70040 research outputs found
Sort by
Breaking the concrete wall: The untapped potential of technology to increase the participation of women in the construction industry
Full text linkMultiscale assessment of fracture, fluid flow and geothermal activity at the Heyuan Fault, South China
Full text linkThis thesis investigates the paleo-hydrothermal flows of the Heyuan Fault, South China, from the Mesozoic to the present day. This 'greenfield' site is largely undocumented, with initial investigations unveiling abundant hot springs and a quartz reef >75 m wide and 40 km long. Three research problems are addressed with the aid of multiscale and multidisciplinary analyses: i) characterisation of the Heyuan Fault geothermal zone; ii) understanding the formation of the quartz reef, and iii) understanding the system’s evolution from quartz reef to present-day geothermal system.
Field studies, structural modelling and fracture analyses revealed several fracture generations that link changes in the tectonic stress regime across the fault zone. The quartz reef was shown to have formed in an extensional framework initiated in the Mesozoic, while a change to compressional stress in the Cenozoic led to the development of cross-cutting, strike-slip faults and associated vertical fracture networks. Present-day hydrothermal fluids channel upwards through this fracture network, permeating the seal overlying the quartz reef, outflowing as hot springs located at fault intersections along the Heyuan Fault.
Microstructural analysis identified a physico-chemical ‘quartz-reef window’ of formation occurring at ~200–350°C, linking quasi-static criteria (accommodation space, massive fluid sources and a cap rock/seal) and dynamic mechanisms (episodic–dynamic permeability, brittle–ductile cycles and fluid injection through a brittle–ductile equivalent of Sibson’s ‘fault-valve’ behaviour). This two-time scale cycle fits a conceptual cnoidal wave-driven fracture model. The micro-macro geochemical fingerprint correlates with structural observations to reveal increasing quartz-vein frequency towards the core and depletion of K2O and Na2O in tectonites during alteration from the host granite; a reaction potentially sourcing SiO2 for the quartz reef. Additionally, widespread Cr was discovered, corresponding locally with trace Au, indicating a potential deep fluid source and wider prospects for mineral resources. Present-day hydrothermal fluid circulation at ~180°C results from a combination of radioactive decay of Mesozoic granite plutons, elevated heat flux due to thermal convection, and deep hydrothermal circulation in the fault.
This work provides the first multiscale study of the Heyuan Fault geothermal zone, revealing valuable information on the quartz reef's genesis, its evolution and geothermal potential
Exploring Potential Interactions of Small-Molecule Reaction Networks at Life’s Origins
Full text linkHow life could have emerged from the chemistry present on early Earth remains one of the most fundamental questions in science. Much progress has been made on the syntheses of compounds proposed to be important for life’s origins under early Earth conditions since the original Miller-Urey experiment in 1953. Nevertheless, it is increasingly apparent that investigating these systems in separation is not only insufficient but also prevents the emergence of more complex behaviours. This thesis aims to investigate chemical reaction networks comprising small molecules proposed to be present on early Earth taking inspiration from nonlinear dynamics and systems chemistry approaches. Chapter 3 bridges well-established prebiotic nucleotide precursor synthesis with the autocatalytic formose reaction. Experimental results revealed that cyanamide defers the exponential phase of the formose reaction and commandeers formose sugar products for nucleotide precursor synthesis. These observations demonstrate how metabolism-like behaviours could have emerged from prebiotic reaction networks. Chapter 4 investigates the formation of pH gradients across phospholipid membranes using the formaldehyde-sulfite reaction. The concentration of (bi)sulfite in the running buffer significantly affected the lumenal pH of the vesicles after purification by size-exclusion chromatography, which was attributed to the leakage of (bi)sulfite as SO2 . In addition, the lumenal pH of vesicles purified with and without (bi)sulfite in the running buffer also behaved differently upon the addition of formaldehyde. These results demonstrate how different environments can affect protocells encapsulating similar chemistries and lead to different responses to the same stimulus. Chapter 5 studies H2-driven carbon fixation catalysed by iron sulfides (FeS) in simulated hot spring environments using an anaerobic flow reactor connected to a GC-FID. Results showed that FeS-catalysed carbon fixation was primarily driven by high temperatures (80–120 °C) and could be further enhanced via irradiation with UV–visible light. Among the synthesised FeS samples, pure and Mn-doped FeS showed the highest catalytic activities. Mechanistic studies using in situ DRIFTS and DFT calculations revealed that the reduction occurred via a reverse water–gas shift, which is comparable to the mechanism of the enzymatic acetyl-CoA pathway. The observations herein indicate that FeS-catalysed carbon fixation likely occurred in early Earth hot springs, the mechanism of which may provide insights into the emergence of early metabolic pathways
Thermoresponsive ionic liquids and hydrogels for thermally driven forward osmosis water treatment systems
No full textForward osmosis (FO) is an alternative water treatment approach which relies on the osmotic force of concentrated solution, referred to as draw solution (DS), to extract water from relatively lower concentration solution. However, a lack of framework for designing FO DSs and their implementation in FO processes requiring DS regeneration has hindered FO development. This study proposes a framework for design of FO DSs especially thermally responsive DSs and their energetic assessment in FO water treatment process to mitigate high specific energy consumption (SEC). This work used multi-criteria analysis based on physicochemical characteristics of DS alone in contrast with earlier DS assessments to identify promising draw candidates and quantified desirable DS performance metrics. Promising thermally responsive DS were found to typically have osmotic density 40 kWh/m3 while achieving recovery ratio of >0.6 and containing residual DS of <15 wt.%. Thermally responsive hydrogels were found to minimally improve water quality of recovered water by reducing residual DS to ~10 wt.%. Therefore, a nanofiltration post-treatment process is recommended to produce potable quality water. The thermally responsive DS performance metrics and the developed modeling framework will aid in development of FO water treatment process prototypes that utilize low-grade thermal energy
Growth and Characterization of Lead-Free Copper-Based Metal-Halide Perovskites for Optoelectronic Applications
No full textOver the past decade, metal halide perovskites (MHPs) have emerged as a new class of semiconductors with favorable physical and chemical characteristics in contrast to traditional semiconductors. The electrical and optical characteristics of the nanostructures of MHPs surpass those of traditional semiconductors and even the bulk form of MHPs making them suitable for diverse applications in various fields including photonics, electronics, and optoelectronics. However, the environmental instability and toxicity of Pb-based MHPs are two major obstacles to the fruitful progress of the field. This thesis is concerned with the development of environmentally friendly Pb-free MHP nano-/microstructures. It involves synthesizing novel MHP nano-/microstructures, investigating their detailed structural, morphological, and optical properties using state-of-the-art experimental techniques, and exploring their applications in optoelectronics.
The synthesis of single-crystalline blue-emitting 0D Cs3Cu2I5 nano-/microwires is demonstrated by using a facile solution self-assembly technique, namely saturated vapor-assisted crystallization (SVAC) method. The morphology of the Cs3Cu2I5 growth product is significantly influenced by the variations in precursor concentration, growth temperature, growth time, and growth rate in the SVAC method. The significant influence of these growth parameters offers a great deal of flexibility in customizing the shapes of Cs3Cu2I5 crystals to meet particular device specifications. In Cs3Cu2I5 MHP, the PL emission does not directly arise from the band gap transition but rather a defect-mediated transition involving self-trapped excitons (STEs). Finally, humidity-induced degradation transforms the surface of 0D Cs3Cu2I5 microwires into 1D CsCu2I3 and CsI phases, compromising material stability but offering the potential for creating 0D/1D hybrid perovskite heterostructures.
Finally, epitaxial growth of layered 2D Ruddlesden-Popper (RP) like Cs2CuBr4 perovskite NWs at near-ambient conditions is demonstrated through a facile solution method. The van der Waals epitaxy of Cs2CuBr4 NWs is achieved on muscovite mica substrate. The solution epitaxial growth presented here is not only energy-efficient but also faster and easier when compared to vapor phase epitaxy. Epitaxial growth mechanism, structural, morphological, and optical properties of Cs2CuBr4 NWs are also investigated in detail. This study represents a significant advancement in achieving facile solution epitaxy of high-quality perovskite NWs that opens up a new avenue to revolutionize the field of MHP optoelectronics
A Deep Learning Framework Based on Knowledge Distillation for The State of Health Estimation of Lithium Batteries in Electric Vehicles
Full text linkThe development of electric vehicles (EVs) is receiving increasing attention as energy and environmental issues intensify. Lithium batteries are now becoming the first choice for new energy vehicles, thanks to their high energy density, long lifespan and high safety factors. However, the usage of batteries after their end of life (EOL) may create safety hazards. The Battery Management System (BMS) collects various information from the battery to analyse its status, then managing the battery pack. The battery degradation degree can be indicated by State of Health (SOH). BMS requires accurate SOH information to make decisions. However, traditional estimation methods have low generalisability and flexibility, and they require substantial computational resources, which is not conducive to applications in EVs systems. Now, with the constant development of hardware, artificial intelligence (AI) technology has also received a widespread attention due to its flexibility and generalisability. This study uses deep learning-based methods to estimate SOH and compares the results of various models. The computational resources of BMS are limited and the battery degradation data have significant noise. This thesis proposes a knowledge distillation (KD) framework for model compression to estimate SOH for EVs. In this framework, the Wasserstein distance (WD) is used as part of the distillation loss. The boundedness and symmetry of the WD addresses the extreme value issue. Additionally, to address the issue caused by noise on model performance, the reversible instance normalisation (RevIN) method is used to filter out unstable information in the data. The framework has been tested by NASA open-source battery dataset. The results show the advantages of the framework in simplification and performance improvement of the model in SOH prediction tasks
Randomised trials in critical care medicine: characterising recruitment challenges and developing solutions
Full text linkCritical care emerged as an independent medical specialty due to the necessity for large-scale mechanical ventilation created by the polio epidemic in 1952. Substantial reductions in mortality were achieved through the application of mechanical ventilation in dedicated Intensive Care Units (ICU). Early advances and research in critical care focussed on pathophysiological mechanisms and potential therapies to normalise abnormal pathophysiology. By the 1960’s, the randomised, controlled, trial (RCT) was firmly established as the “gold standard” in evidence-based medicine. Since the conduct of the first RCT in critical care, a single-centre, 100-patient RCT evaluating use of antacid therapy in the ICU, there has been an ever-increasing numbers of critical care RCTs published.
Along with the numerical increase, the sample sizes and number of participating centres have increased, the quality has improved, and the methodologies employed have evolved. However, a sizeable proportion of critical care RCTs fail to complete recruitment or are plagued by slow recruitment rates, and some are ceased early. Thus, recruitment remains a key challenge faced by the critical care triallist.
In this PhD thesis, I explore the rise of RCTs in critical care, with the intention of answering the following questions: (1) How have the recruitment and methodological characteristics of critical care RCTs changed over time? (2) What are the factors which influence recruitment of patients into RCTs? (3) How does recruitment vary between sites within RCTs? (4) What impact do consent models and established clinical trials groups have on recruitment of patients into RCTs? (5) How did the COVID-19 pandemic impact on RCT conduct?
The systematic review of critical care RCTs with mortality as primary outcome demonstrated that the number of RCTs conducted has increased over time, with some decline in the 2014-2020 period. Certain quality markers were improved including reporting of sample size calculations, prospective trial registration, data safety monitoring presence and lower overall risk of bias. Sample sizes have increased, but this occurred mainly through increased number of participating centres and longer recruitment periods. Recruitment rates per centre did not increase over time, nor did the proportion of RCTs which reached their calculated sample size.
A second systematic review showed that a variety of patient-, family-, clinician-, resource- and trial-related factors may influence consent and patient recruitment into critical care RCTs. Several key themes were identified: (1) Anxiety and distress among family members was a driver of missed recruitment of eligible critically ill patients into RCTs (2) Clinician concerns about trial interventions led to eligible patients not recruited into RCTs (3) Availability of experienced research staff enabled higher consent and recruitment rates. Potential strategies to address these issues include: (1) Adequate psychosocial supports for relatives of critically ill patients, community engagement prior to RCT commencement, innovative consent models (e.g., “consent to continue”) and co-enrolment may all address aspects of anxiety and distress among family members. (2) Clinician engagement and education prior to RCT commencement. (3) Promoting funding, career development and retention pathways for experienced clinical research staff.
A novel application of the Pareto Principle was used to quantify and visualise the variability in patient recruitment across participating centres within a sample of large, multi-centre RCTs. The technique I have described may be used to evaluate, monitor, and track distribution of recruitment across participating centres.
Reasons that may explain why such variability existed, with regards to clinicians’ choice to decline enrolment of patients, were evaluated within the largest ever RCT of septic shock. Factors such as tertiary status of the participating centre, volume of admissions, resourcing for research staff and author affiliations with participating centre were not significantly associated with clinicians declining to enrol patients. Thus, future research into this matter should focus on clinician-specific factors rather than site-level or resourcing factors.
Critically ill patients are often unable to give prospective, or a priori, consent for participation in RCTs due to the nature of critical illnesses and the treatments required (e.g., sedation for mechanical ventilation). Families are faced with significant stress, uncertainty, and anxiety, and are frequently not able to participate in informed consent discussions on behalf of their loved one. Yet, to improve patient outcomes, it is also imperative for researchers to be able to test time-critical interventions in RCTs. Thus, alternatives to a priori consent have been developed. My systematic review of consent models for critical care RCTs with mortality as primary outcome demonstrated that there were 141 RCTs with an a priori consent model, and 45 RCTs with alternate models which included consent to continue (or “deferred” consent- this terminology is historic, and not used anymore) and waiver or exemption from informed consent. After multivariable adjustment, it was found that RCTs with alternate consent models had significantly faster recruitment rates (+6.78 patients per week, 95% CI 3.30 to 10.26, p<0.001).
An analysis of critical care RCTs conducted by established clinical trials groups found that these RCTs had significantly higher recruitment rates than those conducted by other groups (mean difference= +7.47 patients per week, 95% CI 2.67 to 12.26, p=0.003). Quality metrics such as prospective trial registration and data safety monitoring committee presence were also higher among RCTs conducted by clinical trials groups. These data have important implications for ethics committees, funding bodies, healthcare policymakers, critical care researchers, practicing clinicians, patients, and their families alike. These results suggest that critical care researchers should strongly consider engagement with clinical trials groups when planning RCTs.
In-depth explorations of various challenges faced by critical care RCTs have been presented, along with multi-faceted potential solutions that may be applied at individual-, site-, community- and trial-level. Many of these could be evaluated within a SWAT (“Study within a trial”) framework
ChatGPT and CLT: Investigating differences in multimodal processing
No full textDrawing on construal level theory, recent studies have demonstrated that ChatGPT interprets text inputs from an abstract perspective. However, as ChatGPT has evolved into a multimodal tool, this research examines whether ChatGPT's abstraction bias extends to image-based prompts. In a pre-registered study utilising hierarchical letters, ChatGPT predominantly associated these images with local rather than global letters, suggesting a concrete bias when analysing images. This starkly contrasts human participants who predominantly identified the same images with the global letters, indicating that humans and ChatGPT significantly diverge in image interpretations. Furthermore, while humans generally perceive ChatGPT to be more concrete in image processing, there is a notable discrepancy between this perception and the actual level of concreteness exhibited by ChatGPT in handling image-based tasks. These findings provide insights into the distinct cognitive behaviours of LLMs compared to humans, contributing to an emerging understanding of LLM cognition in the context of multimodal inputs
Sustainability with Limited Data: A Novel Predictive Analytics Approach for Forecasting CO2 Emissions
Full text linkUnforeseen events (e.g., COVID-19, the Russia-Ukraine conflict) create significant challenges for accurately predicting CO2 emissions in the airline industry. These events severely disrupt air travel by grounding planes and creating unpredictable, ad hoc flight schedules. This leads to many missing data points and data quality issues in the emission datasets, hampering accurate prediction. To address this issue, we develop a predictive analytics method to forecast CO2 emissions using a unique dataset of monthly emissions from 29,707 aircraft. Our approach outperforms prominent machine learning techniques in both accuracy and computational time. This paper contributes to theoretical knowledge in three ways: 1) advancing predictive analytics theory, 2) illustrating the organisational benefits of using analytics for decision-making, and 3) contributing to the growing focus on aviation in information systems literature. From a practical standpoint, our industry partner adopted our forecasting approach under an evaluation licence into their client-facing CO2 emissions platform
Measuring Stress in Surgeons
Full text linkIntro: Work related stress is one of the most important contributors to physician burnout. Surgical physicians are exposed to complex clinical task-based activities that place them in high stress states. Although it is recognised that surgical environments can increase physician stress there is a paucity of clear data to suggest how surgeon stress should be measured, what constitutes normal adaptive stress, and what may be considered maladaptive stress. The primary aim of this thesis is to determine, what if any, measures of stress are most appropriate to measure in surgeons during surgery.
Methods: Initial review of the literature (chapter 1) identified numerous approaches to measuring stress, divided by biological measures or participant reported outcomes measures. Current biological measures of stress were assessed through systematic review (Chapter 2) and a scoping review was of participant reported measures (Chapter 5). Subsequently, a cohort study of measuring stress in surgeons in live surgery was undertaken with biological measures (Chapter 3) and participant reported outcome measures (chapter 6). A cohort study measuring stress in surgical simulation setting was also undertaken measuring biological (Chapter 5) and participant reported outcome measures of stress (Chapter 7).
Summary of findings: Heart rate, heart rate variability, cortisol, and electrodermal activity had previously been reported with no clear measure appropriate. Changes in all biological measures of stress demonstrated variation in the direction and magnitude of change from pre-surgery to during surgery with less affect in the environmentally isolated simulation study. State trait anxiety inventory, Likert scale, and NASA task load index were previously reported although the relationship between stress, anxiety, and workload unclear. Stress by visual analogue scale was moderately correlated to state anxiety prior to surgery and strongly during surgery with greater differences in simulation settings. Workload scores in simulation were strongly correlated to pre- and post-simulation reported stress but not state anxiety.
Conclusions: Biological measures of stress demonstrate intra-participant variability blunting results measured across a cohort and not currently appropriate to measure in surgical studies. Measure of anxiety prior to surgery, stress during surgery, and workload are reasonable measures of the impact of surgery but the creation of specific questionnaire would be beneficial