In today's textbooks, analyzing prestressed members and bridges takes a two-dimensional approach. Two-dimensional analysis is the only way to analyze prestressed bridges and members because otherwise the hand calculations are extremely difficult. Skewed bridges, however, need to be analyzed and designed three-dimensionally. Based on engineering inspection, it is possible to tell how a non-skewed symmetric bridge will behave. However, the knowledge of how a skewed bridge will behave cannot be obtained by inspection only. Finite element analysis can be used to model a bridge and discover how the bridge will react to dead loads and post-tensioning forces. When a bridge is built on a skew, the acute corners of the bridge support much less concrete weight than the obtuse corners of the bridge. If the post-tensioning force causes a decrease in load at the acute corners of the skewed bridge and if the skew of the bridge is great enough, there is a concern that there could be uplift at these acute corners of the bridge. Uplift at any corner of the bridge should not be allowed. The objective of this study is to investigate a simple span skewed box girder bridge to see if any uplift occurs at the acute corners of the bridge due to post-tensioning forces. After careful study of a skewed simple span cast-in-place post-tensioned box girder bridge, it was found that the post-tensioning force actually transfers more downward force into the acute corners of the bridge. Based on this study, the post-tensioning force will not cause uplift in the acute corners of the skewed bridge.

Master of Science / Department of Architectural Engineering and Construction Science / Darren D. Reynolds / Highly expansive soils can severely damage the foundations which they support. These damages create unnecessary maintenance cost to the owner and can be detrimental to the building superstructure. Post-tensioned ribbed mat foundations are commonly used in light commercial construction in areas in the United States that have highly expansive soils. Mild reinforced ribbed mat foundations are rarely used in these areas. This report investigates why post-tensioned ribbed mat foundations are more common in these areas than mild-reinforced ribbed mat foundations. The approach to this investigation is a design example which designs and compares the two foundation types. The design example is a typical 2-story office building located in Dallas, Texas, which is an area that has highly expansive soils. First, a post-tensioned ribbed mat foundation is designed for the office building. Next, a mild-reinforced ribbed mat foundation is designed for the same building. A comparison is done between the two foundations based on serviceability, strength requirements and construction costs. The findings in the comparison is that post-tensioning is a more economical and constructible method. Using mild-reinforcement requires the use of shear reinforcement in the ribs which is not typical in foundation design and construction and is less economical, and additional reinforcement in the slab is needed to resist bending stresses which is also less economical. The finding of the report is that of the two foundation types, the post-tensioned ribbed mat foundation is the better design based on the three areas of interest listed above. The use of a mild-reinforced mat foundation would require construction procedures that are not typical and would be less economical.

Unreinforced masonry (URM) structures have historically been regarded as structurally unsound in response to seismic events. The tendency for URM walls to collapse out-of-plane in a brittle manner is continually cause for concern. Retrofit of these walls is necessary in order to prevent severe damage and injury to occupants. This paper is concerned with the retrofit of unreinforced masonry (URM) walls in response to out-of-plane loading. A retrofit design was developed and verified through structural testing. The selected retrofit technique incorporates vertical coring of URM walls to allow for the insertion of a single post-tensioning (PT) tendon. Tendons are spaced at a regular interval and anchored at the top of the wall parapet and at the lower diaphragm level. Tensioning of the tendons imparts a compressive stress to the wall that effectively increases the wall cracking moment strength, ultimate moment strength and displacement capacity. Additionally, the insertion of PT tendons allows the wall to behave in a ductile manner in response to out-of-plane ground motion. Extensive research was conducted in order to accurately portray the material properties and construction methods of unreinforced masonry walls in San Luis Obispo, California. Various mortar mix designs were generated and tested so that a mix design could be selected to best reflect the target URM structures. Seismic parameters were generated to reflect a URM structure in San Luis Obispo. An unreinforced masonry wall was constructed by a professional mason using the established mortar mix proportions and salvaged bricks from the 1920 era. Having a pin-pin unsupported height of 11 feet, the wall constructed for testing was reflective of the configuration of URM walls in many downtown San Luis Obispo structures. The wall was loaded in the out-of-plane direction by 4 equal point loads mimicking a uniformly distributed load. The testing program consisted of cycling the wall through target internal moments and target displacements. It was verified through testing that post-tensioning tendons can be successfully introduced in URM walls to resist out-of-plane bending. Testing showed that the addition of PT tendons significantly increased the wall’s cracking moment capacity, giving it the elastic strength to resist twice the forces imposed by the design-level ground motion. PT tendons also increased the nominal strength of the wall, allowing the wall to achieve large displacements without collapse. It was also found that PT tendons provided a restoring force to the wall returning it to almost no residual displacement after each displacement cycle. Thus, the URM wall retrofitted with PT tendons demonstrated significant integrity as a structural system, providing for improved strength and ductility with no residual displacement.

A bridge with precast bridge deck panels was built at the Virginia Tech Structures Laboratory to examine constructability issues, creep and shrinkage behavior, and strength and fatigue performance of transverse joints, different types of shear connectors, and different shear pocket spacings. The bridge consisted of two AASHTO type II girders, 40 ft long and simply supported, and five precast bridge deck panels. Two of the transverse joints were epoxied male-female joints and the other two transverse joints were grouted female-female joints. Two different pocket spacings were studied: 4 ft pocket spacing and 2 ft pocket spacing. Two different shear connector types were studied: hooked reinforcing bars and a new shear stud detail that can be used with concrete girders. The construction process was well documented. The change in strain in the girders and deck was examined and compared to a finite element model to examine the effects of differential creep and shrinkage. After the finite element model verification study, the model was used to predict the long term stresses in the deck and determine if the initial level of post-tensioning was adequate to keep the transverse joints in compression throughout the estimated service life of the bridge. Cyclic loading tests and shear and flexural strength tests were performed to examine performance of the different pocket spacings, shear connector types and transverse joint configurations. A finite element study examined the accuracy of the AASHTO LRFD shear friction equation for the design of the horizontal shear connectors. The initial level of post-tensioning in the bridge was adequate to keep the transverse joints in compression throughout the service life of the bridge. Both types of pocket spacings and shear connectors performed exceptionally well. The AASHTO LRFD shear friction equation was shown to be applicable to deck panel systems and was conservative for determining the number of shear connectors required in each pocket. A recommended design and detailing procedure was provided for the shear connectors and shear pockets. / Ph. D.

A set of experimental tests were performed at Virginia Tech to investigate transverse joints and blockouts on full depth concrete bridge deck panels. The joints were designed on a deck replacement project for a rural three span continuous steel girder bridge in Virginia. Two cast-in-place and four post-tensioned joints were designed and tested in cyclical loading. Each joint was tested on a full scale two girder setup in negative bending with a simulated HS-20 vehicle. The blockouts were built as hollow concrete rings filled with grout and left to shrink under ambient conditions. Thirteen combinations of different surface conditions and grouts were designed to test the bond strength between the materials. The strain profile, cracking patterns, and ponding results were measured for all specimens. A finite element analysis was performed and calibrated with the laboratory results. The cast-in-place joints and the two post-tensioned joints with 1.15 MPa (167 psi) of initial stress experienced cracking and leaked water by the end of the tests. The two post-tensioned joints with 2.34 MPa (340 psi) initial stress kept the deck near a tensile stress of 1.5√(𝑓'c) and performed the best. These transverse joints did not leak water, did not have full depth cracking, and maintained a nearly linear strain distribution throughout the design life. Full depth deck panel may be effectively used on continuous bridges if a sufficient amount of post-tensioning force is applied to the transverse joints. The finite element model provides a design tool to estimate the post-tensioning force needed to keep the tensile stresses below the cracking limit. The blockouts with a roughened surface or an epoxy and a grout equivalent to Five Star Highway Patch grout had the highest bond stresses, did not leak water, and had smaller cracks at the grout-concrete interface than the control samples. A minimum bond strength of 2.5√(𝑓'c) was maintained for all of the specimens with a grout equivalent to Five Star Highway Patch. A pea gravel additive in the grout reduced shrinkage and reduced the bond strength. The finite element model provides a design tool to estimate cracking at the grout-surface interface. / Ph. D.

Currently, NCHRP Report 356 has published an equation to estimate the ultimate strength of the local zone in normal weight concrete. The local zone is the area of concrete directly ahead of the bearing plate. The equation can be broken into two distinct parts: unconfined bearing strength of concrete enhanced by the A/A_b ratio and the enhancement of strength due to the presence of confining. Research has shown that the strength enhancement of the A/Ab ratio and confining reinforcing is less in lightweight concrete than in normal weight concrete. To determine the strength of the local zone in lightweight concrete 30 reinforced prisms, 2 unreinforced prisms, and concrete cylinders were tested. The dimensions of the prisms were 8 in. x 8 in. x 16 in. and the cylinders were 4 in. x 8 in. cylinders. The simulated reinforcing in the prisms extended only through the top 8 in. of the prism and consisted of either ties or spirals with different spacing or pitch, respectively. To determine the effect of the A/A_b ratio for each spacing or pitch arrangement of the reinforcing, one of two different size bearing plates were used. From the testing performed in this research and other research, it is apparent that the NCHRP equation is unconservative when estimating the ultimate strength of the local zone in lightweight concrete. By modifying both parts of the NCHRP equation it is possible to conservatively predict the ultimate strength of the local zone in lightweight concrete. Also investigated in this thesis are equations to predict the splitting cylinder strength and modulus of elasticity of lightweight concrete. For a sand-lightweight concrete, as defined by ACI 318-05 Building code and Commentary, the splitting tensile strength can be accurately predicted by multiplying the square root of the compressive strength by 5.7. / Master of Science

As the aging bridges and infrastructure within the US continue to deteriorate, traffic delays due to construction will become more and more common. One method that can reduce delays due to bridge construction is to use precast deck panels. Precast deck panels can significantly reduce the overall length of the construction project. The panels can be manufactured ahead of time, and with higher quality control than is possible in the field. One of the reasons precast deck panels are not widely accepted is because of a lack of research concerning the required post-tensioning force, shear stud pocket placement, and proper joint design. In a recent dissertation (Swenty 2009) numerous recommendations were made for joint design, shear stud pocket design, and post-tensioning force for full-depth precast deck panel bridges. Design drawings were included for the replacement of a bridge located in Scott County, Virginia. The research in this report focuses on the short-term and long-term testing of this bridge. The short-term testing involved performing a live load test in which two trucks of known weight and dimensions were positioned on the bridge in order to maximize the negative moment at the joints over the piers and document strains and deflections at a number of other critical locations. The long-term testing involved monitoring the strains within the deck and on one of the six girders for a number of months in order to document the changes in strain due to creep and shrinkage. The results of these tests were compared to 2D beam-line models and to the parametric study results of Bowers' research on prestress loss within full-depth precast deck panel bridges. It was determined that the bridge was acting compositely and that the post-tensioning force was sufficient in keeping the joints in compression during testing. / Master of Science

Os efeitos causados pela radiação de raios X em transistores MOS com canal n e p, de porta tripla, com e sem tensionamento mecânico foram estudados teórica e experimentalmente. Após a irradiação de raios X, com doses totais acumuladas de 150Mrad e 3Mrad, em 12 dispositivos diferentes, foi constatada uma grande variação em algumas de suas características, como por exemplo um aumento de 40mV/dec na inclinação de limiar nos dispositivos mais largos, devido a criação de cargas positivas nos óxidos, tanto no enterrado quanto de porta. As cargas geradas no óxido enterrado apresentaram grande influência na condução de corrente pela segunda interface (silício/óxido enterrado), uma vez que o óxido enterrado é espesso (145nm), que por sua vez afetou significativamente diversos outros parâmetros. A formação de cargas positivas no óxido enterrado causou variações distintas no valor da tensão de limiar referente à segunda interface, em função do canal do transistor. No caso dos transistores MOS com canal n, foi observado um aumento na condução parasitária proveniente da segunda interface, enquanto que nos dispositivos MOS com canal p essa corrente foi minimizada devido aos efeitos da radiação, gerando uma melhora do desempenho destes dispositivos (é possível observar uma redução da inclinação de sublimiar de 60 a 80mV/dec no caso dos dispositivos mais largos). Entretanto, como o óxido de porta é mais fino que o óxido enterrado, não foram observadas grandes variações causadas pelas cargas formadas no mesmo. Quando comparada com a literatura, a dose total acumulada foi muito alta, então foi realizada uma nova rodada de radiação em que a distância entre a amostra e o feixe e a taxa de exposição foram alterados. Com isso, foi possível observar o mesmo efeito descrito anteriormente, porém com uma dose total acumulada cerca de 50 vezes menor. Em ambos os casos a exposição foi realizada sem a presença de polarização nos dispositivos, usada normalmente para reduzir a recombinação de portadores aumentando assim o efeito da radiação. Para realizar a polarização dos dispositivos durante a exposição de radiação, é preciso que os dispositivos estejam encapsulados, porém a etapa de encapsulamento poderia danificar os dispositivos utilizados. / The effects caused by the X-ray radiation in n and p channel triple gate transistors, with and without mechanical tensioning, were theoretically and experimentally studied. After X-ray irradiation on the devices, it was observed a large variation in some of their characteristics, such as an increase in the slope of 40mV/dec threshold in larger devices, due to the creation of positive charges in the oxides, both for the buried oxide as for the gate oxide. The charges generated in the buried oxide had great influence on the current conduction through the back interface (silicon / buried oxide), since the buried oxide is thicker (145nm), which in turn significantly affected many other parameters. The formation of positive charges in the buried oxide caused distinct variations in the threshold voltage on the second interface, depending on the channel type of the transistor. In the case of n-channel MOS transistors, it was observed an increase in parasitic conduction from the second interface, whereas for p-channel MOS devices the current was minimized - due to the effects of radiation - generating an improvement in performance of such devices (it is possible to observe a reduction of subthreshold slope of 60 to 80mV/dec in the case of larger devices).. However, as the gate oxide is thinner than the buried oxide, large variations caused by the generated charges were not observed. When compared to the literature, the total accumulated dose was very high, therefore, a new round of radiation were performed, in which the distance between the sample and beam, and the exposure rate were changed. Thus, it was possible to observe the same effect described above, but with a total accumulated dose about 50 times lower. In both cases, the bias was not used during the radiation exposure, normally used to reduce the recombination, thereby increasing the the radiation effect.

Os efeitos causados pela radiação de raios X em transistores MOS com canal n e p, de porta tripla, com e sem tensionamento mecânico foram estudados teórica e experimentalmente. Após a irradiação de raios X, com doses totais acumuladas de 150Mrad e 3Mrad, em 12 dispositivos diferentes, foi constatada uma grande variação em algumas de suas características, como por exemplo um aumento de 40mV/dec na inclinação de limiar nos dispositivos mais largos, devido a criação de cargas positivas nos óxidos, tanto no enterrado quanto de porta. As cargas geradas no óxido enterrado apresentaram grande influência na condução de corrente pela segunda interface (silício/óxido enterrado), uma vez que o óxido enterrado é espesso (145nm), que por sua vez afetou significativamente diversos outros parâmetros. A formação de cargas positivas no óxido enterrado causou variações distintas no valor da tensão de limiar referente à segunda interface, em função do canal do transistor. No caso dos transistores MOS com canal n, foi observado um aumento na condução parasitária proveniente da segunda interface, enquanto que nos dispositivos MOS com canal p essa corrente foi minimizada devido aos efeitos da radiação, gerando uma melhora do desempenho destes dispositivos (é possível observar uma redução da inclinação de sublimiar de 60 a 80mV/dec no caso dos dispositivos mais largos). Entretanto, como o óxido de porta é mais fino que o óxido enterrado, não foram observadas grandes variações causadas pelas cargas formadas no mesmo. Quando comparada com a literatura, a dose total acumulada foi muito alta, então foi realizada uma nova rodada de radiação em que a distância entre a amostra e o feixe e a taxa de exposição foram alterados. Com isso, foi possível observar o mesmo efeito descrito anteriormente, porém com uma dose total acumulada cerca de 50 vezes menor. Em ambos os casos a exposição foi realizada sem a presença de polarização nos dispositivos, usada normalmente para reduzir a recombinação de portadores aumentando assim o efeito da radiação. Para realizar a polarização dos dispositivos durante a exposição de radiação, é preciso que os dispositivos estejam encapsulados, porém a etapa de encapsulamento poderia danificar os dispositivos utilizados. / The effects caused by the X-ray radiation in n and p channel triple gate transistors, with and without mechanical tensioning, were theoretically and experimentally studied. After X-ray irradiation on the devices, it was observed a large variation in some of their characteristics, such as an increase in the slope of 40mV/dec threshold in larger devices, due to the creation of positive charges in the oxides, both for the buried oxide as for the gate oxide. The charges generated in the buried oxide had great influence on the current conduction through the back interface (silicon / buried oxide), since the buried oxide is thicker (145nm), which in turn significantly affected many other parameters. The formation of positive charges in the buried oxide caused distinct variations in the threshold voltage on the second interface, depending on the channel type of the transistor. In the case of n-channel MOS transistors, it was observed an increase in parasitic conduction from the second interface, whereas for p-channel MOS devices the current was minimized - due to the effects of radiation - generating an improvement in performance of such devices (it is possible to observe a reduction of subthreshold slope of 60 to 80mV/dec in the case of larger devices).. However, as the gate oxide is thinner than the buried oxide, large variations caused by the generated charges were not observed. When compared to the literature, the total accumulated dose was very high, therefore, a new round of radiation were performed, in which the distance between the sample and beam, and the exposure rate were changed. Thus, it was possible to observe the same effect described above, but with a total accumulated dose about 50 times lower. In both cases, the bias was not used during the radiation exposure, normally used to reduce the recombination, thereby increasing the the radiation effect.

A protensão não-aderente caracteriza-se pela liberdade de deslocamento relativo entre o cabo de protensão e a fibra de concreto adjacente. A tensão na armadura de protensão no estado limite último é de difícil obtenção, não sendo dependente apenas das deformações em uma determinada seção transversal, mas sendo função de todas as deformações que ocorrem no concreto adjacente ao perfil de protensão. Para que seja obtida a tensão última, é necessária a integração das curvaturas ao longo de todo o elemento a fim de se obter o alongamento no cabo de protensão, o que se consegue com precisão apenas recorrendo-se a ferramentas numéricas, devido às não-linearidades físicas envolvidas no problema. O método construtivo com protensão não aderente vem sendo cada vez mais utilizado na execução de edifícios no Brasil. O principal sistema de protensão não aderente é o que utiliza a mono-cordoalha engraxada e plastificada, que alia os benefícios da protensão e a simplicidade necessária às obras moldadas no local. Apesar disso, não há no Brasil um volume de pesquisa sobre o comportamento dos elementos com protensão não aderente, compatível com a demanda da indústria da construção civil. Visando contribuir para o desenvolvimento das formulações nacionais de projeto dos elementos com protensão não aderente, o presente trabalho é a continuação de uma pesquisa que vem sendo realizada no Programa de Pós-Graduação em Engenharia Civil – PPGEC/UFRGS, a qual foi iniciada com a implementação de um modelo numérico capaz de analisar elementos com protensão não aderente e seguida de uma análise paramétrica sobre as principais variáveis que influenciam na tensão última na armadura de protensão. O trabalho que ora se apresenta traçou uma correlação entre os resultados não-aderentes obtidos da análise paramétrica com resultados aderentes, os quais são de mais fácil obtenção, pois se baseiam na compatibilidade de deformações na seção transversal. Foram feitas análises numéricas e analíticas com aderência dos mesmos protótipos estudados anteriormente sem aderência. Os resultados obtidos mostraram incrementos de tensão maiores no caso aderente, bem como maiores capacidades portantes. Estudou-se também uma metodologia capaz de computar a tensão última na armadura não aderente com análises do tipo compatibilidade de deformações, com a utilização de um coeficiente Lo/L redutor de aderência. Os resultados iniciais mostraram-se bons para carregamento nos terços, mas insatisfatórios para carregamentos distribuído e concentrado. Através de um ajuste no coeficiente Lo/L baseando-se nos dados da pesquisa precedente, bons resultados foram obtidos para todos os tipos de carregamento. A metodologia apresentada foi validada pela comparação com diferentes protótipos analisados numericamente, bem como protótipos ensaiados experimentalmente por diversos autores. / Unbonded post-tensioning is characterized for allowing relative displacement between the tendon and the concrete adjacent fiber. The ultimate stress in the unbonded tendon is difficult to be obtained, because it is not only dependent on the deformed shape of a cross section, but on the whole deformations occurring in the tendon profile adjacent concrete. To evaluate this ultimate stress, it is necessary to integrate all curvatures along the whole element, in order to obtain the total tendon elongation. This can only be precisely obtained using numerical tools, due to the non-linear factors involved in the problem. Unbonded post-tensioning is becoming ever more used in Brazilian building construction. The main unbonded post-tensioning system uses the unbonded mono-strand, that joins the benefits of prestressing with the necessary building simplicity. Nevertheless, there is no research effort in Brazil compatible with the construction industry demand. This study seeks to contribute to the development of the national formulations design of unbonded posttensioning elements. This work is a continuation of a research being undertaken at PPGEC/UFRGS. A numerical model able to determine the behavior of unbonded posttensioned elements was first developed, followed by a parametric study about the main parameters influencing the ultimate stress in unbonded tendons. The present work traced a correlation between the results for unbonded elements, obtained from the parametric analysis, with results for bonded elements. The latter are easily obtained, because they are based on the compatibility of strains in the cross section. Numerical and analytical calculations were performed for the same prototypes studied in the previous research, but now considering the existence of bond in the tendons. The results showed greater increments in stress for the bonded cases and also greater ultimate resistance. It was also devised a calculation method able to determine the ultimate stress in the unbonded tendon using an analysis similar to strain compatibility, but with a reduction bond coefficient Lo/L. The initial results obtained for a third-point loading showed a good agreement, but that not happened for the results of distributed and concentrated loads. However, by an adjustment of the Lo/L coefficient, based on the data of the previous research, good agreement was observed for all loading types. The presented methodology was validated by comparing results with other prototypes analyzed by the numerical model, and also results of experimental studies carried out by several authors.