DR-NTU (Digital Repository of NTU)
Not a member yet
116018 research outputs found
Sort by
Advanced multilevel converters for grid integration of renewable energy
No full textMultilevel converters are becoming increasingly important in modern power electronics due to their advantages such as reduced harmonic distortion, lower electromagnetic interference, and enhanced voltage handling capability. This project aims to explore and develop advanced topologies, control methods, and modulation schemes for multilevel converters, with a focus on improving efficiency, power quality, and system reliability.
This report presents a detailed parameter design of a 3L-ANPC converter, with emphasis on topology selection, voltage balancing methods, and loss distribution optimization. The control system is developed to enhance the converter’s response speed and dynamic stability, ensuring seamless operation under varying grid conditions.
Simulation results from PLECS verify the effectiveness of the proposed 3L-ANPC converter in harmonic reduction, voltage regulation and power efficiency improvement. The results demonstrate that the 3L-ANPC converter has significant potential for large-scale renewable energy integration, offering promising prospects for future research and practical applications.Bachelor's degre
Colonial statues in shaping postcolonial understanding of contemporary national identity
No full textThis paper explores how the debates surrounding the preservation of colonial statues shape contemporary
understandings of national identity through qualitative analysis of focus group discussions, media
commentaries, social media discussions, and surveys. Despite global movements advocating for the
eradication of colonial statues and monuments, Singapore has taken an alternative approach in preserving
and erecting new colonial statues to align with state narratives.
Our findings revealed that while many Singaporeans demonstrated general indifference to colonial
statues, there is an increasing preference for the recontextualization of statues. Participants also expressed
concerns and frustration with the state’s one-sided approach to colonial history in Singapore’s education
curriculum and suggested shifting national identity representation towards local figures and modern
landmarks.
Drawing on governmentality and collective memory, this study argues that while state-led narratives play
a pivotal role in shaping national identity, Singaporeans also negotiate tensions between public debates
and state narratives.Bachelor's degre
Image processing on a SoC-based CGRA hardware accelerator for IoT applications
No full textCoarse-grain-reconfigurable-array (CGRA) architectures are the latest emerging computer architectures surpassing the efficiencies of Field Programmable Gate Arrays (FPGAs) and often achieving near efficiencies near the levels of Application-Specific Integrated Circuits (ASICs). In image processing and computer vision, edge detection is a very foundational step, specially by using convolution-based filters such as Sobel and Laplacian across image inputs. This application and performance evaluation-based final-year project aimed to explore and characterize the extent of parallelism extracted from a RISC-V system-on-chip (SoC) enhanced with an integrated CGRA accelerator designed in-house at the Agency for Science, Technology, and Research (A*STAR) Institute of Microelectronics (IME). The efficiency of executing two-dimensional convolution implementations of Sobel and Laplacian filters on 4 varying image array sizes is compared across CPU-only and CGRAaccelerated environments as part of the experiments. This comparison is presented in terms of execution time and latencies, arithmetic operation counts, and power consumption across all implementations and image sizes, detailing software optimizations that were necessary for this project and analyzing the effectiveness of the given printed circuit board (PCB). The results strongly demonstrated that the CGRA simulated implementation of the two filters processed the images significantly faster than its CPU counterpart – up to 5.08× and 10.41× in terms of energy and latency respectively for Sobel filtering of 32×32 image size, with an average of 31.7× speedup in latency across all sizes. Future works can attempt interfacing this CGRA accelerated PCB with a camera module and LCD display for real-time edge detection which has also been discussed with all the challenges that prevented full hardware integration and testing in time.Bachelor's degre
Profiling the cognitive abilities of children in learning support programmes in Singapore: exploring the use of executive functions in classifying learning difficulty subtypes
No full textWhile research on executive functions (EF) as predictors of literacy and numeracy
achievement has been generally robust, it is still limited in translating the findings to contexts
pertaining to the identification of learning difficulties (LD) in school children. Hence, this
study examined the profiles of EF in 86 Primary 1 children aged 6- to 7-years enrolled in
Learning Support Programmes for reading and mathematics in Singapore and explored how
differences in EF can distinguish between LD subtypes. Multivariate group differences in EF
between LD subtypes was found. Post-hoc tests showed that group differences were
significant for visuospatial working memory. Insufficient evidence precluded a further
discriminant analysis. Secondary analyses applying normative data on reading, spelling and
number skill abilities between LD subtypes found significant impairments of children with
LD on all abilities compared to typically developed peers. There were no significant
differences in all abilities between children with LD in reading only and children with LD in
mathematics only. Results suggest that children in the Learning Support Programmes may
have differing profiles in EF largely explained by visuospatial working memory, and that
initial academic screeners may not be sensitive to future underachievement. Future research
on EF profiles including typically developing children could inform the use of EF in
supplementing current screeners of LD in the local education system.Bachelor's degre
Scaling up purcell-enhanced self-assembled nanoplasmonic perovskite scintillators into the bulk regime
No full textScintillators convert high-energy radiation into detectable photons and play a crucial role in medical imaging and security applications. The enhancement of scintillator performance through nanophotonics and nanoplasmonics, specifically using the Purcell effect, has shown promise but has so far been limited to ultrathin scintillator films because of the localized nature of this effect. This study introduces a method to expand the application of nanoplasmonic scintillators to the bulk regime. By integrating 100-nm-sized plasmonic spheroid and cuboid nanoparticles with perovskite scintillator nanocrystals, nanoplasmonic scintillators are enabled to function effectively within bulk-scale devices. Power and decay rate enhancements of up to (3.20 ± 0.20) and (4.20 ± 0.31) folds are experimentally demonstrated for plasmonic spheroid and cuboid nanoparticles, respectively, in a 5-mm thick CsPbBr3 nanocrystal-polymer scintillator at RT. Theoretical modeling also predicts similar enhancements of up to (2.26 ± 0.31) and (3.02 ± 0.69) folds for the same nanoparticle shapes and dimensions. Moreover, a (2.07 ± 0.39) fold increase in light yield under 241Am γ-excitation is demonstrated. These findings provide a viable pathway for utilizing nanoplasmonics to enhance bulk scintillator devices, advancing radiation detection technology.Ministry of Education (MOE)Published versionAll authors acknowledge research funds from the National Science Center, Poland, under grant OPUS-24 no. 2022/47/B/ST5/01966. M.M. acknowledges the funds from the National Science Center, Poland, under grant Miniatura-8 no. 2024/08/X/ST5/00980 for consumable purchase. F.M. and C.D. acknowledge the financial support from the Ministry of Education, Singapore, under its AcRF Tier 2 grant (MOE-T2EP50121-0012). F.A.A.N. acknowledges funding from the Indonesian Endowment Fund for Education (LPDP) on behalf of the Indonesia Ministry of Education, Culture, Research and Technology, which is managed by Universitas Indonesia under INSPIRASI Program (Grant No PRJ-61/LPDP/2022 and 612/E1/KS.06.02/2022)
Multi-robot communication for collaborative mapping in large-scale environments
No full textMulti-robot collaboration plays a vital role in various applications. A critical challenge lies in ensuring wireless communication among multiple robots. However, current ROS-based solutions often fail to simultaneously meet these requirements under dynamic and large-scale deployment conditions.
This dissertation proposes a complete multi-robot wireless communication framework that addresses the above issues:
1. A custom hardware system was designed and built: The system consists of three mobile robots and one base station, each equipped with 5G routers and optimized for reliable field deployment.
2. A software framework was designed and implemented: The framework bridges ROS 1 and ROS 2 while replacing DDS with Zenoh to eliminate multicast traffic.
3. A collaborative multi-robot mapping task was conducted to validate the system: Field trials over a campus site confirm that the system achieved a median one-way latency of 42ms, an upstream rate of 9Mb s−1 per robot, and 95% link reachability.
Collectively, this integrated solution lays the groundwork for more efficient, resilient, and scalable multi-robot systems in large-scale deployments.Master's degre
New “sober” design of earth observation satellite constellations
No full textEarth Observation (EO) satellites play an important role in each of our daily lives but there is a
lack of revisit times around the equatorial belt. Most of the high-performance EO satellites
launched by big companies are launched into Sun-Synchronous Orbit which does not provide a
good enough revisit time for equatorial region. In response, private companies are preparing to
launch more EO satellites, specifically CubeSats, and is predicted to triple over the next decade.
This poses an environmental issue of resources consumptions for materials and launches as well
as space environmental issues such as overcrowding in the Low-Earth Orbit (LEO). Is there a
method to optimize what is available in space through cross-calibration while providing good
revisit time for equatorial region and reduce environmental impacts as well?
This project explores in synergizing near equatorial satellite, CubeSats, together with near polar
satellite, current high-performance satellite by big companies, using cross-calibration through a
phenomenon called Simultaneous Nadir Overpass (SNO). Cross-calibration improves near
equatorial CubeSats data collection while it provides good revisit time around equator. This study
will explore and understand the opportunities of SNO occurrences as well as providing a method
to predict SNO.Bachelor's degre
Research on financing selection and decision making of green supply chain under retailer financial constraints
No full textThe rapid development of the global economy has led to increasingly serious environmental damage, and governments and related industries around the world are seeking solutions to this problem. Governments around the world are taking action to combat this issue by pushing for greener product manufacturing, fostering an eco-friendly economy, enhancing the natural environment over time, and even facilitating green loans to help firms with their capital needs while they upgrade their technology. Nevertheless, for a green supply chain to be built, it is not enough for upstream producers to use green manufacturing inputs; downstream retailers must also do their part to promote green ideas and sell green products. However, small and medium-sized shops face challenges in obtaining adequate financial backing and covering green-related expenditures. These challenges include a lack of a large production scale and a poor credit rating.
Against this background, this dissertation focuses on the financing issues in the green supply chain, and discusses the options for retailers with funding constraints under different financing models and their impact on operational decisions. The focus of this dissertation is a two-part green supply chain that includes a manufacturer and a store with limited capital. Under three different financing scenarios—no financing, trade credit financing, and bank loan financing—this dissertation builds a green supply chain operating choice model using game theory, operations research, and other theories and methods. This dissertation examines the optimal financing model selection and operational decisions from the dual perspectives of manufacturers and retailers. It compares and analyzes the core decisions and profit changes of all parties in the supply chain under different financing models. The goal is to provide a reference for optimizing capital-constrained green supply chains.Master's degre
Decentralized inertia emulation in hybrid single-phase AC/DC microgrids
No full textOne of the most significant approaches to alleviate the global energy crisis is to incorporate renewable energy sources (RESs) into microgrids. Inertia is a critical aspect of microgrids, reflecting the power system’s ability to regulate frequency during transient events (e.g., load demand changes). However, the significant integration of RESs interfaced via power electronic converters may lead to a reduction in system inertia. To address this issue, inertia emulation can be implemented using distributed generation units (DGUs) in hybrid microgrids. A key challenge is the transmission of frequency information, which typically relies on communication between grid-tied inverter and DGUs or central controllers.
To overcome these challenges, a decentralized inertia emulation strategy for hybrid single-phase AC/DC microgrids is proposed. In single-phase system, the second-order ripple on the DC side inherently reflects the grid frequency, allowing DC-side units to locally extract this information for inertia emulation. These DC-side units can then utilize the extracted frequency data to implement power compensation. The objectives of this project mainly involve designing control algorithms for single-phase inverter and DC-side DGU interface Boost converter, based on theoretical analysis and simulation validation.Master's degre
Hardware implementation of partial discharge detection
No full textThis dissertation presents a hardware-software co-designed FPGA-based system for real-time partial discharge (PD) detection, addressing critical limitations of traditional software-centric approaches in latency, scalability, and industrial applicability. Partial discharge, a key indicator of insulation degradation in high-voltage equipment, necessitates rapid and reliable detection to prevent catastrophic failures. While existing methods rely on computationally intensive software pipelines with poor real-time performance, this work leverages FPGA parallelism and hardware optimisation to achieve low latency. The proposed architecture integrates three core stages: signal preprocessing with a 256-point FFT module, energy spectrum-based feature extraction, and K-means clustering for PD pattern classification.
Key innovations include a ping-pong buffering mechanism for zero-wait data transfer, fixed-point arithmetic for resource efficiency, and automated file generation for seamless MATLAB-to-FPGA workflow integration.Master's degre