Currently, out-of-autoclave (OOA) technology is being used to design and manufacture composite structural components at lower costs. OOA technology enables composites to be produced using only vacuum pressure, eliminating the cost of purchasing and operating an autoclave. The key to OOA prepreg is that they are specially designed to remove air that is entrapped during the lay-up process. The in-plane and through thickness permeability of the prepreg were characterized to determine which bagging configuration would produce the best honeycomb sandwich structure. The bagging configuration that produced the lowest skin porosity was determined to be one ply of non-perforated release film with edge breathing around the perimeter of the panel. The resin content of the prepreg is such that any resin loss from the skin will create dry spots, pinholes, and porosity. The edge breathing allows the air inside the prepreg and core to be removed by the vacuum, and t he non-perforated release film presents resin starvation and subsequently reduces porosity. Caution should be used when debulking out-of-autoclave sandwich panels, since removing all the air from the core reduces the amount of skin compaction available during cure. The skin compaction is needed to suppress void growth, and the optimal internal core pressure was experimentally determined to be between 35-55 kPa. The resin was fully characterized such that when the internal core pressure has reached the optimal range, the resin can flow and close off the air passages. Five representative panels were manufactured using two low temperature cure film adhesives. The optimal curing temperature to minimize porosity of the composite skin was determined to be 100°C. Climbing drum peel tests were used to evaluate the mechanical performance of the panels. / Présentement, la technologie hors autoclave (OOA) est utilisée pour la conception et la fabrication à moindre coût des composants structuraux en composite. La technologie OOA permet de produire des composites en utilisant uniquement la pression générés lors des procédés d'ensachage sous vide, en éliminant par le fait même les coûts relatifs à l'achat et à l'exploitation d'un autoclave. L'avantage des matériaux pré-imprégnés destinés à des cuissons hors autoclave est qu'ils sont spécialement conçus pour éliminer les bulles d'air emprisonnées au cours de la préparation. Afin de déterminer la configuration d'ensachage optimale pour la production de structure en sandwich avec noyau en nid d'abeille, l'épaisseur du matériel pré-imprégné ainsi que la perméabilité à travers le plan ont été caractérisées. Les conditions d'ensachage ayant menées à la plus faible porosité correspondent à l'utilisation combinée d'une couche non perforée de pellicule antiadhésive et d'une bordure permettant l'extraction des produits gazeux. La pellicule antiadhésive non perforée permet de prévenir les pertes de résine, tandis que la bordure permettant l'extraction des produits gazeux permet d'évacuer l'air contenu dans le noyau et les couches pré-imprégnés. L'utilisation d'une pellicule non perforée s'est avérée nécessaire puisque la teneur en résine du matériel pré-imprégné est telle que toute perte de résine crée des régions dites sèches (non imprégnée), produit des trous ou défauts à la surface du laminé et augmente la porosité. Il convient d'être prudent lors du dégazage des panneaux en sandwich hors autoclave puisque la suppression de tout l'air contenue dans le noyau réduit le niveau de compaction disponible durant la cuisson. L'application d'une force de compaction adéquate doit être effectuée lors de la cuisson afin de réduire la formation de vide

An investigation of resistance welding of thermoplastic composite double-lap shear joints is presented. Double-lap shear specimens consisting of carbon fibre/poly-etherether-ketone (PEEK/CF), carbon fibre/poly-etherketone-ketone (PEKK/CF), carbon fibre/poly-ether-imide (PEI/CF) and glassfibre/poly-ether-imide (PEI/GF) were resistance-welded using a stainless steel mesh heating element. The objective of this work was to study the mechanical performances of the double lap shear resistance-welded joints and to compare them with the single lap shear resistance-welded joints. The welded specimens were analyzed using static and dynamic lap shear tests and optical and scanning electron microscopy. Lap shear strengths of 53 MPa, 49 MPa, 45 MPa and an extrapolated value of 34 MPa were obtained for PEEK/CF, PEKK/CF, PEI/CF and PEI/GF double-lap joints, respectively. Infinite fatigue lives were obtained at 30% for PEEK/CF and PEKK/CF, 25% for PEI/CF and 20% for PEI/GF. Resistance-welded double-lap joints were found to have equivalent static and fatigue mechanical properties compared with single-lap joints, for all materials tested. / Ce travail présente une étude des joints à recouvrement double de matériaux composites à matrice thermoplastique assemblés par soudage par résistance. Des échantillons de joints à recouvrement double constitués de fibre de carbone/polyéther éther cétone (PEEK/CF), fibre de carbone/polyéther cétone cétone (PEKK/CF), fibre de carbone/polyéther imide (PEI/CF) et fibre de verre/polyéther imide (PEI/CF), ont été assemblés pas soudage par résistance à l'aide d'un élément chauffant en acier inoxydable. L'objectif de ce travail est d'étudier les performances mécaniques des joints à recouvrement double soudés pas résistance et de les comparer avec celles des joints à recouvrement simple. Les échantillons soudés ont été analysés à l'aide de tests de chargement statiques et dynamiques, ainsi que pas microscopie optique et par microscopie électronique à balayage. Une résistance au cisaillement de 53 MPa, 49 MPa, 45M Pa, et une valeur extrapolée de 34 MPa ont été obtenues pour les joints à recouvrement double de PEEK/CF, PEKK/CF, PEI/CF et PEI/GF, respectivement. Une durée de vie indéterminée en fatigue de 30% du chargement statique pour les joints de PEEK/CF et PEKK/CF, de 25% pour les joints de PEI/CF, et de 20% pour les joints de PEI/GF ont été obtenues. Pour chacun des matériaux testés, les joints à recouvrement double soudés par résistance ont donné lieu à des propriétés mécaniques en statique et en fatigue équivalentes à celles des joints à recouvrement simple.

During underground, long-hole drilling operations, the bending of the drill string and vibrations exert significant effects on hole deviation and penetration rates. Little research has been conducted in the area of the bending and vibrations associated with ITH (In - The - Hole) drills. This thesis deals with the modelling of the static and dynamic properties of the ITH drill system. / The bending of the drill string is the result of the string becoming unstable. The critical length and feed force of the ITH drill string before being unstabilized and its characteristics after unstabilization are studied. By analyzing the static and dynamic properties of the drill string, a continuous system mathematical model has been developed. Two different boundary conditions have been discussed and applied through computer simulations. / Measurements of downhole parameters and vibrations while drilling were acquired by a high-rate telemetry and data acquisition system from field tests of a CMS CD-90B long-hole ITH drill and prototype shock absorber in an underground nickel mine. Determining the inter-relationships between the drilling parameters and the vibrations is also a principal objective of the thesis. The performance of a prototype shock absorber developed by a third party company was also evaluated. The results of the simulations and analyses of the vibration data indicate that a prototype shock absorber effectively reduces the amplitude of the vibrations. A mathematical model using a complex boundary condition has been developed which will be of use in further research relating to the development of a vibration-based control system for ITH drills. By decreasing the vibration levels acting on the drill head and rod, the instrumentation for an automated drill can be protected to ensure maximum life and proper performance during hammer operation. (Abstract shortened by UMI.)

This thesis presents a theoretical study of the stability of pinned-clamped and clamped-pinned coaxial cylindrical shells conveying internal and/or annular incompressible viscous fluid flow. / In the present analytical model, fluid viscous effects are taken into consideration. Generally, the viscous nature of the fluid results in both steady and unsteady viscosity-related loads being exerted on the shells, the latter of which are approximated by their inviscid counterpart in this thesis. Upstream pressurization of the flow (to overcome frictional pressure drop) and skin friction on the shell surfaces are taken into account, generating time-averaged normal and tangential loads on the shells. In this model, the shell motions are described by Flugge's shell equations, suitably modified to incorporate the time-averaged stress resultants arising from viscous effects. The unsteady fluid-dynamic forces in these equations are formulated from potential flow theory: the perturbation pressures on the shells are determined from the perturbation velocity potentials via the unsteady Bernoulli equation; those velocity potentials are governed by the Laplace equation, which is solved by the Fourier transform technique. / For the clamped-pinned system, since the downstream end of the shell is simply supported, a so-called out-flow model is utilized in modelling the decay of flow perturbations beyond the pinned end. / Comparison is made with the existing results for clamped-clamped and clamped-free cases. / Finally, future work is suggested with regard to setting up a new analytical model with the unsteady viscous effects taken into account.

This thesis proposes a quantitative model of fatigue crack initiation and early propagation based on a plastic strain intensity factor $ Delta{K} sb{p}$. In this model, the crack growth rate in the initiation region is simplified as a straight line on the log-log scale of da/dN verses $ Delta{K} sb{p}$ curve. A turning point which divides crack growth into the initiation and propagation regions is introduced as the plastic crack propagation threshold, $ Delta{K} sb{pth}$, whose value is determined by the plastic crack propagation threshold in the $ Delta{J}$-integral form, $ sqrt{ Delta{J} cdot{E}} sb{pth}$. The crack propagation style in the initiation region up to the plastic crack propagation threshold is Stage I propagation. This thesis also proposes a Stage I crack propagation model for cyclindrical specimens which consists of a model of sub-microcrack propagation prior to the sub-threshold and a model of the subsequent microcrack propagation. Crack profile changes and closure effect in the initiation region are also qualitatively proposed. The results of an experimental study carried out on smooth cylindrical specimens of polycrystalline OFHC copper at room temperature are compared with the proposed models. The coated and shadowed replicas viewed by SEM traced a fatal or main surface crack in a specimen back to a size of less than one micron. A combination of the replication technique and post-mortem specimen surface and fractographic surface examination by SEM provided good information of crack initiation and Stage I propagation. The results support the proposed models. The integration of the crack growth rate as a function of the plastic strain intensity factor yields a fatigue crack initiation life estimate whose error is less that 5%. For comparison, the experimental data presented in this paper is also discussed on the basis of $ Delta{K}$ and $ Delta{K} sb{eff}$ parameters proposed by other researchers.

Experimental results of scalar dispersion from a concentrated source in an inhomogeneous turbulent flow field are presented in this thesis. / The flow field is fully developed channel flow---one of the simplest inhomogeneous turbulent flows. A fine line source is used to inject the scalar (temperature). The temperature injection does not affect the flow field, hence it is passive. The line source is in the spanwise (z) direction of the channel, and is located at transverse (i.e., wall-normal) locations: y/h = 0.067, 0.17, and 1.0 (where h is the channel half-width). Measurements of the resultant thermal plume for each source location are taken for two Reynolds numbers: 10400 and 22800 (where Re = Uch/nu: U c is the centerline velocity and nu is the kinematic viscosity). Hot-wire anemometry and cold-wire thermometry are used to acquire velocity and temperature data, respectively. / It is shown that the downstream decay of the mean temperature rise is less than that of isotropic grid-generated turbulence and homogenous turbulent shear flow. The peak RMS temperature fluctuation also decays at a slower rate. For the near-wall source locations, the peak of the transverse RMS temperature profile drifts toward the channel centerline with increasing downstream distance from the source. Also, for the near-wall source locations, the scalar PDF is quasi-Gaussian, indicating improved mixing in that region. As the plume is traversed toward the centerline, the PDF evolves into a very positively skewed shape. / An extensive database of the thermal plume is compiled, which consists of the mean, RMS, and skewness profiles of the temperature field, PDFs of the temperature field, and temperature-velocity correlations. This database can be used to test numerical and/or theoretical models, highlighting their strengths and weaknesses.

An experimental study of spatially periodic fully developed flows of air in straight rectangular ducts with interrupted-plate inserts is presented. These flows have features similar to those produced in plate-fin passages of compact heat exchanger cores. The formulation of mathematical models and numerical solution methods for the prediction of these flows continues to be a challenging and largely unattained goal. The main goal of this work is to contribute experimental results suitable for testing and refining such models and solution methods. Airflows in nine different interrupted-plate rectangular ducts were considered. Reynolds numbers considered (based on Kays and London's definition of hydraulic diameter and the maximum average streamwise velocity) ranged from about 2,000 to 65,000. Time-mean static pressure distributions along the axial centreline of the top wall of the ducts, and module friction factor versus Reynolds number data, all in the spatially periodic fully developed region, were obtained.

This thesis develops the concept of qualitative model-based fault diagnosis technique and applies it to extrusion blow molding. The extrusion blow molding process is reviewed along with theories of qualitative reasoning, from which process-based qualitative models are developed. Based on the models, a fault diagnosis system is developed consisting of two parts: a fault reasoning system based on a simulation of the EBM process and a process monitor to measure process variables and to adapt the diagnosis to changing environments.

Computer simulations play an important role in the design and verification of space robotic operations since on-orbit tests are impossible to conduct before launch. Thus, accurate computer modelling and simulation of space robotic tasks is essential. Of particular difficulty are space manipulator operations, which involve constrained or contact tasks. Here, the contact dynamics capability in the modelling tools becomes critical for high fidelity simulation. This in turn implies a need for accurate determination of contact parameters, which are used as inputs to contact dynamics simulation. In this work, the identification of contact dynamics parameters based on sensor data obtained during robotic contact tasks is considered. / The contact parameter estimation problem is addressed for simple and complex contacting geometries using the SPDM Task Verification Facility Manipulator Test-bed (SMT) at the Canadian Space Agency. The SMT is a space-representative robotic simulation facility. Single- and multiple-point contact parameter estimation software toolboxes were developed and used with SMT experiments. Single point SMT contact experiments were performed with six different payloads. The single point toolbox was used as part of the process of identifying payload stiffness from SMT experimental data. / Multiple point contact parameter estimation experiments with the SMT were conducted using a mock-up of an International Space Station Arm Computer Unit (ACU) as payload. The multiple point toolbox was used to generate contact stiffness, damping and friction estimates. An evaluation of the sensitivity of the parameter estimation algorithm to mismatches in ACU physical dimensions and ACU geometry files was conducted.

The complex musculoskeletal system of a running animal on horizontal surfaces act essentially like a simple pogo stick, and can be modeled as a hopping spring-mass model known as the Spring-loaded Inverted Pendulum (SLIP) model. The SLIP model has been extensively used as a reduced-order model in analysis and control of running legged robots. By contrast, the SLIP model itself has never been implemented in a robot and validated experimentally. This thesis addresses the development and validation of a robotic SLIP, a planar one-legged hopping robot with only one actuator. A feasibility study was performed using numerical simulation. The experimental platform was designed and built based on SLIP-model features. A hopping controller that conceptually reproduced the self-stability property of the SLIP model was implemented. Running was achieved at 6.7 leg lengths per second, which is, to date, the fastest dimension-less speed for a single-legged robot. Simulation and experimental data demonstrated periodic and robust stability. The SLIP model was qualitatively validated for a particular gait in simulation and experimentation.