The research presented herein is an effort to support computational modeling of ultra-high molecular weight polyethylene (UHMWPE) composites. An effort is made to characterize the composites and their constituents. UHMWPE, as a polymer, is time and temperature dependent. Using time-temperature superposition (tTSP), the constituent properties are studied as a function of strain rate. Properties that are believed to be significant are fiber tensile properties as a function of strain rate, as well as the through-thickness shear behavior of composite laminates. Obtaining fiber properties proved to be a challenge. The high strength and low surface energy of the fibers makes gripping specimens difficult. Several different methods of fixturing and gripping are investigated, eventually leading to a combination of friction and adhesion approaches where a fiber was wrapped on an adhesive coated cardboard mandrel and then gripped in the test fixture. Fiber strength is estimated using tTSP to equivalent strain rates approaching 10^6 sec^-1. Punch-shear testing of UHMWPE laminates is conducted at quasi-static strain rates and the dependence of the results on thickness and test geometry is investigated. / Master of Science

Les systèmes de communication par satellite d’aujourd’hui reposent principalement sur le multiplexage temporel pour optimiser leurs performances. Chaque utilisateur utilise le canal pendant une fraction de temps connu. Pendant cette période, la modulation et le taux de codage sont choisis de manière à transmettre le plus d’information possible. En pratique, ce schéma est facile à mettre en œuvre ce qui justifie sa popularité. Cependant, il est désormais bien connu que la répartition temporelle n’est pas optimale en termes d’efficacité spectrale offerte aux récepteurs. En effet, la stratégie qui consiste à superposer des données offre de meilleures performances que le multiplexage temporel. C’est dans ce contexte que s’inscrit la problématique de cette thèse. Le travail réalisé propose des applications du codage par superposition dans le domaine des communications par satellite. Tout d’abord, nous étudions la modulation hiérarchique qui est une implémentation du codage par superposition au niveau de la modulation. Les performances de ce type de modulation sont évaluées d’un point de vue théorique et pratique. Dans un deuxième temps, nous quantifions l’amélioration en termes d’efficacité spectrale que peut apporter la modulation hiérarchique pour les systèmes de communication par satellite. Les standards de diffusion par satellite DVB-SH et DVB-S2 fournissent un cadre pratique. Nous montrons que des gains non négligeables sont envisageables selon la configuration du système. Le dernier point abordé concerne un système où des utilisateurs communiquent entre eux à l’aide d’un satellite qui sert de relais. Nous proposons un schéma de communication où plusieurs utilisateurs émettent en même temps en coordonnant leur puissance de transmission. Ainsi, les signaux vont naturellement se superposer. Les récepteurs utilisent deux mécanismes pour le décodage des signaux : le codage réseau couche physique et la démodulation de constellations superposées. Finalement, les gains de performance obtenus dans les différents domaines par le codage par superposition ouvrent des perspectives pour des travaux futurs. / Modern satellite communication systems mainly rely on time sharing to optimize the throughput. Each receiver uses the channel during a given fraction of time. During this period, the transmission parameters (i.e., the modulation and the coding rate) are chosen in order totransmit as much information as possible. The scheme is easy to implement which explains its popularity. However, it is today well established that time sharing is not optimal in terms of spectrum efficiency offered to the receivers. Indeed, the scheme that consists in sending superposed data offers better performance than the time sharing. This thesis investigates the application of superposition coding in satellite communication systems. First of all, we study the performance of hierarchical modulation which is an implementation of superposition coding at the modulation level. We propose a performance evaluation method for such modulations. We also compare the performance of hierarchical and non hierarchical modulations in terms of spectrum efficiency and link unavailability. These two criteria are very important for broadcast system and we show that hierarchical modulations often offer better performance than non hierarchical modulations.Then, we study the performance improvement in terms of spectrum efficiency when using hierarchical modulation in satellite communication systems. Two issues are addressed. The first one is how to group the receivers in pairs in order to transmit data with a hierarchical modulation. The second issue is the computation of the spectrum efficiency. We show that significant gains are possible depending on the system configuration. The last part considers a system where multiple users communicate through a satellite. The satellite acts as a relay in our scenario. We propose a communication scheme where several users emit at the same time with appropriate transmitting power. Thus the signals naturally superpose and generate interference. The receivers use two mechanisms for decoding the signals: physical layer network coding and demodulation of superposed constellations. Finally, we explain how the performance improvements obtained by superposition coding in several scenarios open perspectives for future work.

Nous nous intéressons à des formules de la logique du premier ordre où certaines constantes sont interprétées dans un domaine défini inductivement, comme les entiers. Le problème de la validité n'est pas semi-décidable pour ces formules. Le but de cette thèse est donc d'accroître les capacités des procédures de preuve les plus efficaces pour la logique du premier ordre (fondées sur le calcul de résolution et de superposition) afin de tenir compte de ces constantes particulières. Pour cela, nous adaptons le calcul de superposition en ajoutant notamment un mécanisme de détection de cycles qui simule une forme d'induction mathématique. Nous étudions dans un premier temps le cas particulier des entiers, puis nous généralisons certains des résultats obtenus au cas où les constantes inductives sont définies à l'aide de constructeurs monadiques (des mots). Nous présentons des classes syntaxiques pour lesquelles nous pouvons assurer la complétude et/ou la décidabilité. Nous décrivons un outil appelé SuperInd, fondé sur le démonstrateur Prover9, implémentant les résultats précédents. Enfin, nous décrivons certaines expérimentations et procédons à des comparaisons avec d'autres approches. / We consider first order formulas where some constant symbols are defined in an inductive domain. The validity problem is not semi-decidable for these formulas. This work aims to increase the capabilities of the usual first order proof procedures (usually based on superposition and resolution calculus) to handle these particular constant symbols. Thus, we adapt the superposition calculus using a loop detection mechanism encoding a form of mathematical induction. We first consider the particular case of natural numbers, then we generalize some of these results to the case where the inductive constant symbols are defined with monadic constructors (words). We present some syntactic classes for which we can ensure completeness and/or decidability. We describe a new tool named SuperInd, based on the theorem prover Prover9, implementing our previous results. Finally we describe some experimentations and some comparisons with other approaches.

Resilin is an almost perfect elastic protein found in many insects. It can be stretched up to 300% of its resting length and is not affected by creep or stress relaxation. While much is known about the static mechanical properties of resilin, it is most often used dynamically by insects. Unfortunately, the dynamic mechanical properties of resilin over the biologically relevant frequency range are unknown. Here, nearly pure samples of resilin were obtained from the dragonfly, Libellua luctuosa, and dynamic mechanical analysis was performed with a combination of time-temperature and time-concentration superposition to push resilin through its glass transition. The tensile properties for resilin were found over five different ethanol concentrations (65, 70, 82, 86 and 90% by volume in water) between temperatures of -5°C and 60°C, allowing for the quantification of resilin's dynamic mechanical properties over the entire master curve. The glass transition frequency of resilin in water at 22°C was found to be 106.3 Hz. The rubber storage modulus was 1.6 MPa, increasing to 30 MPa in the glassy state. At 50 Hz and 35% strain over 98% of the elastic strain energy can returned each cycle, decreasing to 81% at the highest frequencies used by insects (13 kHz). However, despite its remarkable ability to store and return energy, the resilin tendon in dragonflies does not act to improve the energetic efficiency of flight or as a power amplifying spring. Rather, it likely functions to passively control and stabilize the trailing edge of each wing during flight. / Master of Science

Resilin is a cuticular protein found in a variety of insects. It can stretch up to 300% of its natural length without any creep or relaxation. Further, it operates across a wide frequency range from 5 Hz in locomotion to 13 kHz in sound production. Both the protein sequence and composition of natural resilin as well as the dynamic mechanical properties vary substantially across species. This suggests that mechanical properties may be evolutionarily tuned for specific functions within an insect. Here, samples of resilin obtained from the tibia-tarsal joint of the cockroach, Periplaneta americana, were tested using a custom built dynamic mechanical analyzer. The material properties in compression are obtained from the rubbery to glassy domain with time-temperature superposition (-2C to 55C) and time-concentration superposition (0 % to 93% ethanol by volume in water). At low frequency the storage modulus was found to be 1.5 MPa increasing to about 5 MPa in the transition zone. The glass transition frequency at 23C in complete hydration was found to be 200 kHz. The data shows that cockroach resilin is less resilient than dragonfly resilin at low frequencies, returning about 79% of the elastic strain energy at 25 Hz compared to 97% for dragonfly resilin. However, at the glass transition (200 kHz) the material returns about 47% of the elastic strain energy compared to 30% in dragonfly (2MHz ). The resilin pad in cockroach is a composite structure, acting as a compressive spring to passively extend the tibia-tarsal joint during cockroach locomotion. Its mechanical properties are more similar to the composite locust pre-alar arm than to the pure resilin dragonfly tendon, suggesting that macroscopic structural influences may be as important as molecular sequence differences in setting properties. / Master of Science

The last three decades have seen significant advances in the wireless communication field. As the data rates of wireless systems are increasing, the demand for mobile services also is increasing rapidly. Like other natural resources, radio spectrum suitable for mobile wireless communications is also limited. In order to keep up with this ever increasing demand, there is a requirement of new signal processing algorithms. Diversity is a technique used in wireless systems to combat the effects of fading and thereby improve reliability of data transfer. There are many ways in which algorithms can exploit diversity in wireless channels. Hybrid-automatic repeat request (H-ARQ) schemes and relaying mechanisms are two such diversity extracting techniques. Even though these diversity achieving techniques have been well understood in theory, there are many ways in which one can optimize these techniques for specific application scenarios.In this thesis, we focus on improving the performance of retransmission schemes and relaying systems. In the first part of the thesis, we improve the performance of H-ARQ schemes in the 3GPP- long term evolution (LTE) system by improving the performance of feedback signaling. We employ complex-field coding to extract the inherent frequency diversity available in the resources. Next, we provide a sub-optimal solution to the outage-optimal power allocation problem in incremental redundancy based H-ARQ system, whose performance is practically the same as that of the optimal solution. In the later part of the thesis, we propose a retransmission scheme based on superposition coding (SPC) for the symmetric relaying scenario. We provide packet error probability (PEP) expressions and solutions for the optimal fraction of power allocated for the partners' data. Finally, we study the optimal bits-to-symbol mappings for SPC and its effect on an H-ARQ scheme and the symmetric relaying scenario using SPC.

none

This dissertation presents several new techniques for the analysis of the long-term production performance of tight gas wells. The main objectives of this work are to determine pressure response function for long-term production for a the slightly compressible liquid case, to determine the original gas in place (OGIP) during pseudosteady state (PSS), to determine OGIP in the transient period, and to determine the effects of these parameters on linear flow in gas wells. Several methods are available in the industry to analyze the production performance of gas wells. One common method is superposition time. This method has the advantage of being able to analyze variable-rate and variable-pressure data, which is usually the nature of field data. However, this method has its shortcomings. In this work, simulation and field cases illustrate the shortcomings of superposition. I present a new normalized pseudotime plotting function for use in the superposition method to smooth field data and more accurately calculate OGIP. The use of this normalized pseudotime is particularly important in the analysis of highly depleted reservoirs with large change in total compressibility where the superposition errors are largest. The new tangent method presented here can calculate the OGIP with current reservoir properties for both constant rate and bottomhole flowing pressure (pwf) production. In this approach pressure-dependent permeability data can be integrated into a modified real gas pseudopressure,m(p), which linearizes the reservoir flow equations and provides correct values for permeability and skin factor. But if the customary real-gas pseudopressure, m(p) is used instead, erroneous values for permeability and skin factor will be calculated. This method uses an exponential equation form for permeability vs. pressure drop. Simulation and field examples confirm that the new correction factor for the rate dependent problem improves the linear model for both PSS and transient period, whether plotted on square-root of time or superposition plots.

Techn. University, Diss., 2003--Braunschweig.

With tougher emission standards for heavy duty trucks, good control of the engine is of importance. By taking into consideration the torsional vibrations occurring in the crankshaft, the engine control can be improved. This could be done by implementing a torsion model that would give feedback to the engine control unit to reduce the cycle-to-cycle variations in the cylinders, which is partly due to the torsion in the crankshaft. It is therefore of interest to determine if a torsion model can be developed with a reduced computational complexity and still have sufficient accuracy. In this thesis a model was developed in Matlab for estimation of torsion in the crankshaft of an inline 6 cylinder diesel engine. By applying Newtons second law, the equations of motion that describe the torsional vibrations in the crankshaft were set up. The equations of motion were transformed using modal analysis and with the use of mode superposition it was investigated how reducing the number of vibrational modes in the model, affected the models accuracy in its estimation of torsion. Two model reductions were evaluated where the first three and four vibrational modes were used to calculate the torsional displacement between the flywheel and the cylinders. Using measured pressure curves for a number of operating points of the engineas input to the model, results were produced for different crank angle intervalsshowing deviations between the developed torsion models and a reference models estimation of torsion. Due to the damping approximation used in the model, high initial deviations could be found at certain operating points beforereducing the number of vibrational modes. These initial deviations weregreatest for the first two cylinders. Results from the model reductions showed that using the first three vibrational modes in the torsion model, is sufficient for estimating the torsional displacement between the flywheel and all cylinderswith an accuracy of within 0.1 degrees, with the exception of the first two cylinders for the early and late combustion interval.