Scholarly Works, Electrical and Computer Engineering

Permanent URI for this collection

https://hdl.handle.net/10919/23145

Research articles, presentations, and other scholarship

Browse

Now showing 1 - 20 of 988

Allocation of Cost of Reliability to Various Customer Sectors in a Standalone Microgrid System
Nallainathan, Sakthivelnathan; Arefi, Ali; Lund, Christopher; Mehrizi-Sani, Ali (MDPI, 2025-06-20)
Due to the intermittent and uncertain nature of emerging renewable energy sources in the modern power grid, the level of dispatchable power sources has been reduced. The contemporary power system is attempting to address this by investing in energy storage within the context of standalone microgrids (SMGs), which can operate in an island mode and off-grid. While renewable-rich SMGs can facilitate a higher level of renewable energy penetration, they also have more reliability issues compared to conventional power systems due to the intermittency of renewables. When an SMG system needs to be upgraded for reliability improvement, the cost of that reliability improvement should be divided among diverse customer sectors. In this research, we present four distinct approaches along with comprehensive simulation outcomes to address the problem of allocating reliability costs. The central issue in this study revolves around determining whether all consumers should bear an equal share of the reliability improvement costs or if these expenses should be distributed among them differently. When an SMG system requires an upgrade to enhance its reliability, it becomes imperative to allocate the associated costs among various customer sectors as equitably as possible. In our investigation, we model an SMG through a simulation experiment, involving nine distinct customer sectors, and utilize their hourly demand profiles for an entire year. We explore how to distribute the total investment cost of reliability improvement to each customer sector using four distinct methods. The first two methods consider the annual and seasonal peak demands in each industry. The third approach involves an analysis of Loss of Load (LOL) events and determining the hourly load requirements for each sector during these events. In the fourth approach, we employ the Technique for Order of Preference by Similarity to the Ideal Solution (TOPSIS) technique. The annual peak demand approach resulted in the educational sector bearing the highest proportion of the reliability improvement cost, accounting for 21.90% of the total burden. Similarly, the seasonal peak demand approach identified the educational sector as the most significant contributor, though with a reduced share of 15.44%. The normalized average demand during Loss of Load (LOL) events also indicated the same sector as the highest contributor, with 12.34% of the total cost. Lastly, the TOPSIS-based approach assigned a 15.24% reliability cost burden to the educational sector. Although all four approaches consistently identify the educational sector as the most critical in terms of its impact on system reliability, they yield different cost allocations due to variations in the methodology and weighting of demand characteristics. The underlying reasons for these differences, along with the practical implications and applicability of each method, are comprehensively discussed in this research paper. Based on our case study findings, we conclude that the education sector, which contributes more to LOL events, should bear the highest amount of the Cost of Reliability Improvement (CRI), while the hotel and catering sector’s share should be the lowest percentage. This highlights the necessity for varying reliability improvement costs for different consumer sectors.
Standalone Operation of Inverter-Based Variable Speed Wind Turbines on DC Distribution Network
Amini, Hossein; Noroozian, Reza (MDPI, 2025-04-10)
This paper discusses the operation and control of a low-voltage DC (LVDC) isolated distribution network powered by distributed generation (DG) from a variable-speed wind turbine induction generator (WTIG) to supply unbalanced AC loads. The system incorporates a DC-DC storage converter to regulate network voltages and interconnect battery energy storage with the DC network. The wind turbines are equipped with a squirrel cage induction generator (IG) to connect a DC network via individual power inverters (WTIG inverters). Loads are unbalanced ACs and are interfaced using transformerless power inverters, referred to as load inverters. The DC-DC converter is equipped with a novel control strategy, utilizing a droop regulator for the DC voltage to stabilize network operation. The control system is modeled based on Clark and Park transformations and is developed for the load inverters to provide balanced AC voltage despite unbalanced load conditions. The system employs the perturbation and observation (P&O) method for maximum power point tracking (MPPT) to optimize wind energy utilization, while blade angle controllers maintain generator performance within rated power and speed limits under high wind conditions. System operation is analyzed under two scenarios: normal operation with varying wind speeds and the effects of load variations. Simulation results using PSCAD/EMTDC demonstrate that the proposed LVDC isolated distribution network (DC) achieves a stable DC bus voltage within ±5% of the nominal value, efficiently delivers balanced AC voltages with unbalanced levels below 2%, and operates with over 90% wind energy utilization during varying wind speeds, confirming LVDC network reliability and robustness.
Personality Emulation Utilizing Large Language Models
Kolenbrander, Jack; Michaels, Alan J. (MDPI, 2025-06-12)
Fake identities have proven to be an effective methodology for conducting privacy and cybersecurity research; however, existing models are limited in their ability to interact with and respond to received communications. To perform privacy research in more complex Internet domains, withstand enhanced scrutiny, and persist long-term, fake identities must be capable of automatically generating responses while maintaining consistent behavior and personality. This work proposes a method for assigning personality to fake identities using the widely accepted psychometric Big Five model. Leveraging this model, the potential application of large language models (LLMs) to generate email responses that emulate human personality traits is investigated to enhance fake identity capabilities for privacy research at scale.
An Application of Explainable Multi-Agent Reinforcement Learning for Spectrum Situational Awareness
Perini, Dominick J.; Muller, Braeden P.; Kopacz, Justin; Michaels, Alan J. (MDPI, 2025-04-10)
Allocating low-bandwidth radios to observe a wide portion of a spectrum is a key class of search-optimization problems that requires system designers to leverage limited resources and information efficiently. This work describes a multi-agent reinforcement learning system that achieves a balance between tuning radios to newly observed energy while maintaining regular sweep intervals to yield detailed captures of both short- and long-duration signals. This algorithm, which we have named SmartScan, and system implementation have demonstrated live adaptations to dynamic spectrum activity, persistence of desirable sweep intervals, and long-term stability. The SmartScan algorithm was also designed to fit into a real-time system by guaranteeing a constant inference latency. The result is an explainable, customizable, and modular approach to implementing intelligent policies into the scan scheduling of a spectrum monitoring system.
Microgrid Reliability Incorporating Uncertainty in Weather and Equipment Failure
Nallainathan, Sakthivelnathan; Arefi, Ali; Lund, Christopher; Mehrizi-Sani, Ali (MDPI, 2025-04-17)
Solar photovoltaic (PV) and wind power generation are key contributors to the integration of renewable energy into modern power systems. The intermittent and variable nature of these renewables has a substantial impact on the power system’s reliability. In time-series simulation studies, inaccuracies in solar irradiation and wind speed parameters can lead to unreliable evaluations of system reliability, ultimately resulting in flawed decision making regarding the investment and operation of energy systems. This paper investigates the reliability deviation due to modeling uncertainties in a 100% renewable-based system. This study employs two methods to assess and contrast the reliability of a standalone microgrid (SMG) system in order to achieve this goal: (i) random uncertainty within a selected confidence interval and (ii) splitting the cumulative distribution function (CDF) into five regions of equal probability. In this study, an SMG system is modeled, and loss of load probability (LOLP) is evaluated in both approaches. Six different sensitivity analysis studies, including annual load demand growth, are performed. The results from the simulations demonstrate that the suggested methods can estimate the reliability of a microgrid powered by renewable energy sources, as well as its probability of reaching certain levels of reliability.
A Fast Transient Response Distributed Power Supply With Dynamic Output Switching for Power Side-Channel Attack Mitigation
Liu, Xingye; Ampadu, Paul K. (IEEE, 2024-09)
We present a distributed power supply and explore its load transient response and power side-channel security improvements. Typically, countermeasures against power side-channel attacks (PSCAs) are based on specialized dc/dc converters, resulting in large power and area overheads and they are difficult to scale. Moreover, due to limited output voltage range and load regulation, it is not feasible to directly distribute these converters in multicore applications. Targeting those issues, our proposed converter is designed to provide multiple fast-responding voltages and use shared circuits to mitigate PSCAs. The proposed three-output dc/dc converter can deliver 0.33-0.92 V with up to 1 A to each load. Comparing with state-of-the-art power management works, our converter has 2× load step response speed and 4× reference voltage tracking speed. Furthermore, the converter requires 9× less inductance and 3× less output capacitance. In terms of PSCA mitigation, this converter reduces the correlation between input power trace and encryption load current by 107×, which is 3× better than the best standalone work, and it only induces 1.7% area overhead and 2.5% power overhead. The proposed work also increases minimum traces to disclose (MTDs) by 1250×. Considering all the above, our work could be a great candidate to be employed in future multicore systems supplying varying voltages and resisting side-channel attacks. It is the first work bridging the gap between on-chip power management and side-channel security.
Practical Federated Recommendation Model Learning Using ORAM with Controlled Privacy
Liu, Jinyu; Xiong, Wenjie; Suh, G. Edward; Maeng, Kiwan (ACM, 2025-03-30)
Training high-quality recommendation models requires collecting sensitive user data. The popular privacy-enhancing training method, federated learning (FL), cannot be used practically due to these models’ large embedding tables. This paper introduces FEDORA, a system for training recommendation models with FL. FEDORA allows each user to only download, train, and upload a small subset of the large tables based on their private data, while hiding the access pattern using oblivious memory (ORAM). FEDORA reduces the ORAM’s prohibitive latency and memory overheads by (1) introducing 𝜖-FDP, a formal way to balance the ORAM’s privacy with performance, and (2) placing the large ORAM in a power- and cost-efficient SSD with SSD-friendly optimizations. Additionally, FEDORA is carefully designed to support (3) modern operation modes of FL. FEDORA achieves high model accuracy by using private features during training while achieving, on average, 5× latency and 158× SSD lifetime improvement over the baseline.
Stramash: A Fused-kernel Operating System For Cache-Coherent, Heterogeneous-ISA Platforms
Xing, Tong; Xiong, Cong; Wei, Tianrui; Sanchez, April; Ravindran, Binoy; Balkind, Jonathan; Barbalace, Antonio (ACM, 2025-03-30)
We live in the world of heterogeneous computing. With specialised elements reaching all aspects of our computer systems and their prevalence only growing,we must act to rein in their inherent complexity. One area that has seen significantly less investment in terms of development is heterogeneous-ISA systems, specifically because of complexity. To date, heterogeneous- ISA processors have required significant software overheads,workarounds, and coordination layers, making the development ofmore advanced software hard, and motivating little further development of more advanced hardware. In this paper, we take a fused approach to heterogeneity, and introduce a new operating system (OS) design, the fused-kernel OS, which goes beyond the multiple-kernel OS design, exploiting cache-coherent shared memory among heterogeneous-ISA CPUs as a first principle – introducing a set of newOS kernel mechanisms.We built a prototype fusedkernel OS, Stramash-Linux, to demonstrate the applicability of our design to monolithic OS kernels.We profile Stramash OS components on real hardware but tested them on an architectural simulator – Stramash-QEMU, which we design and build. Our evaluation begins by validating the accuracy of our simulator, achieving an average of less than4%errors.We then perform a direct comparison between our fused-kernelOSand state-of-the-art multiple-kernel OS designs. Results demonstrate speedups of up to 2.1×onNPBbenchmarks. Further,we provide an in-depth analysis of the differences and trade-offs between fused-kernel and multiple-kernel OS designs.
The Challenges of EEG in Coma: The Potential of Recent Discoveries
Hbibi, Bechir; Mili, Lamine M. (IntechOpen, 2025)
The utilization of electroencephalography (EEG) has profoundly enriched our comprehension and monitoring of patients, especially those in intensive care units (ICUs), over the past decades. EEG, a method of recording electrical brain signals, is employed to explore a variety of neurological disorders such as epilepsy, dementia, and brain injuries that may affect unconscious patients. In recent years, EEG has also been used to monitor sedation levels, examine the quality of patients’ sleep, and track patient recovery during periods of coma. Groundbreaking findings, derived from EEG recordings in intensive care using various techniques and methodologies, have unveiled new avenues to aid these patients and improve physicians’ understanding of their condition and needs. Innovations such as the examination of sleep quality, the assessment of pain and stress, and the classification of vigilance states represent some of the promising advancements in ICUs, all of which are based on EEG. Recent discoveries stemming from EEG signal analysis have indicated numerous potential enhancements in improving comfort, fostering a better understanding of the situation, and reducing the administration of drugs for ICU patients. In this chapter, we will discuss some new EEG findings for intensive care unit patients and the possible applications that could be revealed based on other investigations on human subjects outside the ICU.
Model-Free Cyber-Resilient Coordinated Inverter Control in a Microgrid
Beikbabaei, Milad; Larsen, Caroline; Mehrizi-Sani, Ali (IEEE, 2024-09-20)
The increasing number of inverter-based resources (IBR) in the grid introduces new challenges due to the fast transient response and low inertia of IBRs. Set point automatic adjustment with correction enabled (SPAACE)–based techniques smoothen the transient response of an IBR already installed in a grid by modifying its set point without accessing its internal parameters in a model-free approach. Coordinated SPAACE (CSPAACE) further enhances SPAACE performance by incorporating communication links to exchange tracking error values between IBRs; however, this creates openings for cyberattacks. This work adds a detection and mitigation algorithm for both denial of service (DoS) and false data injection (FDI) attacks on the communication channels. Long short-term memory (LSTM) detects anomalies in the inputs received from other inverters, and bidirectional LSTM (BiLSTM) mitigates the adverse effect of attacks on the voltage and frequency stability of a microgrid. A hybrid co-simulation platform is developed using a computer running PSCAD/EMTDC software, a network switch, and two Raspberry Pi computers, where the cyberattacks are conducted on the network switch using one of the Pis. The testbed is used to study the effectiveness of the proposed detection and mitigation method under DoS and FDI attacks and various grid transients.
Sub-surface thermal measurement in additive manufacturing via machine learning-enabled high-resolution fiber optic sensing
Wang, Rongxuan; Wang, Ruixuan; Dou, Chaoran; Yang, Shuo; Gnanasambandam, Raghav; Wang, Anbo; Kong, Zhenyu (James) (Springer Nature, 2024-08-31)
Microstructures of additively manufactured metal parts are crucial since they determine the mechanical properties. The evolution of the microstructures during layer-wise printing is complex due to continuous re-melting and reheating effects. The current approach to studying this phenomenon relies on time-consuming numerical models such as finite element analysis due to the lack of effective sub-surface temperature measurement techniques. Attributed to the miniature footprint, chirped-fiber Bragg grating, a unique type of fiber optical sensor, has great potential to achieve this goal. However, using the traditional demodulationmethods, its spatial resolution is limited to the millimeter level. In addition, embedding it during laser additive manufacturing is challenging since the sensor is fragile. This paper implements a machine learning-assisted approach to demodulate the optical signal to thermal distribution and significantly improve spatial resolution to 28.8 μm from the original millimeter level. A sensor embedding technique is also developed to minimize damage to the sensor and part while ensuring close contact. The case study demonstrates the excellent performance of the proposed sensor in measuring sharp thermal gradients and fast cooling rates during the laser powder bed fusion. The developed sensor has a promising potential to study the fundamental physics of metal additive manufacturing processes.
Gate Driver Design and Device Characterization for 3.3 kV SiC MOSFET Modules
Gutierrez, Bryan; Hou, Zhengming; Jiao, Dong; Hsieh, Hsin-Che; Chen, XingRou; Liao, Hsuan; Lai, Jih-Sheng; Yu, Ming-Hung; Chen, Kuan-Wen (IEEE, 2023-11-02)
A compact and 25-kV isolated gate driver with various protective features has been implemented for 3.3-kV SiC power modules. The gate driver design procedure for circuit production is described in detail. With the implemented gate driver, double pulse test (DPT) is performed to evaluate the switching energy losses and parasitic loop inductance at different voltage conditions. Additionally, the DPT together with Ansys Software are used to characterize the stray inductances.
Wide Bandgap Semiconductor Device Fundamentals and Applications
Lai, Jih-Sheng (2022-10-11)
Short Course presentation
Harnessing Meta-Reinforcement Learning for Enhanced Tracking in Geofencing Systems
Famili, Alireza; Sun, Shihua; Atalay, Tolga; Stavrou, Angelos (IEEE, 2025-01-20)
Geofencing technologies have become pivotal in creating virtual boundaries for both real and virtual environments, offering a secure means to control and monitor designated areas. They are now considered essential tools for defining and controlling boundaries across various applications, from aviation safety in drone management to access control within mixed reality platforms like the metaverse. Effective geofencing relies heavily on precise tracking capabilities, a critical component for maintaining the integrity and functionality of these systems. Leveraging the advantages of 5G technology, including its large bandwidth and extensive accessibility, presents a promising solution to enhance geofencing performance. In this paper, we introduce MetaFence: Meta-Reinforcement Learning for Geofencing Enhancement, a novel approach for precise geofencing utilizing indoor 5G small cells, termed "5G Points", which are optimally deployed using a meta-reinforcement learning (meta-RL) framework. Our proposed meta-RL method addresses the NP-hard problem of determining an optimal placement of 5G Points to minimize spatial geometry-induced errors. Moreover, the meta-training approach enables the learned policy to quickly adapt to diverse new environments. We devised a comprehensive test campaign to evaluate the performance of MetaFence. Our results demonstrate that this strategic placement significantly improves tracking accuracy compared to traditional methods. Furthermore, we show that the meta-training strategy enables the learned policy to generalize effectively and perform efficiently when faced with new environments.
Estimation of Global Illumination using Cycle-Consistent Adversarial Networks
Oh, Junho; Abbott, A. Lynn (Springer, 2024-10-15)
Synthesis of realistic virtual environments requires careful rendering of light and shadows, a task often bottle-necked by the high computational cost of global illumination (GI) techniques. This paper introduces a new GI approach that improves computational efficiency without a significant reduction in image quality. The proposed system transforms initial direct-illumination renderings into globally illuminated representations by incorporating a Cycle-Consistent Adversarial Network (CycleGAN). Our CycleGAN-based approach has demonstrated superior performance over the Pix2Pix model according to the LPIPS metric, which emphasizes perceptual similarity. To facilitate such comparisons, we have created a novel dataset (to be shared with the research community) that provides in-game images that were obtained with and without GI rendering. This work aims to advance real-time GI estimation without the need for costly, specialized computational hardware. Our work and the dataset are made publicly available at https://github.com/junhofive/CycleGAN-Illumination.
Hybrid Multi-Level Inverter
(United States Patent Office)
This disclosure provides systems, methods, and apparatus for multi-level inverters. A hybrid binary cascaded multilevel inverter (BCMLI) is discussed that includes a plurality of H-bridge cells connected in a cascaded formation. DC input voltages of some of the H-bridge cells are provided by DC voltage sources. But inputs of other H-bridge cells coupled with capacitors instead. The H-bridge cells are operated to provide an AC output voltage at the output terminals of the inverter. One or more floating capacitor voltage controllers are used to vary one or more switching instances of the H-bridge cells such that a desirable level or charge is maintained across the one or more capacitors coupled with the input terminals of the H-bridge cells.
Hybrid Modulation Controlled DC-to-AC Converters
Aspects of hybrid modulation control for DC-to-AC converters are described. In one embodiment, a hybrid modulation pattern is generated. The hybrid modulation pattern separates switch gating control into multiple control regions for a half cycle of the waveform. A first control region modulates according to a first modulation technique and a second control region modulates according to a second modulation technique. The switches of a resonant converter are controlled according to the hybrid modulation pattern to generate the waveform.
Seamless Distributed Power Management of DC Microgrid under Cyber Attacks with Bidirectional Power Flow
Tran, Dat Thanh; Kim, Kyeong-Hwa; Lai, Jih-Sheng (IEEE, 2025)
A seamless distributed power management for the dc microgrid (DCMG) with bidirectional power flow is presented in this paper to achieve the overall system stabilization even under severe cyber attacks. First, a distributed secondary control (DSC) based on the V*-P droop curves is presented to ensure the power and voltage regulation for the DCMG system consisting of the electric vehicle (EV), wind turbine, battery, load, and utility grid agents under uncertain conditions. The proposed scheme automatically adjusts the utility grid droop curve to minimize electricity expenditure under electricity price change. To eliminate the negative effect of severe cyber attacks such as false data injection (FDI) and denial-of-service (DoS) in the distributed DCMG system, a resilient DSC based on the compensation term is utilized in the proposed scheme. In addition, by modifying the DSC structure of each power agent, the proposed scheme reduces the overshoot of the dc bus voltage even in the presence of the electricity price change, agent power variation, sudden utility grid disconnection, or critical state-of-charge (SOC) levels. Furthermore, the proposed distributed DCMG system utilizes only the unidirectional digital communication links (DCLs) to cut down the system cost and communication burden, which greatly simplifies communication structure. The efficiency and feasibility of the proposed distributed power management are validated by simulation and experimental results under various conditions.
Design of a 15kW High-Efficiency and High Power Density Bidirectional TCM Buck/Boost Converter
Hou, Zhengming; Jiao, Dong; Gutierrez, Bryan C.; Lai, Jih-Sheng; Chen, Po-Li (IEEE, 2024)
A non-isolated buck/boost converter features bidirectional power flow capability and a wide output voltage range. The zero-voltage switching (ZVS) can be achieved under triangular current mode (TCM) operation to achieve high power density and high efficiency. Most research has been conducted on TCM control strategies or low-power applications but rarely on the design of high-power TCM buck/boost converters. In this paper, a simplified inductor design methodology for the power conversion between a common dc bus and a wide-range variable dc voltage TCM buck/boost converter is proposed. A 15kW prototype is designed to regulate the voltage from 150V-1000V from a 1.1kV dc bus. The prototype demonstrates 16.36kW/L power density and 99.81% peak efficiency.
Wide-bandgap semiconductors and power electronics as pathways to carbon neutrality
Zhang, Yuhao; Dong, Dong; Li, Qiang; Zhang, Richard; Udrea, Florin; Wang, Han (Springer, 2025-01-21)
Energy supply and consumption account for approximately 75% of global greenhouse gas emissions. Advances in semiconductor and power electronics technologies are required to integrate renewable energy into grids, electrify transport and the heating and cooling of buildings, and increase the efficiency of electricity conversion. This Review outlines the opportunities for carbon neutrality in the energy sector enabled by synergistic advances in wide-bandgap (WBG) semiconductors and power electronics. First, we present advances in WBG power devices, converter circuits and power electronics applications and their implications. For example, WBG materials have a high critical electric field and thermal stability; therefore, WBG devices can operate at higher temperatures and frequencies than silicon devices, enabling higher efficiency and reducing the number of passive components and cooling systems required in converter circuits. We then discuss advances in renewable energy systems, electric vehicles, data centres and heat pumps enabled by WBG devices, and their potential to reduce carbon emissions through electrification and increased energy conversion efficiency. We also consider the implications of the carbon footprint of WBG device manufacturing being larger than that of silicon manufacturing. Finally, we discuss research gaps that must be addressed to realize the potential of WBG semiconductors and power electronics for carbon neutrality.

Browse

Recent Submissions