UMSS21 Engineering and Information Sciences
UMSS21 Presentations by Category
UMSS21 Social Sciences and Humanities
UMSS21 Physical and Mathematical Sciences
UMSS21 Interdisciplinary Research
UMSS21 Business, Education, and Art
Links provided will take viewers to UMaine’s Kaltura video gallery. Presentations available for viewing by UMaine users only are noted (*).
0501. Production of Renewable Fuels from Biomass-Derived TDO Oil
Graduate Presentation
Author(s): Matthew Kline, Sampath A. Karunarathne, M. Clayton Wheeler Mentor(s): M. Clayton Wheeler
Abstract: As governments around the world place more restrictions on the production of energy from fossil fuels, the demand for fuels from renewable sources is projected to increase. One of the most promising materials to convert into bio-oil is cellulose, which is advantageous because residue from logging, farming, or even municipal waste are sufficient feedstocks. Several pathways have been developed to convert cellulose into bio-oil, one of them being Thermal DeOxygenation (TDO), which shows promise. It is a process developed at the University of Maine that converts organic acids from cellulose hydrolysis and dehydration into a low-oxygen bio-oil with a broad boiling point distribution.
While the chemical compounds in TDO oil could be directly used to produce gasoline, they lack some of the characteristics of petroleum middle distillates, such as jet fuel and diesel. In order to replace these fuels, TDO oil must be upgraded using catalytic hydrogenation and ring opening steps.
This presentation discusses the chemistry and reactions required to produce renewable gasoline, diesel, and jet fuels for commercial use. Past, present, and future work on upgrading TDO oil will be discussed, as these extra steps are integral in producing an array of transportation fuels. The presentation will also discuss my current research focused on the hydrogenation and ring opening of bicyclic molecules to produce fuels with better diesel combustion characteristics.
0502. Professorly
Undergraduate Presentation.
Author(s): John Baker Mentor(s): Jon Ippolito
Abstract: From the perspective of a student, getting to connect with teachers was one of my favorite parts about in person learning. No matter the college, subject, or course, every single one of my teachers had awesome experiences that I was happy to hear about. My project was created with the goal to reignite the connection between professors and students that is often absent from online learning. This platform would allow professors to create personalized profiles that display things like interesting research experiences, photos, hobbies, or anything else they think would be useful to future/current students to learn about them. We also want students to be able to discover new classes that fit their interests that they normally never would have, by searching for keywords/categories. Finally, we want to add a review system that improves on something like RateMyProfessor.com, by promoting positivity and encouraging constructive feedback. One very useful scenario we saw our platform being used would be during syllabus week, where there often is no homework except for students to pour over a pdf of the course structure. Professors could bring a lot of life to their syllabus and create a connection with students by including a link to their profile in their syllabus. We plan on adapting to professor and student feedback as we iterate our project, but we are very happy with what we’ve already created so far.
0503. Thermoplastics 3D Printing Using Fused Deposition Modeling on Fabrics
Graduate Presentation.
Author(s): Maxwell Blais, Scott Tomlinson, Bashir Khoda Mentor(s): Bashir Khoda
Abstract: The creation of large objects by additive manufacturing is something that is desired, but often is unachievable due to the size of the object and capacity of the 3D printer used. Beside additive manufacturing multifaceted single objects will produce a non-homogeneous performance in the object due to directional anisotropic of additive manufacturing process. To address this issue research has be done are part segmentation with various techniques, such as origami, geometric segmentation, and segmentation with manufacturability. However, joining or connecting those segmented or discretized additive manufactured parts can become an issue. Mechanical fasteners, glue, interlocking parts, or hybrid technique are commonly suggested for connecting multiple smaller parts to create a large object. Issues with these methods are that they take additional time and may add some additional weight and be cumbersome to transport and assembly. In this paper we propose to use fabric as a flexible joint and segment carrier when creating larger objects by additive manufacturing. Specifically, segmented parts of the desired large object will be additive manufactured on top of a fabric as to adhere the two. To do this, first we apply a segmentation algorithm to discretize a large object considering manufacturability, mechanical isotropy, weight, fabrication time, and printing envelope. Once segmented a machine-readable file is constructed suitable for open source 3D printer. Treated and untreated fabrics are prepared simultaneously so that parts could be created on top of them. Segments are created and adhesion force between the segment and the fabrics are tested with Instron adhesion tests. We found that untreated cotton duck cloth had an average 78% higher adhesion than other samples with the applied polymeric material, PLA plastic, from additive manufacturing a part on the samples. We also found that bed and nozzle temperature and distance between fabric and nozzle during printing played a significant role in the measured adhesion force. Adhesion forces desired between fabric and 3D printed part could be tailored as needed per specific large object as needed using this information. This material is based upon work supported by the US Combat Capabilities Development Command Soldier Center (DEVCOM Soldier Center) under Contract No W911QY-18-C-0101. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the DEVCOM Soldier Center. Approved for Public Release –DEVCOM SC PAO U1-357
0504. Electrical Power System for CubeSats
Undergraduate Presentation.
Author(s): Joseph Patton Mentor(s): Ali Abedi
Abstract: MESAT1 is a CubeSat that was proposed by the University of Maine in response to NASA’s CubeSat Launch Initiative, and was selected by NASA to fly aboard a NASA rocket in 2022. One of the most important factors in prolonging the lifetime of a CubeSat satellite is the design and testing of the electrical power subsystem (EPS). The EPS is a microprocessor-based device that tracks and controls power generation and distribution for all subsystems and payloads of the spacecraft. This study covers the design, building, and testing of a low-cost EPS system that will safely and efficiently power the MESAT1 for the entire duration of the mission.
0505. Joining Methods For Continuous Fiber Reinforced Thermoplastic Composites in Structural Applications
Graduate Presentation.
Author(s): Andrew Moran Mentor(s): Roberto Lopez-Anido
Abstract: The potential of joining methods for continuous fiber reinforced thermoplastic (CFRTP) composites in structural applications is investigated. The joining methods considered are: adhesive bonding, secondary forming and mechanical fasteners. Carbon fiber-Polyphenylene sulfide (PPS), a type of engineered thermoplastic, unidirectional composite tape was selected to consolidate plates with quasi-isotropic layups. Both composite-composite and aluminum-composite joints were examined using experimental lap shear evaluations. The experiments serve to characterize the mechanical properties required for structural design using the proposed joining methods: adhesive bonding strength, secondary forming interlaminar strength and fastener bearing strength. In addition, a comprehensive program of standard tests for material characterization of the CFRTP composite selected were conducted to generate properties for structural analysis. A structural model of a connection in a structure was developed using finite element analysis. The experimental findings on joining methods, and recommendations on design optimization of connections with CFRTP composites are discussed.
0506. Magnetorquer Design for Small Satellites
Graduate Presentation.
Author(s): Travis Russell, Ashanthi Maxworth Mentor(s): Ali Abedi
Abstract: Cube satellites, or CubeSats, are small satellites that are commonly developed and used by universities and private companies to carry out scientific missions in Low Earth Orbit (LEO). These satellites include various subsystems that work together harmoniously to ensure the success of a mission. One such subsystem is called the Attitude Determination and Control System (ADCS), which is responsible for sensing and maintaining a desired orientation of the satellite while it is in orbit. It is often the case that both the radio communications system and the success of the science mission depend on proper orientation, making the reliability and fidelity of this subsystem of utmost importance. This research covers the design and testing of low cost magnetic actuators, called magnetorquers, which can control the orientation of a CubeSat by their interaction with Earth’s magnetic field.
0507. Operation and Preliminary Energy Balance of a Portable Kelp Dryer in Maine
Graduate Presentation.
Author(s): Tuqa Al-Asadi, G. Peter van Walsum Mentor(s): G. Peter van Walsum
Abstract: Seaweeds are important commodities as food or raw material for additives and biopolymers. One popular species native to the Maine coast is Sugar kelp.,. It is a brown alga, rich in fiber, minerals, and antioxidants. Seaweeds are highly perishable due to their high moisture content and will spoil quickly if not preserved, therefore a common practice is to dry the product to prolong storage life and to minimize the cost for transportation. A drying method using warm air will extend seaweed shelf life and retain many of the valuable bioactive components that are heat sensitive. Thus, a controlled drying environment can best retain product value and prevent degradation. A prototype commercial scale dryer, housed in a portable 40’ shipping container, was developed to provide this environment. My research is focusing on examining the drying rate,, temperature and humidity, and energy efficiency of the process. Experiments were run using wet towels or Sugar kelp in the dryer. Examination of the drying runs shows that as air passes through the drying chamber it decreases in temperature and gains humidity but does not reach saturation. The temperature of the internal air was affected by the degree of air recycle and furnace activity. Simultaneously, as the drying progressed the exiting air showed a steady trend of rising temperature and decreasing RH. Analysis of the run data suggest that improving interior air ducting and insulating the walls of the container (except on bright sunny days) will likely be cost effective.
0508. Search and Rescue Sled Suite
Undergraduate Presentation.
Author(s): Jesse Bosdell, Liza White, Caroline Bray Mentor(s): David Neivandt
Abstract: Existing rescue sleds lack the equipment necessary to assess and treat patients while communicating with first responders during transport. There have been some minor improvements to existing sleds to protect the patient and provider from the weather, however, current solutions inhibit the ability to treat the patient. The purpose of the current project is to modernize the existing, commonly used rescue sled to allow providers to assess and easily treat the patient while also being able to communicate with outside agencies to further enhance the overall patient outcome. In order to accomplish the project objectives, subsystems were created to target patient care, comfort, and communication. Specifically, a vital sign monitoring system was designed to allow the providers to access and track trends in the patient’s health. Additionally, a suspension harness system was created to reduce patient vibration and impact transmission during sled movement; the suspension system incorporated heating elements to greatly improve patient comfort. Finally, headphones and a microphone were added to the existing Motorola communication system to allow for safe and effective transmission of patient condition and location. Once the subsystems were created, the entire unit was tested to validate patient safety, effectiveness, and functionality. Future tests will assess the safety for the medical provider, the capabilities of the communication system, and the effectiveness of the hardware and software that is under development to facilitate the automated uploading of patient vital signs to an online database. The work to data has demonstrated that existing rescue sleds can be modernized to greatly improve patient care, communication, and ultimately, patient medical outcomes.
0509. Investigation of Bioaerosol Catch-and-Release Dynamics via Liquid Net Filtration
Graduate Presentation.
Author(s): Dan Regan, Junie Fong, Justin Hardcastle, Caitlin Howell Mentor(s): Caitlin Howell
Abstract: Materials and methods of collecting bioaerosol for performing analysis on potential biological threats is at an all-time high after over a year of battling COVID-19. Currently, filter-based air samplers capture the highest yield and ratio, while liquid impingers offer the best chance to maintain pathogen viability. In this body of work, we present a novel approach to filter-based filtration by fabricating liquid nets. Infusing polytetrafluoroethylene filters with a perfluoropolyether oil resulted in improved rate of release of captured Escherichia coli aerosols onto culture plates. Furthermore, wetting meltblown polypropylene filters with the same perfluoropolyether oil improved the rate of release, as well as the overall colonies transferred and cultured. The efficiency of bioaerosol transfer demonstrated by the liquid-infused filters could provide a better representation of the airborne pathogens within an environment, enabling informed actions by infectious disease experts and medical professionals.
0510. Nanocellulose Conduits for Enhanced Regeneration of Peripheral Nerve Injuries
Graduate Presentation.
Author: Nicklaus Carter Mentor(s): Dr. David Neivandt
Abstract: Peripheral neuropathy is estimated to afflict 20 million people in the United States. Most cases of neuropathy result from physical injuries and trauma arising from automobile accidents and war. Peripheral nerves have the intrinsic ability to regenerate over time, bridging the injury gap resulting from trauma. Current methods utilized to assist in the regeneration of peripheral nerves include nerve grafts and neural conduits. Nerve autografts are regarded as the most effective method but require a second surgical site to access a donor nerve or a nerve donation from another individual. Current available conduits have equal or lower success rates relative to nerve grafts with specific issues including immune response and stability insufficiencies. It has been proposed that a biocompatible material such as cellulose nanofiber may serve as a viable alternative conduit material. Preliminary studies have shown that cellulose nanofiber conduits are successful in aiding neural regeneration and further, that conduit length has an impact on efficacy in murine models; it is hypothesized that the length dependence may be related to modified diffusion distances of key cellular nutrients and waste metabolites. The present work investigates the concentration profile of a key nutrient, glucose, within the conduit. A finite element analysis of the conduit system has been established using COMSOL Multiphysics. Variations in the physical dimensions of the conduit were investigated to determine the impact on glucose concentration profiles. The resultant information is being used to aid in the development of improved conduit designs to optimize functional recovery of peripheral nerve injuries.
0511. Dynamic visualization of past and future forest change within the Maine Forest Ecosystem Status and Trends Web Application
Undergraduate Presentation.
Author(s): Scott Audet, Eric Schessler, Ethan Esber, Devin Merrow, Domonic Pacelli Mentor(s): Kasey Legaard
Abstract: The Lorax are among a handful of computer science capstone software development teams for the 2020 – 2021 academic year. Our team consists of Scott Audet, Ethan Esber, Devin Merrow, Eric Schessler and Domonic Pacelli. For our project, we elected to work on a web application to support the Center for Research on Sustainable Forests, led by UM faculty Kasey Legaard and Erin Simons-Legaard. The application is titled the Maine Forest Ecosystem Status and Trends (ForEST) App. The ForEST App is an online web mapping application that allows users to monitor the spruce budworm outbreak tracking out of Canada into Maine. Within the application, users can view current measures from budworm trap sites and maps of vulnerable forest resources across the state. This vital information is provided to Maine’s foresters and forest stakeholders to help mitigate the effects of the outbreak and to protect the state’s natural resources. Our team has been tasked to expand upon the existing web application in order to improve the delivery of information to its users. The purpose of a capstone project in computer science is to expose students to professional software development practices. Our team has learned to deliver high quality software in a remote environment using continuous integration techniques such as agile development. The team plans to complete testing and deliver the enhanced application toolset by the conclusion of the semester. These tools will allow users to obtain greater information on the outbreak, including predictive models displaying future impact on Maine’s forests. The Lorax hope this suite of features will improve upon Kasey and Erin’s efforts and benefit future users of the ForEST App.
0512. Precision Actuator for Small Scale Robotics
Undergraduate Presentation.
Author(s): Patrick Murphy Mentor(s): Bashir Khoda
Abstract: Recent developments have made robotics more affordable, like the >$500 actuator developed at MIT. However most robotic actuators are inaccessible to individuals due to the still relatively high cost. In this work, fused deposition modeling 3D printing along with inexpensive brushless dc motors are studied to determine the viability of such methods for the creation of a precision small scale robotic actuator. There is little information about the use of fused deposition modeling for drive components. To overcome the limitations in using such a manufacturing method the gear reduction utilizes a cycloidal gear drive, known for it’s low backlash and compact size. The profile of such a gear is modeled as a parametric curve. Inexpensive ‘hobby’ garde brushless dc motors offer high efficiency and torque density, but operate at higher rpm requiring gear reduction . Initially a 300Kv 42mm by 20mm motor was selected. The open-source, closed-loop brushless dc driver known as ODirve was utilized. Following development and simulation of 5 ‘path-finding’, collective findings were put into the design of Prototype 1. This is the first interaction that had the intention of being built and tested. Motion studies and static load simulations showed the design should perform within expectations, maintaining a safety factor >2. Upon review of these findings Prototype 1 was prepped for testing. The cycloid drive performed as designed when driven by hand. During preliminary testing and setup it was noted that the motor was running hot. While current was limited to 50% the motor underwent a thermal event while trying to reach 500rpm. This resulted in the motor burning out and the deformation of the surrounding plastic fused deposition modeling parts. This led to a major redesign Prototype 2, with a larger motor with a 260% increase in rated current. The assembly was designed to be easily fabricated with a desktop 3d printer and is expected to deliver precise control of motion. Total cost for the actuator is under $120 which can be a great option for the diy community as the design will be open source and available for download form various repositories. In future, a study on the friction coefficient between FDM parts would need to be performed before such a motion study, to ensure high fidelity of the time consuming operation. Contingent on the outcome of experimental static load testing, further research into improving the fidelity of FDM CAD models is also proposed.
0513. High Shear, High Residence Time Drying of Cellulose Nanofibrils
Graduate Presentation.
Author(s): Brandon Dixon, Amos Cline Mentor(s): David Neivandt
Abstract: Cellulose nanofibrils (CNFs) are biomass materials on the nanoscopic scale derived from wood. CNFs are hydrophilic and are typically produced as aqueous slurries. The CNF slurry can be employed in its aqueous form or dried in two- and three-dimensional forms for a variety of applications. The structure of UMaine produced CNF is highly hierarchical with interconnected branches of decreasing size ranging from macroscopic diameter pulp fibers, down to nanoscale diameter fibrils. Upon conventional drying of CNF hydrogen bonding between the fibers and fibrils results in a very large degree of aggregation, and concomitant loss of the nanoscale elements of the material. Supercritical drying employing a solvent change and use of carbon dioxide is capable of drying CNF without aggregation, however it is a challenging technique to scale up. Similarly, spray drying limits the CNF aggregation to create micron scale beads that may have some utility, but again scale up is problematic. The goal for the current work is to create a novel, efficient, and cost-effective method of rapidly drying and separating CNF into individual fibril forms. Work to date has employed high shear generated from a rotor and stator system, coupled with gaseous over-pressurization and release, and modest heating to provide energy and separation force to dry CNF while minimizing fiber aggregation. Prototypes constructed thus far have shown promise, but have indicated that higher shear and longer residence times in the rotor/stator are likely required. The dried CNF powder will be incorporated into wood polymer composites for large-scale 3D printing.
0514. Do Virtual Flavors influence Liking and Emotions?
Graduate Presentation.
Author(s): Meetha Nesam James Mentor(s): Nimesha Ranasinghe
Abstract: Flavors of food and beverages are closely associated with a lot of our daily activities. Food researchers have worked on changing the chemical formula/composition, calorie configuration, the ingredient list of the food to enhance the sensory liking and preference of the food being consumed. By altering those features, we lose the real experience of the food along with health side effects as well. Hence, the challenge is to simulate these sensory experiences without altering food chemically. Virtual flavors refer to flavors that are stimulated externally without chemically or physically altering the food but by digitally augmenting it with the help of different electric stimuli for taste, aroma dispersion for smell, light stimuli for color, thermal stimuli for heat and cold perceptions. However, in HCI (Human-Computer Interaction) research, little is known about the influence of various sensory modalities such as taste, smell, color, and thermal towards perceiving simulated flavor sensations and on people’s emotions, liking, and feeling of pleasure. To evaluate this white space, we present 1) an interactive technology to simulate different flavor sensations by overlaying taste, smell, color, and thermal sensations on plain water externally, and 2) a set of experiments to investigate a) the influence of different sensory modalities on the perception and liking of virtual flavors and b) varying emotions mediated through virtual flavor sensations. Our findings reveal that the participants perceived and liked various stimuli configurations and mostly associated them with positive emotions while highlighting important avenues for future research.
0515. Techno-Economics and Life Cycle Analysis of a Biofuel Blend-stock Process
Graduate Presentation.
Author(s): Aysan Najd Mazhar, Thomas Schwartz, Sampath Reddy Mentor(s): M. Clayton Wheeler
Abstract: Techno-Economics and Life Cycle Analysis of a Biofuel Blend-stock Process Aysan Najd Mazhar, Thomas Schwartz, Sampath Gunukula, M. Clayton Wheeler Abstract The most significant sources of energy for many countries are fossil fuels and their derivatives. Due to environmental impacts of these nonrenewable fuels, finding substitutes for them are of high interest. Fast pyrolysis process has the potential to offer high efficiencies to produce liquid transportation fuels. This process is considered as a direct thermochemical procedure that can liquefy solid biomass into liquid fuels such as gasoline and diesel blend-stocks. The process includes rapid heating of biomass particles in the absence of air at approximately 500°C with the production of non-condensable gases, bio-oil or pyrolysis oil, and char. The pyrolysis oil can be upgraded via integrated mild hydrogenation and etherification processes to a high energy density fuel that can be blended with diesel for transportation use. In this study, a process model to upgrade low energy density forest residues to a high energy density fuel using integrated fast pyrolysis, mild hydrogenation, and etherification processes will be created. The ASPEN plus simulation is used to determine the material and energy balances of process model. Thermodynamic properties of the hydrogenation and etherification model compounds were estimated by using Density Functional Theory (DFT). The feedstock (sawdust) processing capacity of 2000 dry metric ton per day and capacity factor of 0.9 were assumed for process simulations. The process equipment is sized using the simulated material and energy balances. The discounted cash flow analysis with an internal rate of return of 10% is selected to assess the minimum selling price of high energy density fuel. The life cycle analysis (LCA) has been conducted to assess environmental sustainability for producing a renewable high energy density fuel. The ASPEN simulations have shown that the hydrogen requirements of the mild hydrogenation of bio-oil is met through the steam reforming of non-condensable gases of fast pyrolysis. The amount of high energy density fuel produced is 54 million gallons per year and the assessed minimum selling price is $2.63 per diesel gallon equivalent. The LCA well-to-wheels analysis, performed using Argonne’s GREET 2019 software, indicates that the renewable diesel blend-stock would have a 92% reduction in fossil fuel use and a 93% reduction in greenhouse gas emissions relative to petroleum diesel.
0516. A Birefringence Semantic Segmentation of Muscular dystrophy in Zebrafish Using Deep Neural Networks
Graduate Presentation.
Author(s): Ahmed Almaghasilah Mentor(s): Clarissa Henry
Abstract: In any typical lab that studies the molecular level of zebrafish, such as investigating the integrity of the dystrophin glycoprotein complex and its impact on muscle health and development, it can generate over hundreds of Birefringence images on a monthly basis. Manual segmentation of muscle segments, called myotomes, is a tedious task, full of errors and biases. In addition, it demands collaborators to be familiar with the software tools, as well as the anatomy of the zebrafish – which can be significantly hard to distinguish especially when the muscle fibers sustain a severe muscle degeneration. The severe muscle degeneration tends to obscure collaborators’ judgment and increase the error rates, because the myotendinous cannot be phenotypically observed. Deeplab v3+ network, a convolutional neural network (CNN) designed for semantic image segmentation, with a pre-trained ResNet-50 architecture delivers consistent and superior segmentation outcomes. The CNN makes a prediction of muscle segments with 95% accuracy, compared to the manual segmentation annotated by various collaborators. The CNN classifies each individual pixel and renders a fully segmented overlay of a Birefringence image in less than a minute. The prediction produced by the deep neural network accelerates the process of quantification, analysis and validation streamline of wet-lab experiments faster than before.
517. Tuning CNF Fibril Orientation for Tissue Integration Applications
Author(s): Josh Hamilton, Sahar Roozbahar, Omar Alsamsam, Dr. Michael Mason Mentor(s): Dr. Karissa Tilbury
Abstract: Internal injuries can require multiple surgeries creating extra risk for the patient and increasing recovery times. Absorbable scaffolds for wound repair may mitigate the need for multiple surgeries, however, there is a lack of low-cost solutions in the market that meet high force applications. A readily produced fibril material at the University of Maine known as cellulose nanofibers (CNF) fulfills both the mechanical and biological properties required for high force applications such as bone fixation devices. The purpose of this work is to develop a methodology to tune and quantity CNF mechanical properties thereby mimicking different tissue environments for future tissue integration applications. To do this, a novel orientation strategy using a constant unilateral force on CNF films drying on a flexible substrate was developed. Changing the force applied allows adjustment of CNF anisotropy which directly affects its mechanical properties. A quantification methodology was also implemented using the natural birefringence of CNF. An orientation index based off film birefringence was used to quantify CNF. Maximum applied force films had a birefringence orientation index (BOI) that was 80% higher than non-force applied films. Another imaging tool is needed to resolve individual fibers and determine if the BOI is a viable tool for elucidating fiber orientation. To accomplish this, SEM images of both force dried, and non-force dried films were taken, and fiber orientations in these images were analyzed using a FIJI plugin called OrientationJ. This determined the BOI is an effective and readily available tool to determine CNF fiber orientation.
0518. Design and Manufacturing of Printing Head for Aligning Short Fibers
Undergraduate Presentation.
Author(s): Miao Hu, Akhter Zia Mentor(s): Yingchao Yang
Abstract: 3D printing has been widely adopted to manufacture a broad of composites. If feed materials are polymer, particles or 1D fibers, individuals can be with any orientations after landing on the print bed. The obtained composites will be isotropic. For the fiber reinforced composites, generally, the fibers are expected to be aligned in order to maximumly inherit the mechanical performance of the fibers. However, the commercial printer is unable to print fibers with controlled alignment but random distribution only to date. Our project aims at combining the shear force theory with the 3D printing head to constrain fibers to parallel each other. The project includes three major tasks. (1) The printing head will be fabricated with rifling which is inspired by the rifle barrel imparting spin to the projectile along the longitudinal axis to stabilize projectile flight; (2) Regarding that the rifling is very challenging to be fabricated inside of the printing head, we plan to cut the printing head into halves and use laser to sculpture the designed refiling; and (3) A stationary aligner will be designed to have all fibers align along certain direction when the fibers come out from the printing head. Before fabrication of the printing head, the Solidworks software will be used to design a set of rifling modals and apply 3D printer to print the rifling. The printing head will be then used to print short carbon fibers, where the alignment of the fibers will be evaluated by an optical microscope and scanning electron microscope. If necessary, the length, pitches, depths, shape of the rifling will be updated depending on the alignment of the printed fibers.
0519. Characterization of electrically-conductive filament for FDM 3D-printing
Undergraduate Presentation.
Author(s): Mackenzie Ladd Mentor(s): Brett Ellis
Abstract: With the recent introduction of electrically-conductive filaments for fused depositional modeling (FDM), many new avenues have opened for novel design methodologies and processes. An example of such a filament is the ProtoPasta Conductive PLA, which is a polylactic acid polymer substrate infused with conductive carbon black particles, thus imbuing conductivity. The biggest barrier for more widespread use of this new material is a lack of characterization. This research seeks to address this problem by characterizing an electrically-conductive filament and comparing empirical results with manufacturer-produced data and the limited amount of data available in literature. The effects of print geometry and orientation were studied by testing specimens of varying designs and printing methods, and probed using a multimeter to gather resistivity data for each of the studied methods. An emphasis will be placed on contact resistances, as no previous study was identified that analyzed the influence of contact pressure on the resistivity at the probe-tip to artifact divot interface. This will be characterized using a load frame to apply pressure using multimeter leads, thereby defining a force-resistance relationship. This research will facilitate the optimization of 3D-printed designs to achieve preferred material properties. Additionally, such data will enable new design methodologies, including, but not limited to implementing conductive segments within compliant mechanisms, thus adding even more capabilities to compliant mechanisms’ enhanced wear characteristics and reduced manufacturing costs.
0521. Flight Computer Design for Satellites
Undergraduate Presentation.
Author(s): Steele Muchemore-Allen Mentor(s): Ali Abedi
Abstract: In any Cube satellite, or CubeSat, the center of operation is the on-board flight computer. This project covers the design and implementation of a microprocessor-based device to function as the flight computer and data handling unit for the MESAT-1, Maine’s first Cubesat to be launched in 2022. The central flight processor monitors and maintains functionality of the CubeSat as a whole, ensuring each subsystem is operating as designed. Another important function of the flight computer is to collect and manage data from various on-board sensors and the specific science payload. The flight computer will collect and sample data, and respond to any necessary subsystems accordingly. The aim of this research study is to develop a system that will allow the satellite to function properly for the entire duration of the mission.
0522. Relay-Assisted Wireless Energy Transfer Scheduling with Dual Data-Energy Channel Models
Graduate Presentation
Author(s): Sonia Naderi Mentor(s): Ali Abedi
Abstract: Relay energy assisted communication where a transmitter is powered by an energy source through both direct and relay links is considered in this project. Data transmission based on stochastic models for data and energy channels is considered. Various static, mobile and highly scattered channel models are studied to cover a wide array of applications where battery replacement is not feasible. A threshold on required transmission energy and channel quality to decide whether the transmission is beneficial or risky is derived for all possible scenarios. All theoretical results are validated by numerical simulations and verify the effectiveness of energy relaying and proposed energy efficient scheduling method in reducing the outage probability of the system.
0523. Bioinspired Liquid-Infused Membranes Reduce Biofouing
Graduate Presentation
Author(s): Justin Hardcastle, Daniel Regan, Junie Fong, Jessica D. Schiffman, Rushabh Shah Mentor(s): Caitlin Howell
Abstract: Polymer membranes that are widely used for water treatment and wastewater reuse are prone to fouling or the accumulation of particulates and bacteria. Fouling significantly reduces process productivity and additionally, to clean fouled membranes, a combination of physical and chemical processes are used that increase process downtime and degrade the membranes. The purpose of this work is to use surface science inspired by the pitcher plant to create a new approach to membrane design that will reduce the adhesion of particulates on membranes, providing a transformative new way to prevent membrane biofouling. First, we tested membranes with two different chemistries and pore structures, polytetrafluoroethylene (PTFE) and polyvinylidene fluoride (PVDF) for their ability to maintain a stable, continuous liquid coating of a perfluoropolyether fluorinated liquid (DuPont KrytoxTM). Specifically, we measured the continuity of the surface liquid layer by testing how easily a water droplet could begin moving on a surface and the anti-adhesion properties of the surface liquid layer through quantifying the speed of droplet movement at a static angle of inclination. The results indicate that the PTFE membranes could sustain a more consistent liquid layer, with an approximately 75% lower sliding angle and 70% faster droplet movement at a fixed angle than the PVDF. Qualitatively, the liquid coated PVDF membranes show signs of visual dewetting in air over time. Analyses also showed a difference in the quality of the liquid layer depending on the amount of residual water present in the membrane at the time of coating. Interestingly, dead-end pure water permeability (PWP) experiments conducted at an applied pressure of 1.5 bar indicated the liquid coated membranes had a statistically equivalent PWP of 2827 ± 323 L/m2-h-bar over 10 cycles. In static biofilm formation assays, infused PTFE reduced biofilm formation by approximately 98%, while liquid coated PVDF reduced bacterial attachment by approximately 25%. These results are the first steps toward a transformative new approach to prevent filtration membrane biofouling.
0524. Solid Particles Transfer by Dipping from Heterogeneous Mixture
Graduate Presentation
Author(s): S M Abu Naser Shovon Mentor(s): Bashir Khoda
Abstract: Dip coating, a wet deposition method is an effective and straightforward way of thin-film or layers formation. It is extensively used as a coating method due to its simplicity, low cost, and reasonable control over the thickness. In this work, the physical phenomenon of dip coating process with negatively buoyant, non-Brownian micro-particles from density mismatching mixture is investigated. The inorganic particles provide spherical geometry with an average diameter of 5.69 µm. A liquid carrier system (LCS) is prepared by a polymer-based glue and an evaporating solvent. The concentration of polymer glue was varied between 2.5% to 10.5%. The rheology of LCS shows a Newtonian characteristic throughout the shear rate ranging from 0.2 to 1000 s-1. Three different volume fractions (20%, 35% and 50%) of inorganic particles are added to LCS solution and a pseudo liquid carrier particle suspension is prepared by using a magnetic pellet moving at minimal suspending speed from 500 to 700 rpm. The cylindrical AISI 1006 mild steel wire substrate is dipped at different withdrawal velocity ranging from 0.01 mm/s to 20 mm/s.
Once the cylindrical substrate is extracted from the heterogeneous mixture, the coating thickness is measured in our lab. The average thickness as well as the surface packing coverage by the particles is increasing with higher volume fractions of particles in the mixture. Moreover, for same volume fraction of particles, the coating thickness is increasing with high polymer-glue concentration and for the substrate withdrawal speed. The coverage of inorganic particles in the coated substrate is measured at 20% to 35% at lower volume fractions. This coverage referred to the heterogeneous regime with low suspension viscosity, and cluster of particles covered randomly almost all the coated area which also follows closely with Landau-Levich-Derjaguin (LLD) law. In this regime the capillary number, Ca remains ≤ 3×〖10〗^(-3). More inorganic particles are entrained in the substrate at volume fraction of 50%, and the thickness (>35 µm) becomes almost uniform through the entire coated area. This high-volume of particles is referred to the effective viscous regime with capillary number, Ca over 4×〖10〗^(-3). The results present the influence of volume fraction and substrate withdrawal velocity on the solid transfer from the heterogeneous mixture. Solid transfer technology from mixtures is gaining ever-increasing attention from both materials scientists and production engineers due to their high potential in near net shape production of cost-effective engineering components. The result of this research may help understanding the high-volume solid transfer technique and develop a novel manufacturing process.
0525.Developing Software for Aerial Imagery Handling and AI Analysis Management
Undergraduate Presentation.
Author(s): Devin Christianson, Nick Kania, Jacob Morin, Alex Feren, Kyle Walker Mentor(s): Dr. Cynthia Loftin
Abstract: The BirdSpotter web application is designed to integrate machine learning into the analysis workflow for estimating the population of nesting birds using aerial imagery. BirdSpotter is built for use by biologists, to allow rapid population measurement without the added overhead of manual categorization. It allows users of the app to view data on a map as well as in the aggregate, and also allows for the importing or exporting of processed datasets. To automatically categorize the behavior of nesting birds, the project uses a machine learning algorithm created by members of the University of Maine Graduate program. BirdSpotter is being developed using agile methods in a modified V development model, with frequent client involvement from initial requirements specification to prototyping and validation. We learned a lot during the initial design process, in trying to solve one of the most common dilemmas in software: how to design a system that meets the client’s needs, without introducing unnecessary complexity or overhead. During development we had to overcome several unique challenges, such as handling large files, managing sensitive data, and interfacing effectively with external systems. Testing and validation was also a challenge when a system interfaces with other external systems, and has high performance requirements, but a comprehensive testing plan with a strong continuous integration and delivery pipeline helped keep this validation overhead lower.
0526. MLA Survey Analyzer
Undergraduate Presentation
Author(s): Benjamin Patashnik, Alex Gay, Mark Sullivan, Jeremy Thiboutot Mentor(s): Erin Vinson
Abstract: Our project, the MLA Survey Analyzer, will be a tool designed to improve the Faculty Course Modification Incentive Grant – Maine Learning Assistant (FIG-MLA) program survey process at the University of Maine. Currently, surveys administered to students to track the progress of the FIG-MLA program are manually modified, sanitized, and analyzed by employees at the Maine Center for Research in STEM Education (RiSE Center) who have requested a software solution to automate the process and enhance the efficiency of survey analysis. The MLA Survey Analyzer will be able to quickly and automatically de-identify, organize, analyze, and store survey responses.
This project is being developed by four senior students enrolled in the Computer Science Major at the University of Maine. Team members are working with and deriving system requirements from a contact at the RiSE Center. Separate elements of the project have been assigned to different team members, such as database integration, data analysis, and user interaction. The team is developing the system using MySQL for database management and the Python programming language for program functionality and the user interface. Program development adheres to a mixture between an agile methodology and the V-model.
The expected outcome will consist of an executable program that will be accessed by the user. The user will be able to import new survey data into the database, or view data stored on the database. The program will automate the process of importing data from surveys, while simplifying the process used to sort and analyze data. Finally, the program will help generate reports for feedback used to improve the RiSE Center’s FIG-MLA program.
0527. Electrical Impedance Spectroscopy Based Food Recognition to Facilitate Food Journaling
Graduate Presentation
Author(s): Chamath Lochana Amarasinghe Hiti Hamillage, Nimesha Ranasinghe Mentor(s): Nimesha Ranasinghe
Abstract: The rate of chronic health conditions has increased over recent decades due to unhealthy food choices and diets high in processed sugar and sodium. At present, people are more concerned about their food intake and maintaining food journals to monitor their diet content. These food journals help individuals self-manage chronic diseases and their doctors to understand eating habits and answer critical questions to make informed decisions. Some of these key questions are associated with the type of the food, quantity, and the time users are eating. Thus, recognizing the food and its internal attributes is necessary for automatic food journaling. As a solution, contemporary research focuses on recognizing food based on image processing, where AI algorithms analyze images to identify and record food items. However, current solutions are ineffective in identifying internal attributes and quantification of food and beverages.
This work’s broad aim is to study the effectiveness of Electrical Impedance Spectroscopy (EIS) combined with AI algorithms to develop novel non-visual food and ingredient detection technologies to enable new possibilities in the domain of Human-Food Interaction. A smart mug that recognizes the liquid (i.e., a beverage), the sugar content of the liquid, and its volume were the initial study’s centerpiece. It employs two stainless steel electrodes connected to an impedance analyzer to measure the liquid’s impedance. A classifier that employs Support Vector Machines analyzes the impedance data to identify the liquid, and an Artificial Neural Network predicts the volume and sugar content.
0529. A Smart Wheelchair Cushion to Reduce the Incidence of Pressure Sores in Low-Mobility Populations
Undergraduate Presentation.
Author(s): David Flewelling, Malak Helal, Michael Gray Mentor(s): Michael Mason
Abstract: Pressure sores are a common issue for patients who spend extended periods on a wheelchair or lying down. Wheelchair cushions attempt to alleviate this problem by distributing pressure, but do not perform effectively for patients whose immobile periods are extended. This work aims to develop a reasonably cost-effective cushion replacement that distributes pressure to reduce or eliminate pressure sore formation. The proposed design incorporates a memory foam top and an air cushion with an adjustable and self-filling pressure system. The device also includes a smart electronic system which can automatically determine when the patient needs to self-adjust or be moved in the seat by an attendant. The entire device will be encased in a waterproof nylon case to maintain the cleanliness and usability of the device in the case of incontinent users, as well as to protect electronic components. Prototypes were assessed by the impact of the device on comfort, the frequency of self-adjustment by the user, and averaging user-reported comfort with the device. These results confirm that this device can measurably reduce the formation of pressure sores by decreasing the pressure exerted on tissues, and making it easier for medical attendants to recognize when a user needs care to prevent sore formation.
0530. Binary Maze Game
Undergraduate Presentation.
Author(s): Ethan Mason, Jeffrey Edrington Mentor(s): Andrew Sheaff
Abstract: This report describes the design of a microprocessor-based device that takes user input and
to navigate a sixteen step maze was designed, built, and tested. The device features seven
available inputs. One for turning the device on or off, one to reset the current state, one
to change the level, and four to navigate the game. Initially all indicators are on and the
goal is to turn the indicators all off. Each navigational input corresponds to an indicator.
The user may only interact with the input if its current indicator is on. When an indicator
in the on position is interacted with in this way the device changes to a state where the
interacted indicator is off but the other indicators can change or not change depending on
the predetermined maze. The users goal is to get the state of the device from the starting
state of all indicators on to the final state of all indicators off.
0531. Smart Tourniquet: A New Approach to Existing Technology
Undergraduate Presentation
Author(s): Darren Hanscom, Emily LeClair, Tian Morrison, Chris Toothaker, Lisa Weeks Mentor(s): Michael Mason, Robert Bowie,
Abstract: Tourniquets have been used for thousands of years with little change to their design and function; though the ability of tourniquets to save lives and limbs is undeniable, prolonged application leaves patients at risk for severe nerve and tissue damage. Further, they lack the ability to provide feedback to first responders regarding duration of application, heart rate, and pressure applied to the limb. Therefore, the purpose of this design is to augment current tourniquet designs to include electronics capable of providing such information so emergency care providers can make more informed treatment decisions. To accomplish this, an inflatable tourniquet will be assembled and tested to validate its functions; the series of tests will ensure that the device has consistent ability to inflate and maintain applied pressure, stop blood flow, and store and send data via Wi-Fi, Bluetooth, and cable connection. Along with the ability to prevent limb loss, a successful design will fill gaps in current tourniquet functionality and safety, benefitting both patients and emergency care providers. A successful device will also be applicable in any emergency, including battlefield wounds, wilderness survival, and emergency clinical care.
Acknowledgments:
This work was supported by the Department of Biomedical Engineering at the University of Maine. The authors would like to thank Dr. Robert Bowie, Dr. Michael Mason, Dr. David Neivandt, Dr. Lisa Weeks, and Angela Hildreth for their advice and support throughout the duration of this project.
0532. What Personal Information is Your App Collecting?
Graduate Presentation.
Author(s): Vijayanta Jain Mentor(s): Sepideh Ghanavati
Abstract: Developers must provide privacy notices to the users when their application collects, processes, or shares their personal information. However, creating concise and consistent privacy notices is a challenging task for developers. A recent survey found that there are ~2.89 inconsistencies between the privacy notices and the personal information processed in each Android application. Such violations under regulations such as the European General Data Protection Regulation and California’s Consumer Privacy Act can lead to large fines up to millions of dollars for developers. To help them create consistent privacy notices, previous work has focused on creating notices through a questionnaire or predefined templates. Since developers may not be aware of all the personal information processed in their application, especially in third-party libraries, questionnaires can lead to inaccurate notices. Similarly, a notice template will lead to a generic privacy notice that does not contain detailed information. We propose to create privacy notices directly from the source of the Android applications where personal information is used and processed. In this work we present our approach that can identify and extract source code that processes personal information. We create a tool called PDroid that implements this approach. With this tool we can identify which personal information is accessed, how it is used in the application, and help developers create concise and consistent privacy notices using this information. We use this tool to analyze ~80,000 Android applications and extract ~300,000 code segments that use and process personal information.
0533. Investigating 3D-Printability of a Maine-Based Bio-Ink
Undergraduate Presentation
Author(s): Jordyn Judkins Mentor(s): Bashir Khoda
Abstract: Biofabrication is the process of creating complex biologic products, such as artificial tissues, from raw materials such as living cells, biomaterials, and molecules. This can be done using 3D printed bio-ink, which is a combination of biomaterials and cells. However, the bio-ink must be a shear thinning fluid to allow for high-resolution and continuous printing (printability), but also demonstrate post-printing mechanical integrity to self-support the structure (shape fidelity). Achieving printability and shape fidelity simultaneously is a challenging issue due to how the chosen biomaterials interact with each other. If the interactions are too strong, the printability will be low, but if the interactions are too weak, the shape fidelity will be low. The research conducted here investigates how to improve the mechanical functionality of bio-ink using additives available in Maine. Chitosan, which can be extracted from shrimp, sodium alginate, which comes from algae, and TEMPO nanocellulose fibers, which is a forest biproduct, were chosen as the candidate biomaterials due to their biocompatibility. The printability of the bio-ink can be determined by considering the rheological properties and printing parameters, such as pneumatic pressure and nozzle size, for numerous mixtures. This research focuses on how the mixture ratio affects the printability of the bio-ink, while also investigating the individual material contributions. To determine how chitosan changes the behavior of the ink, the rheological data of four ink compositions will be compared, and a “design of experiments” approach will be used to determine which hydrogel ink produced the smallest filament width when printed, therefore revealing which ink composition will print with best quality. The four ink compositions used are 2:1:0.1 w/v%, 2:1:0.5 w/v%, 2:1.5:0.1 w/v%, and 2:1.5:0.5 w/v% of Alginate:TEMPO-NFC:Chitosan. A flow curve, amplitude sweep, and rotational thixotropy test are conducted for each ink to gather viscosity and modulus values data. Each ink is then printed with a 3D printer using air pressure at 10 psi to force the ink through a 410 micron needle, with a printing height of 0.3 mm. The ink is printed and sprayed with a crosslinker of calcium chloride over a 5 by 5 cm area, then a picture of the ink is taken and uploaded into ImageJ, where the filament width can be determined. Based on the analyzed data thus far, increasing the amount of chitosan in the bio-ink decreases the filament width of the printed bio-ink. Looking at the rheological data, increasing the chitosan content of the bio-ink increases the ink’s viscosity, which makes sense with the corresponding filament width decrease. This means increasing chitosan content leads to a bio-ink that has higher printability.
0534. A Custom Centrifuge Tube and for the Collection of Microliter Volume from a Microneedle Skin Patch
Undergraduate Presentation.
Author(s): Cole Perry, Caleb Berry, Scott Collins Mentor(s): Rosemary Smith
Abstract: Microfabricated, silicon needle arrays are being developed at UMaine for the extraction of dermal interstitial fluid (ISF) for off-chip compositional analyses. The microneedle patch collects a minute volume of approximately 1 µL of ISF into an on-chip microfluidic reservoir. ISF removal from the microneedles has been achieved using centrifugation, whereby the liquid is expelled from the patch and pools at the bottom of the tube from where it is pipetted out. Commercially available centrifuge tubes do not permit optimal positioning of the microneedle patch inside the tube and pipetting results in a significant loss of fluid. Using a 3D computer-aided design (CAD) software, a custom centrifuge tube and cap were rendered to accommodate and position the microneedle patch. Additionally, the custom centrifuge tube is designed to enable extraction in the tip. This is achieved by puncture of a 50-micron membrane at the tip. The custom cap features a housing that positions the microneedle patch at an optimal 45° angle. A hole in the housing beneath the microneedle patch allows ISF to pass through and into the tip during the centrifugation process. While the tube and cap are designed to work together, the cap is designed to also be compatible with commercial centrifuge tubes. The tubes and caps are 3D printed at the UMaine IMRC CORE facility. The custom tubes are being tested with patches loaded with 1 µL of water in a centrifuge to ensure instrument compatibility and to assess improvements in efficacy of collection through design optimization.
0535. Compact VO2 Max Testing Device
Undergraduate Presentation.
Author(s): Zane Kwiatkowsky, Quinn Ferguson, Shanna Scribner, Nathan Loranger Mentor(s): Michael Mason
Abstract: VO2 max is the maximum oxygen consumption of a person during strenuous physical activity and is valuable for tracking athletic progress however, devices used for testing are expensive, lack ergonomics, and only cater towards a narrow market composed mainly of competitive athletes at the highest level. Personal VO2 max testing machines remain costly being upwards of $5000 making it evident that companies have yet to develop a cheap alternative to meet the demand. The team designed a new compact VO2 max device with the intentions of addressing the market’s needs and to provide affordable at home testing to athletes at every level. We have conceptualized and built the device under the target budget of $300. Initial testing of the device indicates the ability to accurately measure heart rate via the heart rate sensor incorporated within the device. Additionally, the device is able to measure the oxygen air flow out of a patient while the oxygen air flow in is estimated as the amount does not vary significantly from patient to patient with the exceptions of outliers. The oxygen sensor was 99.7% accurate with the measurement being around ±0.05% of the actual oxygen concentration. A series of additional tests are to be conducted verifying the level of comfort, durability, and portability of the device. Right off the bat, it was clear that using both a Raspberry Pi and Arduino may not be feasible as the combined size does not work well with the given size constraints. This promising new device addresses the main issues with the current devices in the VO2 max market which, in theory, should massively increase the overall availability for individuals that are interested in purchasing a VO2 device
0537. Fabrication of Metallic Organic Frameworks on Carbon for Water Treatment
Undergraduate Presentation.
Author(s): Aidan McGlone Mentor(s): Yingchao Yang
Abstract: Metal-organic frameworks (MOFs) are a scientifically compelling and functionally evolving class of meso-, micro- and ultra-microporous materials. The crystalline MOFs have a large and permanent molecular-scale porosity, which is up to 90% free volume. The MOFs may have an ultrahigh surface area up to 7,000 m2/g. All these unique characteristics bring MOFs a broad potential application, such as gas storage, chemical separation, catalysis, and energy conversion. The focus of this research is to grow selected MOFs nanostructures on porous carbon which is converted from balsa and bass wood sheets. The hybrid structures are expected to be used to remove organic compounds in water. To increase the nucleation and enhance the growth of MOFs in the porous carbon, the carbon has been pretreated by a piranha acid solution to get a hydrophilic surface. The chemicals to synthesize MOFs include copper nitrate trihydrate and benzene tricarboxylic acid mixed in varying amounts in different solutions. A hydrothermal method is employed to complete the reaction at relatively low temperature and high pressure. Currently, two morphologies, particles and fibers, have been successfully synthesized. The hybrid structures are being tested to absorb methyl orange (MO) under visible and UV light for different period. The fiber-like MOFs has a reaction with the MO that turned the color to a neon orange.
0538. Protecting Privacy in the Heterogeneous Internet of Things Undergraduate Presentation
Author(s): Sanonda Gupta, Aubree Nygaard; Stephen Kaplan; Vijayanta Jain Mentor(s): Dr. Stepideh Ghanavati
Abstract: Internet of Things (IoT) is a term used when two or more devices (or “things”) embedded with sensors are deployed in a physical world and are capable of collecting and transferring data over the internet. The increasing growth of IoT devices raises a wide range of privacy concerns, such as inconsistencies between the application and its privacy policy, inference of the personal information of the individual, and lack of detailed privacy notice mechanisms for the users. Previous research attempt to address some of these privacy concerns, however, very few work focus on heterogeneous IoT environment (i.e., smart home as an example). To address this research gap and to protect individuals’ privacy while using IoT devices, we propose a privacy protection framework for the heterogeneous IoT network. Our proposed framework aims at identifying potential privacy risks associated with a new IoT device and generate a detailed privacy notice for the user. We evaluate the functionalities of our framework with four IoT applications from four different categories. The preliminary analysis shows that our proposed framework can help the user to understand potential privacy risks associated with a new IoT device and make an informed decision before the device installed within the network.
0539. Wavelet-Based Automatic Breast Segmentation from Mammograms
Undergraduate Presentation.
Author(s): Basel White, Andre Khalil
Mentor(s): Andre Khalil
Abstract: As a pre-processing step to the computational analysis of thousands of mammograms, the pectoral muscle and image background need to be excluded from each mammogram. This task is currently done manually and takes an unacceptable ~3 minutes/mammogram. Thus, an automated breast segmentation algorithm has been developed, based on an adaptation of the 2D Wavelet Transform Modulus Maxima (WTMM) segmentation method through the use of the program Xsmurf and programming language Tcl/Tk. The output of this algorithm is two binary images: one representing the region of the mammogram bounded by the pectoral muscle, and another image region bounded by the entire breast region. These two images are then used to create the automatic mammographic mask. Currently, the algorithm consists of user thresholds needed to be set for each mammogram. Thus, a cost function has been developed to analyze what combination of size and mean modulus threshold percentile produce the most optimal binary mask, and in turn, allow for the implementation of a global percentile threshold. In order to evaluate mammographic masking efficiency, the Sorenson-Dice Overlap Similarity Coefficient is incorporated into the cost function.
0540. Cellulose Nanofibrils Aerogels and Iron Oxide-based Nanoparticles for Arsenic Removal from Drinking Water
Graduate Presentation
Author(s): Md Musfiqur Rahman Mentor(s): Islam Hafez
Abstract: A novel arsenic adsorbent was prepared from cellulose nanofibrils (CNF) and Mg-doped iron oxide nanoparticles (IONP). Using nanoparticles as adsorbents in the treatment of drinking water at a commercial scale is challenging due to the lack of a well-established and sustainable system to contain them. In addition, the used materials are often non-biodegradable. The IONP was incorporated into CNF aerogels by simultaneously freeze-drying the synthesized IONP and CNF suspension. The synthesized IONP had a higher specific surface area of 165 m2/g compared to other iron oxide-based nanoparticles, which is essential for better arsenic adsorption. Scanning electron microscopy images of the aerogels revealed a uniform distribution of the IONP onto the fibrils. The adsorbent showed a high affinity towards arsenic (III) and (V) (up to 99% removal) following a pseudo-second order kinetic model.
0541. Development and Optimization of Droplet Forming Dipping Process
Graduate Presentation.
Author(s): Aimee Co, Karissa Tilbury Mentor(s): Bashir Khoda
Abstract: Dip coating, a wet deposition method is extensively used due to its simplicity, low cost, and reasonable control over the thickness. During the withdrawal of the substrate, the mixture velocity is directed downward and a layer is entrained. The film thickness results from a balance between the viscous drag, and the capillary rise. The thickness of the layer depends upon various parameter of the liquid and dipping process (i.e., withdrawal speed, viscosity, surface tension etc.). Above a threshold thickness, hydrodynamic instability will trigger formation of droplet whose topology can be controlled for various application such as filtration, material transfer, 3D cell spheroid etc.
The goal of this study is to develop and optimize a new method of droplet or spheroid formation on fiber through a hydrogel dipping process. The adjustment of fiber and hydrogel characteristics allows for control over droplet characteristics. The factors currently being considered for the optimization of the droplet are the fiber diameter and fluid viscosity while the dipping process mechanism parameters are kept constant. By keeping the parameters of dipping speed and submersion time constant, the relationship between the concentration of the chosen hydrogel, alginate, and the diameter of the chosen fiber, PLA, is being observed through the calculation of the resulting droplet volume. The volume calculations of the droplets-on-fiber are highly dependent on the droplet peak diameter, wetting length, and contact angle. Measurements taken of the droplets post-dipping showed that the droplet size increased with diameter as well as hydrogel concentration.
0542. Developing the SMART Cane for the Visually Impaired
Undergraduate Presentation
Author(s): Jordan Miner, Kora Kukk, Natalie St. Louis, Jennifer Triana Mentor(s): Caitlin Howell
Abstract: Currently there are 285 million visually impaired people in the world, and in the US, the number of blind and visually impaired (BVI) individuals is expected to double from 12 million to 24 million by the year 2030. [1] Although there are many existing technologies to help the BVI population, there is still a need for a commercially available device which can reliably combine multiple technologies to give users and, if needed, their health-care providers with navigational and biometric feedback. To develop a solution, our team built the Safe Mobile Aide with Reliable Technology (SMART) Cane that intuitively combines multiple pre-existing technologies through a simple, easy to use interface that aids users in mobility and orientation. The four major components of the device include: navigational assistance, overhead obstacle awareness, fall detection, and haptic and auditory feedback to the user. All components have undergone extensive testing to ensure the device is reliable. The GPS chip in the cane was found to accurately measure the coordinates corresponding to the user’s location. The cane’s proximity sensor was found to accurately measure objects within 6 feet. Lastly, the accelerometer was over 90% accurate in detecting if the cane fell. Future tests will be used to understand how all the components interface together into one seamless design. The overall goal of the SMART Cane is to help to address the ever-growing needs of the aging population and assist in providing the BVI population with more options for navigating daily life.
0543. Concentrating Viruses From Sewage for SARS-CoV-2 Monitoring
Undergraduate Presentation.
Author(s): Kettie Rose Cormier, Audrie French, Harrison Cyr Mentor(s): Jean MacRae
Abstract: Wastewater based epidemiology (WBE) is an approach to health monitoring that uses detection of pathogens in wastewater to track the incidence of public health issues. It has the advantage of integrating the results of the entire community, so it is anonymous, and is less expensive than testing individuals. The goal of my research was to develop a simple, inexpensive lab test for the detection of the SARS-CoV-2 virus in wastewater. I compared methods for concentration of the virus from wastewater samples, and extraction of RNA from the concentrate to try to develop simpler, ideally field-deployable methods that use only equipment that is inexpensive and readily available at wastewater treatment plants. Concentrated virus preparations produced from surrogate virus-spiked wastewater samples were obtained using a Polyethylene Glycol- 8000 precipitation method, and a Nanotrap Magnetic Bead concentration method. RNA from the concentrated samples was either extracted and purified using an IDEXX magnetic bead extraction kit, or released by simple heat treatment. The concentrated viral nucleic acids were tested using two different detection methods: the Reverse Transcriptase Loop-Mediated Isothermal Amplification (RT-LAMP) method, which can be done without access to expensive instrumentation and is a colorimetric qualitative method, and the Quantitative Polymerase Chain Reaction (qPCR) method, which uses more expensive instrumentation, but is a proven method that provides a quantitative reading of viral copies in each sample. Comparisons between the results will be made so that the relative best method can be developed for a given sample type and management goal.
0544. Wireless Soil Moisture Sensor Network
Undergraduate Presentation.
Author(s): Thayer Whitney, Tori Nicholas, Sonia Naderi, Ali Abedi
Mentor(s): Dr. Ali Abedi
Abstract: Forest ecosystem monitoring with high spatial-temporal resolution is of paramount importance for development of accurate prediction models. Current systems are bulky, use high power, and are costly to build and maintain. By using a low cost and power efficient wireless sensor network for soil moisture monitoring, it is possible to circumvent these pitfalls while providing increased resolution.
0547. Strain Sensor for Harsh Environments
Graduate Presentation.
Author(s): David Leff, Mauricio Pereira da Cunha
Mentor(s): Dr. Mauricio Pereira da Cunha
Abstract: Strain sensing in high-temperature, hard-to-reach locations is needed in powerplants, aerospace, and advanced industrial applications for structural health monitoring, condition-based maintenance, increasing process efficiencies, safety, and maintenance cost reduction. Sensing challenges under these conditions include mounting, sensor stability, and data acquisition. Surface acoustic wave (SAW) sensors respond to these demands, as they can operate above 1000°C, are sensitive to temperature and strain, compact, operate wirelessly, and operate without a battery. Wireless strain sensing has been performed at UMaine in high-temperature furnaces up to 400°C. It was found through these tests that after heating, the SAW sensitivity to strain decreased due to cracking of the sensor crystal /adhesive/metallic part interfaces. The differences in the coefficient of thermal expansion between the adhesive, the sensor crystal, and metallic part are believed to be at the center of the identified adhesive failure under excursions between room temperature and high temperature. In this work, modeling in COMSOL Multiphysics was done to study the effects of strain in the mounted SAW due to thermal expansion differences between a metallic surface where strain is being monitored, the adhesive, and the SAW sensor. Simulations with different adhesives and different adhesive shapes were performed to identify alternative mounting methods to minimize the strain at the metallic part/adhesive/sensor interfaces due to thermal expansion while still transferring strain from the metallic part to the SAW sensor. The simulations revealed that the strain on the mounted sensor changes with adhesive shape and type of adhesive used, which can be used to eliminate cracking.
0548. Compact VHF and UHF antennas for integration with SAW devices in harsh environments
Graduate Presentation.
Author(s): Sri Lekha Srimat Kilambi, Mauricio Pereira da Cunha
Mentor(s): Mauricio Pereira da Cunha
Abstract: A Normal Mode Helical Antenna (NMHA) is a combination of short electric dipole and short magnetic dipole. Due to the reduced size with respect to the wavelength, the NMHA structure has a low value of radiation resistance when compared to conventional antennas used around 300MHz frequency range. The NMHA is reported to operate and provide gain in close proximity to a metallic reflecting surface, which can be an advantage in complex industrial and power plant environments. This type of coil antenna structure is used in car smart keys to turn on the car from a distance.
To study the properties of the NMHA, the coil structure has been connected to microstrip transmission lines of different lengths. It has been identified that the size of the nearby microstrip ground plane affects the NMHA structure impedance, resonant frequency, and the radiation properties. In addition, the impedance of NMHA coil structure also depends on the orientation of the coil with respect to the microstrip ground plane. The unbalanced radiating structure was connected to a coaxial cable through a balanced-to-unbalanced (BALUN) device in an attempt to characterize its input impedance. The experimental results are compared with the simulation results from WIPL-D and HFSS. This antenna structure is planned for integration with wireless SAW sensors targeted to operate in harsh environments.
0549. Particles Sorting from Poly-disperse Mixture using Dip-coating
Graduate Presentation.
Author(s): Md Ibrahim Khalil Mentor(s): Bashir Khoda
Abstract: When a substrate is submerged in liquid solution with a relative velocity (rotational speed) between the object and liquid bath, a thin layer of liquid is observed over the object. The thickness of the layer, in submerged condition, depends on the properties of the liquid, especially the percentage of binder used in the solution. When the substrate is fiber, the radius of the substrate has a significant impact on the thickness of the liquid film over the fiber. If the liquid solution contains particle in it, particles entrain into the polymer layer. The entrainment of the particles in the polymer layers is dictated by the trade-off between the viscous force between the layer and the particle and centrifugal force on the particle due to rotation of fluid. Moreover, the size of entrained particles depends on the concentration of the binder used in the solution. When the concentration of binder is low, only small particles entrain into the polymer layer. However, at low binder concentration the large particles remain in the solution as the layer thickness is insufficient to hold the larger particles. Using this phenomenon, an effective filtration process of the particle from poly-disperse particle mixture is shown in this work. The fiber has greater control over the filtration process as well as the size of the entrained particle over the substrate. In this study, the filtration mechanism is observed in submerged condition by dipping a fiber (cylindrical rod) into a solution containing particles. As fiber is used, different regimes are observed at various concentration of binder in the solution. The regimes are classified based on the size of the entrained particles over the substrate. In this work, three different entrainment regimes are found with the variation of binder concentration for the sorting of particles from the poly-disperse particle mixture. Low particle regime, where the diameter of the particles entrained into the polymer layer is smaller than the other regimes. On the other hand, the medium particle regime is characterized by particle sizes that are both greater and equal to those of the previous regime. Finally, the large particle regime is where particle of all sizes (small diameter to large diameter) entrained into the polymer layer from the suspension. Due to the entrainment of particles with a range of diameters, this process leads to a superior filtration process of the poly-disperse particle mixture over the capillary filtration processes. Keywords: liquid bath, binder, viscous force, centrifugal force, filtration, poly-disperse
0550. Lighter Than Air Technology for Remote Sensing Applications
Graduate Presentation.
Author(s): Hamza Azzam Mentor(s): Alex Friess
Abstract: Over the past decade, there has been renewed interest in lighter-than-air (LTA) airship technology, as they have the unique ability to stay aloft with negligible, or in some cases without, energy supply. This property makes them significantly cheaper for certain applications. Therefore, the work proposed in this abstract focus on designing a small-scale airship suitable for remote sensing applications. The closed-loop design algorithm contains multiple modules: such as shape module, aerodynamics module, buoyancy module, weight module, structure module, control module, and stability module. The following modules have been completed to date: A shape module that generates the shape inside SolidWorks, this includes creating the conventional “Gertler shape” and defining and creating a new hybrid shape family, which is composed of elliptical sections driven by a parametric flattened and truncated ellipsoid and a symmetric NACA airfoil cross-section, that allows simple shape adjustments. An aerodynamic module that includes generating the drag, lift, and moment forces and coefficients for different shapes through CFD simulations inside Ansys Fluent 2020 R-1. A buoyancy module that computes the hydrostatic lift that can be obtained from the contained gas in the hull volume. A weight module that computes the total weight (payload, system, and the envelope weight) that has to be balanced by the aerodynamic and the aerostatic lift. Once the algorithm is complete, a specific shape, with specific dimensions and specific structure, will be designed for the required remote sensing mission.
0551. Mechanical Performance of 3D Printed Lattice Structures: Assembled Vs Direct Print
Graduate Presentation.
Author(s): Adeeb Alam, Keith Berube Mentor(s): Bashir Khoda
Abstract: To fill a three-dimensional space cellular solids or, lattice structures can be used instead of solid materials, to get high mechanical performance with lower density. These low-density high-performance structures are used in applications like mechanical, phononic, thermal, and biological fields. However, the possibilities of lattice structures are often unaffordable due to design and manufacturing limitations. Extrusion based 3D printing, which can be defined as incremental deposition of one or two dimensional forms of raw materials, has the potential to be used in fabricating porous structures. Here, an extensive study is done on the fabrication of lattice structures using extrusion based 3D printing. From Maxwell’s criteria, we have bending and stretch dominated lattice structures which depend on the nodal connectivity of the struts. Stretch dominated structures generally show significantly higher mechanical performance than bending dominated structures. Primitive cubic and Octet, which are representatives of bending and stretch dominated structures, were 3D printed by applying both direct printing and assembly methods. To compare the mechanical properties, lattice structures were manufactured with variable density. The density was varied by altering the cell size or strut thickness. Assembly based designs can give us much versatility in printing lattice structures with lower density using less manufacturing time, because of continuous toolpath and ability to print without support. A comprehensive study is done to find mechanical properties of both Cubic (bending dominated) and Octet (Stretch dominated) structures manufactured by both direct and assembly methods. Mechanical strengths of these porous structures were compared by using the power law given by Gibson and Ashby.
0552. Contact Tracing via Bluetooth Beacons
Undergraduate Presentation
Author(s): Jacob Mealey
Mentor(s): Michael Scott
Abstract: Over the past year, the world has been facing the Covid-19 pandemic, forcing the University of Maine to take precautionary measures. In an effort to find better such measures, Asap Media Services (ASAP) has conducted research on various forms of contact tracing since April 2020. At ASAP, we have built a low-power Bluetooth device that will be placed in every room broadcasting a unique identifier, along with a companion smartphone application that scans for Bluetooth devices. This will track where every member of the community is at all times so that, should someone test positive for Covid-19, we can easily get in contact with anyone who may have contracted the virus from that person. ASAP is conducting a test on a small sample of students who are working in the office.
0553. Sailing Velocity Prediction Program
Graduate Presentation.
Author(s): Andrew Arter Mentor(s): Alex Friess
Abstract: To whom it may concern: Sailing velocity Prediction Program. The Sailing Velocity Prediction Program focuses on creating a computer program to simulate the velocities that a sailing craft will reach under different variable conditions. Over the past hundred years, sailing has transitioned from a major mode of transportation into a luxury activity and a very competitive sport. As the sport of sailing has progressed, sailboats have been reaching higher velocities, which is due to the increase in the understanding behind the physics of sailing. A program to predict the velocity of sailing crafts is crucial to the design stage of the vessels. The program balances the forces on the craft to generate a velocity polar comparing the boat speeds to the boat’s angle to the wind. Where this program diverges from current commercial programs employed by designers is that it focuses on sailing crafts that use hydrofoils. The project then examines the ways to optimize the hydrofoils to maximize potential velocity for a given set of parameters. The parameters include size, shape, and camber of the foil. The end results would be compared to the most recent iterations of the most advanced sailing crafts that use hydrofoils.
0544. Exploration of Olfactory Interventions as a Synergistic Emotional and Cognitive Recovery Technique
Undergraduate Presentation.
Author(s): Leah Parrish Mentor(s): Nimesha Ranasinghe
Abstract: Work-stress is a significant health concern in today’s society, and with COVID-19 forcing us to adapt to even more stressful situations, mental health self-care and recovery from work stress are quickly becoming more common priorities. Our goal is first, to investigate how the quality of recovery from online work stress varies based on how the individual recovers from that stress. Secondly, to investigate how these different recovery techniques can be integrated to develop more effective recovery experiences. Trials consist of participants going through a series of work and relaxation tasks, tracking their emotional state and work productivity through the study. We will be presenting our preliminary findings using a combination of digital games and scent stimuli for work-stress recovery. Hopefully this will be an ongoing and expanding investigation.
0555. Implementation of a structural health monitoring system for Carbon Fiber Composite Strands (CFCS) in the Penobscot Narrows Bridge
Graduate Presentation.
Author(s): Braedon Kohler, Roberto Lopez-Anido, Andrew Goupee, Keith Berube Mentor(s): Lopez-Anido
Abstract: The goal of this project is to further advance our knowledge regarding the durability of carbon fiber composite cable (CFCC) strands used in transportation infrastructure. For this reason, we aim to improve the structural health monitoring capabilities of CFCC strands placed within the stays of the Penobscot Narrows Bridge. The bridge is a “living laboratory” in which there are already sensors in place in six strands at two locations throughout the bridge that monitor strand response that were installed in 2007. CFCC strands have high corrosion resistance, low thermal expansion coefficients, high modulus to weight ratio, and high strength to weight ratio. Due to the corrosive environments of highway bridges these strands have the potential to outperform steel cables when used for post-tensioning concrete decking and in cable-stayed bridges. The tasks of this project are: a) to implement a wireless system in the bridge to allow remote and continuous data acquisition at the Penobscot Narrows Bridge; b) to evaluate CFCC durability related to long-term response of the strands and anchorages; c) to demonstrate the value in collecting continuous CFCC strand force data; d) to develop a thermoelastic model for numerical simulation of the CFCC response and validate it with collected data; and e) to apply the model to predict CFCC strands long-term response. Using an array of sensing equipment the forces and strains in the strands, along with the ambient temperature have been monitored for the past twelve years. Using a thermoelastic model, the acquired temperature data at the bridge will be used to predict the long-term CFCC strand response. In an effort to streamline this process, a wireless data acquisition system will be implemented such that data will be collected and processed remotely. In addition, a new weather station will be installed to acquire temperature, wind speed, humidity, and barometric pressure.
0556. Slimming Neural Networks using Adaptive Connectivity Scores
Undergraduate Presentation.
Author(s): Dawsin Blanchard, Madan Ravi Ganesh, Jason J. Corso, Salimeh Yasaei Sekeh Mentor(s): Salimeh Yasaei Sekeh
Abstract: There are two broad approaches to deep neural network (DNN) pruning: Weight-based methods which determine importance scores via the weight matrices, and probabilistic methods which aim to capture the flow of information in the network. Each approach’s advantage supplements the missing portions of the alternate approach yet no one has combined and fully capitalized on both of them. Further, there are some common practical issues that affect both such as manually analyzing sensitivity and setting pruning limits. In this work, we propose Slimming Neural networks using Adaptive Connectivity Measures (SNACS), as an algorithm that uses a probabilistic framework for compression while incorporating weight-based constraints at multiple levels to capitalize on both their strengths and overcome previous issues. We also propose a hash-based estimator of Adaptive Conditional Mutual Information (ACMI) to evaluate the connectivity between filters of different layers, which includes a magnitude-based scaling criteria that leverages weight matrices. To reduce the amount of unnecessary manual effort required to set the upper pruning limit of different layers in a DNN we propose a set of operating constraints to help automatically set them. Further, we take extended advantage of weight matrices by defining a sensitivity criteria for filters that measures the strength of their contributions to the following layer and highlights critical filters that need to be protected from pruning. We show that our proposed approach is faster by over 17x the nearest comparable method and outperforms all existing pruning approaches on three standard Dataset-DNN benchmarks: CIFAR10-VGG16, CIFAR10-ResNet56 and ILSVRC2012-ResNet50.