Sydney eScholarship
Not a member yet
31878 research outputs found
Sort by
Mechanistic Modelling and Parameter Optimisation of Polymer Pyrolysis
No full textThe ever-growing crisis of solid plastic waste (SPW) accumulation in the natural environment has prompted significant industrial and academic enterprise. Pyrolysis has emerged as a principal recycling technology in this fight against SPW accumulation, demonstrating capacity to liberate high value chemical feedstocks from plastic wastes. Despite its conceptual maturation, there exists a sparsity of kinetic data and models describing the complex degradation networks involved in polymer pyrolysis. This lacuna is attributed to both the difficulties in constructing models that appropriately consider the extreme size of common polymers, and the functionally infinite number of ways these polymers decompose and rearrange via free-radical reactions. As a result, industrial-scale SPW pyrolysis is hindered by the costly and time-intensive construction of pilot plants to determine viability and optimise valuable product yield.
This thesis successfully produced highly detailed mechanistic models describing polystyrene and polyethylene pyrolysis. These stochastic models were validated against experimental yield data obtained from literature using a novel parameter optimisation methodology that utilised an artificial neural network to efficiently traverse a high dimensional parametric response surface. Parameter optimisation was enabled due to novel developments in the kinetic Monte Carlo (kMC) stochastic modelling framework, specifically the reduction of necessary simulated volume and the hybridisation of the kMC model with a Markov Chain model. This resulted in between a 4-7 order of magnitude reduction in simulation time, allowing for the rapid simulation of the decomposition of polymers with a high degree of polymerisation, while maintaining full accounting of all chain information. These models provided insight into outstanding questions in the literature regarding the role of competing highly degenerate reaction pathways on low molecular weight product yield
The Intersection of Branding and Architecture: The Brandscape of Joy City Mall in China
No full textThis study explores the complex interplay between branding and architecture. I engage the term brandscape to demarcate this zone of interplay and consider brandscapes to be the point at which branding and architecture become more blended (and less discernible) in contemporary consumer cultures. I extend the understanding of the concept of brandscape through a detailed study of retail architecture in the Chinese Joy City mall series. The focus on Chinese shopping malls relates to the unique position of mall architecture in China, manifesting as a brand in its own right and responding to the fast rise of consumerism. The branded Chinese shopping mall series Joy City exemplifies the conspiration of branding and architecture, characterised by identity-formation, semiotics and phantasmagorical features.
Given the nature of this focus, the research involves architectural inquiry bolstered by a critical methodology, encompassing theorised and multi-disciplinary studies; research methods include literature surveys, case studies and visual studies.The working definition and the existing scholarship related to “brandscape” are extended by exploring particular instances of the complex interplay of branding and architecture. The particularity of the relation is framed through interdisciplinary insights, socio-economic and cultural studies. By identifying the interwoven characteristics of branding and architecture in the brandscape of the Chinese Joy City malls, I make an original contribution to the realms of architectural theory related to retail cultures and concerning the semi-autonomous nature of architecture itself. By emphasising the deepening interplay and particularities of the relation between branding and architecture, the study significantly broadens architects’ understanding of the imperative and characteristics of brandscapes and, in turn, demonstrates the significance of both theoretical and practical implications
Deep-Learning-Based 3D Medical Image Segmentation and Registration
No full textUnderstanding the importance of image segmentation and registration in clinical diagnosis and treatment is critical for advancing precision medicine. Medical image analysis involves complex data processing tasks, particularly in scenarios with limited annotated data and multi-modality feature inconsistencies. Recent deep learning-based methods have provided breakthroughs in image segmentation and registration, especially in modalities like MRI and CT. However, challenges such as one-shot learning, incremental learning, and efficient deformable registration still persist. This thesis introduces a series of innovative deep learning frameworks, focusing on three one-shot segmentation settings and three Segment Anything Model (SAM) -assisted registration methods, to address the issues of annotation scarcity and feature misalignment in medical imaging. For segmentation, we develop three methods for one-shot, class-incremental, and novel-class tract segmentation. These incorporate uncertainty estimation, knowledge distillation, and voxel-level contrastive learning to enhance performance and generalization in one-shot settings. For registration, we propose three SAM-guided frameworks that leverage segmentation masks and anatomical prompts, as well as attention mechanisms, prototype learning, and contour-aware losses. These approaches achieve state-of-the-art accuracy in unsupervised deformable registration. Overall, our methods yield substantial improvements in segmentation and registration accuracy across diverse datasets and demonstrate strong potential for clinical application
Novel Integrated Approaches to Greenhouse Gas Sensing
No full textThe effects of anthropogenic climate change pose significant risks to the environment and human
property, with the most serious being loss of life. Addressing this issue requires accurate monitoring
of all greenhouse gases, including trace gases like methane and nitrous oxide, which are difficult to
detect at low concentrations. Studies suggest human activities release more of these gases than
previously estimated, creating an urgent need for improved, lightweight, portable detectors that can
cover large areas while maintaining high selectivity and sensitivity.
Absorption spectroscopy enables precise gas detection, with modulation spectroscopy offering
sensitive, low-cost solutions. This thesis explores two applications of modulation spectroscopy for
"fingerprint" gas detection—highly sensitive and selective molecular species identification and
monitoring. The first focuses on cross-correlation using a complex aperiodic fiber Bragg grating
(FBG) that mimics acetylene’s P-band absorption features. We demonstrate, for the first time, that
complex FBGs can selectively identify different gas concentrations while maintaining high selectivity.
Simulated data reinforce these findings, highlighting the potential for selective gas detection.
The second part investigates wavelength modulation spectroscopy through quartz-enhanced
photoacoustic spectroscopy (QEPAS). QEPAS generates sound waves from modulated light
interacting with the target gas, with a quartz tuning fork (QTF) converting these waves into electrical
signals. This thesis presents dedicated voltage amplifiers to enhance the QTF’s piezoelectric signal
and two custom-built QEPAS units using 3D-printed materials for methane detection. These
chambers incorporate chemical annealing for improved material finish. A direct amplifier comparison
demonstrates methane detection with a minimum detection limit of 6.4 ppm, showcasing a practical,
customizable approach for sensitive gas detection
Stuck in place: the cognitive and neural bases of inflexibility in ageing and dementia
No full textFlexible thinking and behaviour are critical for navigating the uncertainties of daily life, enabling individuals to adapt, problem-solve, and respond to dynamic challenges. The capacity for flexible thinking and behaviour is diminished in ageing and dementia; however, the neurocognitive mechanisms underlying these changes remain poorly delineated. This thesis explores changes in flexibility in healthy ageing and dementia syndromes across cognitive, behavioural, and neural levels, leveraging machine learning approaches to predict rigid behaviours in dementia. Across four experimental chapters, the thesis aims to provide a comprehensive understanding of the multidimensional nature of inflexibility in healthy and pathological ageing, with a view to informing
strategies to manage these symptoms.
Chapter 2 examines cognitive flexibility in healthy ageing and reveals associations with reduced neural flexibility within specific functional brain networks. Chapter 3 investigates cognitive and behavioural inflexibility across dementia syndromes, identifying marked cognitive rigidity in semantic dementia (SD) and pronounced behavioural rigidity in behavioural-variant frontotemporal dementia (bvFTD). Chapter 4 applies connectome-based predictive modelling to assess whether structural connectivity in the frontoparietal, salience, default mode, and limbic networks predicts behavioural rigidity in dementia. Chapter 5 explores alterations in neurotransmitter systems in FTD using a novel approach, demonstrating that reduced serotonin transporter density is associated with behavioural rigidity in bvFTD.
Collectively, the work presented in this thesis offers an integrated framework for understanding inflexible thoughts and behaviours in ageing and dementia. By combining cognitive, neural, and behavioural insights, it informs predictive and management strategies to improve patient care and quality of life
The Odyssey of an Ileostomy: A Journey of Challenges and Resilience
No full textUndergoing an ileostomy is a major life transition requiring medical, psychological, and social adjustment. This thesis examines the multifaceted experience of patients undergoing ileostomy formation and reversal, focusing on postoperative complications—particularly incisional hernia and renal impairment—and the often-overlooked psychosocial impact on patients and caregivers. Using Homer’s Odyssey as a framework, it explores resilience and adaptation throughout this journey.
A mixed-methods approach was employed. A retrospective cohort of 224 patients who underwent ileostomy reversal was analysed to identify predictors of complications, with emphasis on incisional hernia. A systematic review following PRISMA guidelines synthesised evidence on renal impairment after ileostomy in rectal cancer surgery. Qualitative data were obtained through structured interviews and validated instruments, including the Stoma Quality of Life (SQOL) questionnaire and Caregiver Burden Scale (CBS), to assess long-term psychosocial outcomes.
Results showed that while surgical outcomes have improved, significant complications persist. Incisional hernias occurred in 5% of cases, with obesity a major risk factor. Renal dysfunction, both acute and chronic, was linked to fluid imbalance, comorbidities, and chemotherapy. Patients often reported greater life satisfaction after ileostomy, though this came with increased caregiver demands, leading to moderate stress.
This thesis demonstrates that the ileostomy journey involves ongoing clinical risks and psychosocial consequences. Surgical advances have improved care, but hernias and renal complications remain, underscoring the need for better risk assessment, optimisation, and follow-up. The interdependence of patient outcomes and caregiver burden highlights the importance of multidisciplinary, patient-centred care and supportive policies. Like Homer’s Odyssey, the ileostomy journey is one of trial, resilience, and growth
Understanding and Treating Family Factors in Child and Adolescent Obsessive Compulsive Disorder
No full textObsessive compulsive disorder (OCD) is a debilitating mental health condition affecting approximately 1-4% of children and adolescents. Despite the effectiveness of gold-standard treatment, cognitive behaviour therapy (CBT) that involves exposure and response prevention, a significant number of young people show limited treatment response. Emerging evidence suggests that family factors may play a role in maintaining OCD symptoms and affecting treatment outcomes. In support of this, recent family-enhanced CBT interventions have shown promising results. However, the specific family factors that contribute to improved treatment outcomes, and their relative importance for OCD symptom change, remain poorly understood. Further investigation of family factors in child and adolescent OCD is warranted to inform theory and improve treatment. As such, this thesis aims to investigate the relevance and role of a range of family factors in child and adolescent OCD, its treatment, and OCD symptom change. The current thesis presents four studies to address these aims, a systematic-review and meta-analysis, and three empirical studies, specifically a treatment trial and two observational studies. Overall, the current thesis presents novel findings that highlight the importance of family factors in child and adolescent OCD and its treatment, including for OCD symptom change. Findings suggest that family factors characterising paediatric OCD families may be contextual, and demonstrate the importance of addressing these family factors in OCD treatment to improve treatment outcomes. Findings support the integration of a range of family factors into models of OCD maintenance and the direct targeting of these family factors during treatment to optimise outcomes for children and adolescents with OCD
The proteomic, immune, and antigenic profiling of colorectal cancer for the design of CAR-T cell therapies
No full textColorectal cancer (CRC) is forecast to become the leading cause of cancer and cancer-related deaths by 2040. Although it predominates in older patients, there has been a rise in young onset disease. Efforts have been made to improve prognostication methods; however, none have been able to standalone independent of the AJCC classification system. Chimeric antigen receptor T cell (CAR-T cell) therapies have significantly improved survival in haematological malignancies, and if rationally designed have the potential to improve outcomes for solid organ cancers including CRC. The aim of this thesis was to design a signature that predicts relapse in CRC and to characterise the antigenic, immune and proteomic landscape of CRC for the purpose of rational CAR-T cell design.
Utilising a large cohort of early-stage CRC between 2009-2020 from a tertiary referral centre, 495 patients treated for stage II and III CRC, and 145 patients treated for locally advanced rectal cancer (LARC) were identified. High-risk groups were identified including Tumour stage 4 (T4), mucinous histology, young onset rectal cancer (yRC), MSI-H LARC and mucinous histology.
A PILOT project using SWATH proteomics and Imaging Mass Cytometry identified immunosuppressive pathways and proteins in CRC for potential targeting, with CAR-T cells against EPHA2 and CEACAM5 offering therapeutic options. A signature combining EPHA2, CEACAM5 and GDF-15 for targeting with next generation CAR-T cell therapies could prove effective and improve durable activity, whilst eliminating life-threatening toxicities.
A 19-protein signature was discovered and was prognostic for Relapse Free Survival (hazard ratio (HR) 2.7, 95%CI: [2.24 – 3.18]; p < 0.0001) and overall survival (HR 1.6, 95%CI [1.40, 1.76). Significant heterogeneity in pathways and proteins was seen in this cohort of stage II and III CRC, paving the way forward for novel CAR-T cell design and rational clinical trial designs for the high-risk groups
Enhancing Autonomous Science: Efficient Robotic Informative Sampling and Environment Hypothesis Testing
No full textRobotic systems can be crucial in automating information gathering tasks in natural environments. However, sample collection can be expensive and damaging to the environment, making data sparse and resource-to-information efficiency of utmost priority. Moreover, effective data collection as well as appropriate models can fundamentally improve scientific understanding. This thesis uses hypothesized models to accelerate information gathering while simultaneously validating those models.
First, a Multi-Quantity Gaussian Process is introduced that links multivariate Gaussian distributions and linear models, allowing to derive parametric models between multiple quantities from sparse, non-collocated data. The technique also yields coefficients of determination, providing goodness-of-fit measures for those hypothesized models. The learned inter-quantity relationships adapt the model's predictions such that good hypotheses are utilized and poor ones are ignored.
Second, an objective function is employed that balances exploration and exploitation to reduce both the mean and maximum model errors. New sample locations are optimally chosen based on global and local uncertainties through the Gaussian Process variance and a nearest-sample prediction difference. Sample spread is maintained while actively choosing regions that test the proposed quantity relationships.
Third, the informative objective function is scaled by the travel-distance to reduce travel costs and maximize energy-to-information efficiency. This naturally self-adjusts to different search areas and quantity variations without any manual parameter tuning. Furthermore, computation is saved by dropping poor hypotheses from consideration in future learning iterations.
Finally, the full algorithm is showcased directing an autonomous ground vehicle in a farming application characterizing pasture abundance, demonstrating collection and mapping of plant heights and evaluation of correlation with prior elevation data
The effect of piezocision and mechanical vibration on orthodontically induced root resorption of first premolars following the application of orthodontic forces
No full textIntroduction
The aim was to compare the effect of piezocision and mechanical vibration on orthodontically induced inflammatory root resorption (OIIRR) of human first premolars following application of orthodontic forces.
Methodology
Forty-one orthodontic patients (aged 13–18) requiring first premolar extractions were treated with a 225g buccal force on both maxillary first premolars for four weeks. Premolars were divided into control (C), piezocision (P), vibration (V), or a combination (PV) through randomization. A total of 82 premolars were extracted, and OIIRR crater volumes were measured using X-ray microtomography and Fiji software to compare OIIRR across the groups.
Results
OIIRR crater volumes were analyzed using a cube root transformation (crt) by treatment (C, P, V, PV) and location. While the mean crt for PV (0.75mm) was slightly higher than for C (0.65mm), V (0.71mm), and P (0.65mm), there were no statistically significant differences among treatments (p=0.33). The location analysis showed statistically significant differences (p<0.001), with the highest OIIRR on the buccal middle (0.44mm) and the lowest on the palatal cervical (0.07mm).
Conclusions
The study found no significant differences in mean crtRR volumes between treatments (C, P, V, PV), suggesting piezocision and vibration may not impact OIIRR following the application of heavy orthodontic forces. OIIRR levels varied by location, with the highest on the buccal middle root surface