Filters
Availability
Publication Year
Resource Type
Faculty
School/Institute
Publication Search Results
Sort
Results per page
1 - 10 of 66
-
(2020)ThesisThe demand for high frequency, high power density and high efficiency on power converters has been increasing through the years. Solid state transformer (SST), which is typically formed by three stages, one being DC/AC inverter, which is controllable, one being medium-frequency transformer isolated converter and the other being AC/DC rectifier, has been developed quickly during past few decades to replace traditional 50Hz transformer. In SST, the use of planar transformer on print circuit boards (PCB) working at tens to hundreds of kilohertz can provide isolation and save space. In this thesis, research activities to develop a planar transformer isolated DC/DC converter for solid-state transformer are presented. To further increase efficiency, resonant converter with series LC and shunt LC filter is adopted to realize soft-switching, and to eliminate high order harmonics hence reducing unnecessary core loss and copper loss. For the designs of planar magnetics, PCB based litz wire air-core planar inductor and PCB based litz wire planar transformer are proposed, analyzed, simulated and tested, which shows better performance than traditional solid rectangular winding ones. By applying the novel litz wire inductor and transformer in the proposed DC/DC converter prototype, high order harmonics are eliminated and the efficiency of 90% is reached from experimental results, which can be further developed by an integrated system.
-
(2021)ThesisA bipartite graph is a graph with two layers such that vertices in the same layer are not connected, which is widely used to model the relationships among two types of entities. Examples of bipartite graphs include author-paper networks, customer-product networks, and ecological networks (e.g., the predator-prey network and the plant-animal network). In bipartite graphs, cohesive subgraph computation is a fundamental problem that aims to find closely-connected subgraphs, which can be applied to group recommendations, network visualization, and fraud detection. In this thesis, we propose a novel cohesive subgraph model called τ -strengthened (α, β)- core (denoted as (α, β)τ -core), which is the first work to consider both tie strength and vertex engagement on bipartite graphs. An edge is a strong tie if contained in at least τ butterflies (2 x 2-bicliques). (α, β)τ -core requires each vertex on the upper or lower level to have at least α or β strong ties, given strength level τ. To retrieve the vertices of (α, β)τ -core optimally, we construct index Iα,β,τ to store all (α, β)τ -cores. Effective optimization techniques are proposed to improve index construction. To make our idea practical on large graphs, we propose 2D-indexes Iα,β, Iβ,τ, and Iα,τ that selectively store the vertices of (α, β)τ -core for some α, β, and τ . The 2D-indexes are more space-efficient and require less construction time, each of which can support (α, β)τ -core queries. As query efficiency depends on input parameters and the choice of 2D-index, we propose a learning-based hybrid computation paradigm by training a feed-forward neural network to predict the optimal choice of 2D-index that minimizes the query time. Extensive experiments show that (1) (α, β)τ -core is an effective model capturing unique and important the proposed techniques significantly improve the efficiency of index construction and query processing.
-
(2021)ThesisRelation prediction is a fundamental task in network analysis which aims to predict the relationship between two nodes. Thus, this differes from the traditional link prediction problem predicting whether a link exists between a pair of nodes, which can be viewed as a binary classification task. However, in the heterogeneous information network (HIN) which contains multiple types of nodes and multiple relations between nodes, the relation prediction task is more challenging. In addition, the HIN might have missing relation types on some edges and missing node types on some nodes, which makes the problem even harder. In this work, we propose RPGNN, a novel relation prediction model based on the graph neural network (GNN) and multi-task learning to solve this problem. Existing GNN models for HIN representation learning usually focus on the node classification/clustering task. They require the type information of all edges and nodes and always learn a weight matrix for each type, thus requiring a large number of learning parameters on HINs with rich schema. In contrast, our model directly encodes and learns relations in HINs and avoids the requirement of type information during message passing in GNN. Hence, our model is more robust to the missing types for the relation prediction task on HINs. The experiments on real HINs show that our model can consistently achieve better performance than several state-of-the-art HIN representation learning methods.
-
(2021)ThesisPhotovoltaic (PV) energy is one of the most prominent renewable energy sources today. Traditionally, PV modules are connected in series to form a PV string, interfacing with PV inverter for grid connection. Since all PV modules are operating at the same current, the energy yield of the system is limited by the underperforming modules. PV optimizers, a concept which enhances maximum power point tracking (MPPT) of PV systems using power electronics, has been studied over the last few years. An emerging technique, submodule differential power processing (DPP) is proposed to improve the efficiency and enhance MPPT to a finer granularity. By diverting the differential current between submodules, DPP optimizers only process a fraction of the power the system produces, improving the overall efficiency. This thesis aims to research PV-to-PV DPP systems and its PV-to-Serial-Port variant, including modelling, control techniques, and MPPT strategies. The inverting buck-boost converter is the preferred topology for optimizers, as it is one of the simplest topologies which supports bidirectional power flow and both step-up and step-down voltage conversions. A small-signal model is derived as the basis of controller tuning involved in this thesis. A novel pairing between control inputs and outputs based on relative gain array analysis is proposed, mitigating coupling effects between optimizers, and simplifying the controller design. Furthermore, a double-loop outside-in exact MPPT strategy is presented, improving the tracking speed. Simultaneous voltage regulation of multiple submodules and exact MPPT strategy have been verified by experiments. Voltage equalization MPPT techniques are investigated as they eliminate the communication requirements between modules. Voltage equalization is investigated for both MPPT and flexible power point tracking (FPPT) applications, leading to the discovery of an unstable operation mode caused by linearization with the differential resistance method. Experiments validate that, with proposed tuning techniques closed-loop equalization is stable in regions where the power is sufficiently low to perform FPPT. A comparison of MPPT performance between open-loop and closed-loop equalizations shows that closed-loop equalization has better energy yield under severe mismatch, as it vastly reduces steady-state errors. Finally, a PV-to-Serial-Port variant, aiming to improve the practicality of DPP techniques is presented. With only one inter-module power connection and no communication requirements, this new architecture is more suitable for modular integration than its PV-to-PV counterpart. A novel topology of PV-to-Serial Port architecture which supports voltage equalization is proposed. A clear contribution that thoroughly analyses and compares the inductor sizing of practical PV-to-Serial-Port topologies considering arbitrary PV current mismatch is presented. It is validated by simulation that, with the combination of PV-to-Serial-Port and voltage equalization, MPPT can be performed at module-level autonomously and can handle sudden irradiance changes. A high-level comparison between systems in three core chapters is given in the conclusion chapter, outlining the techniques utilized in each chapter and their limitations. Experiments in this thesis are undertaken on two identical inverting buck-boost converter prototypes, rated at 100W, switching at 200kHz, and interfacing with a TI controlCARD. Control techniques and MPPT strategies are implemented digitally and can be applied to existing inverting buck-boost optimizers if power ratings and sensory requirements are met. The research outcome of this thesis improves the practicality of DPP techniques by simplifying control methods and eliminating communication requirements. Furthermore, it establishes the FPPT capability of DPP systems for providing grid support. Lastly, it provides a modular-integrable topology, enhancing the scalability of DPP systems.
-
(2021)ThesisWith the current increase in demand for cleaner energy sources and natural gas in particular, the requirement for better understandings of fluid transportation in porous media is also on the rise. This thesis is focusing on the sorption and diffusion process of hydrocarbon in different formations, a process not well understood in the oil and gas industry. In the first major section, this thesis addresses the main disadvantages of the current models for the sorption/diffusion process, which is the independence of diffusion rate with respect to time and the saturation of gas in the adsorbent. Furthermore, these models assume all the pore sizes within the coal are of constant size, which is not representative of the real rock. We propose a new model based on the reaction-diffusion theory to improve upon the popular unipore model. The model separates the adsorption process from the effective diffusivity by coupling in a reaction term. Moreover, this model also describes the dependence of the gas transportation rate on temperature, activation energy, and gas concentration rate. Together with the new model, a hypothetical experimental and analysis procedure is presented to validate this modification. The new findings of the latter part of the thesis call, however, for an extension of this approach. In the second part of the thesis, Transmission Electron Microscopy (TEM) and Guinier Analysis on Small-Angle Neutron Scattering (SANS) data were performed, revealing the complex nanopore structure of the silica aerogel at 3 different sizes: 3 nm, 9 nm and 0.18 nm. Contrast-match (CM) SANS was employed to investigate the sorption behaviour of methane in these pore regions, using CD4 with fluid pressure up to 1 kbar. The CM-SANS experiment discovered the following sorption behaviour of CD4 in the pore region of 3 nm and 9 nm: (1) all the pores are accessible to CD4, (2) CD4 pressure within the pore is equal to the bulk CD4 pressure, and (3) no adsorption layer on the pore-matrix interface was found. Analysis of SANS data for the 0.18 nm pores indicates capillary condensation is the major factor controlling the CD4 sorption behaviour. After the pressure cycling process to up to 1000 bar, when returning to vacuum, the silica structure at this scale is permanently damaged due to the invading gas
-
(2020)ThesisCyanobacteria, also referred to as blue-green algae, are aquatic and photosynthetic, that is, that live primarily in fresh water and salt water. In addition to unattractive colour and smelly odour, abundance of cyanobacteria worsens water-quality and generates toxins that can harm humans and animals alike. Lack of enough data, its independence across multiple sampling time steps, as well as the presence of more than one causative factors, each with different levels of influence on the response, has resulted in limited progress in the development of generalized Cyanobacteria modelling and prediction frameworks. In this thesis, using a few key dominant factors, relatively practical and universally applicable two models for predicting the cyanobacterial bloom have been developed. The first model is a binary model and forecasts the occurrence/non-occurrence of cyanobacterial bloom at a given time step conditional on the dominant environmental variables and the cyanobacteria concentration at the preceding time step. The bacterial growth dynamic to the model is included by defining the weight functions which quantify the importance assigned to the key environmental variables namely, temperature, velocity and nutrients. A probabilistic model can yield a distribution of possible outcomes and therefore helps not only to understand the degree of outcome but also to make a relevant solution with uncertainty. Following this, a two-stage probabilistic model has been developed. In the first stage, cyanobacteria occurrences are generated using a first-order conditional Markov model. The conditioning vector includes the cyanobacterial count on the preceding time step and a few environmental variables. On occasions where the first stage model predicts the occurrence of cyanobacteria, the second stage model generates cyanobacterial cell counts using a nonparametric kernel density approach assuming first-order Markovian conditional dependence. As a final stage of this thesis, a few scenarios for controlling cyanobacterial bloom growth are assessed in terms of changes in the environmental variables and financial implications. Both developed models provide promising results and offer the capability of applying them to other areas. The suggested countermeasure provides an interesting and economically feasible solution to deal with cyanobacterial bloom issue
-
(2022)ThesisEmerging modes of air transport such as autonomous airport shuttle and air taxi are potentially efficient alternatives to current transport practices such as bus and train. This thesis examines bus shuttle service within an airport and air metro as two examples of network design. Within an airport, the bus shuttle serves passengers between the terminals, train stations, parking lots, hotels, and shopping areas. Air metro is a type of pre-planned service in urban air mobility that accommodates passengers for intra- or inter-city trips. The problems are to optimise the service, and the outputs including the optimal fleet size, dispatch pattern and schedule. Based on the proposed time-space networks, the service network design problems are formulated as mixed integer linear programs. The heterogeneous multi-type bus fleet case and stochastic demand case are extended for the airport shuttle case, while a rolling horizon optimisation is adopted for the air metro case. In the autonomous airport inter-terminal bus shuttle case, a Monte Carlo simulation-based approach is proposed to solve the case with demand stochasticity, which is then further embedded into an "effective" passenger demand framework. The "effective" demand is the summation of mean demand value and a safety margin. By comparing the proposed airport shuttle service to the current one, it is found that the proposed service can save approximately 27% of the total system cost. The results for stochastic problem suggest estimating the safety margin to be 0.3675 times of the standard deviation brings the best performance. For the second case, the service network design is extended with a pilot scheduling layer and simulation is undertaken to compare the autonomous (pilot-less) and piloted service design. The results suggest that an autonomous air metro service would be preferable if the price of an autonomous aircraft is less than 1.6 times the price of a human-driven one. The results for rolling horizon optimisation suggest to confirm the actual demand at least 45 minutes prior to departure. Based on data from the Sydney (Australia) region, the thesis provides information directly relevant for the service network design of emerging modes of air transport in the city.
-
(2021)ThesisDegenerate microwave amplifiers offer the potential to perform a measurement of weak mi- crowave signals free from any degradation to their signal-to-noise ratio, a benefit afforded by the quantum mechanics of their phase sensitive gain [1]. The degenerate parametric amplifier may even be used to enhance the signal-to-noise ratio of a weak measurement by squeezing the vacuum fluctuations below the standard quantum limit. Consequently, parametric amplifiers have featured in a number of recent sensitivity breakthroughs across multiple fields. These include the detection of gravitational waves [2], the search for dark matter [3], and high sensitivity electron spin resonance spectroscopy [4]. In this thesis, we present a novel microwave degenerate parametric amplifier based on the non-linear superconducting phenomenon of kinetic inductance. The device is a simple and robust quantum limited phase sensitive amplifier, which is well described by theory and shows potential as a highly-effective tool for the production of squeezed microwave light.
-
(2021)ThesisThe Arabian basin was subject to several tectonic events, including Lower Cambrian Najd rifting, the Carboniferous Hercynian Orogeny, Triassic Zagros rifting, and the Early/Cretaceous and Late/Tertiary Alpine orogenic events. These events reactivated Precambrian basement structures and affected the structural configuration of the overlying Paleozoic cover succession. In addition to a 2D seismic array and several drill well logs, a newly acquired, processed 3D seismic image of the subsurface in part of the basin covering an area of approximately 1051 km2 has been provided to improve the understanding of the regional tectonic evolution associated with these deformation events. In this study, a manual interpretation is presented of six main horizons from the Late Ordovician to the Middle Triassic. Faults and folds were also mapped to further constrain the stratigraphic and structural framework. Collectively, this data is used to build a geological model of the region and develop a timeline of geological events. Results show that a lower Paleozoic sedimentary succession between the Late Silurian to the Early Permian was subject to localised tilting, uplift, and erosion during the Carboniferous Hercynian Orogeny, forming a regional unconformity. Subsequent deposition occurred from the Paleozoic to the Mesozoic, producing a relatively thick, conformable, upper succession. The juxtaposition of the Silurian rocks and Permian formations allows a direct fluid flow between the two intervals. Seismic analysis also indicated two major fault generations. A younger NNW-striking fault set with a component of reverse, east-side-up displacement affected the Lower Triassic succession and is most likely related to the Cretaceous and Tertiary Alpine Events that reactivated the Najd fault system. These fault structures allow vertical migration that could act as conduits to form structural traps. Manual mapping of fault structures in the study area required significant time and effort. To simplify and accelerate the manual faults interpretation in the study area, a fault segmentation method was developed using a Convolutional Neural Network. This method was implemented using the 3D seismic data acquired from the Arabian Basin. The network was trained, validated, and tested with samples that included a seismic cube and fault images that were labelled manually corresponding to the seismic cube. The model successfully identified faults with an accuracy of 96% and an error rate of 0.12 on the training dataset. To achieve a more robust model, the prediction results were further enhanced using postprocessing by linking discontinued segments of the same fault and thus, reducing the number of detected faults. This method improved the accuracy of the prediction results of the proposed model using the test dataset by 77.5%. Additionally, an efficient framework was introduced to correlate the predictions and the ground truth by measuring their average distance value. This technique was also applied to the F3 Netherlands survey, which showed promising results in another region with complex fault geometries. As a result of the automated technique developed here, fault detection and diagnosis were achieved efficiently with structures similar to the trained dataset and has a huge potential in improving exploration targets that are structurally controlled by faults.
-
(2021)ThesisIn the past decade, the discovery of the CRISPR/Cas system launched a new era of genome editing and has rapidly become a universal research tool. The specific recognition between guide RNA and target nucleic acids enables CRISPR/Cas systems to exhibit high specificity compared to gene editing systems such as zinc finger or transcription activator-like effector nucleases. In this thesis, the CRISPR/Cas12a and CRISPR/Cas13a systems were designed to modify DNA or RNA associated with the HSPG2 gene which encodes perlecan, the major extracellular heparan sulphate (HS) proteoglycan in basement membranes. Perlecan has essential roles in organ development and contributes angiogenesis in pathological processes including cancer. The Cas12a and Cas13a nucleases were expressed in transformed E. coli and purified via His-tag affinity followed by size exclusion chromatography. The purified Cas12a nuclease exhibited high activity and specificity in a collateral activity assay in which the reporter sequence was only cleaved in the presence of the Cas12a protein and target ssDNA. Similarly, Cas13a nuclease exhibited high activity and specificity in a collateral activity assay with target ssRNA. A CRISPR/Cas12a gRNA was designed to target exon 2 of HSPG2 and was able to cleave amplified genomic DNA extracted from human melanoma cell line, MM200. Additionally, the collateral activity assay revealed that Cas12a nuclease dose-dependently cleaved the reporter ssDNA when used with target HSPG2 DNA. Similarly, a CRISPR/Cas13a gRNA was designed to target exon 2 of HSPG2 RNA and was able to cleave the target RNA extracted from MM200 cells and was active in the collateral activity assay when used with target HSPG2 RNA. Modification of HSPG2 nucleic acids in both MM200 cells and human umbilical vein endothelial cells (HUVECs) was also established. The CRISPR/Cas12a system resulted in up to 39 and 24 % reduction in HSPG2 gene expression in MM200 and HUVECs, respectively. Moreover, the CRISPR/Cas13a system achieved up to 69 and 99% reduction in HSPG2 RNA in MM200 and HUVECs, respectively. The HSPG2 mRNA modification in both MM200 and HUVECs resulted in decreased expression of FGF2 and VEGF-A, genes involved in the perlecan signalling pathway networks and associated with angiogenesis. The established CRISPR/Cas12a and CRISPR/Cas13a systems provide novel and efficient nucleic acid editing tools to further study the functions of perlecan in vitro and potentially in vivo. In addition, the LbCas12a or LwCas13a-based collateral cleavage assay enabled efficient and specific detection of the HSPG2 genome or transcripts, suggestive of its potential in perlecan-related disease diagnoses, such as cancer or genetic disorders.