RISE Technical Abstracts
Through the RISE program, mechanical engineering undergraduate students leverage our state-of-the-art facilities working alongside internationally-renowned faculty to tackle cutting edge projects that impact our society. The Mechanical Engineering Undergraduate Symposium (MEUS) is the accumulation of the students’ work.
During the day, seniors in ME 490 will present their RISE projects in 20 minute presentations. Everyone is welcome to attend these sessions and ask probing questions!
Sophomores and juniors conclude their projects with a poster session during the evening reception, where students will be available to discuss their projects in detail. The posters will also be on display during the day, if you are unable to attend the reception.
The public is invited to peruse the posters, attend presentations, and interact with the students throughout the day.
Complete poster and presentation schedule details can be found in the Symposium Schedule. RISE project technical abstracts can be viewed below.
Posters (ME 290 and 390)
Bulk Modulus of Compressibility of Different fuels
Student: Jiawei Song Instructor: Andre Boehman
Student: Jiawei Song Instructor: Andre Boehman
Abstract: The isothermal bulk modulus of compressibility is one of the properties affecting engine fuel injection process. In this study, we analyze the variation in the chemical compositions of different fuels, including the conventional jet fuel (JP-8), alcohol to jet, HEFA, and biodiesel. The analysis asserts that a less rigid molecule structure leads to a higher compressibility, which means a lower bulk modulus. The diesel fuel has longer chain alkanes and a larger quantity of aromatics than the jet fuel. Among the jet fuel, the ATJ (alcohol to jet) is primarily composed of branched isomers of dodecane and hexadecane. The HEFA is a kind of bio-oil with a high level of unsaturation. The bulk modulus of these fuels are also measured. The results supports our analysis that the biodiesel has the highest bulk modulus of compressibility and the ATJ has the lowest bulk modulus. The present work confirms that a higher bulk modulus of compressibility results in a higher speed of sound in the fuel blend. This causes an advance in injection timing with in-line, pump-line-nozzle-type fuel injection systems. The next step of this study is to test the injection timing of each fuel and compare them with the bulk modulus data.
Keywords. Diesel, Jet fuel, Bulk modulus, Fuel, Engines,
Thin-film Piezoelectric Microactuators for Into-tissue Imaging by Endoscopic Microscopy
Student: Issac Loo Instructor: Kenn Oldham
Student: Issac Loo Instructor: Kenn Oldham
Abstract: The project objective is to design, fabricate, and test a 2-axis angle adjustment fixture for a thin-film lead-zirconate-titanate micro-actuator for fast vertical cross-section imaging of tissue by multi-photon imaging. This micro-actuator is currently under an intermediate step of a hand held design and needs exact vertical motion to keep the laser beam for scanning aligned. The goal is to create an angle calibration fixture for the micro-actuator that has a stage that is less than 2 cm in width, length, and height and has an angle adjustment of at least 2.5 degrees. This fixture is designed through CAD modeling, tested using Oldham Lab’s in-house 3D printer, revised through solid mechanics and fracture calculations, and fabricated through an outsourced high-quality 3D printer. The final angle calibration fixture is composed of a 3D printed bi-level stage and 18-8 Stainless Steel Cup-Point Set Screws and Brass Flanged Screw-to-Expand Inserts for Plastics that were purchased from McMaster-Carr and has an angle adjustment of approximately 3 degrees and is less than 2 cm in width, length, and height.
Micro-actuator, multi-photon, angle, calibration, 3D printing, microscopy
Design of a Reconfigurable 3D Printer
Student: Justin Joseph Instructor: Chinedum Okwudire
Student: Justin Joseph Instructor: Chinedum Okwudire
Abstract: Currently, the default way to increase the build volume or functionality of a 3D printer is to buy a larger and more expensive printer. This project focuses on designing a reconfigurable 3D printer comprised of various modules that allow the build volume and functionality of a printer to be varied at low cost. This will allow consumers to purchase one printer that can be adjusted and upgraded to fit their changing needs. Existing Cartesian and Delta 3D printers were analyzed to generate a functional decomposition of 3D printers. Each desired function for our 3D printer was grouped into interlocking, reconfigurable hardware modules. By adding modular functionality to 3D printers, we were able to design a Delta printer that can easily improve build volume and functionality by combining various modules.
Redesign of Experimental Set-up for Dual Beam Laser Welding
Student: Quinton Ho Instructor: Elijah Kannatey-Asibu Jr
Student: Quinton Ho Instructor: Elijah Kannatey-Asibu Jr
Abstract: This project involves designing and manufacturing a new experimental welding setup for a PRC 3000/2 dual beam laser generator. This setup is to be used to conduct future research in areas such as laser cutting, dual beam laser welding and laser forming. Using the methods taught in the Mechanical Engineering Manufacturing/Design curriculum, a new experimental setup was designed. This fixture consists of an adjustable test stand to house the beam bending components along with a chiller system to cool the optics. The parameters that may be varied are the distance between the beams, the power of each beam, and the focal point of each beam. The setup will be manufactured in the ME Machine Shop and is estimated to be installed by the end of 2017.
Manufacturing/Design, Welding processes, Laser Welding
Modeling bi-directional trust in semi-autonomy for improved system performance
Student: Beth Beindit Instructor: Dawn Tilbury
Student: Beth Beindit Instructor: Dawn Tilbury
Abstract: Autonomous and semi-autonomous vehicles have the potential to help drivers successfully and safely complete many military missions while providing drivers with the flexibility to address other pressing issues not possible while driving. Autonomy can best be described as the technical capability that allows vehicles some latitude to drive. This project will examine factors that impact a driver’s trust in vehicle autonomy in order to improve the overall system performance. This will be measured by subject questionnaires (such as propensity to trust, mood, and after-use trust in automation), driving inputs, driving performance, and physiological data measured in real time (such as gaze behavior and heart rate). These factors will be used to predict the initial and thereafter, dynamic trust the driver will have toward the vehicle’s autonomy and the trust the vehicle’s autonomy will have toward the driver.
My project this semester supports the above mentioned research by collecting and analyzing physiological data, specifically gaze behavior through the use of an eye-tracker. As mentioned above, this data will be used to develop a model of trust of a semi-autonomous vehicle. This eye-tracker will be used in experiments to capture data such as gaze position, pupil position, pupil diameter, confidence, and blink detection. Research areas pertinent to this supporting project include the eye-tracker capture and analysis functions as well as the accuracy of the instrument. Through this research conducted this semester, a final document will be created as a framework for performing data collection, extraction, and analysis. The collection of this data from the experimentation this project supports will allow us to develop a model of real-time driver trust. This work will be used in many if not all ongoing studies related to this project.
Key Words: Semi-Autonomous Vehicles, Trust, System Performance, Eye-Tracker
Non-Linear Control System for Lightweight and Portable Transportation Device
Student: Benjamin Eu Instructor: Ram Vasudevan
Student: Benjamin Eu Instructor: Ram Vasudevan
Abstract: Currently, strollers are four-wheeled, bulky baby transportation products which are a bane for families living in crowded urban communities such as New York City and Singapore. To develop a baby transportation product that is lightweight and portable, research and development was conducted to explore the ability of a non-linear control system to enable the baby transportation product to be easier for parents to close and carry and more intuitive for parents to maneuver. Firstly, user discovery interviews and surveys were conducted with 100 parents in New York City, Singapore and Ann Arbor to determine the top needs of parents in a baby transportation product. Next, the mechanical structure was modelled and 3D printed to include the ability to be closed and carried in one step. Finally, the electronic system and a differential drive algorithm was developed to improve the maneuverability of the baby transportation product. The survey results indicate that top needs of the parents are the ease of maneuvering, ease of opening and closing, and ease of carrying, which validates our hypothesis. The mechanical structure is 0.53m in height and 0.4m in width when closed, representing an average decrease of 47% and 30% in height and width compared to three other products most similar to this. The differential drive algorithm allows parents to effectively maneuver the baby transportation product with one hand, compared to two previously, freeing the parents’ other hand to complete other tasks. The non-linear control system improves the convenience of parents and should be included in the further development of baby transportation products.
Paper Sessions (ME 490)
1. Thermal and Fluidic Systems
CFD Modeling of Electrohydrodynamic Jet Printing
Student: Maxwell Wu Instructor: Kira Barton
Student: Maxwell Wu Instructor: Kira Barton
Abstract: Electrohydrodynamic jet (e-jet) printing is an additive manufacturing method that has emerged in recent years as a novel printing process to rival similar technologies such as inkjet printing. By utilizing electrostatic forces to induce fluid flow, e-jet has shown the ability to fabricate high resolution features (<10μm) with a wide variety of materials. This versatility has generated a growing interest in e-jet due to its potential applications in the fields of microelectronics and biotechnology among many others. However, as a relatively new subject of research, the capabilities of e-jet printing are still being explored. In particular, the printing dynamics of many different types of inks are not completely understood, as existing knowledge of the process relies almost entirely on empirical methods which can be time intensive and difficult to implement for materials that are hazardous or not readily available. Fortunately, with constantly improving computing power and refined modeling techniques, Computational Fluid Dynamics (CFD) has become an increasingly popular way to model fluid systems such as those used to implement e-jet printing methods. This paper discusses procedures for modeling both the electrical properties of the e-jet system such as the electric field and potential, and the resulting fluid dynamics within a CFD solver. Further discussion has been provided to show completed CFD simulations of typical e-jet dynamics and to present existing limitations of the model.
Keywords: Electrohydrodynamic jet printing, e-jet, computational fluid dynamics, CFD, modeling
Ignition Experiments on Cetane Reference Fuels
Student: Archit Gupta Instructor: Andre Boehman
Student: Archit Gupta Instructor: Andre Boehman
Abstract: Exhaust Gas Recirculation (EGR) has taken the spotlight in the automotive world as a technology that can reduce emissions and increase fuel efficiency of modern engines. To determine the effect of EGR on the combustion process, studies on ignition and combustion delay of transportation fuels can be used to provide important information, allowing for advancement in research topics relevant to modern IC engines. Additionally, the analysis of other quantities including pressure rise rate and apparent heat release help to further characterize specific fuels. Recognizing these considerable impacts this work investigates blends of n-heptamethylnonane and n-hexadecane with an exhaust gas recirculation (EGR) sweep from 0-55%. The experiment was conducted in a modified Cetane Ignition Delay (CID 510), which contains an optically accessible constant volume combustion chamber (CVCC). All the experiments were conducted at a constant pressure (20 atm). For specific blends ignition delay was further analyzed to determine physical and chemical delay. This was done using a photomultiplier tube (PMT) system accessed in the bottom of the chamber. The PMT voltage signals that were obtained from the chemiluminescence of excited radicals (CH2O*, OH*, CH* and C2*) were used to quantify time values of chemical and physical ignition delay periods. Trends obtained indicate that an increase in EGR extends ignition delay, with lower cetane fuels not experiencing combustion at high EGR values. To this end, this research provides a benefit for CFD modelers to optimize more efficient IC engines with the alteration of various fuels and modification of combustion processes.
Analysis of Spray Geometry
Student: Evan Harris Instructor: Andre Boehman
Student: Evan Harris Instructor: Andre Boehman
Abstract: This project on the Analysis of Spray Geometry was setup to both rebuild a spray chamber previously used by the EPA, but also to begin obtaining and analyzing data on spray geometry using high speed phenolic, or “light field” cameras. Spray geometry is very important in the study of combustion, and even small improvements in efficiency of spray can have large consequences when used on a national or international level. This project began by reconstructing the spray chamber, assisting students sponsored by the Chinese Scholarship Council. Due to scheduling issues, the rebuild took most of the semester. Upon completing the mechanical rebuild, as well as adding on to previous work with a safety cage, the electronically PID needed to be reconfigured and re-designed, this is the current state of the use of the chamber. When that is completed, the chamber will be used for spray optimization with Bosch injectors.
Microfluidics/Nanoelectronics-Integrated biosensors with mechanical flexibility
Student: Jung Hyuk Kim Instructor: Xiaogan Liang
Student: Jung Hyuk Kim Instructor: Xiaogan Liang
Abstract: One of the most popular methods of identifying a specific substance in a chemical sample has been the enzyme-linked immunosorbent assay (ELISA). However, numerous researches recently have shown that biosensors comprised of MoS2(molybdenum disulfide) can overcome some of the disadvantages of the ELISA method. Compared to the ELISA method, the MoS2 biosensors are less labor and time intensive in procedure and also benefits from a higher sensitivity of 10 fM, compared to 600 fM of ELISA method. In this study, the MoS2 biosensors are manufactured in a resistor configuration, instead of a field-effect transistor configuration used in other researches. Benefits of the resistor configuration include an even simpler manufacturing process and an easier detection procedure. The aim of this study is to manufacture and to test the MoS2 biosensors in resistor configurations for a cycle-wise time-dependent correlation between the different concentrations of streptavidin and the resistance of the biosensor. This was done by submerging the biosensors in different concentrations of streptavidin solutions for a time ranging from 5 to 20 minutes and applying a voltage across the sensor ranging from -1V to 1V to obtain the average resistance value. A successful establishment of streptavidin concentration and biosensor resistance correlation can lead to a wide array of applications, including a sensor that can detect a patience’s health by merely being held in mouth for just a few minutes.
Key Words: Biosensor, ELISA, MoS2, Resistor, Field-Effect Transistor, Streptavidin
2. Smart Materials Devices and Structures
3D-printed Self-Locking Origami Metamaterials with Piecewise Stiffness
Student: Shihcheng Chu Instructor: Kon-Well Wang
Student: Shihcheng Chu Instructor: Kon-Well Wang
Abstract: In recent years, origami techniques have been explored in engineering to achieve extraordinary kinematical and mechanical properties that are unseen in traditional structures and materials. Previous research has shown that certain origami structures cannot be folded flat, but will instead self-lock at a pre-specified configuration owing to the binding of origami facets. Such facet-binding would induce a sudden increase of the structure stiffness, giving rise to a piecewise stiffness profile. Previous theoretical and experimental studies on this phenomenon were conducted based mainly on the assumption of rigid-foldability, where the creases are abstracted as ideal hinges and the facets are assumed to be flat and rigid. This research aims at developing a self-locking origami metamaterial with elastic material through 3-D printing techniques. In such case, the origami structure will no longer be rigid-foldable because of the significant facet/crease thickness and facet deformation. In detail, this research focuses on the locking-induced stiffness properties of the 3D-printed origami metamaterial through finite element analysis (FEA) and experimental testing. We find that self-locking still exist in this non-idealized scenario, not only in a single-cell but also multi-cell settings. The structure stiffness experiences significant jumps owing to the contacting of materials and exhibits a piecewise profile. In addition, FEA results manifests that it is possible to incorporate certain actuation methods into the origami metamaterials to actively lock a specific cell and tailor the structure’s stiffness. This research lays the foundation for future research and development on origami metamaterials with piecewise stiffness.
Keywords: Origami, piece-wise stiffness, metamaterials, finite element model, 3D-printing
Piezoelectric Polymer REM Sleep Sensor
Student: Andrew Holmes Instructor: Kenn Oldham
Student: Andrew Holmes Instructor: Kenn Oldham
Abstract: Rapid eye movement (REM) sleep is considered to have a beneficial impact on patient recovery in critical care, but there are concerns regarding how to monitor REM sleep by existing methods given complexity of physiological signals in an intensive care unit. The purpose of my research is to combat this problem by understanding whether a piezoelectric sensor could be used to measure the physical eye movement of a patient experiencing REM sleep. A sleep mask was manufactured to accommodate the piezoelectric sensor. Patients underwent in lab tests where baseline measurements, blink test measurements and eye movement test measurements were recorded. Mechanical movements were translated through the sensor and read as output voltages by an attached oscilloscope. As hypothesized, the blink test readings were stronger than eye movement readings which were stronger than the baseline readings. We could validate these measured voltages by using a piezoelectric voltage equation. The equation calculated an expected output voltage based on known input stresses. Stress calculations were performed on the sleep mask. The equation was used to calculated an expected output voltage. Despite the calculated voltage and the measured voltage not equating, the values were within a tenth of one another. I believe with more technology and better estimated values for the stress analysis, the difference between these values will diminish and I will be able to say with confidence the developed sleep mask can measure a patient's eye movement in REM sleep.
Variably Ventilated Veneers
Student: Brian Kalinowski Instructor: Diann Brei
Student: Brian Kalinowski Instructor: Diann Brei
Abstract not yet available.
3. Acoustics
Detection of Transients in Noise
Student: David Lectka Instructor: David Dowling
Student: David Lectka Instructor: David Dowling
Abstract: In order to detect the presence of a transient signal in a noisy environment, I compared the performance of two different types of time-frequency analysis, spectrograms and wavelet transforms, and their effectiveness in detecting the presence of different types of transient signals. This project was completed under the guidance of Prof. David R. Dowling at the University of Michigan. The research was sponsored by Naval Sea Systems Command (NAVSEA) through the Naval Engineering Education Consortium (NEEC).
This project was motivated by the naval applications of being able to detect transient signals despite the environment, such as a naval ship, being very noisy. NAVSEA desires information such as time and frequency characteristics of transients that are on the order of 100 ms in duration. One of the most commonly used methods for such types of detection is the spectrogram, or short-time Fast Fourier Transform (FFT). Another method is the wavelet transform, which compares the signal to a known transient, known as a wavelet, at different “shift” (variations in time) and “scale” (variations in frequency).
The goal of this project is to compare the two methods of detection using statistical analysis. To compare the two, I analyzed spectrogram and wavelet transform results of three different signals: a Gaussian-enveloped sine wave, an exponentially decaying sine wave, and a recorded transient of a balloon pop. For detection in a noisy environment, I analyzed each signal at four different noise levels: 10 dB, 0 dB, -10 dB, and -20 dB signal to noise ratio (SNR).
To determine the effectiveness of each method of detection, I created a threshold detector using MATLAB. The detector calculates a “metric,” which is the maximum amplitude of the spectrogram or wavelet transform divided by the average amplitude, for many different realizations of the same signal in an environment of added white Gaussian noise. It also finds the location of the maximum value and compares it to the time and frequency information that is known about the transient. The detector then compares each metric to the current threshold level, determining that a transient exists if the metric value is above the threshold value and the time and frequency locations of the maximum amplitude is sufficiently close to the known values. The detector gradually increases the threshold, constructing a receiver operating characteristic (ROC) curve at each SNR value for each signal and detection method. The results of the ROC curves show that the spectrogram used in this project performs slightly better than the wavelet transform, most notably at low SNR (around -10 dB).
Measurement of the Theorized Frequency-Difference Autoproduct Field
Student: Jessica Lipa Instructor: David Dowling
Student: Jessica Lipa Instructor: David Dowling
Abstract: An acoustic autoproduct is a quantity formed by multiplying two complex signal amplitudes from different frequencies contained in a signal. Current autoproduct theory proposes that autoproducts allow out-of-band information to be extracted from an acoustic signal having finite bandwidth, a novel conjecture (Worthmann and Dowling, 2017). However, to verify the theory and promote wider acceptance of autoproducts, they must be measured experimentally. The purpose of this project was to measure one type of autoproduct, the frequency-difference autoproduct, as a field quantity in an underwater acoustic environment with one reflecting surface. To do so, sound was broadcasted in a water tank, and the acoustic pressure was measured at different depths using a hydrophone. Simulations of the experiment were performed in the time and frequency domains, and the directionalities of the hydrophone and source were investigated. Then, the autoproduct was computed at measurement locations, allowing the autoproduct field to be measured along a vertical line within the tank. Fields were constructed for hydrophone depths of approximately 0-40 cm with the sound source broadcasting from approximately 20 cm depth with 70 kHz center frequency. Difference frequencies of 5-30 kHz were used in forming the autoproducts. Autoproduct fields were found to closely match predicted fields, especially when corrections were made for geometry and non-ideal broadcasting behavior. In addition, the fields showed similarities to true acoustic fields at the below-bandwidth difference frequencies, supporting the claim that acoustic signals carry out-of-band information that autoproducts can extract.
Keywords: acoustics, autoproduct, Lloyd’s mirror, out-of-band signal information
Relationship between tail-pipe bevel angle and reflection coefficient at low acoustic frequencies
Student: Stefano DeBellis Instructor: David Dowling
Student: Stefano DeBellis Instructor: David Dowling
Abstract: Sound travelling in a pipe reflects at the open end back into the pipe. While this phenomenon is well understood for flat-ended pipes, it has not been studied for bevel-tipped pipes, such as the angled tailpipes one would find on an automobile’s exhaust system. The goal of this project was to setup a pipe and low-frequency sound driver apparatus with replaceable tips so that an array of bevel angles may be applied at the end, and to analyze the relationship between the reflection coefficient and the bevel angle. I wrote a Matlab script that would send a Gaussian-modulated signal at a specified frequency to an acoustic driver and record the output with a B&K microphone. The output was recorded in two different positions: one position approximately 5 inches from the driver, and the other position approximately 9 inches from the open end of the pipe. The distance between these two positions varied depending on the frequency of the sound analyzed and the corresponding pipe chosen for the analysis. An FFT was performed on the recorded data to analyze the sound pressure in the pipe at the frequency of interest at the two locations. The complex pressure values were used to calculate the reflection coefficient that was then plotted for an array of bevel angles to highlight the relationship between the two.
Keywords: acoustics, reflection coefficient, sound in pipe, bevel tip pipe, low frequency acoustics
4. Actuators and Sensors
Investigation into the energetic effects of various spinal morphologies in quadrupedal robots
Student: William Yang Instructor: C. David Remy
Student: William Yang Instructor: C. David Remy
Abstract: Studies have shown that a flexible spine can be used in conjunction with multiple gaits to improve the energetic economy of quadrupedal robots. However, there is no consensus on the optimal spine morphology for quadrupedal robots. The multi-segmented spines in quadrupedal animals seem to suggest some advantage for adopting a more complex spine morphology. However, existing studies in simulation and optimization have so far been limited to single-joint spines. This paper seeks to build upon those studies and examine the energetic effects of various, more complex spine morphologies on quadrupedal robot locomotion. To answer this question, we developed a planar quadrupedal model with a 3-joint spine capable of capturing complex spine morphologies. The model features mass in the feet, legs, and torso as well as series elastic actuators that include realistic stiffness, damping, and limits on actuator torque and speed. Using optimal control methods, we determine the energetically optimal motion for the 3-joint model and then compare the energetics of the rigid, single-joint, and 3 joint models across various gaits and speeds. We use the results from this comparison to gain insight into the optimal spine morphology.
Knitting Active Materials
Student: Sumayya Atmeh Instructor: Diann Brei
Student: Sumayya Atmeh Instructor: Diann Brei
Abstract not yet available.
Cardiovascular Sensing-Finger Phantom Construction
Student: George Tsirukis Instructor: Kenn Oldham
Student: George Tsirukis Instructor: Kenn Oldham
Abstract: Peripheral arteries in the human body are key physiological indicators for both acute and chronic medical conditions. Peripheral arterial constriction or dilation is also the dominant factor in determining the body’s systemic vascular resistance, which is the resistance felt by the heart in forcing blood through the arterial system. The Microdynamics lab has been developing a piezoelectric, polyvinyl fluoride (PVDF) sensor that measures the artery’s radial deformation. The project this semester entails construction of a human finger phantom, which closely resembles an actual human finger. The Micro dynamics lab needed a replica human finger that matches the properties of an actual human finger, in order to be able to model the fluid flow with the developed sensor. Using the PVDF sensor that has been developed, the goal of this project is to be able to model the fluid flow through the phantom human finger that has been constructed. The construction of the finger will involve computer aided design tools (SolidWorks CAD), as well as 3D printing. The finger will include a central bone structure, as well as a silicone artery running axially next to the bone. After fabrication, the sensor will be used to measure different pressures of water flowing through the artery. After conducting various experiments, a fluid model of the system, for various input flows, has been created. Therefore, the sensor can be used for real life situations for patients who are suffering from acute and chronic medical conditions.
5. Sensing and Controls
Software-defined Control of Smart Manufacturing Systems
Student: Vincent Salpietro Instructor: Dawn Tilbury
Student: Vincent Salpietro Instructor: Dawn Tilbury
Abstract: Software-Defined Control (SDC) is a novel framework that aims to leverage techniques from computer networks to improve overall equipment efficiency in large manufacturing systems through dynamic reconfiguration. SDC uses a global view of the entire manufacturing system to dynamically modify resource allocation and routes in the logic controllers in response to anomalies, faults, cyber-attacks, or changes in production demand. In this project, a table-top manufacturing testbed was proposed as a cost effective and easily reconfigurable tool to test reconfiguration algorithms for logic controllers. The testbed consists of infrared sensors, motors, Raspberry Pi microcontrollers, and network routers that mimic the machines, programmable logic controllers, and large network infrastructures that are common to manufacturing facilities. Additionally, a Human Machine Interface (HMI) was created for this testbed with the ability to detect faults and anomalies, and determine the productivity of the testbed. This information will be used by the software defined controller to determine when it is necessary to reconfigure the production system and the most adequate reconfiguration. The primary goal of this testbed is to act a realistic test platform for reconfiguration methods before they are deployed in real manufacturing settings.
Biometric Sensor Integration into Automotive Systems
Student: Syed Mahdi Instructor: Diann Brei
Student: Syed Mahdi Instructor: Diann Brei
Abstract not yet available.
Development and Validation of a Toolkit for Time Series Data Feature Selection
Student: Anne Gu Instructor: Kira Barton
Student: Anne Gu Instructor: Kira Barton
Abstract: Classification of time series data can be cumbersome due its high-dimensional and noisy nature. Feature selection is a necessary step to find relevant data to improve robustness of a classifier and to reduce processing time with a lower-dimensional data set. Though there are numerous feature selection methods, there are trade-offs between computation complexity and optimality. Filter is a fast feature selection algorithm that uses only the inherent data characteristics to evaluate features. For this project, we created a feature extraction and selection repository that is generalizable to any input time series signal data and classification task. Specifically, this repository contains an efficient feature extraction function and three filters that use a common strategy called min-redundancy max-relevancy (mRmR) for feature selection. We tested this repository for lower back electromyography (EMG) data collected during a dynamic lifting task. Forty time-domain and frequency-domain features were extracted and ranked to maximize the accuracy of classifying three loading conditions. The optimal features were validated using a Linear Discriminant Analysis (LDA) classifier.
Keywords: Time Series Data, Feature Selection, Filter, mRmR, EMG Signals, Lower Back, and Dynamic Lifting
6. Device Design and Testing
Stress/Strain Testing on Patient Ovarian Cancer Cells
Student: John Bohenick Instructor: Krishna Garikipati
Student: John Bohenick Instructor: Krishna Garikipati
Abstract: According to the National Cancer Institute, this year over 22,000 women will be diagnosed with ovarian cancer, and over 14,000 women will die from the disease. Only 46.2% of women survive 5 years with ovarian cancer. Creating more effective treatment strategies for ovarian cancer requires understanding how different microenvironmental factors influence the disease progression and chemoresistance. Mechanical stimulation, including shear stress, plays an important role in carcinogenesis, but it is not yet completely understood. I designed and fabricated a 3D bioreactor capable of mimicking the dynamic physiological cellular microenvironment by applying a pulsatile shear stress on ovarian cancer primary patient cells. My hypothesis is shear forces promote increased cancer metastasis, cell invasion, and increased chemotherapy resistance in patient ovarian cancer cells. The 3D bioreactor applies a pulsatile shear stress for 24 or 72 hours to Pt224 ovarian cancer cells suspended in agarose-collagen IPN (interpenetrating network) hydrogels. I modeled the shear profile within these IPN gels using COMSOL finite element modeling. The next step for this project is to create a qPCR array to analyze the gene modifications between the sheared and control groups. Then, immunohistochemistry staining will be performed to confirm the qPCR results and to investigate proliferation differences, along with viability assays for drug treatment effects using alamarblue.
Device Design; Device Manufacture; Ovarian Cancer; Materials Testing; Bioreactor; Comsol; Finite Element Analysis; Cell Research; MECHE; MSE; BME; Chemoresistance
Design, modeling, and experimental evaluation of non-invasive methods for bio-logging sensors
Student: Riley Doherty Instructor: Kira Barton
Student: Riley Doherty Instructor: Kira Barton
Abstract: The purpose of this project is the experimental evaluation and modeling of suction cups that will be used to secure a biologging tag for cetaceans. In order to prevent early tag removal, reduce the potential impact of skin loading, and limit changes in animal behavior, a new bio-logging tag design was proposed. Biologging tags are an important tool for tracking animal behavior. Experimental analysis must be performed on the suction cup under different conditions (wet surface, dry surface, soft surface, etc.) to determine if the suction cup will be sufficient for implementation of the final design. After designing and constructing an experimental interface for testing, experiments were performed on two different suction cups using an Instron machine and the displacement, load, pressure, and area of contact measurements were then used to compare the suction cups’ response under different test conditions. After comparing the results of the experiments, it was found that varying the wetness of the surface did not greatly impact the behavior of the cup. The total displacement and average pressure remained approximately the same; however, the amount of load the cup could withstand was slightly greater for a dry surface. Additionally, the material of the two suction cups that were compared was proven to have an insignificant effect on the results of the experiments. Analysis of the results allows us to conclude that when a normal force load is applied, the suction cup behaves as a nonlinear spring system. In the future, it will be important to conduct experiments that test the shear force loading that a suction cup can withstand and the response of a cup when the surface is soft, resembling the material of marine mammal skin.
Veinless Vents
Student: Mary McMeekin Instructor: Diann Brei
Student: Mary McMeekin Instructor: Diann Brei
Abstract not yet available.
Inflatable Attachments
Student: Jesse Velleu Instructor: Diann Brei
Student: Jesse Velleu Instructor: Diann Brei
Abstract not yet available.
Software-defined Control of Smart Manufacturing Systems
Student: Michael Murray Instructor: Kira Barton
Student: Michael Murray Instructor: Kira Barton
Abstract: Towards anomaly detection for software defined control of smart manufacturing systems
Software Defined Control (SDC) is a novel approach which seeks to improve overall equipment efficiency in large manufacturing systems through dynamic reconfiguration. SDC dynamically reallocates resources and reroutes parts in the production system in response to anomalies, faults, cyber-attacks, or changes in production demand. A crucial component of the framework is the timeline detection and isolation of anomalies and faults. We propose an anomaly detection framework for a tabletop smart manufacturing testbed which will be used for software defined control studies. The closed-loop operation of the table-top testbed was formalized with a set of timed automata. A separate model was created for each manufacturing cell. In this study, timed automata were used as references for the detection of anomalies in temporal behavior and controller faults. The framework followed in this work will be applied in more challenging scenarios (e.g., larger testbeds, industrial systems) in future studies.
7. Fluids and Gases
Development of an Aerosol Deposition System for Battery Prototyping
Student: Kenny Van Instructor: Jeff Sakamoto
Student: Kenny Van Instructor: Jeff Sakamoto
Abstract: In order to meet the growing demand for a safer, more compact, faster charging, and more durable battery, advanced battery technologies and alternatives must be developed. One possible solution is the solid-state battery. The University of Michigan Energy Institute is currently focusing on developing solid-state battery prototypes. A major hindrance to the development of the battery is its difficult and tedious manufacturing process. Due to this complicated process, potential electrolytes cannot be efficiently characterized, leading to an undesirable development timeline. One promising method to solve this problem is Aerosol Deposition. Aerosol Deposition is a pre-commercialized ceramic film formation technique that uses a pressurized inert gas to spray and deposit a film of a fluidized powder material onto a substrate material at room temperature. Presently, the Aerosol Deposition method has been used to create ceramic films and protective coatings. Aerosol Deposition will be a viable technique for solid-state battery manufacturing if it shows the ability to develop films from powdered electrolytes. If the method is successful, the film of electrolytes will be able to be easily and efficiently tested for feasibility. This study will explore the development of an Aerosol Deposition prototype for use in solid-state battery research.
The Effects of Diffusion of non-Condensable Gases on Cavitation in Soft Tissue
Student: Daniel Knister Instructor: Eric Johnsen
Student: Daniel Knister Instructor: Eric Johnsen
Abstract: Acoustically controlled cavitation is used in therapeutic ultrasound techniques to ablate undesired tissue such as tumors. Over many oscillation cycles, it is expected that there may be some appreciable mass transfer of non-condensable gases dissolved in the surrounding media. The goal of the present work is to look at how the diffusion of non-condensable gases affects the dynamics of the bubble as well as the surrounding tissue. The first step in this approach was to look at the full convection-diffusion equation for dissolved gas concentration in the surrounding media through a numerical scheme in MATLAB. In order to lessen the significant computational cost involved in finding the mass transfer through this method, the next approach was to look at simplified models for mass diffusion and see if these could be appropriately incorporated into the overall bubble model. At this stage, the numerical model of the full convection-diffusion equation is still being validated and reviewed, and the simplified models are being added into an existing overall bubble model.
Ultraviolet Cleaning of Diesel Fuel
Student: Benjamin Golder Instructor: Andre Boehman
Student: Benjamin Golder Instructor: Andre Boehman
Abstract: Hydrocarbon fuels that are stored in large tanks are often subject to bacteria and fugal growth due to dark and damp conditions that are favorable for such growth. Bacteria and fungal growth can cause problems in engine performance if the fuel is not properly cleaned. The purpose of this study is to research the practicality of using ultraviolet light to clean diesel fuel. This research study attempts to identify the optical properties of different diesel fuels that are often stored in large storage tanks. From these properties, the practicality of a filtration system will be assessed. Experiments were performed by first radiating germicidal ultraviolet light through different diesel fuels. Each fuel was tested with multiple light path lengths. The intensity of the light at the end of the light path was measured. From these numbers, the Beer-Lambert law was used to find the light properties of diesel fuel such as the proportionality coefficient of absorbance.
Keywords: Diesel, Fuel, Ultraviolet, Cleaning, Hydrocarbon, Optical, Properties, Germicidal
Edge Detection Method to support differential approach to CBR
Student: Xi Chen Instructor: Bill Schultz
Student: Xi Chen Instructor: Bill Schultz
Abstract: Professor William Schultz and Dr Louise McCarrol are working on a differential approach for Capillary Breakup Rheometry (CBR) to determine fluid surface tension to viscosity ratio. An important step of this approach is to analyze the fluid filament photos taken by a high-speed camera and derive the fourth-order derivative of fluid surface profile. A new edge detection technique combining Canny edge detection and the sum of gray level pixels is proposed. This technique consists of two steps: the first step is to apply Canny edge detection scheme to find the approximate pixel-level edges, and draw a mask around the edges in the gray level figure; the second step is to sum gray levels over columns to determine edges. For images with low noise intensity, the technique works effectively to detect subpixel resolution. Tests results are presented show that the technique can withstand blurriness and offset of the color space.
Keywords: Edge detection; Subpixel edge detection; Curve fitting; Image segmentation