MENG 472/474 Projects - 2018

In MENG 472/474, students work on independent projects that cover a wide range of topics, from traditional mechanical engineering topics (e.g., mechanical device design, fluid flow, and materials analysis) to interdisciplinary topics at the interface between mechanical engineering and other branches of engineering such as biomedical, chemical, electrical, or environmental engineering. Under the supervision of faculty advisers, students investigate physical phenomena through experimental measurement and/or numerical simulation, and they design and construct functioning prototypes to solve engineering problems. The majority of the faculty advisers come from within the mechanical engineering department, with the remaining advisers distributed across the University (and occasionally outside the University). Funding for projects is generously provided by the Yale SEAS Dean's Office and, in some cases, through the faculty advisers. The students were asked to write the following short summaries two-thirds of the way through the semester, when they still had a few weeks to go on their projects. All projects are represented here, except for those that cannot be publicized due to information of a proprietary nature.

Investigating the feasibility of generating Janus particles using the electrospray technique


Saeed Arafeh
Adviser: Prof. Alessandro Gomez, Department of Mechanical Engineering & Materials Science

A Janus particle is a special type of nanoparticle whose surface has two or more distinct physical properties. This is a result of its composition – two liquids or polymers merge into a single droplet, and solidify before they have a chance to diffuse. Various new discoveries about Janus particles have enabled progress in nanotechnology, which has allowed for the creation of water-repellent fibers and new and precisely designed materials in the biological sciences. Electrospray ionization is a technique that allows for the production of uniformly-sized nanodroplets. Attempting to generate Janus nanoparticles, we ran the electrospray using a dye alongside a transparent polymer emerging from two adjacent, charged emitters (see microscope image). Next, we will use a single, theta-shaped emitter that will allow us to run two adjacent liquids in a more controlled manner. If we succeed in preventing diffusion of the liquids, we will ascertain the nature of the particles by examining their residues using microscopy techniques. We will then identify conditions and constraints of electrospray ionization in this research area, and subsequently some potential applications of the technique in medicine and nanotechnology.

Particle contamination in microfluidics


Sierra Jackson and Richard Hicks
Advisers: Dr. Sara Hashmi, Chemical & Environmental Engineering and Prof. Corey O’Hern, Mechanical Engineering & Materials Science

In this project, we are determining if a cheap, mechanical device in microfluidics can be designed to detect the presence and concentration of harmful particles in water, oil, and water and oil emulsions in place of expensive chemical tests currently used for detection. Currently, testing for contamination requires samples to be sent to off-site laboratories, which often have a turnaround time of over 48 hours. In the case of water safety tests, besides this process being expensive logistically, it doesn’t allow for the results of testing to be returned in time for the water to be stopped from entering the main system. The device itself is a polydimethylsiloxane mold on a glass slide that we insert the sample fluid through by way of a syringe and small tubing. Each device has approximately 200 channels, which allowing for 200 largely independent tests in a small device. We vary different particle characteristics in the solution (such as concentration and particle size) in order to determine whether these microfluidic devices will clog at predictable times with the presence of contaminants. So far we have worked on creating our own microfluidic devices as well as creating solutions for the tests with different variables. We have performed a preliminary test with the devices and seen clogs due to the particulate matter, indicating that these devices are capable of detecting particles. We hope to create a matrix that holds different probabilities for clogging times dependent on particle size, shape, and concentration. This matrix can then be applied in conjunction with a simple timer in order to determine the presence and type of particles in unknown solutions.

Correlating heat rejection and flow rate in thermoGreenWalls


Kevin Koste and Greg Campbell
Adviser: Prof. Alexander Felson, Schools of Architecture and Forestry & Environmental Studies

Buildings conventionally use energy-intensive cooling towers to reject heat from air conditioning systems in warm climates. We are researching a potential alternative to cooling towers called thermoGreenWalls, which are green infrastructure systems that use evaporation and plant transpiration to reject this heat. In thermoGreenWalls, water flows through the roots of wetlands plants, which are embedded in a porous felt material, and the plants cool the water as it flows. In addition to environmental conditions such as solar radiation and wind speed, the amount of water flowing through the thermoGreenWall each minute also affects the overall heat rejection potential. This experiment investigated the relationship between flow rate and heat rejection for an indoor thermoGreenWall system. Our experimental setup included two 3’ x 8’ thermoGreenWall panels operating in a closed-loop system including a reservoir, pump, and heater. The water distribution system was redesigned to support a maximum flow rate of 6 gallons per minute, which was ten times greater than the historical maximum. Each panel was also fitted with instrumentation allowing researchers to modulate flow rate while measuring the conditions of the water entering and exiting each panel. These measurements will be used to calculate the overall heat rejection of each panel to determine the optimal operating conditions. To test an additional variable, one panel included plants while the other did not. The heater ensured that the water entering each panel remained at a constant temperature of approximately 35°C. The results of this study will determine the feasibility of integrating a thermoGreenWall system with the air conditioning system of a campus building. thermoGreenWalls are a promising green infrastructure concept with a high potential for reducing the energy consumption of the built environment.

Dobsonian telescope


Betsy Li
Adviser: Prof. Marshall Long, Mechanical Engineering & Materials Science

The purpose of this project was to design and build a complete Dobsonian telescope using SolidWorks and tools available in a general makerspace setting, and to make the final product open source for educational purposes. After reading The Dobsonian Telescope: A Practical Manual for Building Large Aperture Telescopes, by David Kriege and Richard Berry, SolidWorks was used to design around the following constraints: an eight-inch diameter parabolic mirror (primary) with an f-ratio of six, a 50-millimeter diagonal mirror (secondary), under 45 pounds (compact, lightweight, and transportable), and with the idea in mind that the design could be constructed with tools found in a makerspace. The final product utilized a laser cutter for constructing the main structure, and although 3D-printed parts were used for a majority of the prototyping for connectors and the focuser, some parts were replaced by stronger materials because the material strength of polylactic acid (PLA) was determined to not be strong enough for the loads expected. With a completed telescope, the next steps for this project will be to introduce methods to motorize the telescope. A Dobsonian telescope was designed, built, and will be documented for its process and design considerations for interested makers through this project.

Dimensional synthesis and fabrication of a Stewart platform-inspired robotic hand


Connor McCann
Adviser: Prof. Aaron M. Dollar, Mechanical Engineering & Materials Science

In the field of robotic grasping, researchers have attempted to mimic the structure of the human hand with the aim of replicating its dexterity and reliability in a robotic system. That said, no current robotic hand has ever been able to robustly duplicate this level of performance. This can largely be attributed to the high mechanical complexity of the human hand, which can be difficult to translate into a robot. In our prior research, we explored an alternate design approach, drawing inspiration from the far simpler “Stewart platform” mechanism. The novel hand we developed was able to perform far more dexterous motions than existing hand designs, despite having a simpler mechanical structure. In this current work, we have sought to optimize the performance of the hand, determining (in simulation) the optimal design parameter values for particular applications. We now seek to validate these results experimentally, constructing an improved prototype hand that will allow us to investigate the impact of the different design parameters on performance. Eventually, by introducing a novel design approach for robotic hands, this work will lay the groundwork to improve hand performance across multiple robotic disciplines, pushing the boundaries of robotic hand capabilities.

Constructing a time-series Input-Output model for U.S. emissions


Lea Rice
Adviser: Prof. Edgar Hertwich, School of Forestry & Environmental Studies

The Yale Center for Industrial Ecology is developing a model of U.S. greenhouse gas emissions that incorporates the carbon footprints of capital assets. Capital assets include such things as roads, factories, and intellectual property, which are used long after the year in which they were first produced. The new model includes these products’ impacts over the course of their lifetimes, and attributes appropriate emissions intensities for a given product based on the year that it was produced. This project uses Input-Output analysis, a “top-down” approach to carbon footprint accounting that utilizes trade data as a proxy for material flows between industries. We incorporated time-series data from the EPA Greenhouse Gas Inventory, along with several other government databases, into the existing model. This allows us to quantify the emissions intensities for an industry for a particular year. Currently, we are determining which model of capital assets’ emission depreciation is most accurate; for example, how much of a road’s emissions do we count after ten years, as opposed to after thirty? The ability to more accurately calculate industry-level greenhouse gas emissions will provide a basis for determining more effective mitigation measures, and empower decision-makers to enact policies to help mitigate climate change.

To move or not to move: Principal curvatures of articular surfaces

Aaron Michael West
Adviser: Prof. Madhusudan Venkadesan, Mechanical Engineering & Materials Science

When recreating species that no longer exist, anatomists and paleontologists aim to put together bone structures and try to determine how that species may have moved. We aim to quantify their work by producing a geometric theory that can determine how a limb moves based on the surface geometry at its joint. To do so, we have designed an experiment that will allow us to predict movement capabilities at a joint based purely on the principal curvatures of the articular surfaces. Our experiment will hold one end of a chicken bone connected to a load cell. On the other end, the chicken bone will have a small displacement applied to it, and based on statics, the stiffness of the joint in that particular orientation can be determined. By comparing our data and CT scans of surface geometry at the joint to the data of in vivo motion, a geometric theory can be produced and validated. We have tested one prototype of this experiment. We are currently building the final version of this experiment. We hope to do preliminary tests prior to the end of the semester.