Mechanical and Mechatronics Engineering

Permanent URI for this collection

https://uwspace.uwaterloo.ca/handle/10012/9912

This is the collection for the University of Waterloo's Department of Mechanical and Mechatronics Engineering.

Research outputs are organized by type (eg. Master Thesis, Article, Conference Paper).

Waterloo faculty, students, and staff can contact us or visit the UWSpace guide to learn more about depositing their research.

Browse

Now showing 1 - 20 of 1524

A New Tightly-Coupled Dual-VIO for a Mobile Manipulator with Dynamic Locomotion
(University of Waterloo, 2024-09-12) Xu, Jianxiang
This thesis presents a novel approach to address the challenges encountered by a mobile manipulator engaged in dynamic locomotion within cluttered environments. The proposed technique involves the use of a dual monocular visual-inertial odometry (dual-VIO) strategy, which integrates two independent monocular VIO modules, one at the mobile base and the other at the end eﬀector (EE). These modules are intricately coupled at the low level of the factor graph to provide a robust solution. The approach leverages arm kinematics to treat each monocular VIO as a positional anchor in relation to the other, thereby introducing a soft geometric constraint during VIO pose optimization. This mechanism eﬀectively stabilizes both estimators, mitigating potential instability during highly dynamic locomotions. The performance of the proposed approach has been rigorously evaluated through extensive experimental testing, directly comparing it to the concurrent operation of independent dual Monocular VINS (VINS-Mono). The envisaged impact extends beyond the speciﬁc application, as the approach may lay the groundwork for multi-VIO fusion and enhanced system redundancy within the realm of Active-SLAM (A-SLAM).
Design, Fabrication, and Testing of Assistive Mobility Solutions for Health Care Facilities
(University of Waterloo, 2024-09-10) Braimah, Maltiti
The movement of people, medication, and equipment in health care is a crucial part of the workflow, however, it is also an area that causes significant physical strain to the workers. This thesis proposes a design which is used to retrofit existing medical beds and carts with motorized modules that reduce the amount of force required for a worker to move equipment within hospitals and long-term care homes. The unique design makes use of mechatronic design principals to offer reliable and effective solutions to mobility problems in health care environments. Though available motorized beds and carts exist, they tend to be expensive as they are integrated in the bed and cart designs, thus requiring new equipment to be purchased. This is not always possible as it requires a large initial investment, while also introducing a foreign piece of equipment that workers must learn to use. The designed solution solves this by considering the existing design of medical equipment and adding technology to its frame to vastly reduce the costs. Adding onto the established medical designs allows for the electromechanical assembly to be the focus of the retrofit, coupling its additional functionality to familiar equipment. In this thesis, modules for improving beds and carts are designed, fabricated, and tested. The mechatronic design of the bed establishes a set of components and parameters that are successful in allowing for locomotion in indoor environments without the need for significant muscle exertion. The thesis scope details the mechanical design and electric and pneumatic component selection of the bed for a platform created to allow for manual, assistive, and autonomous controls. Using the experience and observations made while designing the bed platform, a more general assistive cart design is created. The cart offers a streamlined module that offers longitudinal torque assistance and can be paired with a wider range of medical equipment. An assistive handle allows for a seamless control experience on the assistive cart, wherein the assistive torque provided by the motorized module can be scaled by sensing the force exerted on the handle. This is observed to reduce the force needed to push a cart, and significantly decreases the effort needed to maneuver a heavy cart. The cart is also proven to be effective in its intended environment through limited prototype use in Grand River Hospital. An alternative cart is also designed, which can couple and detach from carts. This solution allows for a large fleet of similar carts to be powered using one product, rather than requiring a module for each cart.
Design and Enhancement of Fiber Optic Evanescent Wave LSPR-Based Biosensor Fabricated via Self-Assembled Silver Nanoparticles for Salivary Cortisol Detection
(University of Waterloo, 2024-09-05) Awad, Omar
There are currently no point-of-care means accessible for non-invasive, real-time monitoring of cortisol levels. All available home test kits for cortisol testing require the patients to send out their samples to laboratories after collection and wait days or weeks in some instances to receive their results. Cortisol is an important stress biomarker that is at the core of many mental and physical disorders, and early detection of abnormal cortisol levels is crucial for their prevention. This work aims to fabricate a highly accessible, inexpensive, sensitive, and non-invasive point-of-care salivary cortisol sensor, utilizing a fiber optic evanescent wave sensor (FOEW) enhanced by the effect of localized surface plasmon resonance (LSPR). The development of this sensor presents an advanced analysis on the parameters affecting the performance of FOEW LSPR sensors fabricated via self-assembly of silver spherical nanoparticles. The analysis is based on a combination of numerical and analytical models validated with experiments. A numerical model is first used to solve Mie’s theory to study the absorption of light interacting with nanoparticles due to LSPR using ANSYS LUMERICAL. Nanoparticles ranging from 10 to 100 nm diameter were considered due to their LSPR properties at this range. Simulations show that silver spherical nanoparticles with 30 nm diameter exhibit the highest absorption efficiency, and that the magnitude of absorption correlates positively with the number of nanoparticles. An analytical model is then used to describe the adsorption kinetics and formation of a nanoparticle monolayer on the longitudinal surface of the fiber’s core based on diffusion transport process. The analytical adsorption model suggested that smaller nanoparticles result in higher final surface density. The performance of the FOEW LSPR sensor signal was modeled utilizing smaller nanoparticle sizes (10 nm, 20 nm, and 30 nm) using COMSOL MULTIPHYSICS. Results show that 30 nm-based FOEW LSPR have the highest absorption signal and refractive index sensitivity. The electromagnetic field decay length was modeled in ANSYS LUMERICAL as well, and the results showed that 30 nm nanoparticles have electromagnetic field decay length that engulfs the conjugated ligand and analyte, which is optimum for cortisol biosensing. The sensor response was also modeled by integrating the ligand-analyte interaction model with the FOEW LSPR sensor model. The modeling results agree with experiments performed on FOEW LSPR sensors fabricated using 30 nm diameter nanoparticles, which show enhanced LSPR signal and refractive index sensitivity compared to results obtained using 20 nm, and 10 nm diameter nanoparticles. Characterizations performed on the prepared fiber sensors using Atomic Force Microscopy (AFM) and Scanning Electron Microscopy (SEM) show significantly higher nanoparticle coverage for sensors fabricated using smaller nanoparticles, which validated the modeling results. A working cortisol biosensor was fabricated using the 30 nm based FOEW LSPR sensor by functionalizing the nanoparticles surface with Cysteamine Hydrochloride, allowing for the bioconjugation of anti-cortisol IgG antibodies. This sensor was reproducible with a sensitivity of 0.0128 nm/nM, a limit of detection of 0.1125 pM and a limit of quantification of 0.3712 pM, covering the range cortisol found in saliva. The findings reported in this Thesis therefore present a promising technology for the fabrication of highly sensitive and cost-effective point-of-care cortisol monitoring devices. Which can be widely accessible and can help with the early detection of abnormal cortisol levels, allowing for the prevention of many mental and physical disorders.
Predicting Energy Savings Associated with Falling-Film Drain Water Heat Recovery Systems in Residential Buildings
(University of Waterloo, 2024-09-03) Manouchehri, Ramin
Falling-film drain water heat recovery (DWHR) systems are heat exchangers utilized for recovering thermal energy from water travelling down shower drains in residential buildings. There are many commercially available DWHR heat exchangers on the market, and there exist a large variety of shower conditions that change from one house to another. For instance, the showerhead fixture and its associated flowrate, shower temperature and the plumbing system in a house are variables that directly impact energy savings, which can vary significantly between different houses and occupants. Clearly, reliable modeling procedures must be created to account for the large variety of operating conditions for DWHR heat exchangers. The main goals for this project were to create and validate robust models to predict energy savings during steady-state and transient conditions, and to use these models in a building simulation software to identify the significance of plumbing configuration on energy savings. To this end, a thorough review of the literature concerning heat exchangers was done to identify the best procedure for modelling DWHR systems. For the steady-state analysis, a novel procedure was devised to derive correlations for the heat transfer coefficients in terms of heat transfer fundamentals. A semi-empirical correlation was also devised to allow faster calculations in cases where correlations for heat transfer coefficients are not readily available. Both methods were validated and were shown to have a mean absolute error smaller than 4%. Additionally, expressions for the transient behaviour of DWHR systems were derived based on heat exchanger theory and it was shown that these heat exchangers can be represented as first-order systems. Experimental data showed that the time constants associated with the typical operation of DWHR systems were generally on the order of a few seconds. Next, the steady-state and transient models were programmed into TRNSYS (Transient System Simulation Tool) to perform monthly simulations to predict energy savings. For simulation purposes, five distinct plumbing configurations were considered, which covered all possible methods a DWHR heat exchanger can be incorporated into the plumbing system for a house. Identical draw schedules were devised and applied to all simulations, and the overall energy consumptions for all simulations were estimated and compared with a base case (i.e. a plumbing system without a heat exchanger). The results showed that small changes in plumbing configuration can have significant impact on the energy savings that can be achieved using DWHR heat exchangers. For instance, the results for one of the simulated heat exchangers showed that the energy savings can range from 25% to 35% depending on the plumbing configuration. These findings were then used as a basis to formulate how DWHR heat exchangers should be modeled, and perhaps more importantly, to highlight the urgent need to update the current building codes and standards to better reflect the actual performance for this particular heat exchanger technology.
Bespoke Binder Jetting System and Powder Bed Temperature Assessment
(University of Waterloo, 2024-08-30) Chai, Matthew
Binder jetting is an additive manufacturing process in which powdered material is bound together via a printhead that jets binder, similar to an inkjet printer. Parts are created from powder and binder, built up layer-by-layer. The parts produced are called green parts. These green parts are often post-processed by sintering in a furnace to fully solidify and densify them to achieve desired mechanical properties and final geometric characteristics. New and continuing developments in additive manufacturing as well as research requirements motivates the development of an open-architecture binder jetting system that can be updated modularly, is capable of producing complex parts at reasonable scale, and which provides full control of the software and hardware to researchers and users. The development of such a system is described herein. To demonstrate the research ability of the system and provide information on binder jetting parameters, the thesis also details the use of thermocouples and a thermal camera to gather data on the effects of infrared heating on powder bed temperature, used for in-situ curing to improve green part properties. The heating rate at the surface of the powder bed can exceed 8°C/s, even for relatively modest heater powers (400 W), and heater parameters can be directly related to the peak temperatures achieved in the bed, consistently producing peak temperatures of over 100°C for linear energy densities of 70 J/mm, providing data for evaporation in aqueous binders. The presence of binder caused substantial differences compared to dry powder, producing consistently lower heating rates until solvent evaporation, but having minimal effect on long-term peak temperatures. The thermal camera was able to produce qualitative data on the surface of the build bed, indicating that the temperature differences across the surface can be significant, on the scale of 10°C to 20°C.
Optimizing Heat and Algal Biomass as Renewable Energy Sources in Building Façade Systems
(University of Waterloo, 2024-08-23) Elmalky, Adham
The growing demand for sustainable design strategies in buildings has underscored the potential of microalgae photobioreactors (PBRs) as a renewable source for heat and bioenergy generation. In cold climates, PBR modules integrated in double-skin façades (DSFs) can mitigate snow deposition impact and minimize adverse effects of heat loss to low ambient temperatures. This research presented a two-part study to develop a heat transfer model of DSFs and establish optimal chemical kinetics for PBRs functionality in the system. In the first study, hourly heat transfer analysis was performed to assess PBRs in terms of energy production, panels’ efficiency and tilt angle, DSF cavity width, and heat gain profiles in cellular and multistory typologies. For this purpose, shading analysis was carried out to evaluate the solar radiation received by the PBRs. The system’s productivity in terms of heat and bioenergy generation was maximized using exploratory and multi-objective optimization algorithms, including graphical and Pareto search methods. In the second study, the chemical model was experimentally validated by applying non-invasive biomass prediction methods using RGB image processing and Neural Networks to predict biomass production from solar energy. The chemical model additionally evaluated the impact of variable flow rate of fresh medium and optimized PBR integration into different façade surfaces, ranging from flat to folded and free-form geometries. In terms of heat transfer in DSFs, PBRs in direct contact with building occupied spaces significantly reduced winter heat loss, achieving an average gain of 35.7 W/m2 compared to a loss of 82.1 W/m2 with conventional façades. It was further noticed that multistory DSFs were advantageous by generating 20.4% more thermal energy and 79.5% more biological energy than cellular DSFs. A developed Trigonometric Model for shading prediction was comparable to Polygon Clipping and Pixel Counting techniques by achieving an average error of 5.2% while significantly reducing simulation time by over 40%. Computationally, Neural Network – Aided algorithms substantially reduced optimization time from 17.2 hours to 6.5 minutes on average. In relation to chemical analysis, variable residence time of microalgae increased biomass generation by 28.8% and CO2 extraction by 10.8% in various façade geometries. Overall, this research established a comprehensive framework on the behavior of DSFs with integrated renewable sources that can alleviate the burden of climate change in the built environment.
Pressure-Velocity Coupling in Transpiration Cooling
(University of Waterloo, 2024-08-22) Hillcoat, Sophie
Transpiration cooling is an active thermal protection system of increasing interest in aerospace applications wherein a coolant is effused through a porous wall into a hot external flow. The wall is cooled due to convection within the porous structure and the formation of a coolant buffer layer between the wall and the main flow. In order to design an effective transpiration cooling system, it is necessary to understand the complex interaction between the high-temperature turbulent boundary layer and the pressure-driven coolant flow through the porous wall. However, high fidelity simulations of this interaction are rare and computationally limited to simple problems when accounting for both fluid domains. In the present work, two shallow convolutional neural networks (CNN) were trained on pore-network simulations of flow through a pressure-driven porous wall. CNN were used due to their ability to consider spatial correlations, meaning they can capture the influence of flow between neighbouring pores. The CNN are coupled with direct numerical simulations of a turbulent boundary layer over a massively-cooled flat plate. The coolant flow inside the porous medium is thus indirectly coupled to the near-wall pressure in the boundary layer, allowing the interaction between the two flow domains to be considered. Linear expressions that do not account for flow interactions between neighbouring pores were also coupled with direct numerical simulations in order to investigate the significance of this effect. With the incorporation of the pressure-coolant injection velocity coupling, the streamwise variation in mean pressure was found to have a significant impact on the local coolant injection. Blowing was reduced near the beginning of the transpiration region due to the high pressure region formed by the incoming boundary layer flow encountering the injected coolant. This effect reduced as the flow recovered and eventually reversed as the coolant film accumulated in the boundary layer, lifting it off of the wall. A corresponding trend was observed in the local cooling effectiveness, while the inverse was found in the local friction coefficient. The incorporation of lateral flow between neighbouring pores via the neural network was found to greatly attenuate these coupling effects. In all coupled cases, the turbulent kinetic energy was reduced at the beginning of the transpiration region due to the more gradual introduction of coolant. However, further downstream the rapid increase in coolant injection in the cases coupled using linear expressions resulted in increased turbulent production such that the turbulent kinetic energy was greater than in the uniform injection case at the end of the transpiration region. In the neural network cases, the increase in shear due to increased coolant injection was not significant enough to overcome the modulation of the turbulence due to the coupling. An analysis of the power spectral density of the pressure fluctuations at the wall within the transpiration region revealed that the implemented pressure-coolant velocity coupling only attenuated the largest scales of the turbulence, leaving the smaller scales relatively unaffected.
MPC for Off-Road Vehicle Trajectory Tracking
(University of Waterloo, 2024-08-21) Tuer, Steven
This thesis explores the trajectory tracking control of autonomous off-road vehicles. The research aims to create a suitable trajectory tracking controller for use in industries such as mining, agriculture, and material transport, where enhancing safety and efficiency is paramount. With these applications comes the need to translate across varying off-road conditions. The complexity of off-road environments poses significant challenges, including varying bank and inclination angles, changing traction conditions, vehicle payloads, and complex terrain-tire dynamics. Furthermore, the need for real-time performance mandates that the controller be efficient, as traditional Non-linear Model Predictive Control (NL-MPC) is too computationally intensive for practical and cost efficient use. To tackle these challenges, a coupled controller for longitudinal and lateral control has been developed using a dual-track vehicle with a linear tire model. This physics-based approach incorporates road angles into the formulation to account for significant bank and inclination angles. Simulation results in off-road scenarios indicate that Road Angle Model Predictive Control (RA-MPC) shows potential for improving trajectory tracking. However, in practical applications, accurately estimating these angles remains difficult due to the varying planes on which the tires operate. This direct modelling approach also limits the generalizability of the controller in off-road conditions as other sources of unmodelled dynamics are ignored. To enhance the controller's performance in a broader way, a separate method of compensating for off-road modelling complexities through the use learning methods is explored. Gaussian Process Regression (GPR) is employed to improve tracking performance through data-driven modelling of complex off-road dynamics. While this thesis focuses on the integration and inital proof of concept using learning to augment the MPC formulation, the results demonstrate that GPR-MPC can effectively compensate for path inclination and bank angles as well as other sources of unmodelled dynamics. Notably, GPR-MPC excels in low friction (low mu) scenarios where there are significant parameter mismatches in the physics-driven MPC formulation.
Skill Transfer from Multiple Human Demonstrators to a Robot Manipulator Using Neural Dynamic Motion Primitives
(University of Waterloo, 2024-08-15) Hanks, Geoffrey
Programming by demonstration, also known as imitation learning, has shown potential in reducing the technical barriers to teaching complex skills to robot manipulators. It involves obtaining one or more demonstrations of how to complete a task, often from a human, which are then transferred to a robotic system. Dynamic Motion Primitives (DMPs) are an efficient method of learning trajectories from individual demonstrations using second-order dynamic equations. Research has been done to overcome some of the limitations of DMPs, by generalizing over multiple demonstrations, sequencing multiple primitives to complete goals involving multiple sub-tasks, and adding via-points for increased control over complex motions. However, accomplishing more complex tasks using DMP sequencing and via-points requires task specific knowledge so that the demonstrations can be segmented or annotated, and the breakdown of some tasks may be unintuitive. This can further increase the time and effort required to collect demonstrations beyond the already demanding process of collecting physical demonstrations, decreasing the feasibility of learning from demonstration in certain situations. This thesis applies state of the art Cartesian space DMPs that utilize physically collected and augmentation data to create a framework that can reduce the task specific knowledge and human effort required to teach robots multi-step tasks. DMPs that integrate neural networks are used, not only to generalize over multiple demonstrations from different demonstrators, but also to learn from complete demonstrations without requiring segmentation or annotation. For comparison, sequenced DMPs which require their demonstrations to be segmented into sub-tasks prior to learning are also implemented. Both techniques utilize physically collected demonstrations which are augmented to reduce the time and effort required to collect demonstrations, while ensuring sufficient samples for proper learning. The framework was tested on a pouring task which could be split into sub-tasks, and was tested both in simulation and on a 7 degree of freedom Franka Emika Panda robot manipulator. The task involved reaching for and grasping a container with water, pouring water into another container placed in the workspace, and returning the pouring container to its original location. Both sets of models were tested on their ability to recall trajectories shown in training, and generalize to new inputs. They were then implemented on the physical robotic system, and both methods were successful in completing the task. The trade-offs between the models trained on full and segmented demonstrations are discussed. While the sequenced DMPs were found to have reduced average error and greater flexibility, they required extra work and task knowledge to generate the demonstrations, and were reliant on specific subtasks being defined. It was determined that the models trained from full demonstrations using this framework could be an alternative to sequence primitives for more complex tasks. Despite a higher error between the demonstrations and predicted trajectories when compared to a sequence of DMPs, the full models are able to recall trajectories, generalize to new inputs well enough to complete the task on a physical robot. As such, they have the potential to reduce effort and task knowledge during demonstration preparation, and expand the applicability of imitation learning to a wider range of tasks.
Advancing Applications in Fluid Powered Artificial Muscle Technology Through Artificial Intelligence Modeling and The Development of a Posture Sensing System
(University of Waterloo, 2024-08-14) Savage, Jordan
This thesis aims to improve human posture by exploring the development and integration of Pneumatic Artificial Muscles (PAMs) and intelligent sensing systems. The objective is to not only develop an efficient posture corrector but also to make a meaningful contribution to the fields of PAM and wearable technology research in the format of delivering a tool to enable the design and optimization of PAMs for wearable applications. To achieve these objectives, three primary projects are designed that show a cohesive progression in the creation and application of these technologies. As compared to other actuators, PAMs can output larger forces which are influenced by many parameters such as their geometries and manufacturing methods. Developing a functional tool for designing and optimizing PAMs is not trivial. The first project involves the creation of ForceSight, by leveraging AI advancements, a tool enabling designers to accurately size PAMs based on specific force requirements. ForceSight predicts the force output of various actuator geometries, thereby simplifying the design process and enhancing the customizability of PAMs for diverse applications. Inertial Measurement Units (IMUs) are sensors used in wearable systems to detect body motion using accelerometers, gyroscopes, and magnetometers. In this thesis, IMUs are used to determine the sagittal slouch and shoulder rounding states of a human subject for posture correction. The focus of the second project is to use a Fuzzy Inference System (FIS) is used to classify the IMU data. This system utilizes data from two 9-axis shoulder-mounted IMUs, emphasizing magnetometer data to assess shoulder rounding, and a lumbar IMU to monitor sagittal slouch posture. The final FIS reliably detects compound slouching motions, providing comprehensive posture assessment based on the sensor data. To validate the AI tool for the design and optimization of PAMs for wearable applications, the third project utilizes the ForceSight tool for determining suitable actuator geometries for a posture corrector and employs the FIS to evaluate shoulder rounding. This section also demonstrates the workflow and benefits of using ForceSight, highlighting its open-source nature and the potential for expanding its data pool to enhance prediction accuracy and applicability across various fields. The posture corrector is an example of evaluating the effectiveness of the developed AI tool, which has broad implications across many industries such as aerospace, industrial automation, and wearable devices. It is the hope of the researchers involved that the technologies demonstrated within this thesis can increase the implementation of PAMs in new areas and use cases across the world.
In-Situ Monitoring and Quality Assurance Algorithms for Laser Powder-Bed Fusion using Optical Tomography
(University of Waterloo, 2024-08-13) Ero, Osazee
Conventional methods for evaluating the quality of components produced through laser powder-bed fusion additive manufacturing (LPBF) are often costly and resource-intensive. These post-production techniques involve mechanical testing, detailed metallographic examination, and non-destructive methods like X-ray computed tomography (CT) to identify flaws. Recently, there has been a shift towards employing in-situ monitoring systems, such as optical tomography (OT), which capture near-infrared light emissions to detect defects arising during LPBF. This dissertation introduces innovative approaches for defect detection in LPBF, utilizing OT data alongside machine learning techniques. LPBF processes inherently exhibit random behavior, presenting challenges in developing robust defect detection algorithms adaptable to diverse machine setups and process parameters. The proposed model integrates a self-organizing map (SOM), a fuzzy logic scheme, and a tailored U-Net architecture to detect and predict defect probabilities in LPBF-produced parts using in-situ OT analysis. Specifically, the model is designed to identify common flaws such as lack of fusion and keyhole defects. The effectiveness of the approach was validated through a series of experiments. Initially, the influence of process parameter selection on recorded in-situ optical tomography (OT) data was investigated. This was followed by the intentional and random recreation of process defects to simulate the stochastic nature of real-world manufacturing processes and to gain a deeper understanding of defect formation. The developed model was subsequently evaluated on a complex geometry to assess its performance in a practical setting. Validation of the model was done by comparing its predictions against computed tomography (CT) scans, to achieve this, Dynamic Time Warping (DTW) technique was used to measure the similarity between porosity curves generated by the model and those from CT scans. The developed model effectively predicted porosity resulting from lack of fusion or keyhole defects across various process parameter settings, achieving average Euclidean distance scores of 0.243 for lack of fusion pores and 0.6 for keyhole pores. Additionally, the model successfully detected defects in complex geometries with internal lattice structures. A significant advantage of this developed model is its adaptability. Fuzzy logic allows for the integration of soft decision boundaries and expert rules into the model, which is crucial when dealing with complex phenomena like porosity where the boundaries between the presence of defects in the fabricated part, based on monitoring OT data, are not always clear-cut. Expert knowledge can be encoded into fuzzy rules that mimic human reasoning and decision-making processes. Quality assurance experts can use their experience to provide insights through the application of fuzzy rules, determining how certain visual or measurable features of an image typically correspond to specific types of porosity. They can also adjust the probability threshold for defect detection based on specific quality criteria. This adaptability enhances the approach's utility across diverse manufacturing scenarios, offering flexibility in meeting quality assurance requirements.
An examination of escape times in a mock residential test house by analyzing smoke filling and detection times for couch fires in a living room
(University of Waterloo, 2024-08-08) Senez, Peter
A series of full-scale multi-compartment and multi-storey fire experiments are undertaken to evaluate the potential “available safe escape time” (ASET) in fires fueled by different upholstered furniture types, burning under controlled ventilation conditions in a representative multi-storey residential dwelling. Ten different furniture fire experiments are conducted to fill key gaps in current understanding of fire growth and behavior, smoke filling, detection times, and available escape times in limited ventilation conditions as established within the well-instrumented “burn house” at the University of Waterloo Fire Research Laboratory. Findings for mass loss rate, smoke movement and resultant visibility, oxygen consumption, and the evolution of carbon monoxide along the “escape path” are compared across furniture types. An estimate is made for the available escape time accounting for occupant movement while upright, or when the environmental conditions deteriorate, movement out of the house by crawling. Findings show that the North American furniture fires result in crawl-out available escape times between 96 and 238 seconds whereas the UK furniture fires have significantly longer crawl-out available escape times, between 281 – 1487 seconds. In slower-burning fires, the importance of considering the incipient fire time in comparing detector response highlights the benefit of placing detectors near fuel loads in living areas, which in this study allows for a 127% increase in available escape time. An overall increase in available escape time of 19 – 50% is found for the fast-burning North American furniture. Flame-retardant interliners, combined with flame-retardant treatments, are observed to significantly limit the exposure and participation of the polyurethane foam in the fire, controlling fire growth in one instance and preventing sustained ignition in two other couches. Further examination of key elements of furniture fire behavior illustrates that ignition, incipient fire time, growth period, and peak mass loss rate are critical functional parameters. These can be used to define a scoring system by which to compare the fire performance of different couches. Through detailed analysis of video evidence, several unique patterns of smoke filling are observed. These include encapsulation of the fire in the burn room, a smoke hazing effect, visual evidence of a smoke layer within the smoke layer, and smoke layers that ascend from floor to ceiling. Due to their importance in the determination of visibility through a given escape route, these are worthy of more in-depth investigation in future research. Overall, the research fills key gaps in our understanding of the evolution of fire environments in limited ventilation residential furniture fires as related to available time for occupant escape. As such, it highlights the need for defined fire safety objectives in house design to improve residential fire safety.
Investigation of Brain Response in Canadian Armed Forces Volunteers Subjected to Recoil Force from Firing Long-Range Rifles Using Instrumented Mouthguards and Finite Element Head Model
(University of Waterloo, 2024-07-18) Seeburrun, Tanvi
Mild traumatic brain injury (mTBI) may be caused by occupational hazards military personnel encounter, such as falls, shocks, exposure to blast overpressure events, and recoil from weapon firing. The repeated exposure of Canadian Armed Forces (CAF) members to sub-concussive events during the course of their service may lead to a significant reduction in quality of life. Symptoms may include headaches, difficulty concentrating, and noise sensitivity, impacting how personnel complete their duties and causing chronic health issues. CAF members have reported experiencing symptoms of mTBI, and some studies have associated these symptoms with repeated firing of long-range rifles. However, there is limited physical data on head response resulting from rifle recoil and different rifle configurations. The objectives of this study were to quantify head kinematics for volunteers, assess the head response using kinematic-based metrics, and assess brain response using a detailed finite element head model. Measurements of head motion were recorded in a group of military volunteers using instrumented mouthguards while firing long-range rifles. The head kinematics were then used as inputs in a finite element head model to calculate the brain strains for each firing event and assessed using common response metrics and a Cumulative Strain Volume (CSV) measure to quantify brain deformation resulting from head acceleration. The measured head kinematics and predicted brain deformation among CAF volunteers were lower than those associated with acute injury. The study highlighted the corpus callosum as the primary site of higher strains in the brain, consistent with previous research on head response to acceleration events. Brain deformation was primarily associated with angular velocity rather than linear acceleration. Comparative analysis between different rifle calibers revealed higher values of head kinematics associated with increased rifle caliber, owing to the higher level of energy. The CSV method identified statistically significant differences between rifle configurations and reductions in brain deformation with a recoil mitigation system (RMS), offering a potential solution to reduce long-term symptoms from firing long-range rifles. The results of this study offer important information about the magnitude of kinematics and strains that volunteers experience when firing long-range rifles.
Low-Frequency Acoustic Source Detection and Localization
(University of Waterloo, 2024-07-15) Joshi, Arnav
In aviation, clear-air turbulence (CAT) is a major cause of in-flight injuries. It occurs in cloudless skies and cannot be detected by the onboard weather radar. Studies have predicted the extent of CAT to increase substantially in the next few decades, thus necessitating a method for detecting CAT. With CAT known to generate low-frequency and infrasonic acoustic emissions, acoustic-based methods can potentially be deployed for detection and localization. This thesis studies low-frequency acoustic source detection and localization in the context of CAT. Localizing low-frequency acoustic sources is challenging for acoustic beamforming which suffers from poor resolution at low source frequencies. A deep learning-based method is adopted as an alternative. Deep learning models for two-dimensional and three-dimensional acoustic source localization (ASL) have been built using synthetic data and computationally inexpensive neural network architectures. These models are necessary to prove the viability of deep learning for low-frequency ASL. The thesis then explores the potential of a deep learning-enabled, acoustics-based method for CAT detection in the future through a large-scale, virtual flight case, set up for the detection of a representative CAT source. The flight case tries to predict what an in-flight microphone will detect around a CAT source through a technique known as auralization which simulates the acoustic field of a source by modeling the sound propagation and determining what a receiver would hear. The deep learning models yield promising qualitative and quantitative results that prove the feasibility of using deep learning for low-frequency ASL. Combined with the results from auralization, it can be concluded that there exists considerable scope for deep learning-enabled, acoustics-based detection and localization of CAT. The future work involves expanding the current scale of research with deeper network architectures to process real, in-flight acoustic data.
Anomaly Detection in CNC Milling Machines using Transfer and Incremental Ensemble Learning of LSTM Autoencoder Networks
(University of Waterloo, 2024-07-08) Li, Eugene
Since the industrial revolution, there has been a steady and continued effort to bring more automation and efficiency to the manufacturing process. Milling machines have long been a valuable tool to create precise parts quickly and effectively. CNC Milling machines have been the next evolution of automation in manufacturing and have allowed for complex parts to be produced quickly and with high precision. Although CNC milling machines are able to semi-autonomously produce parts effectively, they still require significant human interaction to operate. This human interaction is especially true since many tasks are completed in open loop configuration with little to no feedback. To try to address this issue there has been significant effort in the literature to develop systems to provide feedback to the machine controller. This work is often focused on detecting anomalies such as chatter, broken tools and other conditions that will impact the surface finish or machine health. A limitation of much of the current work is that it tends to be machine or material specific. These approaches developed in the lab often do not scale well to production as they require custom setups or complex machine dynamics to be studied. To overcome this problem, this thesis proposes a machine learning based solution that leverages deep learning to create a solution that can potentially be quickly and easily transferred to machines in production. In this thesis we demonstrate that by using simple accelerometers mounted on the spindle of a CNC milling machine, we can create an LSTM-Autoencoder to detect anomalies such as chatter. This feat is accomplished by creating an artificial neural network that is trained on sensor data from stable cutting conditions. This network aims to reproduce the original signal with as little error as possible. If the network is provided data from a stable condition it will reconstruct the signal with little error, but if it is presented data from an anomaly condition it will reconstruct with significant error, which indicates that an anomaly is present. In Chapter 4 we show that this approach can also be achieved by implementing what is known as transfer learning. In transfer learning we begin with a network that is trained on one source data set, and then transfer the knowledge to another target data set. We investigate under what conditions this is most feasible and demonstrate that we can train a network from a robust data set on one three-axis CNC machine and then transfer it to another three-axis CNC machine. We also demonstrate that this method works for both chatter detection and broken tool detection. In Chapter 5, we introduce an incremental learning method based on ensemble learning. This approach takes the LSTM-Autoencoder trained previously as a strong learner and has weak learners continually learn as new data is made available. This approach is shown to have comparable results to a large network trained from scratch and improves the performance of a system trained with transfer learning. Taking these transfer learning and incremental learning algorithms, we extend the approach to anomaly detection for five axis CNC milling machines in Chapter 6. This is accomplished by introducing a stacked ensemble learning approach by transferring the encoder from the three axis CNC anomaly detection algorithm and then combining it with an encoder and decoder that is trained from the target data. Incremental learning is then integrated by adding weak learners to this strong learner. These weak learners allow the network the ability to improve the performance of the system to be comparable to a network trained from scratch with a fraction of the data. Lastly, we demonstrate how these approaches can be used for multi-class prediction in Chapter 6. In this chapter, we use the LSTM-Autoencoder to perform dimensionality reduction. We then use this dimensionality reduced output and apply a one-versus-all SVM classifier and Platt scaling to obtain a probabilistic prediction of the classes of interest. This approach allows us the ability to both detect and differentiate cutting with broken tools and chatter conditions. The approaches presented in this thesis demonstrate that this proposed approach is capable of not only detecting chatter in a specific lab setting, but can potentially be used to detect multiple anomalies across a variety of machines and materials. This allows users to potentially scale these approaches to many machines quickly with minimal setup and minimal configuration.
Additively Manufactured Low-frequency Piezoelectric Energy Harvester Design: Modeling, Fabrication, And Experiment
(University of Waterloo, 2024-06-19) Dinh, Minh Hao
With conventional energy sources like fossil fuels becoming increasingly scarce and the widespread adoption of electric vehicles placing growing demands on lithium, the primary material in battery manufacturing, there is a critical need for scientists and engineers to explore alternative energy sources for powering microelectronic devices. Among these alternatives, integrating piezoelectric materials within cantilever beam structures for energy harvesting applications is a promising solution, attributed to its straightforward design and ability to undergo significant deformation under applied loads. However, this technological approach faces notable challenges, including limitations associated with low power density and a high natural frequency due to inherent geometric constraints. These challenges have become a focal point for ongoing research endeavours to enhance the efficiency and applicability of piezoelectric energy harvesting. This thesis delves into a prospective solution for powering microelectronic devices, emphasizing its merits in terms of uncomplicated packaging and advancements in micro-scale power density. A MEMS ring-shaped piezoelectric energy harvesting device was fabricated, utilizing 3D printing for substrate production and precision dicing techniques to achieve the required dimensions of the piezoelectric material. The device's design was modelled using SOLIDWORKS, and its performance was thoroughly simulated in COMSOL to ensure alignment with observations. Inspired by the Vesper microphone's square form, the energy harvester's geometric configuration offers scalability and the potential for incorporating multiple cantilever beams. According to the findings, this energy harvester demonstrates a total power output of 53.46 $\mu$W when subjected to an acceleration of 0.08g, establishing its promising viability relative to other energy harvesting technologies. The study presents a novel approach to energy harvesting and highlights the practical implications and potential advancements in micro-scale power generation for sustainable electronic devices.
Physics-Based Pressure Field and Fluid Forcing Inference for Cylindrical Bluff Body Experiments
(University of Waterloo, 2024-06-17) McClure, Jeffrey
The proceeding work details contributions to the state-of-the-art of velocimetry-based experimental fluid mechanics through the application of novel pressure and force estimation methods to studies in bluff-body aerodynamics and the problem of vortex-induced vibrations. Together, these techniques allow the measurement of fluid velocity and pressure, in space and time, for an area of interest surrounding an immersed body, along with the estimation of the total forcing on the immersed body. Conditions for optimal data sampling from the velocimetry data for the estimation of pressure fields are approximated analytically, and a variety of common pressure integration techniques are compared. The assessed integration techniques are characterized as having similar accuracy, with minor differences in error sensitivity observed. The errors in the estimated pressure fields can be expressed by considering the conformity of the obtained velocimetry data with the governing equations of motion. Accordingly, an analytical framework is developed which propagates the errors in the velocity field measurement through the pressure calculation. A subset of the error terms may be resolved in practical experiments, while others must remain neglected, in the absence of an extended model. Once equipped with the time-resolved pressure field, a control-volume-based analysis then allows the estimation of time-resolved forcing data. The dependence of the time-resolved force estimations on an often neglected three-dimensional term in the planar momentum balance is shown analytically. As a result, specific recommendations are provided for experimental best practises and field of view selection for obtaining accurate time-resolved forcing data from planar velocimetry measurements. Finally, following the previous methodological verification studies, the post-processing techniques are applied to an experiment of a stationary cylinder and that undergoing forced oscillations in a steady free-stream. The three-dimensional flow field surrounding the body is statistically reconstructed along with the pressure estimates in order to resolve the velocity/pressure and force distributions in the volume immediately surrounding the cylinder.
Defect Engineering in Metal Oxide Through Laser Irradiation
(University of Waterloo, 2024-06-14) Zheng, Shuo
Metal oxides such as copper oxide (CuO), titanium dioxide (TiO2) and zinc oxide (ZnO) are one of the most important and broadly-studied classes of semiconductors. Their nano-materials have shown great potential in the development of functional nano-devices. In metal-oxide nano-materials, zero-dimensional point defects are believed to play a central role in the control and optimization of their properties because they would be strongly determined by the nature, concentration and arrangement of these point defects. Therefore, tailoring the properties of metal-oxide nano-materials for targeted applications through engineering these characteristics of defects is of growing interest. Laser irradiation, as one of the emerging technologies, has shown the ability of engineering point defects in metal-oxide nano-materials through. However, the information on the characteristics of these laser-induced defects remains limited and further investigations are highly required to understand the defects related properties and how the defects can be introduced. In this thesis, the following research works were demonstrated. The CuO nanowires (NWs), one of the most popular p-type metal-oxide nano-materials, were prepared by thermal oxidation and irradiated by ns laser. The produced defects resulted in intragap energy levels that narrow the bandgap of CuO, which gives rise to improved absorption in the visible region. The concentration of defect centers after laser irradiation increases electrical conductivity by a factor of two at a forward bias of 15 V and enhances photo-conductivity of the Au/CuO/Au structure, tripling the optical gain of these structures. Besides, the concentration of defects was also successfully tailored in ZnO, an n-type metal oxide. The defect-to-lattice ratio of oxygen species can be tuned in a range between 0.24 and 0.61. The increased concentration of defects in ZnO thin films resulted in narrowed bandgap energies and extended the photo-response of these ZnO thin films into the visible region. Next, the control over the distribution of these defects was explored. CuO NW films were grown and surface defects were introduced through laser irradiation, which were verified by electrochemical measurements. Further control over the arrangement of the defects was demonstrated in ZnO NWs. ZnO NWs with abundant defects locating at the surface regions (within 1.5 nm from the surface) and residing in the region as deep as 6 to 12 nm were obtained, respectively. The surface-to-bulk ratio of defects in ZnO NWs can thus be modulated by tuning the laser fluence and exposure time. ZnO NWs with abundant surface defects showed enhanced photodegradation rate of dye molecules while the ZnO NWs with more bulk defects exhibited less efficiency. Lastly, the type of defects was tailored in Cu2O and ZnO thin films through laser irradiation under different atmosphere conditions. Either oxygen-rich or oxygen-poor ambient conditions were provided during laser irradiation so that corresponding cationic or anionic vacancies can be generated. The formation of V_O in Cu2O and V_Zn in ZnO thin films leads to the abnormal conductivity types in these materials, resulting in n- and p-type doping respectively. Thin film transistors with complementary conducting channels were then fabricated in Cu2O and ZnO thin films with laser induced defects to show the efficacy of this laser doping process. Overall, the investigation of defect engineering in metal-oxide nano-materials through laser irradiation is still limited and requires more effort. Some of the remaining research questions and potential research studies are stated in the last chapter of this thesis to inspire future research activities.
Variable-Speed and Multi-Mode Solar Assisted Heat Pumps: System Design and Controls Development
(University of Waterloo, 2024-06-13) Howarth, Julian Craig Peter
In an era characterized by increasing energy cost and frequent reminders of the climate impact of emissions from combustion, consumers and energy regulators have a heightened interest in solar and other renewable energy sources. This thesis details work that was undertaken to investigate the performance of a Solar-Assisted-Heat-Pump system (SAHP) for Domestic Hot Water (DHW) made novel by the incorporation of a variable capacity heat pump constructed using a 3-phase scroll compressor whose speed can be modulated by a variable frequency drive. The overall purpose of the system being investigated is to meet the DHW load demands of a single-family household, while reducing the annual purchased energy and thereby reducing operating cost and emissions associated with the residence’s DHW consumption. A key characteristic of the system under study is the variable-speed Heat Pump (HP) which is of custom construction for the current research. A modified factorial experiment was designed and completed to characterize and model the HP’s performance under source and load inlet temperatures and flow rates typical of a mains-connected SAHP system. The resulting multivariate polynomial expressions for HP compressor work rate, source-side heat transfer rate, and load-side heat transfer rate were programmed into a custom component “TYPE” model for TRNSYS, a transient system simulation software package suited to thermal systems. This work represents an improvement on the default HP model in TRNSYS which does not offer the required flexibility to modulate the compressor speed of the HP with a continuously variable input. Validation of the HP model was performed with a separate set of data from those used to fit the model. A multi-mode SAHP system was modeled in TRNSYS to match an Experimental Test Unit (ETU) housed in the Solar Thermal Research Lab (STRL) at the University of Waterloo (UW). Components of the ETU are commercially available Solar DHW tanks and hydronic heating components. In parallel, a whole system TRNSYS model and a physical system representing a multi-mode variable-capacity SAHP were constructed with model components configured to match their analogous physical components. The TRNSYS model was validated at the component level for the heat pump, Heat Exchanger (HX), Auxiliary Electric Resistance Heaters (AUX), and storage tank. The model was then validated as a whole-system operating over day-long trials under a simplified control regime. Daily validation trials showed agreement between simulated and experimental results. Annual performance of a variety of configurations, under a temperature-based control scheme consistent with other studies in the literature were studied. The results of these annual simulations showed some performance benefit of the system under study, but highlighted the need for a more advanced control strategy that would make better use of the variable-capacity HP, and correctly decide when the HP should be used over the HX. Poorer performance of the SAHP system than expected was consistent with other studies findings in the literature that also called for more advanced controls. A Predictive Controller for the variable-SAHP was developed using MATLAB and TRNSYS. The controller works through iterative calls to the TRNSYS system model, the results of which feed into a time-series of control signals that the controller stores and feeds back to the system being operated. Annual simulations were conducted using a top-level TRNSYS simulation in place of the system being operated, and through a MATLAB-link, a separate instance of TRNSYS was used for the iterative sub-simulations. These simulations showed a marked improvement in the performance of the system under the new predictive control scheme compared with simulations of temperature-based control. This improved performance is taken to represent an approximation of the maximum performance of the SAHP system being studied because the predictive controller has selected the optimum control series for the system under perfect simulation conditions. It is acknowledged that in order to maintain realistic performance predictions from the annual TRNSYS simulation, future work is needed to address how the controller would handle model prediction error when controlling a real SAHP system. As a final verification and demonstration of the work, the new controller created to control TRNSYS simulations was ported to an instance of LABView running on the ETU. The controller was implemented to operate the equipment in the lab as a form of Hardware-In-The-Loop (HIL) simulation. This exercise demonstrates that the concept of predictive control as implemented in this work is capable of controlling a real system under study with the goal of meeting DHW demands. Some disagreement was noted between simulation and experimental operation of the system and explained within the context of limitations of the ETU to reproduce certain losses, and model timing errors that can lead to missed milestones for collection on some poor solar days. A number of suggestions are offered to address the shortcomings uncovered by these verification trials. These suggestions included model improvements, and changes to the controller itself that would make it more robust and capable of dealing with variation between model inputs and the observed conditions.
Assessing the Tissue-Level Response and the Risk of Neck Pain in Rotary-Wing Aircrew using a Finite Element Model of the Neck
(University of Waterloo, 2024-06-06) Hadagali, Prasannaah
Epidemiological studies report a prevalence of neck pain among rotary-wing aircrew (RWA) potentially associated with head-supported mass (HSM), frequent physiologic motions of head-neck, aircraft vibration, and prolonged time in non-neutral head-neck positions. Experimental studies with human volunteers and computational studies using head-neck models have suggested potential causal pathways for neck pain in RWA, including increased activity in muscles and increased forces in the spinal column. However, additional insight is required to understand the interactions of HSM, which comprises a helmet with optional mounted devices, and non-neutral head-neck positions. The present study aimed to simulate RWA non-neutral head-neck positions with the HSM using a detailed finite element (FE) head-neck model to assess the tissue-level biomechanical response and potential sources for neck pain in RWA. A detailed FE head-neck model (NMM50) was extracted from a full human body model of a 50th percentile male. The NMM50 model was enhanced, verified and validated starting sequentially from the ligamentous upper cervical spine (UCS), full cervical spine, and full head-neck with active musculature for physiologic loading conditions (NMM50-Hill-E). The NMM50-Hill-E model was simulated for non-neutral head-neck positions (flexion and axial rotation) using a conventional boundary condition and a novel active muscle repositioning approach, demonstrating the importance of active muscle repositioning on tissue-level response. Finally, the NMM50-Hill-E model with active muscle repositioning was simulated for non-neutral head and neck positions with HSM. The present study demonstrated that the muscle-based method of repositioning the FE head-neck model improved the head and neck kinematic response by capturing the in vivo flexion and axial rotation positions better than the conventional boundary condition method. In the simulated RWA head-neck positions, tissue-level investigations demonstrated an increase in the muscle force, intervertebral disc (IVD) force, endplate stress and annulus fibrosus(AF) collagen fiber strain with an increase in the HSM in flexion. Similarly, an increase in the magnitude of non-neutral position from flexion to a combined position was shown to increase the ligament distraction along with an increase in muscle force, IVD force, endplate stress and AF collagen fiber strain. The detailed FE head-neck model provided valuable insight by predicting tissue-level biomechanical responses in the RWA neck while providing guidance on factors that may contribute to neck pain risk in the RWA.

Browse

Recent Submissions