Dissertations / Theses on the topic 'Engineering mechanics'

A common technique to improve the performance of shell and tube heat exchangers (STHE) is by redirecting the flow in the shell side with a series of baffles. A key aspect in this technique is to understand the interaction of the fluid dynamics and heat transfer. Computational fluid dynamics simulations and experiments were performed to analysis the 3-dimensional flow and heat transfer on the shell side of an STHE with and without baffles. Although, it was found that there was a small difference in the average exit temperature between the two cases, the heat transfer coefficient was locally enhanced in the baffled case due to flow structures. The flow in the unbaffled case was highly streamed, while for the baffled case the flow was a highly complex flow with vortex structures formed by the tip of the baffles, the tubes, and the interaction of flow with the shell wall.

Interest in thermoelectric devices (TEDs) for waste-heat recovery applications has recently increased due to a growing global environmental consciousness and the potential economic benefits of increasing cycle efficiency. Unlike conventional waste-heat recovery systems like the organic Rankine cycle, TEDs are steady-state, scalable apparatus that directly convert a temperature difference into electricity using the Seebeck effect. The benefits of TEDS, namely steady-state operation and scalability, are often outweighed by their low performance in terms of thermal conversion efficiency and power output. To address the issue of poor device performance, this dissertation takes a multi-faceted approach focusing on device modeling, analysis and design and material processing.

First, a complete one-dimensional thermal resistance network is developed to analytically model a TED, including heat exchangers, support structures and thermal and electrical contact resistances. The purpose of analytical modeling is twofold: to introduce an optimization algorithm of the thermoelectric material geometry based upon the realized temperature difference to maximize thermal conversion efficiency and power output; and to identify areas within the conventional TED that can be restructured to allow for a greater temperature difference across the junction and hence increased performance. Additionally, this model incorporates a component on the numerical resolution of radiation view factors within a TED cavity to properly model radiation heat transfer. Results indicate that geometric optimization increases performance upwards of 30% and the hot-side ceramic diminishes realized temperature difference. The resulting analytical model is validated with published numerical and comparable analytical models, and serves as a basis for experimental studies.

Second, an integrated thermoelectric device is presented. The integrated TED is a restructured TED that eliminates the hot-side ceramic and directly incorporates the hot-side heat exchanger into the hot-side interconnector, reducing the thermal resistance between source and hot-side junction. A single-state and multi-stage pin-fin integrated TED are developed and tested experimentally, and the performance characteristics are shown for a wide range of operating fluid temperatures and flow rates. Due to the eliminated to thermal restriction, the integrated TED shows unique performance characteristics in comparison to conventional TED, indicating increased performance.

Finally, a grain-boundary engineering approach to material processing of bulk bismuth telluride (Bi₂Te₃) is presented. Using uniaxial compaction and sintering techniques, the preferred crystallographic orientation (PCO) and coherency of grains, respectively, are controlled. The effect of sintering temperature on thermoelectric properties, specifically Seebeck coefficient, thermal conductivity and electrical resistivity, are determined for samples which exhibited the highest PCO. It is shown the performance of bulk Bi₂Te₃ produced by the presented method is comparable to that of nano-structured materials, with a maximum figure of merit of 0.40 attained at 383 K.

This dissertation is primarily concerned with the development of an effective methodology for reducing structure-borne sound radiation from an arbitrarily shaped vibrating structure. There are three major aspects that separate the present methodology from all the previous ones. Firstly, it is a non-contact and non-invasive approach, which is applicable to a class of vibrating structures encountered in engineering applications. Secondly, the input data consists of a combined normal surface velocity distribution on a portion of a vibrating surface and the radiated acoustic pressure at a few field points. The normal surface velocities are measured by using a laser vibrometer over a portion of the structural surface accessible to a laser beam, while the field acoustic pressures are measured by a small array of microphones. The normal surface velocities over the rest surface of the vibrating structure are reconstructed by using the Helmholtz Equation Least Squares (HELS) method. Finally, the acoustic pressures are correlated to structural vibration by decomposing the normal surface velocity into the forced-vibro-acoustic components (F-VAC). These F-VACs are mutually orthogonal basis functions that can uniquely describe the normal surface velocity. The weightings of these F-VACs represent the relative contributions of structural vibrations into the sound radiation. This makes it possible to suppress structure-borne acoustic radiation in the most cost-effective manner simply by controlling the key F-VACs of a vibrating structure. The effectiveness of the proposed methodology for reducing structure-borne acoustic radiation is examined numerically and experimentally, and compared with those via traditional experimental modal analyses. Results have demonstrated that the proposed methodology enables one to reduce much more acoustic radiation at any selected target frequencies than the traditional approach.

Methods for modeling structural failure with applications for fluid structure interaction (FSI) are developed in this work. Fracture as structural failure is modeled in this work by both the extended finite element method (XFEM) and element deletion. Both of these methods are used in simulations coupled with fluids modeled by computational fluid dynamics (CFD). The methods presented here allow the fluid to pass through the fractured areas of the structure without any prior knowledge of where fracture will occur. Fracture modeled by XFEM is compared to an experimental result as well as a test problem for two phase coupling. The element deletion results are compared with an XFEM test problem, showing the differences and similarities between the two methods.

A new method for modeling fracture is also proposed in this work. The new method combines XFEM and element deletion to provide a robust implementation of fracture modeling. This method integrates well into legacy codes that currently have element deletion functionality. The implementation allows for application by a wide variety of users that are familiar with element deletion in current analysis tools. The combined method can also be used in conjunction with the work done on fracture coupled with fluids, discussed in this work.

Structural failure via buckling is also examined in an FSI framework. A new algorithm is produced to allow for structural subcycling during the collapse of a pipe subjected to a hydrostatic load. The responses of both the structure and the fluid are compared to a non-subcycling case to determine the accuracy of the new algorithm.

Overall this work looks at multiple forms of structural failure induced by fluids modeled by CFD. The work extends what is currently possible in FSI simulations.

According to Wohlers Report 2014, the worldwide 3D printing industry is now expected to grow from $3.07B in revenue in 2013 to $12.8B by 2018, and exceed $21B in worldwide revenue by 2020. With 3D printing rapidly evolving from a prototype commodity to a means to produce full production items, lattice structures are becoming of great interest due to their superior structural characteristics and lightweight nature. Within design, lattice structures have typically been defined by preset beam configurations within a cube. Certain configurations have been proven analytically to be optimal for certain load functions, but never has there been optimization performed to discover or verify the optimal lattice shapes and sizes within a predefined cubic space. By performing optimization on these cubic cells, a design guideline can be created for designers of lattice structures. In this thesis, several lattice configurations are analyzed both from a micro level (single unit cell) as well as a macro level (a simple series of unit cells). Optimization is performed with respect to stiffness and compliance to identify strategic configurations for bending, torsion, compression and tension. Only cubic base cells are analyzed (i.e. no hexagonal). Knowing optimal lattice configurations from a structural standpoint enables designers to further reduce weight and increase structural efficiencies when designing for additive manufacturing. The results of this study yield a well-defined guideline for design engineers to utilize when lattice structures are incorporated in a structural design. With this design guideline information available to design engineers, further utilization of lattice structures can be exploited by efficiently applying strategic unit cell configurations to the overall design.

Image-based modeling and simulation has become an important analytic and predictive tool for patient-specific medical applications, including large-scale in silico patient studies, optimized medical device design, and custom surgical guides and implants via additive manufacturing. The pipeline for patient-specific modeling and simulation is: image acquisition, image segmentation, surface generation, mesh generation, physics-based modeling and simulation, and clinical application. This research establishes a semi-automatic workflow for these steps, which includes a novel image-based meshing tool Shabaka. The toolchain is demonstrated by modeling the mechanics of a beating human heart based on magnetic resonance imaging (MRI) data.

The Shabaka workflow ensures robust execution of each step of the pipeline. Medical images are processed and segmented using thresholding, region-growing, and manual techniques. Watertight surface meshes are extracted from image masks using a novel Voronoi-based algorithm. For scientific computing purposes, surface meshes are supplied either to tetrahedral meshing routines for conventional finite element approaches, or to a robust polyhedral mesh generation tool for a novel polyhedral finite element approach. A polyhedral finite element code is explored, that retains most of the favorable properties of conventional finite element methods, while reducing the system size by up to an order of magnitude compared to conventional techniques for the same input surface.

In conjunction with a cardiac simulation code, the workflow is utilized to model finite-deformation cardiac mechanics. A quadratic tetrahedral mesh is generated from MRI data of the human heart ventricles. The constitutive law is comprised of an incompressible orthotropic hyperelastic stress response for the myocardium, plus an electrical activation-dependent active stress for the muscle fibers. Muscle fiber orientations are generated using a rule-based approach. Electrical activation times are read from a coupled electrophysiology code. A lumped circulatory model is used to impose time-dependent ventricular volume constraints. Simulation results are presented. The same mechanics are also implemented for the polyhedral finite element mesh, and preliminary verification results are presented.

The toolchain used in performing image-based cardiac mechanics simulations makes important improvements to the speed and robustness of image-based modeling techniques. As efforts continue to mature, so too does the promise for simulation to impact and improve healthcare.

It is proposed that both the adhesive interface geometry and the mismatch of elastic moduli influence the tensile strengths of dental bonds attaching restorative ceramics to dentin. Prior calculations indicate this to be due to peripheral interface stress singularities. A physical testing approach to examine and validate the theoretical conclusions utilizing a microtensile testing system is presented.

Considering the choice of shear versus tensile and then macro versus micro tensile testing, reasons were identified for choosing micro tensile testing. Specimen dimensions and shapes were developed to optimize the adjustment of the interface geometries and information that could be obtained therein. Here a dumbbell structure is best suited to the testing needs.

Dumbbell specimens were first fabricated using the ceramic press technique, and then mini-CNC machining. Specimens fabricated by each technique were examined, showing that the mini-CNC machining methodology was superior.

Significant problems in instrumentation were overcome by the design and fabrication of two testing fixtures: 1. A collet based design with independent upper and lower mechanical grips for each end of the dumbbell, to be used in conjunction with a loading device; 2. A screw based clamping design similar to previous jigs, using two screw clamps on V-channels connected by sliding rods. Testing revealed that the collet-based design shows the most promise because of its distributed gripping load. Further tests that can evaluate the effectiveness of this device for microtensile testing are outlined.