and soil movements in time and depth showed an acceptable performance of the predictions. The approach was then extended to study the influence of future climate projections (2050) on the soil’s physical variables and movements. Three RCP climate change scenarios were considered in this analysis which revealed different possible behavior in both short term and long term. Finally, the developed approach was applied to the French territory by dividing it to six different climatic regions. Different soil parameters were attributed to each of these climatic regions in order to set the reference condition. Thereafter, the influence of different external factors was analyzed on the soil movements over a chosen period. The study finally suggests the adequate actions to take for minimizing the amplitude of the shrinkage and swelling phenome-non in a SVA context.