The work described herein emphasizes a new fiber-based approach to modeling the response of structural frames subjected to realistic fire conditions. The proposed approach involves the development and validation of two finite elements that can be used collectively to simulate the thermal and mechanical response of structural frames at elevated temperatures. To model the thermal response, a special-purpose fiber heat transfer element is introduced. The first of its kind, the fiber heat transfer element uses a combination of finite element and finite difference methods to provide an accurate and highly efficient solution to the three-dimensional thermal problem. To simulate the mechanical response, a flexibility-based fiber beam-column element is used. The element presented here extends the formulation of Taucer et al. (1991) to include thermal effects, geometric nonlinearities, and residual stresses. Both fiber elements are implemented in ABAQUS (2007) using the user-defined element (UEL) subroutine. The element formulations are verified by analyses of benchmark experimental tests and comparisons with traditional finite elements. Results indicate that both elements offer superior accuracy and computational efficiency when compared to traditional methods of analysis. Analyses of structures subjected to non-uniform heating emphasize the advantages of the fiber-based approach. To demonstrate a realistic application of the proposed approach, the work concludes with an investigation of the response of unprotected steel beams subjected to localized fires. Because realistic fires are considered, the treatment of strain reversal upon cooling is also addressed. The analyses are used to demonstrate that the standard fire test is generally unconservative at predicting the time at failure of a structure subjected to realistic fire conditions, since failure depends more on the evolution of temperatures within the steel beams than the duration of fire exposure. The analyses also show that critical temperatures from the standard fire test are conservative and thus offer a better means for predicting failure in steel structures within the scope of the standard fire test. / Ph. D.

For the past several years there has been great effort in the analysis of internal duct cooling. The steady increase in power output and thermal efficiency requirements for gas turbine engines has called for significant advancement in turbine blade internal duct cooling technology. Numerical analysis of turbulent duct flow has been largely limited to Reynolds Averaged Navier-Stokes (RANS) simulations. This is because of the low computational requirements of such calculations relative to Large Eddy Simulation (LES) and Direct Numerical Simulation (DNS). However, the tides have started to turn in favor of LES, partly because of the exponential increase in computer hardware performance in recent years. Three conference papers make up the contents of this thesis. LES is performed for fully developed flow and heat transfer in a staggered 45º ribbed duct in the first paper. The rib pitch-toheight ratio P / e is 10 and a rib height to hydraulic diameter ratio h e / D is 0.1. The Reynolds numberbased on the bulk flow rate and hydraulic diameter is 47,300. The overall heat transfer enhancement obtained was a factor of 2.3, which matched experimental data within 2%. The surfaces of highest heat transfer enhancement were the ribbed walls and the outer wall. Results from LES of an orthogonally rotating 90º ribbed duct are presented in the second paper for rotation numbers: Ro = 0.18, 0.35 and 0.67. The Reynolds number is 20,000. The P / e and h e / D were the same as in the first paper. Turbulence and heat transfer are augmented on the trailing surface and reduced at the leading surface. Secondary flows induced by Coriolis forces, increase heat transfer augmentation on the smooth walls. Finally, the third paper studies the same flow conditions of the second paper and goes further by including effects of centrifugal buoyancy forces using LES. Two buoyancy numbers are studied: Bo = 0.12 and 0.29. Centrifugal buoyancy does not have a large effect on leading side augmentation ratios for all rotation numbers, but increases heat transfer significantly on the trailing side. In all papers, mean flow and heat transfer results compare well with experimental data. / Master of Science

The contribution of present research work is to experimentally investigate the effects of blowing ratio and mainstream Mach number/Reynolds number (from 0.6/8.5X10⁵ to 1.0/1.4X10⁶) on the performance of the fan-shaped hole injected turbine blade and vane. The study was operated with high freestream turbulence intensity (12% at the inlet) and large turbulence length scales (0.26 for blade, 0.28 for vane, normalized by the cascade pitch of 58.4mm and 83.3mm respectively). Both convective heat transfer coefficient, in terms of Nusselt number, and adiabatic effectiveness are provided in the results. Present research work also numerically investigates the shock/film cooling interaction. A detailed analysis on the physics of the shock/film cooling interaction in the blade cascade is provided. The results of present research suggests the following major conclusions. Compared to the showerhead only vane, the addition of fan-shaped hole injection on the turbine Nozzle Guide Vane (NGV) increases the Net Heat Flux Reduction (NHFR) 2.6 times while consuming 1.6 times more coolant. For the blade, combined with the surface curvature effect, the increase of Mach number/Reynolds number results in an improved film cooling effectiveness on the blade suction side, but a compromised cooling performance on the blade pressure side. A quick drop of cooling effectiveness occurs at the shock impingement on the blade suction side near the trailing edge. The CFD results indicate that this adiabatic effectiveness drop was caused by the strong secondary flow after shock impingement, which lifts coolant away from the SS surface, and increases the mixing. This secondary flow is related to the spanwise non-uniform of the shock impingement. / Ph. D.

The purpose of this study is to quantify the effects of heat flux sensor design and interaction with both test article material choice and geometry on heat flux measurements. It is the public domain component of a larger study documenting issues inherent in heat flux measurement. Direct and indirect heat flux measurement techniques were tested in three thermally diverse model materials at the same Mach 6 test condition, with a total pressure of 1200 psi and total temperature of 1188° R, and compared to the steady analytic Fay-Riddell solution for the stagnation heat flux on a hemisphere. A 1/8 in. fast response Schmidt-Boelter gage and a 1/16 in. Coaxial thermocouple mounted in ¾ in. diameter stainless steel, MACOR, and Graphite hemispheres were chosen as the test articles for this study. An inverse heat flux calculation was performed using the coaxial thermocouple temperature data for comparison with the Schmidt-Boelter gage. Before wind tunnel testing, the model/sensor combinations were tested in a radiative heat flux calibration rig at known static and dynamic heat fluxes from 1 to 20 BTU/ft2/s. During wind tunnel testing, the chosen conditions yielded stagnation point convective heat flux of 15-60 BTU/ft2/s, depending on the stagnation point wall temperature of the model. A computational fluid dynamic study with conjugate heat transfer was also undertaken to further study the complex mechanisms at work. The overall study yielded complex results that prove classic methodology for inverse heat flux calculation and direct heat flux measurement require more knowledge of the thermal environment than a simple match of material properties. Internal and external model geometry, spatial and temporal variations of the heat flux, and the level of thermal contact between the sensor and the test article can all result in a calculated or measured heat flux that is not correct even with a thermally matched sensor. The results of this study supported the conclusions of many previous studies but also examined the complex physics involved across heat flux measurement techniques using new tools, and some general guidance for heat flux sensor design and use, and suggestions for further research are provided. / Ph. D.

In the past few decades, the operating temperatures of gas turbine engines have increased significantly with a view towards increasing the overall thermal efficiency and specific power output. As a result of increased turbine inlet temperatures, the hot gas path components downstream of the combustor section are subjected to high heat loads. Though materials with improved temperature capabilities are used in the construction of the hot gas path components, in order to ensure safe and durable operation, the hot gas path components are additionally supplemented with thermal barrier coatings (TBCs) and sophisticated cooling techniques. The present study focusses on two aspects of gas turbine cooling, namely augmented internal cooling and external film cooling. One of the commonly used methods for cooling the vanes involves passing coolant air bled from the compressor through serpentine passages inside the airfoils. The walls of the internal cooling passages are usually roughened with turbulence promoters like ribs to enhance heat transfer. Though the ribs help in augmenting the heat transfer, they have an associated pressure penalty as well. Therefore, it is important to study the thermal-hydraulic performance of ribbed internal cooling passages. The first section of the thesis deals with the numerical investigation of flow and heat transfer characteristics in a ribbed two-pass channel. Four different rib shapes- 45° angled, V-shaped, W-shaped and M-shaped, were studied. This study further aims at exploring the performance of different rib-shapes at a large rib pitch-to-height ratio (p/e=16) which has potential applications in land-based gas turbines operating at high Reynolds numbers. Detailed flow and heat transfer analysis have been presented to illustrate how the innate flow physics associated with the bend region and the different rib shapes contribute to heat transfer enhancement in the two-pass channel. The bend-induced secondary flows were observed to significantly affect the flow and heat transfer distribution in the 2nd pass. The thermal-hydraulic performance of V-shaped and 45° angled ribs were better than W-shaped and M-shaped ribs. The second section of the study deals with the analysis of film cooling performance of different hole configurations on the endwall upstream of a first stage nozzle guide vane. The flow along the endwall of the airfoils is highly complex, dominated by 3-dimensional secondary flows. The presence of complex secondary flows makes the cooling of the airfoil endwalls challenging. These secondary flows strongly influence endwall film cooling and the associated heat transfer. In this study, three different cooling configurations- slot, cylindrical holes and tripod holes were studied. Steady-state experiments were conducted in a low speed, linear cascade wind tunnel. The adiabatic film cooling effectiveness on the endwall was computed based on the spatially resolved temperature data obtained from the infrared camera. The effect of mass flow ratio on the film cooling performance of the different configurations was also explored. For all the configurations, the coolant jets were unable to overcome the strong secondary flows inside the passage at low mass flow ratios. However, the coolant jets were observed to provide much better film coverage at higher mass flow ratios. In case of cylindrical ejection, the effectiveness values were observed to be very low which could be because of jet lift-off. The effectiveness of tripod ejection was comparable to slot ejection at mass flow ratios between 0.5-1.5, while at higher mass flow ratios, slot ejection was observed to outperform tripod ejection. / Master of Science

An experimental study was performed to measure surface Nusselt number and film cooling effectiveness on a film cooled first stage nozzle guide vane using a transient thin film gauge (TFG) technique. The information presented attempts to further characterize the performance of shaped hole film cooling by taking measurements on a row of shaped holes downstream of leading edge showerhead injection on both the pressure and suction surfaces (hereafter PS and SS) of a 1st stage NGV. Tests were performed at engine representative Mach and Reynolds numbers and high inlet turbulence intensity and large length scale at the Virginia Tech Transonic Cascade facility. Three exit Mach/Reynolds number conditions were tested: 1.0/1,400,000; 0.85/1,150,000; and 0.60/850,000 where Reynolds number is based on exit conditions and vane chord. At Mach/Reynolds numbers of 1.0/1,450,000 and 0.85/1,150,000 three blowing ratio conditions were tested: BR = 1.0, 1.5, and 2.0. At a Mach/Reynolds number of 0.60/850,000, two blowing ratio conditions were tested: BR = 1.5 and 2.0. All tests were performed at inlet turbulence intensity of 12% and length scale normalized by leading edge diameter of 0.28. Film cooling effectiveness and heat transfer results compared well with previously published data, showing a marked effectiveness improvement (up to 2.5x) over the showerhead only NGV and agreement with published showerhead-shaped hole data. NHFR was shown to increase substantially (average 2.6x increase) with the addition of shaped holes, with only a small increase (average 1.6x increase) in required coolant mass flow. Heat transfer and effectiveness augmentation with increasing blowing ratio was shown on the pressure side, however the suction side was shown to be less sensitive to changing blowing ratio. Boundary layer transition location was shown to be within a consistent region on the suction side regardless of blowing ratio and exit Mach number. / Master of Science

The topic of heat transfer is traditionally taught as an upper level, lecture-style course to mechanical engineering students. Such courses do not provide students with ways to see and feel the important heat transfer concepts at hand. As a way to overcome this, novel, hands-on workshops have been designed and implemented into a heat transfer class taught to junior level mechanical engineering students. Two types of experimental workshops were created and used in two different years of a section of a heat transfer class. In the first year, twelve workshops were designed which included live demos so that the students could see and feel different modes of heat transfer while taking data and seeing real-time plots of temperature and heat flux in different experiments. The workshop introduced each topic the students would be learning in the lecture and was performed the week before the actual lecture on the topic. Each workshop included easily available materials, thermocouples, heat flux sensors, and data acquisition instruments for the students to use. The workshops also served replacements for what would be the third lecture of the week. Results from a concept inventory test given at the end of the first year showed a significant difference on certain question between an experimental group of students who had the workshops and a control group who took the traditional class lecture. However, there were still concepts and topics that the experimental group did not show improvement. They also showed a lack of improvement in their problem solving skills for quiz and test problems. For the second year of the experiment, the workshops were restructured quite a bit. The original 12 workshop format was cut down to only six in order to focus on the ones the students seemed to have benefited from the most. The workshops were also changed into a video-enhanced format where the students would watch a video of the experiment being done while also having the materials in front of them to place their hands on themselves. The students could therefore see and feel what was physically happening and still perform the experiment while watching real-time, pre-recorded plots of heat flux and temperature without worrying about making sure their setup was right and acquiring good results. The new video-enhanced workshops also included control volume and resistance diagrams for each experiment in order to help the students relate the workshops and concepts back to problems on their quizzes and tests. Results from these workshops seemed to show some statistical significance between the experimental and control groups on concept questions given on quizzes throughout the semester, but there was no difference on any questions from the ten concept questions given on the final exam. However, surveys taken by the students indicate that they believed the workshops did help them to understand the concepts in a real-world sense and that they helped them understand the class material better overall. Aside from the results of the workshops on the students learning, this study concludes with an analysis of important heat transfer concepts and how to test them. There is much debate about the underlying concepts in the topic of heat transfer and a thorough analysis on what specific concepts are important for students to know must be addressed. Many heat transfer concept questions on current concept inventories have more to do with thermodynamics and the mixing of the two topics is itself a misconception. / Master of Science

Accurately measuring the total temperature of a high-speed fluid flow is a challenging task that is required in many research areas and industry applications. The difficulty in total temperature measurement generally stems from attempting to minimize measurement error or accurately predict error so it can be accounted for. Conduction error and aerodynamic error are two very common sources of error in total temperature probe measurements. Numerous studies have been performed in prior literature to account for simple cases of both errors. However, the impacts of a mounting strut and freestream flow angle on conduction error and aerodynamic error have not been previously modeled. Both of these effects are very common in gas-turbine applications of total temperature probes. Therefore, a fundamental study was performed to analyze the impact of mount interference and freestream flow angle on a probe's conduction error and aerodynamic error. An experimental study of aerodynamic error was performed using strut-mounted thermocouples in a high-speed jet at Mach numbers ranging from 0.25-0.72. This study showed that a strut stagnation point can provide aerodynamic error reductions and insensitivity to approach Mach number. An off-angle experimental study of conduction error was also performed using strut-mounted thermocouples at pitch angles ranging from -30° to 30°. High-fidelity Computational Fluid Dynamics (CFD) simulations with Conjugate Heat Transfer (CHT) were performed in conjunction with the experiments to provide key heat transfer information and flow visualizations. It was identified that unshielded total temperature probes have reduced conduction error at off-angles, but are sensitive to changes in the freestream flow angle. A low-order method was developed to account for mount interference and flow angle effects. The developed low order method utilizes a local Mach number for aerodynamic error predictions and a local Reynolds number for conduction error predictions. This developed low-order method was validated against experiment and 3D, CFD results, and was shown to accurately capture flow angle trends, mount interference effects, and the impacts of varying probe geometry. / Master of Science / Accurately measuring the total temperature of a high-speed fluid flow is a challenging task that is required in many research areas and industry applications. Many methods exist for measuring total temperature, but the use of thermocouple based probes immersed into a flow remains a common and desirable measurement technique. The difficulty in using thermocouple based probes to acquire total temperature stems from attempting to minimize or accurately predict the probe’s measurement error. Conduction, convection, and radiation heat transfer between the fluid flow and probe create challenges for minimizing measurement error so that the accurate total temperature can be obtained. Numerous studies have been performed in prior literature to account for simple cases of each error source. However, there are many complex, practical applications in which the influence of each error source has not been studied. The impacts of a freestream flow angle and the total temperature probe’s mounting structure have not been previously modeled. Both of these effects are very common in gas-turbine applications of total temperature probes. This Thesis will present a fundamental study analyzing the impact that freestream flow angle and a probe’s mount have on a total temperature probe’s measurement error. The influence of conduction and convection heat transfer was studied experimentally for numerous probe geometries, and the impacts of a mounting strut and freestream flow angle were analyzed. A low-order method was developed to predict conduction error and aerodynamic error for total temperature probes in offangle conditions with the presence of mount interference. The developed low-order method was shown to accurately capture the effects of a mounting strut, varying probe geometry, and varying flow angle. Additionally, the low-order method was validated against experimental and 3D, CFD/CHT results.

The effects of greenhouse gases have seen a significant rise in recent years due to the use of fossil fuels like gasoline and diesel. Conversion of the energy stored in these fossil fuels to mechanical work is an extremely inefficient process which results in a high amount of energy rejected in the form of waste heat. Thermoelectric materials are able to harness this waste heat energy and convert it to electrical power. Thermoelectric devices work on the principle of the Seebeck effect, which states that if two junctions of dissimilar materials are at different temperatures, an electrical potential is developed across them. Even though these devices have small efficiencies, they are still an extremely effective way of converting low grade waste heat to usable electrical power. These devices have the added advantage of having no moving parts (solid state) which contributes to a long life of the device without needing much maintenance. The performance of thermoelectric generators is dependent on a non-dimensional figure of merit, ZT. Extensive research, both past and ongoing, is focused on improving the thermoelectric generator's (TEG's) performance by improving this figure of merit, ZT, by way of controlling the material properties. This research is usually incremental and the high performance materials developed can be cost prohibitive. The focus of this study has been to improve the performance of thermoelectric generator by way of improving the heat transfer from the exhaust gases to the TEG and also the heat transfer from TEG to the coolant. Apart from the figure of merit ZT, the performance of the TEG is also a function of the temperature difference across it, By improving the heat transfer between the TEG and the working fluid, a higher temperature gradient can be achieved across it, resulting in higher heat flux and improved efficiency from the system. This area has been largely neglected as a source of improvement in past research and has immense potential to be a low cost performance enhancer in such systems. Improvements made through this avenue, also have the advantage of being applicable regardless of the material in the system. Thus these high performance heat exchangers can be coupled with high performance materials to supplement the gains made by improved figure of merits. The heat exchanger designs developed and studied in this work have taken into account several considerations, like pressure drop, varying engine speeds, location of the system along the fuel path, system stability etc. A comprehensive treatment is presented here which includes 3D conjugate heat transfer modeling with RANS based turbulence models on such a system. Various heat transfer enhancement features are implemented in the system and studied numerically as well as experimentally. The entire system is also studied experimentally in a scaled down setup which provided data for validation of numerical studies. With the help of measured and calculated data like temperature, ZT etc, predictions are also presented about key metrics of system performance. / Ph. D.

Master of Science