Doctoral Theses

Permanent URI for this collectionhttps://hdl.handle.net/2292/2

Deposit Your Thesis instructions on ResearchHub
Copyright and Access information
Third Party Copyright information (School of Graduate Studies)
Contact Us

Browse

Now showing 1 - 20 of 7126

Building Information Modelling-based Smart Inspection Data Management for Unmanned Aerial Vehicle-enabled Visual Building Inspection
(ResearchSpace@Auckland, 2024) Zhang, Cheng; Zou, Yang; Dimyadi, Johannes; Chang-Richards, Alice
Unmanned Aerial Vehicles (UAV) equipped with red, green, blue (RGB) and thermal cameras have been growingly used for visual building inspections. However, its full potential remains untapped due to challenges in managing fragmented and distorted images. While Building Information Modelling (BIM) has been envisaged as an effective platform for managing building lifecycle data, its potential for handling UAV inspection data has not been explored. To address this challenge, this doctoral research has investigated the use of BIM to establish a smarter and more efficient method for managing the UAV building inspection data. Firstly, a systematic review was conducted on the state-of-the-art UAV and BIM applications in visual building inspection. The review findings revealed four key research gaps. Secondly, the research conducted a comprehensive investigation in effective UAV thermal image acquisition strategy, focusing on three critical factors: temperature difference between building interior and exterior, ground sampling distance (GSD) of thermal images, and UAV oblique angle. The impact of these factors on inspecting façade anomalies was examined both qualitatively and quantitatively through laboratory and field experiments. Thirdly, the research proposed a BIM-based scheme for managing fragmented and distorted UAV images. An improved Generalised Hough Transform (GHT) method was developed for aligning RGB images with BIM by matching building façade features. However, this method faces certain obstacles when dealing with specific types of façades, such as curtain walls. To overcome this, a Structure from Motion (SfM) method was introduced as a supplementary solution. Additionally, perspective and geometry distortions on UAV images, especially when inspecting single-curved façades, were eliminated by BIM-based 3D surface unwrapping. Moreover, UAV thermal images were calibrated, corrected, enhanced, and registered onto BIM through multi-source image fusion and histogram-based correction. The effectiveness of the proposed scheme has been validated by computer simulations and field experiments, demonstrating its ability to convert fragmented and distorted UAV RGB and thermal images into a distortion-free panoramic image, seamlessly integrable into BIM. Finally, an implementation of a system that consolidates all developed approaches has validated the formulation of an effective UAV image acquisition strategy with efficient inspection data management in BIM.
Contemporary Central Banking - Analytical Perspectives
(ResearchSpace@Auckland, 2024) Haworth, Cameron McFarlane; Gai, Prasanna
This thesis considers the design of contemporary central bank policy from a theoretical perspective. In Chapter 1, we investigate the effects of central bank commitments to alternative monetary policy tools. We present a simple model to characterise central bank forward guidance, large scale asset purchases and yield curve control. Endogenous yield curve reactions to policy shocks are a key determinant of monetary conditions and can support or offset policy intentions. Alternative monetary policy tools allow the central bank to signal central bank private information to investors when the policy rate is at the effective lower bound, shaping yield curve reactions. Commitments to these tools offer policy certainty but can become time-inconsistent if the economy recovers sooner than expected. If commitments become time-inconsistent, endogenous yield curve tightening can offset excess stimulus. The strength of this offset, and the optimality of monetary conditions, improves with the precision of investor inferences of central bank private information. In Chapter 2, we examine the causal link between asset bubbles and wealth inequality in a twoagent macroeconomic model. Bubbles influence wealth inequality through two channels: altering the debt-asset ratio and fuelling speculation. When bubbles grow, they can temporarily decrease wealth inequality if asset prices rise faster than debt. However, when they burst, wealth inequality increases as the debt-asset ratio rises. Steady state wealth inequality is unaffected by bubbles if household types share symmetric speculative timing. Although macroprudential policy, communication, and leaning against the wind can reduce negative bubble effects on aggregate utility, they have a limited effect on wealth inequality. In Chapter 3, we introduce cautious expectations to a macroprudential policy model where average growth is traded off against growth-at-risk (GaR). Policymakers with cautious expectations estimate the optimal weight to apply to risk signals, creating biased, historically dependent crisis forecasts. They optimally downweight the effects of risk and their policy settings on GaR forecasts, decreasing the expected efficiency of the growth-GaR trade-off. This loosens the optimal policy stance, but also causes policymakers to respond more aggressively to changing signals. As policymakers experience additional crises, they better understand the effects of their policy instruments and tighten their stance. When past crises are forgotten, this tendency reverses.
Synthesis of Antiviral Imino-C-Nucleoside Analogues
(ResearchSpace@Auckland, 2023-10) Wong, Chi Hung Andy; Brimble, Margaret
Nucleoside analogues are a diverse class of synthetic drugs that exhibit broad antiviral and anti infective activity. Critically, nucleoside analogues serve as first-line therapeutic agents to treat viral infections of pandemic potential. Despite various clinical successes, nucleoside drug discovery has been hampered by inefficient access to unnatural nucleoside frameworks, such as the iminoribitol scaffold observed in the adenosine analogue, galidesivir (29). With the overall goal of streamlining the synthesis of nucleosides, the modular assembly of imino-C-nucleosides was devised, featuring use of a cycloaddition reaction as the key heterocycle-forming step. The first part of this work involved identification of an appropriate alkyne-bearing intermediate that would be compatible with the proposed cycloaddition reaction and downstream transformations to effect nucleoside synthesis. To this end, divergent access to both imino-C-nucleoside anomers was established from protected D-serine 112. Hemiaminal 130, when subjected to a telescoped cycloaddition and pyrrolidine formation sequence, showed preference for α-nucleoside formation. On the other hand, the analogous cycloaddition and deprotection using alkynyl pyrrolidine 148 led to the exclusive formation of the desired β-nucleoside 139·HCl. The second part of this work elaborated on the cycloaddition strategy as a rapid method to introduce diversity in medicinal chemistry campaigns. To maximise the efficiency of triazole construction, a one-pot method was developed for the in-situ generation of aryl or heteroaromatic azides from the corresponding boronic acids using Chan-Lam-type conditions, that were then captured for concomitant Copper-catalysed Azide Alkyne Cycloaddition (CuAAC). Copper-cyclodextrin complex was identified as the superior catalyst for this transformation. After extensive optimisation, 24 analogues were prepared from advanced alkyne 200 using the developed protocol. Synthetic efforts were supported by structural analysis and antiviral assessment of the flex-imino-C-nucleosides, a subclass of nucleoside analogues characterised by a split arene ring motif. This work provides insight for the continued development of nucleoside analogues, whereby use of a cycloaddition strategy could be broadly implemented for the convenient access to C-nucleosides bearing unique heterocyclic nucleobase surrogates in future diversity oriented syntheses.
Corporate Disclosure Strategies among Chinese Firms: Insights from CSR Reporting, Value-reducing M&As, and the COVID-19 Pandemic
(ResearchSpace@Auckland, 2024) Yin, Fangyi; Cahan, Steven; Chen, Jerry
The overarching theme of this thesis centres on corporate disclosure strategies of Chinese A-share firms in varying contexts. It comprises three studies, each delving into how firms adapt their disclosure practices in response to peer influences, the challenges of investor relations management, and the demands of crisis communication, respectively. The first study (Chapter 2) investigates industry peer effects under China’s selective mandatory corporate social responsibility (CSR) disclosure regime. It reveals that non-mandated firms are more likely to voluntarily disclose CSR information when a higher proportion of industry peers are mandated to do so, exhibiting a positive peer effect. On the other hand, firms that already engage in voluntary CSR disclosures tend to cease these practices as the number of mandated disclosers in their industry increases, indicating a negative peer effect where firms free-ride on peers’ mandatory disclosures. The second study (Chapter 3) examines how firms strategically manage their disclosure tone after value-reducing mergers and acquisitions (M&As). The findings suggest that acquirers tend to employ a negative tone in their CSR reports aimed at institutional investors, while adopting a positive tone on the Interactive Investor Platform (IIP) that is mainly used by retail investors. This inconsistency in disclosure tone between the platforms appears driven by an intent to inform institutional investors and, at the same time, mislead retail investors, consistent with the acquirer “speaking in two tongues” after negative events. The third study (Chapter 4) focuses on firms’ communication strategies during the novel Coronavirus disease (COVID-19) pandemic. A difference-in-difference (DID) analysis reveals that firms subject to complete or partial lockdown shift to a more positive reply tone on the IIP in response to a less positive ask tone from retail investors. This overly optimistic tone, linked to poorer future firm performance, suggests firms’ attempt to mislead retail investors during the lockdown; however, the market reaction tests show that the impact of such opportunistic behaviour is short-lived.
Complicating Vulnerability: Exploring the Emplaced Experiences of Children with Migrant Parents in Rural Southwest China
(ResearchSpace@Auckland, 2024) Wu, Shuang; Kemp, Susan; Bartley, Allen
In China, the wellbeing of children whose parents have migrated to seek employment in cities has been a focus of policy and research concern since the economic reforms of the 1980s. A large number of primarily quantitative studies have explored the impact of parents’ migration on their children’s emotional, behavioural, educational and nutritional wellbeing. As a whole, that body of research presents a view of so-called ‘left-behind children’ as vulnerable-a group at educational, emotional and social risk because of their parents’ migration decisions. While research has begun to recognize the rights and agency of this group, studies from the children’s perspectives remain scarce. Responding to this gap, this study explored how children with migrant parents in rural China engage with and perceive their daily living environment. The study design was grounded in a children’s rights frameworks, the literature on children’s place experiences, perspectives from the sociology of childhood that position children as capable informants, and participatory, place-focused frameworks and research methodologies. Taking a constructivist perspective, the study used photovoice and digital storytelling to explore the perceptions of the everyday lives of 20 children with migrant parents (ages 11-15) from a rural town in Southwest China. The children were invited to photograph local places that evoke happiness or sadness, explain the significance of these photos to peers, and craft a digital story with their chosen images. The COVID-19 pandemic necessitated a shift in the study design to online place-focused methods. The findings of this study highlight the strong attachment of the participating children to their homes, families, and family land, revealing their situated agency in key settings and the influence of place and socio-cultural context on their behaviour and perspectives. They also showcase the effectiveness of virtual participatory methods, including smartphone-supported photography and digital storytelling, as flexible and empowering tools for actively involving Chinese children in research, exploring their place experiences, and enabling them to express their views. This study makes a small but important step toward a broadened understanding of ‘left-behind’ children, shedding fresh light on their lives and experiences and contesting common stereotypes. It also introduces new approaches to comprehending and learning from Chinese children, especially those whose lives span a range of structural, cultural and interpersonal complexities.
Natural convection boundary layer in a stably stratified medium: Stability, transition and turbulence
(ResearchSpace@Auckland, 2024) Maryada, Krishna Reddy; Norris, Stuart
The three-dimensional linear stability, oblique transition and turbulence of a vertical natural convection boundary layer immersed in a stably stratified medium (vertical buoyancy layer) at a constant Prandtl number of 0.71 are investigated using numerical simulations. For linear instabilities, buoyancy and viscosity do not always destabilise the laminar flow but, depending on the Reynolds number, have a non-trivial influence on the linear growth of disturbances. Further, certain three-dimensional oblique wave instabilities have comparable growth rates to two-dimensional instabilities. These three-dimensional oblique waves can cause the vertical buoyancy layer to transition to turbulence without two-dimensional waves. This O-type transition in vertical buoyancy layers drastically differs from that observed in canonical wall-bounded flows. In vertical buoyancy layers, depending on the wavenumber of the initial oblique waves, the flow can transition by forming streaks, two-dimensional waves or a combination of the two. In contrast, canonical wall-bounded flows always transition by forming streaks. At sufficiently high Reynolds numbers, the flow becomes turbulent and new scaling laws are proposed for the turbulent mean flow and the one-point second-order turbulence statistics. The statistical turbulent structure of vertical buoyancy layers is investigated using two-dimensional correlations and one-dimensional energy spectra. It is demonstrated that large-scale streamwise elongated motions (eddies having length scales comparable to the outer length scale of the flow) populate the outer layer of the vertical buoyancy layer, exhibiting coherence across the entire thickness of the boundary layer. Finally, a theory and a phenomenological model are proposed to predict the relative scaling of longitudinal structure functions of streamwise velocity fluctuations. It is shown that the energy of the large-scale eddies is related to the intermittent dissipation field, implying that the large-scale and small-scale eddies are related. The developed theory is not limited to vertical buoyancy layers but is valid for shear-dominated turbulence in general.
On Business Cycle and Asset Pricing Implications of Quantitative DSGE Models
(ResearchSpace@Auckland, 2023) He, Qiaoxian; Dmitriev, Alexandre; Zhang , Haiping
This thesis develops more-data-consistent dynamic stochastic general equilibrium (DSGE)models by incorporating the respective characteristics of two key elements – cyclical capital utilization and composite-good habit formation – that are routinely missing but empirically important. Based on improved DSGE models, this thesis explores channels and mechanisms of these elements that help explain observed salient domestic and international business cycles and asset prices. Chapter 2 incorporates a vector-error-correction-model process for the utilizationadjusted total factor productivity (TFP) into a two-country, one-good model with variable capital utilization, motivated by the empirical evidence supporting cointegration between adjusted TFP processes across countries. The simulation results show that by introducing endogenous capital utilization, the benchmark model succeeds in producing (i) observed positive crosscountry correlations of investment if the cross-country correlation of productivity shocks is moderate and (ii) historical cross-country hours worked, output, and consumption correlations if the cross-country productivity differential is very persistent. Chapter 3 studies the role of consumption-leisure composite habit formation in explaining both the observed procyclical labor supply and the observed high level of equity premium. The composite habit is featured by a free habit-persistence parameter and a free habit-intensity parameter that is empirically and theoretically evident but has not been considered in asset-pricing models. The analytical results suggest that as habit intensity increases, the wealth effect on hours worked dampens; also, the equity premium rises. Habit persistence has the reverse effect. The numerical analysis confirms these analytical steady-state relations when introducing the composite habit into a centralized production-based asset pricing model. However, the effect of habit persistence is marginal for a medium value. Chapter 4 extends the model presented in Chapter 3 to a decentralized setting and estimates this model using Bayesian likelihood techniques to fit four U.S. quarterly macroeconomic quantity data. The estimation results suggest a level of habit intensity near one and a medium level of habit persistence, but the data are not informative well on the latter. By using posterior modes to parameterize composite habit, the estimated model matches the data relatively well, and the non-linearly approximated baseline model can reproduce a realistic mean risk-free rate and equity premium.
Longitudinal insights into the social and economic circumstances of families with children
(ResearchSpace@Auckland, 2024) Grant, Molly; Meissel, Kane; Exeter, Daniel
The social and economic circumstances of childhood play an important role in shaping children’s development. Examining the age children encounter these circumstances, the duration of exposure, and the patterning across childhood can provide insights into these experiences that cannot be ascertained from single-point-in-time data. Existing longitudinal research has investigated social and economic factors across time however, there are remaining gaps in understanding how these circumstances progress over childhood and how varied patterns of exposure differentially affect children’s development. The aim of this thesis is to address these gaps by focusing on the timing, duration, and sequencing of social and economic circumstances across childhood. Drawing on ecological and life course perspectives, and situated within the critical research paradigm, four studies within this thesis address this aim. The initial study is a systematic review that synthesises existing research to identify how the temporality of social and economic factors influences developmental outcomes. The subsequent three quantitative studies are situated within the Aotearoa New Zealand context. These studies draw on longitudinal data from families in the Growing Up in New Zealand study to focus on the economic construct of material hardship. As a measure of poverty, material hardship provides direct insight into the absence of basic necessities. Longitudinal research on material hardship is a growing field and the three quantitative studies in this thesis contribute to this scholarship by focusing on the key areas of: measurement, correlates, and actions aimed at mitigating material hardship. These studies track material hardship across four points in time, between infancy and early adolescence, to examine how material hardship is experienced by families with children. The findings from this thesis demonstrate the distinct understanding that can be gained from tracking social and economic circumstances over time. For the field of material hardship, this research offers important methodological and empirical contributions, as well as context-specific findings for Aotearoa New Zealand where prior longitudinal research is limited. This thesis also offers reflections on why some families face material hardship while others do not, highlighting material hardship as a structural issue of inequity linked to broader systems in society, beyond individual families.
Structural behaviour of cold-formed steel built-up stiffened box sections under bending, axial compression and web crippling
(ResearchSpace@Auckland, 2024) Dai, Yecheng; Raftery, Gary
The use of cold-formed steel (CFS) built-up sections in engineering practice is widely gaining popularity due to their ability to provide sustainable solutions and optimised section design opportunities. The aim of this research is to investigate the structural performance of novel CFS built-up stiffened box sections under bending, axial compression and web crippling. In total, 38 experimental tests were conducted, covering bending tests, axial compression tests, and web crippling tests. The material properties of test specimens were obtained from tensile coupon tests. Nonlinear finite element (FE) models were also developed and validated against the experimental test results, which showed good agreement. The validated FE models were then used to conduct a parametric study involving a total of 1700 FE models to investigate the effects of key parameters on the resistance of such CFS built-up stiffened box sections. Because these new sections are very different from those for which current design guidelines in the standards comprising AISI S100 (2016), AS/NZS 4600 (2018) and EN 1993-1-3 (2006) have been developed, the comparison shows that the design strengths predicted by these standards are not a good predictor of the actual performance for all three loading cases. Based on the experimental and numerical results, modified design equations in the form of the Direct Strength Method were proposed for calculating the resistance of CFS built-up stiffened box sections. A reliability analysis was conducted to evaluate the feasibility of the proposed design equations. The results indicated that the proposed design equations can closely predict the resistance of CFS built-up stiffened box sections under bending, axial compression and web crippling.
Ā Mātou Taonga Pīhoretanga: Halting Intergenerational Familial Childhood Sexual Abuse
(ResearchSpace@Auckland, 2024) Harrison, Nicola; Elizabeth, Vivienne; Le Grice, Jade
Whanaungatanga can halt familial childhood sexual abuse (FCSA) in whānau. There is global agreement that the prevalence of childhood sexual abuse (CSA) is unacceptable (Convention on the Rights of the Child, 1990). There is global acknowledgement that Indigenous communities are disproportionately affected (United Nations Declaration on the Rights of Indigenous Peoples, 2007). Yet CSA continues, despite research that traces intergenerational abuse in Indigenous families to the disruption of Indigenous protective systems (Cavino, 2016; Deer, 2015; Kuokkanen, 2015; Pihama et al., 2016; Rameka, 2018; A. Smith, 2015; D. Wilson, 2016; D. Wilson, Mikahere-Hall, Jackson, et al., 2019). This study reinforces existing research by demonstrating that FCSA-conducive contexts find space to grow as the mechanisms of colonialism disconnect Māori from their bases of power. It extends existing research by demonstrating that resisting the influences of colonialism can heal and halt intergenerational FCSA in whānau. Informed by Kaupapa Māori methodology, this study gathers narratives from 17 survivors of FCSA. The research design and methods are grounded in principles of manaakitanga, kaitiakitanga, and wairuatanga. A pūrākau method is used in analysis to center kaiāwhina lived experience. Modifiers to the term whanaungatanga are proposed. Practices that are infused with colonial attitudes (whanaungatanga hē) can then be distinguished from those that reflect ancestral tikanga9 (whanaungatanga iho). The method of distinction is demonstrated by examining the social products of whanaungatanga enactments as described by kaiāwhina. This project demonstrates that by discerning a distinction between enactments of whanaungatanga and leaning into those that reflect ancestral protocols, FCSA can be healed in whānau and halted across generations.
The Impact of Professional Learning and Development on Teachers’ Efficacy Beliefs and Implementation of the Hangarau Matihiko and Digital Technologies Curricula
(ResearchSpace@Auckland, 2024) Stringer, Lynley Rose; Lee, Kerry; Sturn, Sean; Giacaman, Nasser
To prepare young people for the impending workforce and societal changes caused by advancements in technology, many countries have begun introducing digital technology learning areas into their curricula where students are given opportunities to learn about technology. The New Zealand National Curriculum was updated in 2017 with the inclusion of the Hangarau Matihiko technological area in Te Marautanga o Aotearoa (the Indigenous Māori-medium curriculum) and the Digital Technologies technological area to the New Zealand Curriculum (English-medium curriculum). This need to prepare students to learn about new technologies and ways of working is well known, yet globally, implementation of Digital Technologies curricula has been slow and there are limited studies into the support teachers need to implement these new curricula. The aim of this research project was to gather empirical evidence about the impact training has on teachers’ experiences with these new curricula to inform decision-making and increase implementation. This study examined how three different models of digital technologies professional learning and development influenced 48 New Zealand primary and intermediate school teachers’ efficacy beliefs and implementation with reference to both the Hangarau Matihiko and Digital Technologies curricula. Efficacy beliefs were chosen as a study variable due to the ability to categorise these into value-beliefs (the importance a teacher places on DT), self-efficacy beliefs (a teachers belief in their own DT capabilities) and teacher-efficacy beliefs (teachers beliefs in their capacity to teach DT). The mixed methods investigation compared data from a survey at three points in time (pre-, post- and 6 months-post- completing the training) and involved descriptive and statistical analysis of the quantitative dataset and thematic analysis of the qualitative dataset prior to triangulation. Statistically significant interactions between time (pre- and 6 months post-training) and confidence, implementation and efficacy beliefs were found, yet no statistical significance was found between the three different training models. A range of connections between implementation and each of the efficacy beliefs were discovered highlighting the multifaced impact of training on teachers’ implementation. The research found teachers required additional support both in adopting relevant pedagogies and in planning to implement these curricula. School environments were shown to influence efficacy beliefs, and lack of time both in terms of non-contact time and time in the classroom was found to be a pressing concern for participants. Using the contributions of the study as evidence, the research closes with a discussion of further areas of research needed as well as the support needed to raise teachers' implementation of these important curricula.
Develop Biological Denitrification Strategies to Reduce Greenhouse Gas Emissions under Multi-extreme Conditions
(ResearchSpace@Auckland, 2023) Chen, Jiazhe; Zhuang, Wei-Qin
As a prevalent nitrogen removal process, heterotrophic denitrification plays a crucial role in wastewater treatment plants. Through the process of denitrification, heterotrophic microbes utilize organic carbon to reduce nitrates and nitrites, ultimately releasing nitrogen gas (N2). Notably, incomplete denitrification can also yield nitrous oxide (N2O), a potent greenhouse gas, which is of particular concern due to its substantial impact on the environment. Given the environmental implications and New Zealand's commitment to achieving carbon neutrality by 2050, there is a pressing need for innovative denitrification methods. This thesis explores hydrogenotrophic denitrification, which employs inorganic carbon (i.e., carbon dioxide) and hydrogen gas to reduce nitrates, as an alternative to conventional heterotrophic denitrification. Notwithstanding its environmental merits, hydrogenotrophic denitrification generally presents lower nitrate removal efficacy compared to its heterotrophic counterpart. To address this, the research proposes a mixotrophic denitrification strategy that utilizes both inorganic and organic carbon, potentially offering lower nitrous oxide emissions with almost comparable nitrate removal efficiency. Because New Zealand's stringent biosafety laws hinder the utilization of foreign microbial species in large-scale applications, the study necessitated the use of indigenous inoculants to enrich both heterotrophic and autotrophic denitrification communities. These communities demonstrated the efficacy of using New Zealand indigenous microorganisms to maintain wastewater effluent quality standards and reduce greenhouse gas emissions. In addition, this study investigated wastewater denitrification under challenging conditions that resemble industrial wastewater sources. Specifically, high salinity (3.5%) and extreme multi-conditions (1.5% salinity and pH 5.5) that mirror seafood processing wastewater, scenarios induced by seawater intrusion, and the particularities of dairy wastewater in New Zealand. The latter is characterized by its demanding management due to notable pH fluctuations and high BOD loads (Daly, 2016). The former, seawater intrusion, is a pervasive issue that compromises freshwater aquifers, particularly in coastal regions, by introducing saltwater into freshwater zones, which can be exacerbated by factors like groundwater pumping and reduced freshwater flow (Barlow & Reichard, 2010). The seed inoculum, derived from the heterotrophic activated sludge of a wastewater treatment plant, was subjected to these extreme conditions, aiming to replicate and understand the potential implications and management strategies of such environments on wastewater treatment processes. The results demonstrated that the specific nitrate removal rate of heterotrophic denitrification under high salinity was 55 mg NO3-N /g MLSS/d, while under multi-extreme condutions, it was 58 mg NO3-N /g MLSS/d. Concurrently, the rate for hydrogenotrophic denitrification under multi-extreme conditions was observed to be 50 mg NO3-N /g MLSS/d. Furthermore, the data on N₂O concentration revealed a significant decline in hydrogenotrophic denitrification compared to heterotrophic conditions. While the heterotrophic system accumulated 0.98% N₂O in the reactor headspace, no N₂O accumulation was observed in the hydrogenotrophic system. Lastly, mixotrophic denitrification exhibited also reduced nitrous oxide accumulation potentially due to the relative ratio of nitric oxide reductase (Nir) to nitrous oxide reductase (Nos), a hypothesis corroborated by proteomic results. Future research directions include exploring gene regulation mechanisms, expression patterns, shifts between autotrophic and heterotrophic denitrification, and the interplay between hydrogenotrophic and heterotrophic denitrification. To identify the key species of the hydrogenotrophic denitrification enrichmet cultures, a pure culture of a Paracoccus species was isolated from the hydrogenotrophic mixed culture. This particular species exhibited a notable tendency to accumulate nitrite under the imposed multi-extreme conditions in this study. Furthermore, the genomic data supported its capability to synthesize Polyhydroxyalkanoates (PHA) and carboxysomes, enhancing its adaptability to environments characterized by limited carbon availability.
Transverse Liquid Composite Moulding Processes for Advanced Composite Material Manufacturing
(ResearchSpace@Auckland, 2024) Lee, Jeeeun; Kelly, Piaras; Duhovic, Miro; Allen, Tom
Many recent developments in Liquid Composite Moulding (LCM) processes have focused on the benefits of transverse (through-thickness) flow to reduce the impregnation path. The fluid flows in the same direction as the preform deformation, so the complex coupling between flow and deformation, known as hydrodynamic compaction, leads to a non-uniform fibre volume fraction distribution in the wet region. This type of behaviour occurs in a wide variety of manufacturing processes and experimental characterisation techniques involving through-thickness impregnation of fibrous reinforcements, such as Compression Resin Transfer Moulding (CRTM), permeability measurement systems and Resin Transfer Pressing (RTP). A comprehensive numerical model is developed based on the fully coupled, non-linear, time-dependent governing equations to predict the homogenisation time, fluid pressure, effective stress, and fibre volume fraction distribution for a general transverse resin impregnation/compression process. Detailed analysis of transverse permeability measurement systems is carried out, and the true permeability relation is extracted from measurements of the apparent permeability by removing the effects of hydrodynamic compaction. Commonly made assumptions, such as quasi-steady flow and uniform through-thickness deformation, introduce significant errors in manufacturing process predictions. It is shown that the fully coupled governing equations must be used when modelling processes involving direct contact between the tool and preform. Optimal resin-to-reinforcement ratios are determined using the developed model to minimise the processing time and resin wastage. A Physics-Informed Neural Network (PINN) is used to simulate the simple case of transverse compression of a saturated preform, with solution times of less than 1 ms, to explore the applicability of machine learning to composite manufacturing simulations. Furthermore, viscoelastic compaction models are implemented into simulations to capture the interaction between hydrodynamic compaction behaviour and viscoelastic effects. When a saturated stack is held at a constant thickness, the total load decreases due to the equilibration of fluid pressure and viscoelastic stress relaxation. Viscoelastic simulations were found to significantly improve the accuracy of predictions compared to elastic models. The comprehensive analysis given in this thesis provides an in-depth understanding of various aspects of the flow and deformation in transverse impregnation/compression processes.
Sentiment and Emotion Analysis of Social Media Text: Multilingual Fine-Grained Low-Resource Personalised and Explainable AI Approaches
(ResearchSpace@Auckland, 2024) Li, Yuming; Chan, Johnny; Sundaram, David; Peko, Gabrielle
The surge in the global influence of social media and the complexities of digital communication have underscored the significance of sentiment and emotion analysis within the information systems field. This thesis addresses the nuanced challenges of extending sentiments and emotions across the dynamic landscape of social media, where traditional methods falter against the backdrop of growing data volume and complexity. It identifies five primary research challenges: conducting sentiment and emotion analysis in multilingual contexts; identifying finer-grained and mixed emotions; ensuring analysis efficiency in low-resource settings; personalising analysis for different user groups; and explaining outcomes from deep learning-based models. In pursuit of addressing these challenges, the thesis unfolds across five published papers. The first paper, published in PACIS (2021), introduces a multilingual sentiment analysis framework that leverages shared multilingual word embeddings and pre-trained language models, aiming to standardise sentiment analysis across languages. The second paper, published in Data & Knowledge Engineering (2023), proposes an advanced approach for fine-grained and mixed emotion analysis using a graph-structured storage for improved visualisation and categorisation. The third paper, presented at HICSS (2024), introduces a framework for hateful emotion recognition using knowledge distillation and data augmentation tailored for low-resource scenarios. The fourth paper, presented at AMCIS (2022), develops a system for personalised analysis of harmful emotions based on user profiles. The fifth paper, published in Decision Support Systems (2024), addresses the challenge of interpreting the outputs of deep learning models in sentiment and emotion analysis, offering an explainability framework grounded in emotion theory. In terms of its overarching methodological approach, this thesis is firmly anchored in the design science research methodology. Through the iterative cycles of design, implementation, and evaluation intrinsic to design science research, this thesis not only bridges theoretical gaps but also produces empirically validated artefacts that enhance the capabilities of sentiment and emotion analysis in handling multilingual texts, fine-grained emotions, resource constraints, personalisation, and explainability. The advancements contribute significantly to the academic realm and offer cutting-edge solutions for industry practices, including empathetic AI systems and enriched decision-making processes, marking a crucial evolution towards globally integrated, human-centric social media analysis.
Measuring Fracture Healing with Smart Orthopaedic Implants
(ResearchSpace@Auckland, 2023) Malcolm, Angus; McCormick, Daniel; Budgett, David; Paul, Professor
There is a lack of methods for quantitatively assessing fracture healing. Current practice relies on a subjective mix of observation and indistinct imaging studies, yielding uncertainty and variability in fracture assessment, resulting in conservative clinical management, lost productivity, and increased costs which are exacerbated by delayed union and non-union cases. Previous researchers have shown that fracture union can be tracked by measuring changing strain for known applied loading on fixation plates as healing progresses. However, no devices exist for clinical use. This thesis aims to advance fracture healing assessment using strain-based measurements. Specific areas contributing to a functional implantable device include strain measurement technology, wireless power and communications, and a physical implant construction consistent with obtaining regulatory approval. A proposed solution suggests utilizing piezoresistive strain gauges in combination with a new wireless power technique: conductive transcutaneous energy transfer (CTET). A dual pathway is suggested due to an initial lack of knowledge of the optimal solution, which explores embedding this technology within 1/3 tubular plates as well as cortical lag screws. The electrical power, voltage delivery, safety, and electromechanical design requirements are assessed via an experimentally validated in-silico modelling approach showing 3.9mW power delivery to the smart screw and 17.5mW to the smart plate, providing evidence that these devices can be powered. The strain outputs and sensor requirements are assessed via an in-silico approach, with finite element modelling showing as much as 19μƐ divergence on the plate and 25μƐ on the screw between healing states. The results are confirmed experimentally in cadaver studies showing 22μƐ of strain discrepancy on the plate between fracture states. The experimental data resulted in a final prototype for an instrumented lateral malleolar smart plate implant powered by CTET that communicates wirelessly with an inductive phase shift keying mechanism. The design was constructed and tested in a benchtop saline leg phantom, proving feasibility by predicting fracture healing in a wirelessly powered, full-metallic bodied lateral malleolar smart plate implant, that incorporated miniaturized electronics exhibiting a preliminary 13mm x 6mm x 3.5mm flexible PCB footprint with capacity for further miniaturization and a design consistent with gaining regulatory approval.
Synthetic Studies Towards Purpurolide A
(ResearchSpace@Auckland, 2024-12) Geurts, Alisha; Brimble, Margaret; Furkert, Dan
Purpurolide A (37) is a sesquiterpene lactone isolated from the endophytic fungus Penicillium purpurogenum. Purpurolide A (37) exhibits anti-obesity activity, by the inhibition of pancreatic lipases, with an IC50 value of 2.83 ± 0.52 μM. Thought to be derived from the bisabolyl cation (10), purpurolide A (37) is the first natural product to arise from an oxidative cleavage of the cyclohexene ring. The novel purported biosynthesis, the resulting unprecedented 5/5/5 spirocyclic system, and the promise for an anti-obesity treatment makes purpurolide A (37) an appealing target for total synthesis. This thesis herein describes the evolution of a synthetic strategy towards the asymmetric synthesis of purpurolide A (37). The initial synthetic strategy focused on the construction of the framework from an aldol reaction between 4-methyl-2-furfural (128) and cyclopentene ester 129. Access to ester 155 was first secured from (S)-carvone by a four-step sequence, using a Favorskii rearrangement of 159 to effect the ring contraction. Unfortunately, aldol reaction of 155 with furfural (154) as a model study provided trace quantities of the conjugate addition product. In an effort to prevent the conjugate addition, the aldol reaction was attempted with cyclopentane ester 156 and aldehyde 172, however no product was observed. The aldol reaction was then attempted utilising β-ketoester 201 as an improved nucleophile for the aldol reaction, however this was also unsuccessful. A second revised overall synthetic strategy reviewed several approaches to access cyclopentene 269. The second strategy focused on use of a vinylcyclopropane-cyclopentene (VCP-CP) rearrangement of 320 or 329 to afford 270 or 330, respectively. Cyclopropanes 320 and 329 were planned to be accessed via a sulfur ylide-mediated cyclopropanation of diene 273, that unfortunately could not be effected, despite using a range of ylides. Alternatively, vinylcyclopropane (±)-375 was prepared by a six-step sequence, employing a Knoevenagel condensation which ultimately only gave minute quantities of (±)-375. The subsequent VCPCP rearrangement unfortunately did not afford cyclopentene (±)-400. Next, access to cyclopentene 269 from a bicycle such as 496 was investigated. The Danheiser annulation between allene 404 and lactones 426 or 427 proved unsuccessful. Alternatively, Michael addition of aldehyde 446 with 275 afforded malonate 483, which successfully underwent a Conia-ene reaction to give key intermediate bicycle (±)-489. The synthetic approaches investigated herein have provided important insight for future studies towards the total synthesis of purpurolide A (37). The current synthetic strategy has afforded the cyclopentane moiety with all the required functionalities for later elaboration to purpurolide A (37), which remains an ongoing synthetic pursuit of this research group.
Wind tunnel test and analysis method for segmented-rigid models of super-tall high aspect-ratio buildings
(ResearchSpace@Auckland, 2024) Jiang, Zhenhua; Flay, Richard; Ma, Quincy; Li, Yin Fai
Traditional wind tunnel testing technologies face challenges from super-slender buildings and skyscrapers with large aspect ratios. These buildings are difficult to fit into typical boundary layer wind tunnels with an acceptable geometric scale. Limited research has been conducted into overcoming this issue by combining wind tunnel test results of multiple building sections. This study proposes a more precise and comprehensive sectional model wind tunnel testing and analysing method to address this gap. The high-frequency pressure integration method is selected for its ability to accurately capture sectional model surface pressure distributions. Two super-slender building models—a 2D pressure-tapped circular cylinder and a 2D pressure-tapped rectangular prism—are used to investigate the method. A case study demonstrates the application of the proposed method to an actual commercial building project in an urban setting. Experimental testing includes subjecting the 2D models to uniform flow conditions with varying turbulence intensities and Reynolds numbers. The correlation mechanisms of drag and lift with different separation distances for these two models in diverse flow conditions have been thoroughly investigated. Shear flow conditions, simulating those in skyscraper cities like Shanghai, are also examined. The study introduces a method to synthesise the time series of wind loads along the building height, ensuring the synthetic signal embodies the characteristics of simultaneous sampling. The mathematical derivation of the synthetic wind load (SWL) method is validated through comparisons of wind load cross-spectra and dynamic responses of a hypothetical chimney. Furthermore, the study investigates the effects of structural mode shape, model split position, the number of horizontal pressure-tap lines, and the number of tested sectional models on the SWL method for both representative Reynolds number-dependent and independent models. The SWL method's feasibility is demonstrated through combining sectional wind loads from different tests on a 3D model. In summary, this research pioneers the development and validation of the SWL method, offering a robust solution for wind tunnel testing of very high aspect ratio buildings. The method reproduces the deflection prediction of a high aspect ratio building within 10% accuracy compared to the prediction using a full building model, addressing challenges in segmented-rigid model testing.
Effects of Task Complexity and Individual Differences on EFL Writing Performance
(ResearchSpace@Auckland, 2024) Zhang, Yujie; Zhang, Lawrence Jun; Ormond, Barbara
Grounded in the English-as-a-foreign-language (EFL) learning context, this research investigates the effects of writing task complexity, the predictive power of cognitive and psychological individual differences, and the moderating effect of task complexity on argumentative writing performance through the lens of multidimensional metrics. To differentiate writing tasks demanding different levels of cognitive resources, number of elements in two writing tasks was manipulated. To investigate learner individual differences, this research examines foreign language aptitude, Willingness to Communicate (WTC), and Foreign Language Enjoyment (FLE) for their claimed significance to foreign language learning. Writing performance was measured by syntactic complexity, lexical complexity, phraseological complexity, accuracy, and writing quality. This research adopted a mixed-methods design to combine the merits of both quantitative and qualitative approaches. In total, one pilot study and three main studies were conducted to achieve the research aims. First, 522 post-pubertal EFL learners from China were invited to participate in Study One in three steps to develop and validate an L2 writing WTC scale. Then, the pilot study recruited 20 participants to pilot the procedures, instruments, interviews, coding, and analysing. Finally, 166 intermediate post-pubertal EFL learners from China were recruited voluntarily to participate in Studies Two and Three, in which they were guided to join two cross-sectional quasi-experiments. The results indicate that (1) the newly developed and validated L2 writing WTC scale has a five-factor underlying structure; (2) an increase in number of elements leads to lower accuracy, higher lexical complexity, and higher phraseological sophistication in writing; (3) increasing number of elements leads to longer length of syntactic units, more complex noun-phrase modification, and more complex post-modification; (4) foreign language aptitude cannot predict writing quality; (5) L2 writing WTC, FLE, and multiple indices of complexity and accuracy can predict writing quality; (6) task complexity does not moderate the relationships between the three individual difference variables and writing quality. The conclusions aid previous research in determining the effects of task complexity and learner individual differences on EFL argumentative writing performance. Implications, limitations, and recommendations are discussed for EFL writing pedagogy and assessment.
Bifurcations of Heterodimensional Cycles
(ResearchSpace@Auckland, 2024) Wong, Nelson; Krauskopf, Bernd; Osinga, Hinke M
Mathematical models of real-world systems can exhibit highly complicated phenomena that organize observed behavior or dynamics. In this thesis, we study one such phenomenon known as a heterodimensional cycle. A heterodimensional cycle consists of two saddle periodic orbits that have unstable manifolds of di erent dimensions|together with connecting orbits from one periodic orbit to the other, and vice versa. A system with a heterodimensional cycle is structurally unstable, meaning its dynamics is sensitive to arbitrarily small parameter changes. Nevertheless, the existence of a heterodimensional cycle can be a \robust" phenomenon, in which case it is known to generate highly complex dynamics, also called wild chaos. Heterodimensional cycles are complicated structures, and all the known examples have been constructed abstractly without a realistic application in mind. However, Zhang, Kirk and Krauskopf (2012) found and computed a codimension-one heterodimensional cycle of the Atri model, which is an explicit, four-dimensional vector eld model of intracellular calcium oscillations. This forms our starting point, and we use advanced numerical methods, including Lin's method, to compute new and more complicated heterodimensional cycles. With the continuation software AUTO, we explore a two-parameter region of the Atri model where heterodimensional cycles are found, and we show how the loci (curves) of these new heterodimensional cycles t together in an overall bifurcation structure, which also involves local and global bifurcations of equilibria and periodic orbits. Speci cally, we discover that the heterodimensional cycle found by Zhang et al. undergoes a sequence of geometrical transformations, which implies new dynamical phenomena, such as a strong homoclinic orbit. Furthermore, we nd two novel codimension-two bifurcations: a heterodimensional cycle at a period-doubling bifurcation, and a resonant heterodimensional cycle. The former bifurcation generates new heterodimensional cycles involving the period-doubled orbit, and the latter gives rise to in nitely many families of codimension-one homoclinic tangencies. Finally, we relate our research back to the theory of structural stability and discuss the existence of a robust heterodimensional cycle of the Atri model. We show how heterodimensional cycles can be abundant in the limit of a period-doubling cascade. Moreover, we identify a codimension-two organizing center (called a 3DL bifurcation), whose existence suggests that heterodimensional cycles can exist throughout a large two-parameter region.
Explainable and Automated Scientific Fact-Checking with Neural Networks
(ResearchSpace@Auckland, 2024) Tan, Neset Ozkan; Witbrock, Michael
Fact-checking plays a crucial role in combating misinformation, especially in scientific domains where the stakes are high, and the consequences of false claims can be severe. As experienced during the COVID-19 pandemic, unfaithful claim verifications underscored the need for robust fact-checking systems. This thesis addresses the challenges of verifying claims in the scientific literature using deep learning-based computational linguistics techniques, focusing on faithfully incorporating knowledge from a vast existing literature and addressing the scarcity of appropriate training datasets for robust fact-checking systems. Language models in Natural Language Processing (NLP) are computational models designed to understand language. These models are trained on massive amounts of data to learn statistical patterns and relationships within language. They work by predicting words, sub-words, or characters in a sequence, taking into account the context provided by preceding sequence elements. In recent years, the dominant architecture for language models has been transformerbased models, which can be seen as a milestone in NLP research due to their significant improvements for downstream tasks. However, these models also have limitations, including modelling challenges and dataset challenges. Modelling challenges refer to the limitations and complexities faced by computational models, particularly those based on deep learning and natural language processing techniques. These challenges include difficulties in accurately capturing nuanced arguments, potential generation of false information (‘hallucinations’), constraints on handling lengthy input texts, and limitations in reasoning capabilities. On the other hand, dataset challenges stem from the scarcity of appropriate training data essential for building robust fact-checking systems. The specialized nature of scientific content, coupled with the need for accurate annotations, creates an expertise bottleneck. In this context, developing large-scale, domain-specific datasets becomes crucial to train models effectively. These challenges collectively necessitate innovative methodologies to enhance the capabilities of computational models for accurate scientific fact-checking, addressing both their inherent modelling intricacies and the scarcity of specialised training data. This thesis proposes methods to advance the field of scientific fact-checking by developing approaches that enhance the capabilities of transformer-based language models. In addressing modelling challenges, we present a novel methodology that leverages multiple viewpoints from scientific literature, allowing the assessment of contradictory arguments and implicit assumptions. Our proposed inference method enhances reasoning by distilling information from diverse, relevant scientific abstracts. This approach yields a verdict label that can be weighted based on the article’s reputation and an explanation that can be traced back to sources to avoid hallucinations. Our findings demonstrate that human evaluators perceive our explanation to be significantly superior to off-the-shelf models, enabling faithful tracing of evidence back to its original sources. For the problem of handling lengthy input texts, we introduce a method that utilises the layer-based attention scores of transformers to filter input length. This approach proves efficient for scientific paper topic classification and verdict label prediction tasks, which is critical for effective fact-checking. Regarding dataset challenges, we address the expertise bottleneck limiting the availability of appropriate training data for scientific fact-checking. We propose a pipeline, Multi2Claim, for automatically converting multiple-choice questions into fact-checking data. Using this pipeline, we create two large-scale datasets: Med-Fact for the medical domain and Gsci-Fact for general science. These datasets represent significant contributions as they are among the first large-scale scientific fact-checking datasets. Baseline models developed using each dataset show promising results, with performance improvements of up to 26% on existing fact-checking datasets such as SciFact, HEALTHVER, COVID-Fact, and CLIMATE-FEVER. In conclusion, the proposed methodologies in this thesis contribute to the advancement of scientific fact-checking by addressing modelling intricacies and dataset challenges, offering a promising step towards more accurate and effective systems to combat misinformation in scientific domains.

Browse

Recent Submissions