Search results

Title: Constellation and Detection Design for Non-orthogonal Multiple Access System
Creator: Hao, Xing
Date: 2022
Description: It is well known that the Non-Orthogonal Multiple Access (NOMA) system has the capability to achieve higher spectral efficiency and massive...
Show moreIt is well known that the Non-Orthogonal Multiple Access (NOMA) system has the capability to achieve higher spectral efficiency and massive connectivity. In this thesis, some optimized designs in both code-domain and power-domain NOMA systems are studied. Overall, the main contributions are listed as follows:Firstly, we investigate a NOMA system based on the combinatorial design with a novel constellation design for eliminating the surjective mapping from the linear adding data of multiuser and lowering the complexity of constellation design and Multiuser Detection (MUD). And for further enlarge the connectivity, we propose a Low-Density Codes structure to build a trade-off between the diversity and multiusers in resources by expurgating excessive interference on coding matrices. Therefore, our scheme can not only provide a one-to-one mapping pattern with a sparser multiple access structure but also be adjusted with more flexibility to achieve diversity and transmit a large number of users.Secondly, we proposed a constellation mapping scheme based on sub-optimized signal constellation designs by shaping the receiver’s constellation with a strategy that allows differentiated users by which resolvable points will be received allowing simpler detection and design.Thirdly, a novel NOMA system in uplink with time-delayed symbols is investigated, in which a modified Successive Interference Cancellation (SIC) scheme is used at the receiver side. In conventional SIC, when the transmission power is distributed to one user with trivial shifts to other users, the Bit Error Rate (BER) performance will be decreased significantly. Thus, we evaluated a modified SIC by adding artificial time offsets to the conventional power domain-NOMA (PD-NOMA) between users, which can provide higher degrees of freedom for power allocation of users and reduce mutual interference. And then, the added time offsets can provide additional resources to detect the superimposed signals, then the combination of users’ estimations of overall time slots will be considered to get detection improvements. Numerical results demonstrate that the BER performance of our modified SIC outperforms the PD-NOMA with other SIC-based schemes.Thirdly, we propose a new modulation scheme based on polynomial phase signals (PPS) for downlink and uplink non-orthogonal multiple-access (NOMA) transceivers in both the code and power domains. The PPS leads to outstanding spectral efficiency and bit error rate (BER) performance. We also propose a design criterion for CD-NOMA systems to enable the NOMA system to deploy a large number of users with more flexibility as well as lower design and detection complexity than traditional CD-NOMA systems, such as SCMA and PDMA.
Show less

Title: Fault Detection and Localization in Flying Capacitor Multilevel Converters
Creator: Hekmati, Parham
Date: 2021
Description: This dissertation addresses fault detection, fault localization, and recovery in different topologies of the flying capacitor multilevel...
Show moreThis dissertation addresses fault detection, fault localization, and recovery in different topologies of the flying capacitor multilevel converters to guarantee the safe post-fault operation of the system and maintain load supply. There are multiple contributions of this dissertation, including techniques for device open-circuit fault (OCF) detection in stacked multicell converters (SMCs), a windows detector circuit to track the output terminal voltage levels and current directions, a fast and straightforward active power device OC fault detection and localization technique for the family of flying capacitor multilevel converters (FCMCs), a model-based open circuit fault detection and localization technique for the Buck-FCMC, a new estimator for tracking the voltage of flying capacitors, and fault detection and localization for interleaved converters. Each of these contributions is summarized below.The first contribution of this dissertation proposes a fast and straightforward technique for power device OCF detection in SMCs. The fault detection concept only needs to sense the converter's output terminal voltage and current. The sensed output terminal voltage is compared to a predicted one to detect and localize the OCF. A front-end routing circuit is then added to the SMC to maintain the operation of the converter post fault. The second contribution proposes a window detector circuit to track the output terminal voltage levels and current directions. The window detector circuit detects output terminal voltage level and current direction instead of requiring high sample rates and interrupt loops in the controller.The third contribution proposes a fast and straightforward active power device OCF detection and localization technique for the family of FCMCs, including DC to DC FCMCs, single or multi-phase H-bridge FCMCs, and cascaded H-bridge multilevel converters. This technique only needs to sense voltage and direction of current at the output terminals of the converters to detect and localize the fault. The method compares the measured and the expected terminal voltage while considering the commanded switch states and the terminal current direction. As switches transition to different states, healthy switches are excluded from the set of possible faulty switches until only one faulty switch remains. Coordination of the asynchronous operation of FPGA, DSP, and sensors is addressed for practical implementation. The fourth contribution is a model-based OCF detection and localization technique for the Buck-FCMC using model predictive control. In this technique, state-space equations of the system are developed. Comparison of the measured output inductor with the predicted one from the state-space model is used for the OCF detection and localization. This technique can potentially be used for other converters of the FCMC family. The fifth contribution is a new estimator for tracking the voltage of flying capacitors as the internal states of the FCMC. Using the proposed flying capacitor voltage estimator reduces the number of required sensors compared to the conventional model-based methods. At the same time, the overall technique's robustness to dynamic changes, including startup and load changes, is maintained. The last contribution is open and short circuit switch fault detection and localization for interleaved converters using the harmonic analysis of the output terminal parameters. With this method, monitoring electrical parameters of each leg of the interleaves converters is no longer required for fault detection and localization purposes.
Show less

Title: INTELLIGENT SOLID STATE CIRCUIT BREAKERS USING WIDE BANDGAP SEMICONDUCTORS
Creator: Zhou, Yuanfeng
Date: 2021
Description: Electricity, in its predominant form of alternating current (AC), is at the heart of modern civilization. However, direct current (DC)...
Show moreElectricity, in its predominant form of alternating current (AC), is at the heart of modern civilization. However, direct current (DC) electricity is re-emerging, offering higher transmission efficiency, better system stability, better match with modern electrical loads, and easier integration of renewable and storage resources than AC. DC power is gaining tractions in HVDC or MVDC grids, DC data centers, photovoltaic farms, EV charging infrastructures, shipboard, and aircraft power systems. However, DC fault protection remains a major challenge. Interruption of DC currents is extremely difficult due to the lack of current zero crossings which are naturally available in AC power systems. Conventional mechanical breakers only offer a very limited DC current interruption capability even after significant power derating. Hybrid circuit breakers (HCBs) offers a relatively low conduction loss but a response time too slow to protect many low-impedance DC grids. Solid state circuit breakers (SSCBs) can quickly interrupt a DC fault current within tens of microseconds but suffer from high conduction losses. Furthermore, it is generally difficult for an SSCB to distinguish between a short circuit fault and a normal inrush current condition during the start-up of a capacitive load.The purpose of this thesis is to develop a tri-mode, intelligent solid-state circuit breaker technology using wide bandgap semiconductors (especially Gallium Nitride transistors), referred to as iBreaker. The iBreaker design methodology includes the use of mΩ-resistance GaN and SiC devices, new circuit topology and control techniques beyond the commonly used ON/OFF switch configuration, and integration of intelligent functions without increasing component count. The iBreaker adopts a distinct pulse width modulation (PWM) current limiting (PWM-CL) state in addition to the conventional ON and OFF states to facilitate soft startup, fault authentication, and fault location functions. Key design elements, such as use of wide bandgap (particularly GaN) switches, tri-mode operation, combined digital and analog control, the bidirectional buck topology, variable PWM frequency control and universal hardware/software architecture, are discussed in detail. Multiple iBreaker prototypes, rated at 380 V/20 A and 1000 V/10 A, respectively, are built and tested to validate the proposed SSCB design concept. 99.95% transmission efficiency, passive cooling, and μs-scale response time are demonstrated experimentally.
Show less

Title: INTELLIGENT STREET LIGHTING AND REMOTE POWER UNITS AS CASE STUDIES FOR CITIES TO DECARBONIZE
Creator: Burgess, Patrick G.
Date: 2022
Description: There is a scientific consensus that atmospheric warming caused by the release of emissions will reach critical levels in our lifetime if...
Show moreThere is a scientific consensus that atmospheric warming caused by the release of emissions will reach critical levels in our lifetime if significant efforts are not made to decarbonize our buildings and power grid. The City of Los Angeles is a prime example of the challenges of decarbonizing, balancing global, federal, and state policies and issues and addressing environmental justice. The first research case studies of the details and challenges of decarbonization efforts include the implementation of the first networked light-emitting diode (LED) streetlights in the city of Chicago on IIT’s campus to improve the reliability and economics of its main campus, 2.5 mi south of downtown Chicago. Research shows that these networked LED streetlights greatly reduce a city's rising energy costs, but the CSMART project team has set out to prove the benefits of integrating an intelligent communications and control system with an existing smart grid infrastructure, such as an existing network and supervisory control and data acquisition (SCADA) systems. In addition to assessing the economic and environmental drivers for the intelligent streetlight solution, the project team is dedicated to assessing the potential cybersecurity vulnerabilities of such a system and working to mitigate or eliminate them. The second research case study covers off-grid remote power units providing continuous illumination for safer streets and safer driving that is unaffected by power outages. Thanks to individual lighting control potentially allowing for dimming, blinking, and even color changing, streetlights powered by RPUs can be used as emergency signaling devices, directing traffic during a city evacuation or other emergency. The RPU control and monitoring can be accessed through the cloud, thereby avoiding reliance on local servers.
Show less

Title: Intelligent Battery Switching Module for Hybrid Electric Aircraft
Creator: Kamal, Ahmad
Date: 2022
Description: The growth in world economics, tourism and international cooperation has resulted in significant growth of civil aviation industry. This...
Show moreThe growth in world economics, tourism and international cooperation has resulted in significant growth of civil aviation industry. This growing number of fossil fuel reliant aircrafts will significantly increase waste gas emissions with detrimental impact on the environment. The system efficiency of the aircraft must be substantially improved to reduce the fuel burn and thus waste gas emissions. Therefore, the aircraft industry is pushing towards higher electrification of future aircrafts to increase system efficiency, reduce fuel burn and to lower emissions as well as operational costs. The more electric aircraft (MEA) design concept, commercially realized by Boeing 787 and Airbus A380, increases system efficiency by replacing the mechanical, pneumatic, and hydraulic systems with electrical systems. However, global regulation authorities demand further reduction in waste gas emissions and fuel burn. To meet these stringent demands, the aircraft industry is exploring hybrid electric aircrafts which can significantly reduce fuel burn by electrifying the propulsion train of the aircraft. This higher penetration of electrical energy in the aircraft warrants smart short-circuit protection with ultrafast response time. However, current hybrid aircrafts still use outdated mechanical and thermal short-circuit protection which have historically proven to cause numerous tragedies. Solid-state power controller (SSPC) is an alternate solution which uses semiconductor devices to offer faster response. However, the main drawbacks of SSPCs are their need for active cooling due to higher conduction loss and the use of foldback current limiting approach to limit the inrush current of DC-link capacitor of the powertrain. The foldback current limiting approach degrades the power semiconductor devices used due to excessive heat loss by driving the device near the safe operating area (SOA) limits of the device. This thesis presents a 750V/250A intelligent Li-ion battery switching module (BSM) for hybrid electric aircraft propulsion application. The BSM uses commercially available 1200 V SiC JFET power modules with ultra-low RDSON in parallel to achieve sub-mΩ total on-resistance, comparable to the incumbent mechanical contactor solution. This allows the total nominal conduction power loss of the BSM to be less than merely 23 W, permitting maintenance-free passive cooling. In contrast to the incumbent contactor solution, the BSM has ultrafast response (µs-level) to a fault condition. Which, in conjunction with the reduced fault current stress, significantly improves the operation lifetime of the entire system. The BSM incorporates various intelligent features by implementing a tri-mode operation concept, which allows to pre-charge the DC-link capacitor with a limited charging current in PWM mode. To mitigate single-point failures, several design redundancy measures are implemented to ensure reliability and safety for the aircraft. Design considerations of the circuit and physical design of the BSM are discussed in detail including the design of the custom laminated busbar and thermal analysis. Furthermore, the inherent uncontrolled oscillation phenomenon of the JFET cascode structure is explored and addressed. Finally, the experimental results obtained from the built and tested prototype of the BSM are reported.
Show less

Title: EMAT DESIGN CONFIGURATIONS AND SOFTWARE-DEFINED ULTRASONIC COMMUNICATIONS THROUGH METALLIC CHANNELS IN NUCLEAR FACILITIES
Creator: Huang, Xin
Date: 2022
Description: Nuclear facilities are partitioned into different blocks, and all equipment therein is well-packed for isolation purposes. The primary...
Show moreNuclear facilities are partitioned into different blocks, and all equipment therein is well-packed for isolation purposes. The primary barriers of each block include a thick, reinforced, high-strength concrete wall. The presence of physical boundaries introduces a major challenge to implementing wired or radio frequency (RF) wireless communication. Achieving data communication through the solids channel, especially considering the complex environment in nuclear power plants, is very challenging. Ultrasonic communication is a desirable method for information transfer through solid mediums such as metallic bars or pipes. This thesis is methodologically innovative in the way it seeks the best solution for ultrasonic communications through metallic channels. Therefore, we address the following research areas: 1. The advantages of using electrical-magnetic acoustic transducers (EMATs) as transmitter and receiver; 2. The fundamentals of the EMAT structure and wave generation mechanism for ultrasonic communications; 3. The channel model and behavior of ultrasonic wave propagation in a different structure of solid channels; 4. How to minimize the adverse impact of wave dispersion and reverberation; 5. How to increase the bitrate and decrease the bit error rate (BER) of an ultrasonic communication system; 6. How to utilize the software-defined system-on-chip (SoC) platform for ultrasonic communications; and 7. How to implement secure ultrasonic video transmission through solid channels. In this thesis, we have investigated the feasibility of using Periodic-permanent-magnet electromagnetic acoustic transducers (PPM-EMATs) transmitter and receiver as the information-bearing of ultrasonic waves across the plate channels (shear horizontal waves) and pipe channels (torsional waves). Methods such as time-reversal (TR), pulse shaping, and adaptive equalizer techniques are studied for improving the signal-to-noise ratio (SNR) of ultrasonic communication systems. We also investigated a novel software-defined ultrasonic communication system (SDUC) for real-time video transmission through a highly reverberant and dispersive metallic bar channel. Furthermore, we investigated the feasibility of combining orthogonal frequency-division multiplexing (OFDM) with quadrature amplitude modulations (QAM) for bitrate peak performance. Strategies and guidelines were established for the best solutions to combat intersymbol interference (ISI) caused by the severe reverberation inherent in metallic channels. A practical solution for video transmission, adhering to the Digital Video Broadcasting Terrestrial (DVB-T) standard, was also examined for video streaming transmission of 240p, 480p, and 720p resolutions at 20 frames per second (FPS) across a rectangular aluminum bar (ARB) channel. Through ultrasonic experimental studies for channel analysis, we achieved a peak video transmission rate of 1074 kbps with 3.3×10-4 BER despite reverberation, the multipath effect, and signal fading within the ARB channel.
Show less

Title: Advances in Distribution System State Estimation
Creator: Huang, Jianqiao
Date: 2022
Description: With the increasing penetration of renewable energy in the distribution system, the system states are becoming more volatile. How to maintain...
Show moreWith the increasing penetration of renewable energy in the distribution system, the system states are becoming more volatile. How to maintain the normal operation is an urgent question to the operators. Distribution system state estimation (DSSE) is the key to the monitor and control of distribution systems.Distribution systems feature a larger number of nodes and heterogeneous measurements. Due to these features, directly employing traditional state estimation methods cannot provide fast and accurate estimation results. The existing semidefinite programming based methods show promising for the accuracy, but it is not scalable for a large system. In this thesis, we propose fast and accurate DSSE methods. First, we improve the efficiency of the state-of-art SDP-DSSE method, convex iteration (CI) method. We design a scalable convex iteration method, CDQC, by fully exploiting the radial topology of distribution system. However, the efficiency of CDQC depends on efficient feeder partition solution. It is time-consuming to get a good partition especially when the system is large. Hence, we propose a bus injection based semidefinite relaxation method (SDR-BIM) that fully exploits the radial topology of the network without the need for partitioning the networks. However, SDR-BIM has numerical issue for large scale network. This motivates us to design a branch power model based SDR-DSSE method. The proposed SDR-BPM-DSSE method improves the numerical stability and the increase in the average estimation error of voltage is less than $0.04\%$. To further improve the computational efficiency, we developed a generalized linear power flow model (GLDF) and propose an iterative method to solve the DSSE based on GLDF.Finally, the efficiency and accuracy of the proposed methods are validated on IEEE 13-bus, 37-bus, 123-bus, and 8500-node test feeders.
Show less

Title: Active Load Control in a Synchronized and Democratized (SYNDEM) Smart Grid
Creator: Lv, Zijun
Date: 2021
Description: Smart grid is envisioned to take advantage of modern information and communication technologies in achieving a more intelligent grid in order...
Show moreSmart grid is envisioned to take advantage of modern information and communication technologies in achieving a more intelligent grid in order to facilitate: Integration of renewable resources; Integration of all types of energy storage; Two-way communication between the consumer and utility so that end users can actively participate. The Synchronized and Democratized (SYNDEM) smart grid is regarded as the next generation smart grid. The objective of the SYNDEM smart grid is for all active players in a grid, large or small, conventional or renewable, supplying or consuming, to be able to equally and laterally regulate the grid in a synchronous manner to enhance the stability, reliability, and resiliency of future power systems. In a SYNDEM smart grid, power electronic converters are controlled to behave as conventional synchronous machines. Such converters are called virtual synchronous machines (VSMs).Following the SYNDEM structure, this thesis mainly focuses on developing the VSM technology for the automatic grid regulation at the demand side. The major aim and objective is to achieve active or intelligent loads that can flexibly and automatically take part in grid regulation. Moreover, the active load is expected to have similar grid regulation behavior as other active players in the grid, for e.g., renewable generations. To achieve this, a droop-controlled rectifier is proposed that acts as a general interface for a load to grid. The rectifier is controlled as a VSM so that a load equipped with such a rectifier can take part in grid regulation continuously like a traditional synchronous machine. Such a rectifier has a built-in storage port, in addition to the normal AC and DC ports. The flexibility required by the AC port to support the grid is provided by the storage port. The DC-bus voltage of the storage port is able to fluctuate with in a wide range to exchange energy with the grid.In order to further take use of the energy in the storage port (DC-bus capacitor) of a rectifier more reasonably and increase the support time to grid, an adaptive droop mechanism is proposed. Under such a droop mechanism, the rectifier can automatically change the power consumed according to the grid voltage variations as well as its potential to provide grid support. To achieve this, a flexibility coefficient is introduced to indicate the power flexibility level of the DC-bus capacitor. Then this flexibility coefficient is embedded into the universal droop controller (UDC) to make it adaptive. Hence, the adaptive droop controller has a changing droop coefficient corresponding to the power flexibility of a rectifier, so it can take advantage of the energy stored in its DC-bus capacitor wisely to support the grid. This droop controller can also be applied into connection between two SYNDEM smart grids. To achieve this, a grid bridge (GB) that enables autonomous and equal regulation between two SYNDEM grids is proposed. The real power transferred through a GB has linear relationship with the voltage deviation between the two micro-grids connected. The micro-grid with a higher voltage will automatically provide power to the lower one. Moreover, the power direction of a GB is bidirectional and determined by the grid voltage difference, this makes the two micro-grids equal to each other. The GB is physically a back-to-back converter. In order to achieve autonomous and equal regulation, both sides of the back-to-back converter are controlled under droop controller with the same droop coefficients. The VSM control technology is also developed to control Modular multilevel converters (MMCs) for high voltage applications. Like active loads introduced above, the MMCs can take part in the grid regulation according to the droop mechanism designed. In order to eliminate the circulating current that exists in MMCs, proportional-resonant (PR) controllers are adopted to inject second-order harmonics to the MMCs to suppress the second order circulating current. The dynamics, implementation and operation of the VSM-like MMC are introduced and analyzed. Particularly, how the VSM control algorithm works with the circulating current control in MMCs is presented. An IIT SYNDEM Smart Grid Testbed is built in an aim of achieving a minimize realization of the SYNDEM system. Extensive experiments are done on the system to show the operational scenarios when the proposed active loads are integrated in the system. There are in total eight nodes in the IIT SYNDEM testbed, which contains two utility grids, one AC load, one DC load, two solar farms and two wind farms. All the nodes are connected to a local grid through VSMs, so that they can take part in the local grid regulations in similar ways. The IIT SYNDEM Smart Grid Testbed is described in details and experimental results are provided to show the dynamic and steady performance of the IIT SYNDEM smart grid.
Show less

Title: Machine learning applications to video surveillance camera placement and medical imaging quality assessment
Creator: Lorente Gomez, Iris
Date: 2022
Description: In this work, we used machine learning techniques and data analysis to approach two applications. The first one, in collaboration with the...
Show moreIn this work, we used machine learning techniques and data analysis to approach two applications. The first one, in collaboration with the Chicago Police Department (CPD), involves analyzing and quantifying the effect that the installation of cameras had on crime, and developing a predictive model with the goal of optimizing video surveillance camera location in the streets. While video surveillance has become increasingly prevalent in policing, its intended effect on crime prevention has not been comprehensively studied in major cities in the US. In this study, we retrospectively analyzed the crime activities in the vicinity of 2,021 surveillance cameras installed between 2005 and 2016 in the city of Chicago. Using Difference-in-Differences (DiD) analysis, we examined the daily crime counts that occurred within the fields-of-view of these cameras over a 12-month period, both before and after the cameras were installed. We also investigated their potential effect on crime displacement and diffusion by examining the crime activities in a buffer zone (up to 900 ft) extended from the cameras. The results show that, collectively, there was an 18.6% reduction in crime counts within the direct viewsheds of all of the study cameras (excluding District 01 where the Loop -Chicago's business center- is located). In addition, we adapted the methodology to quantify the effect of individual cameras. The quantified effect on crime is the prediction target of our 2-stage machine learning algorithm that aims to estimate the effect that installing a videocamera in a given location will have on crime. In the first stage, we trained a classifier to predict if installing a videocamera in a given location will result in a statistically significant decrease in crime. If so, the data goes through a regression model trained to estimate the quantified effect on crime that the camera installation will have. Finally, we propose two strategies, using our 2-stage predictive model, to find the optimal locations for camera installations given a budget. Our proposed strategies result in a larger decrease in crime than a baseline strategy based on choosing the locations with higher crime density.The second application that forms this thesis belongs to the field of model observers for medical imaging quality assessment. With the advance of medical imaging devices and technology, there is a need to evaluate and validate new image reconstruction algorithms. Image quality is traditionally evaluated by using numerical figures of merit that indicate similarity between the reconstruction and the original. In medical imaging, a good reconstruction strategy should be one that helps the radiologist perform a correct diagnosis. For this reason, medical imaging reconstruction strategies should be evaluated on a task-based approach by measuring human diagnosis accuracy. Model observers (MO) are algorithms capable of acting as human surrogates to evaluate reconstruction strategies, reducing significantly the time and cost of organizing sessions with expert radiologists. In this work, we develop a methodology to estimate a deep learning based model observer for a defect localization task using a synthetic dataset that simulates images with statistical properties similar to trans-axial sections of X-ray computed tomography (CT). In addition, we explore how the models access diagnostic information from the images using psychophysical methods that have been previously employed to analyze how the humans extract the information. Our models are independently trained for five different humans and are able to generalize to images with noise statistic backgrounds that were not seen during the model training stage. In addition, our results indicate that the diagnostic information extracted by the models matches the one extracted by the humans.
Show less

Title: Video Object Detection using CenterNet
Creator: Mondal, Madhusree
Date: 2021
Description: This thesis investigates the options of video object detection with key-point-based approaches. The problem of recognizing, locating, and...
Show moreThis thesis investigates the options of video object detection with key-point-based approaches. The problem of recognizing, locating, and tracking objects in videos has been a challenging task in the computer vision area. There are few applications on key-point-based object detectors like CornerNet and CenterNet. At the first stage, this work involves the use of the previously proposed CenterNet module as a baseline detector on each frame of the Imagenet Video dataset. Then we apply an RNN module to exploit the temporal information from the past frames for better results.There are challenges in video object detection compared to still image-based object detection. It is not efficient to apply a still-image-based detector on each frame independently because we cannot exploit the temporal contextual information in videos since neighboring frames in a video are highly correlated. Object detection from videos suffers from motion blur, video focus, rare poses, etc. To overcome these issues one way of improving CenterNet for video object detection is to propagate the previous reliable detection results to boost the detection performance.
Show less

Title: DEEP LEARNING IMAGE-DENOISING FOR IMPROVING DIAGNOSTIC ACCURACY IN CARDIAC SPECT
Creator: Liu, Junchi
Date: 2022
Description: Myocardial perfusion imaging (MPI) using single-photon emission computed tomography (SPECT) is a noninvasive imaging modality widely utilized...
Show moreMyocardial perfusion imaging (MPI) using single-photon emission computed tomography (SPECT) is a noninvasive imaging modality widely utilized for diagnosis of coronary artery diseases (CAD) in nuclear medicine. Because of the concern of potential radiation risks, the imaging dose administered to patients is limited in SPECT-MPI. Due to the low count statistics in acquired data, SPECT images can suffer from high levels of noise. In this study, we investigate the potential benefit of applying deep learning (DL) techniques for denoising in SPECT-MPI studies. Owing to the lack of ground truth in clinical studies, we adopt a noise-to-noise (N2N) training approach for denoising in full-dose studies. Afterwards, we investigate the benefit of applying N2N DL on reduced-dose studies to improve the detection accuracy of perfusion defects. To address the great variability in noise level among different subjects, we propose a scheme to account for the inter-subject variabilities in training a DL denoising network to improve its generalizability. In addition, we propose a dose-blind training approach for denoising at multiple reduced-dose levels. Moreover, we investigate several training schemes to address the issue that defect and non-defect image regions are highly unbalanced in a data set, where the overwhelming majority by non-defect regions tends to have a more pronounced contribution to the conventional loss function. We investigate whether these training schemes can effectively improve preservation of perfusion defects and yield better defect detection accuracy. In the experiments we demonstrated the proposed approaches with a set of 895 clinical acquisitions. The results show promising performance in denoising and improving the detectability of perfusion-defects with the proposed approaches.
Show less

Title: Control and Operation of Microgrids and Networked Microgrids
Creator: Sheikholeslami, Mehrdad
Date: 2022
Description: This dissertation presents the practical operation and control of microgrids and networked microgrids, particularly, the networked IIT Campus...
Show moreThis dissertation presents the practical operation and control of microgrids and networked microgrids, particularly, the networked IIT Campus Microgrid (ICM) and Bronzeville Community Microgrid (BCM). Microgrids (MGs) provide a potential solution to accommodating renewable and distributed energy resources (DERs). MGs and the networked form of MGs, i.e., networked microgrids or NMGs, have received significant attention in the past two decades. However, several details are often neglected in the literature that need to be considered for the practical operations of MGs and NMGs. First, there is a need for a step-by-step sequence of operations (SOO) that clearly defines the procedures for changing the operation modes of MGs and NMGs for their reliable and resilient operation. Second, there is a need to develop new control strategies for the centralized and distributed control of MGs and NMGs that are resilient to extreme events and are also more sustainable than the ones available in the literature. Third, there is a need for developing the model of MGs and NMGs in a real-time simulator to safely evaluate the performance of the control and operation of MGs and NMGs. Finally, to close the engineering loop, there is a need to connect the digital and physical layers which are known as digital twins. This dissertation proposes solutions for these four requirements and presents results to evaluate the performance of the proposed solutions. First, an SOO is proposed to enable the reliable and safe transition between different microgrid operation modes. The proposed SOO is adaptable to any MG and NMG with minor modifications. Second, for the centralized control, a DER control model is proposed that allows for the regulated power exchange between networked MGs to ensure information privacy and respect the electrical boundary of each MG. For the distributed control, two control schemes are proposed that are resilient to extreme cases, allow the integration of renewable energy resources (RES), and require the minimum intervention of the operators. Third, several techniques are proposed that can be adopted for developing the real-time models of MGs and NMGs. Finally, as a proof of concept, a digital twin of a microgrid with connections between the physical and digital layers is implemented and tested. The IIT Campus Microgrid (ICM) and Bronzeville Community Microgrid (BCM), as well as their networked form (networked ICM-BCM), are selected as the practical testbeds and are modeled in Real-time Digital Simulator (RTDS). The RTDS model is interfaced with microgrid master controllers (MMC) for real-time data exchange and the performance of the MMCs and the distributed control strategies are tested to illustrate the importance of adopted methods in the real-time control of MGs and NMGs. Finally, a proof of concept for the digital twin of ICM is presented.
Show less

Title: SOLID-STATE SMART PLUG DEVICE
Creator: Deng, Zhixi
Date: 2022
Description: Electrical faults are a leading cause of residential fire, and flexible power cords are particularly susceptible to metal or insulation...
Show moreElectrical faults are a leading cause of residential fire, and flexible power cords are particularly susceptible to metal or insulation degradation that may lead to a variety of electrical faults. Smart Plugs are a type of plug-in device controlling electrical loads via wireless communication for consumer market. However, there is lack of circuit protection features in existing Smart Plug products. Moreover, there is no previous product or research on Smart Plug with circuit protection features. This thesis introduces a new Smart Plug 2.0 concept which offers all-in-one protection against over-current, arc, and ground faults in addition to the smart features in Smart Plug products. It aims at preventing fire and shock hazards caused by degraded or damaged power cords and electrical connections in homes and offices. It offers microsecond-scale time resolution to detect and respond to a fault condition, and significantly reduces the electrothermal stress on household electrical wires and loads. A new arc fault detection method is developed using machine learning models based on load current di/dt events. The Smart Plug 2.0 concept has been validated experimentally. A 120V/10A solid-state Smart Plug 2.0 prototype using power MOSEFTs is designed and tested. It has experimentally demonstrated the comprehensive protection features against all types of electrical faults.
Show less

Title: WIDE BANDGAP FRACTIONAL POWER PROCESSING
Creator: Kundu, Aritra
Date: 2022
Description: The adoption of wide bandgap (WBG) power semiconductors can improve the performance of power converters at the expense of significantly higher...
Show moreThe adoption of wide bandgap (WBG) power semiconductors can improve the performance of power converters at the expense of significantly higher cost than Si at present time. In this thesis, an innovative Wide bandgap Fractional Power Processing (WFPP) design concept is introduced where silicon devices process the base power at a low switching frequency (2kHz or lower) while WBG devices are judiciously used to process only a fraction of the total power at a much higher frequency (tens of kHz). WFPP inverter is a design concept that splits the power processing into a low frequency Si base power processor and a high-frequency WBG fractional power processor. WBG devices are therefore judiciously used to process only a fraction of the total power to achieve both high-efficiency and lower cost than a full-WBG converter design. This thesis investigates an optimization strategy to minimize the total power loss while maintaining a reasonable THD and cost for a hybrid inverter design with comprehensive power loss analysis and calculation on fundamental and harmonics currents. Optimal selection of power sharing between the Si and WBG inverters and switching frequency are discussed in the thesis. The circulating current paths in topology with hybrid switching frequencies are also analyzed and presented in this thesis. Experimental results on a 9kW SiC/Si hybrid inverter prototype with isolated and non-isolated DC power supplies are presented to validate the design concept.
Show less

Title: ANALYTICAL APPROACH TO ESTIMATE ROTOR TEMPERATURE IN SWITCHED RELUCTANCE MOTOR
Creator: Koujalagi, Shweta Manohar
Date: 2022
Description: Motors contribute most of the loads. Motors find major applications in automobile industries, household appliances, industrial equipment, and...
Show moreMotors contribute most of the loads. Motors find major applications in automobile industries, household appliances, industrial equipment, and other areas. With the time, engineers and industries found some of the drawbacks or disadvantages of using induction motors in certain applications. They started developing other types of motors that are more efficient than existing ones. Among those, switched reluctance motor, referred as SRM is the one. SRMs are simple in construction, rugged and highly efficient motors.Even though SRM has higher efficiency, it still contribute some losses in the form of heat which will increase the temperature of SRM. If the temperature increases beyond certain limit, cable insulation fails, degrades rotor capability of aligning characteristics, damages bearings, etc. Therefore, it is important to understand the flow of heat in SRM. This thesis focuses on heat transfer analysis from stator coil to rotor of SRM using analytical method and numerical method such as finite element analysis from available coil temperature without using any kind of sensors. Analytical and FEA models are built separately to obtained rotor temperatures at various coil temperatures and rotor speeds. Finally, analytical results are validated with FEA model results. Therefore, once the rotor temperature is estimated accurately, model can be implemented in automotive and other industrial applications to continuously monitor the rotor temperature. It is important to monitor temperature to avoid damage of SRM by thermal effects.
Show less

Title: Deep Learning Methods For Wireless Networks Optimization
Creator: Zhang, Shuai
Date: 2022
Description: The resurgence of deep learning techniques has brought forth fundamental changes to how hard problems could be solved. It used to be held that...
Show moreThe resurgence of deep learning techniques has brought forth fundamental changes to how hard problems could be solved. It used to be held that the solutions to complex wireless network problems require accurate mathematical modeling of the network operation, but now the success of deep learning has shown that a data-driven method could generate powerful and useful representations such that the problem could be solved efficiently with surprisingly competent performance. Network researchers have recognized this and started to capitalize on the learning methods’ prowess. But most works follow the existing black-box learning paradigms without much accommodation to the nature and essence of the underlying network problems. This thesis focuses on a particular type of classical problem: multiple commodity flow scheduling in an interference-limited environment. Though it does not permit efficient exact algorithms due to its NP-hard complexity, we use it as an entry point to demonstrate from three angles how the learning-based methods can help improve the network performance. In the first part, we leverage the graphical neural network (GNN) techniques and propose a two-stage topology-aware machine learning framework, which trains a graph embedding unit and a link usage prediction module jointly to discover links that are likely to be used in optimal scheduling. The second part of the thesis is an attempt to find a learning method that has a closer algorithmic affinity to the traditional DCG method. We make use of reinforcement learning to incrementally generate a better partial solution such that a high quality solution may be found in a more efficient manner. As the third part of the research, we revisit the MCF problem from a novel viewpoint: instead of leaning on the neural networks to directly generate the good solutions, we use them to associate the current problem instance with historical ones that are similar in structure. These matched instances’ solutions offer a highly useful starting point to allow efficient discovery of the new instance’s solution.
Show less

Title: DEEP LEARNING AND COMPUTER VISION FOR INDUSTRIAL APPLICATIONS: CELLULAR MICROSCOPIC IMAGE ANALYSIS AND ULTRASOUND NONDESTRUCTIVE TESTING
Creator: Yuan, Yu
Date: 2022
Description: For decades, researchers have sought to develop artificial intelligence (AI) systems that can help human beings on decision making, data...
Show moreFor decades, researchers have sought to develop artificial intelligence (AI) systems that can help human beings on decision making, data analysis and pattern recognition applications where analytical methods are ineffective. In recent years, Deep Learning (DL) has been proven to be an effective AI technique that can outperform other methods in applications such as computer vision, natural language processing, autonomous driving. Realizing the potential of deep learning techniques, researchers have also started to apply deep learning on other industrial applications. Today, deep learning based models are used to innovate and accelerate automation, guidance, and decision making in various industries including automotive industry, pharmaceutical industry, finance, agriculture and more. In this research, several important industrial applications (on Biomedicine and Non-Destructive Testing) utilizing deep learning algorithms will be introduced and analyzed. The first biopharmaceutical application focuses on developing a deep learning based model to automate the visual inspection process in Median Tissue Culture Infectious Dose(TCID50). TCID50 is one of the most popular methods for viral quantification. An important step of TCID50 is to visually inspect the sample and decide if it exhibits cytopathic effect(CPE) or not. Two novel models have been developed to detect CPE in microscopic images of cell culture in 96 well-plates. The first model consists of a convolutional neural network (CNN) and support vector machine(SVM). The second model is a fully convolutional network (FCN) followed by morphological post-processing steps. The models are tested on 4 cell lines and achieve very high accuracy. Another biopharmaceutical application developed for cellular microscopic images is the clonal selection. Clonal selection is one of the mandatory steps in cell line development process. It focuses on verifying the clonality of the cell culture. The researchers used to visually inspect the microscopic images to verify the clonality. In this work, a novel deep learning based model and a workflow is developed to accelerate the process. This algorithm consists of multiple steps, including image analysis after incubation to detect the cell colonies, and verify its clonality in day0 image. The results and common mis-classification cases are shown in this thesis. Image analysis method is not the only technology that has been advancing for cellular image analysis in biopharmaceutical industry. A new class of instruments are currently used in biopharmaceutical industry which enable more opportunities for image analysis. To make the most of these new instruments, a convolutional neural network based architecture is used to perform accurate cell counting and cell morphology based segmentation. This analysis can provide more insight of the cells at very early stage in characterization process of cell line development. The architecture and the testing results are presented in this work. The proposed algorithm has achieved very high accuracy on both applications, and the cell morphology based segmentation enables a brand new feature for scientists to predict the potential productivity of the cells. Next part of this dissertation is focused on hardware implementation of Ultrasonic Non-Destructive Testing (NDT) methods based on deep learning, which can be highly useful in flaw detection and classification applications. With the help of a smart and mobile Non-Destructive Testing device, engineers can accurately detect and locate the flaws inside the materials without reliance on high performance computation resources. The first NDT application presents a hardware implementation of a deep learning algorithm on Field-programmable gate array(FPGA) for Ultrasound flaw detection. The Ultrasound flaw detection algorithm consists of a wavelet transform followed by a LeNet inspired convolutional neural network called Ultra-LeNet. This work is focused on implementing the computationally difficult part of this algorithm: Ultra-LeNet, so that it can be used in the field where high performance computation resources (e.g., AWS) are not accessible. The implementation uses resource partitioning to design two dedicated pipelined accelerators for convolutional layers and fully connected layers respectively. Both accelerators utilize loop unrolling, loop pipelining and batch processing techniques to maximize the throughput. The comparison to other work has shown that the implementation has achieved higher hardware utilization efficiency. The second NDT application is also focused on implementing a deep learning based algorithm for Ultrasound flaw detection on a FPGA. Instead of implementing the Ultra-LeNet, the deep learning model used in this application is Meta-learning based Siamese Network, which is capable for multi-class classification and it can also classify a new class even if it does not appear in the training dataset with the help of automated learning features. The hardware implementation is significantly different than the previous algorithm. In order to improve the inference operation efficiency, the model is compressed with both pruning and quantization, and the FPGA implementation is specifically designed to accelerate the compressed CNN with high efficiency. The CNN model compression method and hardware design are novel methods introduced in this work. Comparison against other compressed CNN accelerators is also presented.
Show less

Title: Exploiting contextual information for deep learning based object detection
Creator: Zhang, Chen
Date: 2020
Description: Object detection has long been an important research topic in computer vision area. It forms the basis of many applications. Despite the great...
Show moreObject detection has long been an important research topic in computer vision area. It forms the basis of many applications. Despite the great progress made in recent years, object detection is still a challenging task. One of the keys to improving the performance of object detection is to utilize the contextual information from the image itself or from a video sequence. Contextual information is defined as the interrelated condition in which something exists or occurs. In object detection, such interrelated condition can be related background/surroundings, support from image segmentation task, and the existence of the object in the temporal domain for video-based object detection. In this thesis, we propose multiple methods to exploit contextual information to improve the performance of object detection from images and videos.First, we focus on exploiting spatial contextual information in still-image based object detection, where each image is treated independently. Our research focuses on extracting contextual information using different approaches, which includes recurrent convolutional layer with feature concatenation (RCL-FC), 3-D recurrent neural networks (3-D RNN), and location-aware deformable convolution. Second, we focus on exploiting pixel-level contextual information from a related computer vision task, namely image segmentation. Our research focuses on applying a weakly-supervised auxiliary multi-label segmentation network to improve the performance of object detection without increasing the inference time. Finally, we focus on video object detection, where the temporal contextual information between video frames are exploited. Our first research involves modeling short-term temporal contextual information using optical flow and modeling long-term temporal contextual information using convLSTM. Another research focuses on creating a two-path convLSTM pyramid to handle multi-scale temporal contextual information for dealing with the change in object's scale. Our last work is the event-aware convLSTM that forces convLSTM to learn about the event that causes the performance to drop in a video sequence.
Show less

Title: Defense-in-Depth for Cyber-Secure Network Architectures of Industrial Control Systems
Creator: Arnold, David James
Date: 2024
Description: Digitization and modernization efforts have yielded greater efficiency, safety, and cost-savings for Industrial Control Systems (ICS). To...
Show moreDigitization and modernization efforts have yielded greater efficiency, safety, and cost-savings for Industrial Control Systems (ICS). To achieve these gains, the Internet of Things (IoT) has become an integral component of network infrastructures. However, integrating embedded devices expands the network footprint and softens cyberattack resilience. Additionally, legacy devices and improper security configurations are weak points for ICS networks. As a result, ICSs are a valuable target for hackers searching for monetary gains or planning to cause destruction and chaos. Furthermore, recent attacks demonstrate a heightened understanding of ICS network configurations within hacking communities. A Defense-in-Depth strategy is the solution to these threats, applying multiple security layers to detect, interrupt, and prevent cyber threats before they cause damage. Our solution detects threats by deploying an Enhanced Data Historian for Detecting Cyberattacks. By introducing Machine Learning (ML), we enhance cyberattack detection by fusing network traffic and sensor data. Two computing models are examined: 1) a distributed computing model and 2) a localized computing model. The distributed computing model is powered by Apache Spark, introducing redundancy for detecting cyberattacks. In contrast, the localized computing model relies on a network traffic visualization methodology for efficiently detecting cyberattacks with a Convolutional Neural Network. These applications are effective in detecting cyberattacks with nearly 100% accuracy. Next, we prevent eavesdropping by applying Homomorphic Encryption for Secure Computing. HE cryptosystems are a unique family of public key algorithms that permit operations on encrypted data without revealing the underlying information. Through the Microsoft SEAL implementation of the CKKS algorithm, we explored the challenges of introducing Homomorphic Encryption to real-world applications. Despite these challenges, we implemented two ML models: 1) a Neural Network and 2) Principal Component Analysis. Finally, we hinder attackers by integrating a Cyberattack Lockdown Network with Secure Ultrasonic Communication. When a cyberattack is detected, communication for safety-critical elements is redirected through an ultrasonic communication channel, establishing physical network segmentation with compromised devices. We present proof-of-concept work in transmitting video via ultrasonic communication over an Aluminum Rectangular Bar. Within industrial environments, existing piping infrastructure presents an optimal solution for cost-effectively preventing eavesdropping. The effectiveness of these solutions is discussed within the scope of the nuclear industry.
Show less

Title: Large Language Model Based Machine Learning Techniques for Fake News Detection
Creator: Chen, Pin-Chien
Date: 2024
Description: With advanced technology, it’s widely recognized that everyone owns one or more personal devices. Consequently, people are evolving into...
Show moreWith advanced technology, it’s widely recognized that everyone owns one or more personal devices. Consequently, people are evolving into content creators on social media or the streaming platforms sharing their personal ideas regardless of their education or expertise level. Distinguishing fake news is becoming increasingly crucial. However, the recent research only presents comparisons of detecting fake news between one or more models across different datasets. In this work, we applied Natural Language Processing (NLP) techniques with Naïve Bayes and DistilBERT machine learning method combing and augmenting four datasets. The results show that the balanced accuracy is higher than the average in the recent studies. This suggests that our approach holds for improving fake news detection in the era of widespread content creation.
Show less

repository.iit

Search the repository

Pages

Pages

Enabled Filters

Refine Results

Type

Date

Department

Subject

Creator

Rights