Randall D. Manteufel

Replicated Latin Hypercube Sampling

A new sampling algorithm is presented based on repeated additions of Latin Hypercube Sampling (LHS) where the new samples are drawn at predetermined bin fractions. The method is motivated by a desire to have incremental growth in the sample size and maintain uniform marginal distributions. The method can be coupled with a correlation control algorithm as well as a discrepancy reduction algorithm to ensure a uniform distribution of points in the sample space. The overall size of the sample space increases by a user specified base size (typically 25 to 100 points). At the completion of each base sample, the algorithm can estimate the convergence or error in the simulation to assess when sufficient samples have been executed to achieve a specified tolerance. As presented here, the replicated Latin Hypercube Sampling (rLHS) scheme is commonly coupled with correlation control, and discrepancy control. The preferred correlation control algorithm is by Iman & Conover that can be used to minimize spurious correlation between inputs or to induce a desired correlation. The discrepancy control algorithm is based on a simple accept/reject scheme and is only used in this work to demonstrate the effect of reducing the discrepancy for a number of test cases. The preferred discrepancy measure is based on the wrap-around L2. This discrepancy measure allows the algorithm to clean-up (i.e., reduce the discrepancy) the most important faces of the hypercube based on the importance of each input variable. The algorithm is applied to several application problems where it is observed to yield up to four orders of magnitude lower error than simple random sampling (SRS) when estimating the mean of the response. The improvement in the percentile estimates are less dramatic, yet is often one order of magnitude lower error than for SRS.

Understanding Why Engineering Students Take Too Long to Graduate

There is growing pressure on public colleges and universities to decrease the time students take to earn an undergraduate degree. There are many factors that slow students’ progress towards graduation. For example, urban universities may have a significant number of non-traditional students who don’t take a full load of courses required to graduate in four years. Also, some freshman students interested in engineering may not be prepared for college and are required to take remedial math and science courses. Engineering is a highly-structured program, often with a long sequence of courses requiring one or more prerequisites. If some courses aren’t offered each semester, this can delay progress toward graduation for some students. This paper examines graduating students’ academic records and surveys senior-level mechanical engineering students to identify some of the causes for the increased graduation times. Students provided detailed information such as their full- or part-time status, how many semesters left to graduation, whether they attended summer school, the courses they had difficulty passing, and other issues related to the length of time required to complete their degrees. Feedback from students is essential as universities look to improve graduation rates. The results presented are based on the data for the mechanical engineering program at a public institution in Texas. Although each institution is unique, the findings presented in this paper are expected to apply to similar institutions throughout the nation.Copyright

Sequential Perturbation Uncertainty Propagation in Thermal-Fluid Applications

A challenge in an undergraduate mechanical engineering curriculum is having students demonstrate “an ability to design and conduct experiments, as well as analyze and interpret data” as required for ABET accreditation. It is expected that students be able to identify and quantify sources of uncertainty, propagate uncertainties to intermediate and final results, interpret the relative importance of uncertainty sources, and develop experimental strategies to reduce the uncertainties in the final results. A spreadsheet application is presented that helps students learn these concepts and “see” what drives the uncertainty in the final results. The method known as sequential perturbation is used and shown to greatly reduce the tediousness of the calculations. As presented, the method significantly reduces the complexity of uncertainty analysis by eliminating the need to differentiate relationships between primary measurements and inferred measurements. Differentiation of complex relationships is often tedious and error-prone. The method is applied to three thermal-fluid application problems. Feedback from students is presented and is positive. The method summarized here should help students learn about the propagation of uncertainties and help demonstrate meeting ABET outcome “b”. The method can be extended beyond laboratory classes and is shown to be useful in design of experiments.Copyright

Use of conceptual questions with prompt feedback in engineering thermodynamics

An effective strategy to promote deep understanding in engineering thermodynamics is to increase the use of conceptual questions during lectures coupled with prompt assessment of student responses. A key is to collect responses from all students and provide prompt feedback explaining the correct response. It has been found that conceptual questions are more effective than numerical. Good questions explore if a quantity will increase, decrease or remain unchanged in response to a change in the system. In previous semesters, an instructor would pose conceptual questions during lecture and discuss with those students who participated with the instructor. Using an electronic collection system for student responses, all student responses can be collected and assesses. Results show that (1) it is rare that the entire class is correct even for the simplest of questions, (2) a nearly identical question can be repeated in a subsequent lecture and there will continue to be a incorrect responses, and (3) repeating questions throughout the semester is effective at addressing common conceptual misunderstandings and improving long-term student learning in engineering thermodynamics.Copyright

Experimental Measurement of Vacuum Assisted Drying of Spent Nuclear Fuel

A prevalent issue within extended long term dry storage units for spent nuclear fuel has always been fuel and cask contamination. This contamination can be the result of the helium within the cask leaking into the atmosphere or inadequate vacuum drying techniques. Once the cask integrity has been compromised, the helium starts to leak, and the resulting space once occupied by helium in the casks is replaced with ambient air. One of the other prominent gases found within ambient air besides oxygen is water vapor which can be a result of both helium leaking and poor vacuum drying techniques. Contact between water and the fuel rods/assemblies for a prolonged amount of time can result in corrosion of the fuel cladding, and the canister if exposed. The potential of corrosion of the fuel cladding increases risk of radioactive fission fragments contaminating the environment, increases the radioactive period of spent nuclear fuel, and decreases the potential for fuel rod repurposing within the future if U.S. law permits.With literary findings showing liquid water within the inner cask in a long term storage unit of fifteen years or longer, proper drying techniques have not been fully developed. There are a number of projected theories about how water is entering the cask without an external crack or imperfection within the inner cask walls. This case study aims to solve this issue by inspecting the vacuum drying process of the fuel rods/assemblies from the temporary on-site storage pools to their respective long term dry storage casks.The purpose of this case study is to conduct a laboratory experiment of a scale replica of one dry storage cask and the vacuum drying process before long term storage. The experiment will be focused around the process of applying several cycles of vacuum and backfilling the cask with Helium. The purpose of several cycles of backfilling gas is to simultaneously introduce more of a pressure gradient for water evaporates to depart the pressure vessel and to avoid thermodynamic temperatures that would otherwise freeze the top layer of water. To do this, the vacuuming process must be properly understood, as pulling a vacuum drops pressures instantaneously. There are possibilities of freezing water vapor into its solidified form due to its thermodynamic triple point during this vacuum process. Once water is trapped under a layer of ice within the vessel, water will remain throughout storage time due to restrictions to its own geometries. The importance of developing a scale model and improving the drying process that precedes long term storage of spent nuclear fuel is a necessary solution to existing contamination results for practical future applications within the United States and other countries moving towards long term storage of spent nuclear fuel.Copyright

Exergetic Analysis of a Cross-Flow Microchannel Heat Exchanger for Bleed Air Cooling in Aircraft Gas Turbine Engines

A current issue with high-pressure-ratio compressors found in aircraft engines is the temperature of the air exiting the compressor. The exiting air is used as coolant for engine components found in later stages of the engine such as first-stage turbine blades, and afterburner walls. A viable option for reducing outlet temperature of high-pressure-ratio compressors is to “bleed-off” a fraction of the air which is cooled in a heat exchanger by rejecting heat into the liquid fuel stream and then use the air for cooling critical components downstream. Bleeding off air from the outlet of the compressor has two benefits: (1) air temperature is reduced, and (2) fuel temperature is elevated. Along with reduced air temperatures, the fuel will ultimately receive the heat lost from the air, making the fuel more ideal for combustion purposes. The higher temperature the fuel is received in the combustion process, the greater the work output will be according to the basics of thermodynamic combustion. The objective of this case study is to optimize the efficiency of the cross-flow micro channel heat exchanger, with respect to (1) volume (1.75–2.75 mm3) and heat transfer, and (2) weight (0.15–.25 N) and heat transfer. The optimization of the heat exchanger will be evaluated within the bounds of the 2nd law of thermodynamics (exergy). The only effective way to measure the 2nd law of thermodynamics is through exergy destruction or its equivalent form: entropy generation as a factor of dead state temperature. With relations and equations obtained to design an optimal heat exchanger, applications to high performance aircraft gas turbine engines is considered through exergy. The importance of developing an exergetic analysis for a thermal system is highly effective for identifying area’s within the system that have the path of highest resistance to work potential through various modes of heat transfer and pressure loss. Thus, optimization to reduce exergy destruction is sought after through this design method alongside verifying other heat exchanger methods through effectiveness.Copyright

Consideration of uncertainties in compact cross-flow heat exchanger design for gas turbine engine application

Recent experimental work characterized the performance of a unique cross-flow heat exchanger design for application of cooling compressor bleed air using liquid jet fuel before it is consumed in the gas turbine combustor. The proposed design has micro-channels for liquid fuel and cools air flowing in passages created using rows of intermittent fins. The design appears well suited for aircraft applications because it is compact and light-weight. A theoretical model is reported to be in good agreement with experimental measurements using air and water, thus providing a design tool to evaluate variations in the heat exchanger dimensions. This paper presents an evaluation of the heat exchanger performance with consideration of uncertainties in both model parameters and predicted results. The evaluation of the design is proposed to be reproduced by students in a thermal-fluids design class. The heat exchanger performance is reevaluated using the effectiveness–NTU approach and shown to be consistent with the method reported in the original papers. Results show that the effectiveness is low and in the range of 20 to 30% as well as the NTU which ranges from 0.25 to 0.50 when the heat capacity ratio is near unity. The thermal resistance is dominated by the hot gas convective resistance. The uncertainties attributed to fluid properties, physical dimensions, gas pressure, and cold fluid flow rate are less significant when compared to uncertainties associated with hot fluid flow rate, hot fluid inlet temperature, cold fluid inlet temperature, and convective correlation for gas over a finned surface. The model shows which heat transfer mechanisms are most important in the performance of the heat exchanger.© 2013 ASME

Behavior of Internal Energy and Enthalpy of Fluids Along Isotherms and Isentropic Lines in the Compressed Liquid Region

It is a common practice to approximate the thermodynamics properties of fluids in the compressed liquid regions from their saturation properties. Most thermodynamics textbooks state that the specific volume, specific internal energy, and specific entropy in the compressed liquid region are functions of temperature only and are independent of pressure. Therefore, compressed liquid property tables are not provided for any substance, except for water, and compressed liquid properties are approximated by their saturated liquid properties at a given temperature. Recent examination of current practice in approximating compressed liquid properties has shown that the internal energy of fluids exhibits growing dependency on pressure with increases in temperature. This paper compares the behavior of internal energy and enthalpy four compressed fluids along isotherms with those behaviors along isentropic lines. Water, ammonia, methane, and propane are examined in this study. It is shown that effects of pressure on the internal energy and enthalpy of compressed liquids are much lower along isentropic lines than those along isotherms.Copyright

Thermal Performance of High-Flow Single-Phase Liquid-Cooled Heat Sinks

Experimental measurements are reported for high-flow liquid-cooled heat sinks designed for cooling electronics components such as a CPU. The flow rate is up to 2 GPM with internal flow passage length scales on the order of 0.1 to 1.0 mm in the primary heat transfer region. Of the designs tested, three achieved maximum flow rates with pressure drops of less than 1.5 psi. Two have lower maximum flow rates because of higher internal flow resistance. In the experiments, particular attention is given to sources of experimental uncertainty and the propagation of uncertainty through the calculations to reported thermal resistance, R (°C/W). Analysis includes bias and precision errors for direct measurement of temperature, flow rate, and pressure drop. Additionally, a separate thermocouple calibration test is reported to establish measurement uncertainties for the system. Main emphasis is made to the error propagation in thermal resistance calculations of each heat sink and measurement of heat removal rate from the CPU. Data is used to determine the standard error for R which ranges up to about 0.05 °C/W with the maximum for one heat sink up to 0.07 °C/W. Averaging of repeated measurements at the same flow rate without accounting for the range of the original data will result in lower uncertainties in the reported results.Copyright

Energy use comparison of air distribution systems serving a section of a school building

An optimal air distribution design accomplishes both comfort and ventilation requirements while consuming as little energy as possible. This paper analyzes four different air distribution systems and technologies including single duct variable air volume air handlers, chilled beam cooling systems, total energy recovery wheels, displacement ventilation, and dedicated outside air systems; in an effort to determine the best air distribution system for a representative section of a school in hot and humid climate. The effectiveness of the air distribution systems is evaluated by analyzing how the different technologies take advantage of the natural convective properties of air to create a comfortable environment for the occupied region of the space. Distribution effectiveness and energy consumption must be weighed against considerations such as system complexity and ease of operation.This paper compares several alternative air distribution systems to a baseline single inlet VAV system that is commonly used in new schools designed today. Calculations show that the total energy recovery wheels result in a 16% energy savings over the baseline air distribution system because of the large amount of outside air required in school buildings. Chilled beams are not well suited for schools because of the large amount of outside air required by the space and the sophisticated design and operation needed to prevent condensation from occurring at the chilled beam. The results show that the air distribution system that consumes the least amount of energy is a displacement ventilation system. The system also inherently promotes better indoor air quality as it allows air to naturally rise out and return out of the space with minimal mixing of contaminates that may be recirculated within the room for others to breath. The displacement ventilation system’s overall energy savings of 20% over the baseline is mainly attributed to its total energy recovery wheel and the system’s ability to drastically reduce the cooling load seen by the air cooled chiller by effectively ventilating spaces using less outside air.Copyright