48080 research outputs found
Sort by
Geographies of philanthropy: The intricate, the critical, and the generative
No full textFuentenebro et al.'s (2024) intervention on geographies of philanthropy provides a welcome prompt for geographers to engage more extensively with philanthropy's undeniable socio-political significance. My commentary focuses on the authors’ proposition and disaggregation of ‘the philanthropic complex’. Useful as the proposition is, I argue we will be well advised to resist the urge to allow ‘the complex’ to sediment as a macro-concept. Geographical engagements with philanthropy will benefit from engaging with more fully relational thinking and an insistent focus on practice to understand how philanthropy's powers and capacities are assembled and, more generatively, to identify the potential to disassemble and even to redirect philanthropy's impacts.</p
Political dynamics, environmental concerns, and corporate governance: Evidence in China
No full textCorporate governance remains a critical topic in corporate finance, particularly in emerging markets like China, where firms operate under a distinct regulatory framework, ownership structure, and institutional environment. One key focus is understanding managerial incentives and establishing effective governance mechanisms, as these factors directly influence corporate decision-making and firm performance, and thereby shareholder wealth. Additionally, the increasing emphasis on environmental sustainability has reshaped traditional business models, pushing firms toward more sustainable practices. Based on China's unique hybrid market conditions, institutional environment, and development pathway, this thesis aims to advance understanding of the interplay between political dynamics, environmental concerns, and corporate governance over three chapters of analysis.Chapter 2 examines the political rank of CEOs in Chinese state-owned enterprises (SOEs) and their managerial incentives in pursing innovation. Specifically, it investigates how political promotion influences corporate innovation in SOEs. This chapter proposes that CEOs with political aspirations pursue innovation to advance their political rank, particularly in response to the national innovation policy. The findings reveal that the state is more likely to reward SOE CEOs with political promotions for superior innovation performance, especially when performance metrics are quantifiable. This effect is stronger when CEOs are nearing retirement or are in politically dynamic environments with lower anti-corruption exposure. Therefore, the existence of the political labour market serves as an external governance mechanism that shapes CEOs’ incentives for innovation in SOEs.Chapter 3 explores how firms’ environmental responsibility influences managerial incentives through the examination of CEO pay-performance sensitivity. It argues that CEOs may exploit environmental investments to enhance their personal publicity and exposure at the expense of firm profitability, prompting the board to strengthen the link between CEO compensation and firm performance to mitigate agency conflicts and safeguard shareholder wealth. The evidence shows that firms’ environmental investments significantly increase the marginal effect of firm performance on CEO pay, thereby strengthening the sensitivity of CEO compensation to firm performance, particularly in firms exposed to environmental risk or operate in highly-polluting industries. This relationship is driven by insufficient external monitoring of environmental spending, highlighting the need to strengthen internal governance mechanisms, particularly through the design of compensation contracts. Moreover, the effect is weaker for powerful CEOs who also serve as board chairs and hold firm ownership, suggesting the importance of enhanced governance measures to mitigate their influence.Chapter 4 builds on Chapter 3 by examining how environmental problems shape external governance; specifically, how air pollution affects analyst coverage. Analysts play a crucial role as monitors and information intermediaries between firms and the markets. This chapter demonstrates that firms located in cities with worse air pollution receive greater analyst coverage due to their weaker information environments, thereby increasing investor demand for analyst services. This effect is moderated by state ownership, as non-SOEs face greater operating uncertainty without state capital guarantees, making the demand for analyst services more sensitive to air pollution. Additionally, analysts covering firms in polluted environments tend to have more industry expertise and focus on fewer firms, owing to the challenges posed by weaker information quality. Overall, the findings from this chapter highlight that increased analyst coverage could serve as an external governance mechanism, helping to mitigate information asymmetry for firms in polluted regions.</p
Grid Interaction of Electrolysers and Fuel Cells for Utilisation of Renewable Energy Surplus
No full textThis thesis investigates the grid interaction of electrolysers (ELs) and fuel cells (FCs) combined as green hydrogen storage for utilisation of renewable energy surplus. The integration of renewable energy sources (RESs), such as solar PV and wind, with power grids is becoming increasingly popular to reduce greenhouse emissions and mitigate climate change. However, integrating RESs with power grids can produce challenges, such as the impacts of sudden variations in weather conditions including say cloud passing and wind gusts on PV and wind systems respectively, and sudden variations in load demands. These sudden variations are usually accompanied by high ramp rates, where the power outputs of renewable energy sources or the load demands can ramp up or down at a very high rate. Besides these, the harmonic distortions caused by modern power electronic loads can threaten the power quality and reliability. To minimise above mentioned problems, a hybrid energy storage system (HESS) is considered for the future power grid infrastructure. HESS can mitigate not only the system challenges but also, with sufficient capacity, the fast-rising energy demand, by storing surplus when there is excess generation and discharging the energy when there are insufficient power outputs from renewables to meet the load demand. Traditionally, battery energy storage (BES) is considered to supplement uncertain renewable generation. However, BES alone may have unanticipated environmental and financial consequences. Recently, with the increasing global interest in producing green hydrogen energy (GHE) to transition to a zero-carbon economy, several research papers in the literature have proposed to convert any surplus output from RES (whenever the electricity marginal price is zero or negative) to hydrogen using an EL. The produced hydrogen gas is transferred and stored in a controlled manner. Whenever there is a need for energy say due to unavailability of supply from renewables, the stored hydrogen gas is used to produce electrical energy using a FC to supply to the grid. To make the integrated power system including RESs, and HESS comprising ELs and FCs, an optimal, efficient, reliable, and cost-effective power management control strategy (PMCS) with advanced control decisions needs to be formulated. The main aim is to utilise the surplus RES to generate and store hydrogen gas and convert the stored hydrogen gas into electrical power with the aid of FC. Furthermore, it is important to optimally share the energy between the grid and HESS.This thesis conducts a detailed review on FCs/ELs comprising GHE storage technologies for utilising the RES surplus because the state of the art is still limited in terms of topologies integrated with different types of grids (i.e., AC or DC), multiple types of FCs/ELs, and a variety of power electronic interfaces with appropriate PMCS. The thesis proposes a mathematical and novel electrical equivalent circuit model of proton exchange membrane electrolyser (PEMEL) where the dynamic cell performance has been evaluated by considering the fluctuations in the external current profile. To demonstrate its adaptive feature, the proposed electrical equivalent circuit model is then scaled up to a 1 MW array by series and parallel connections, and the resulting system is validated by comparing it with the other reported experimental results of a 1 MW stack.Furthermore, the dynamic features of the PEMEL stack are used to investigate the impact of introducing the PEMEL stack's control loop into a single-area power system populated with solar PV units, while accounting for the communication time lag during the control signal transfer process. The dynamic performance of the proposed power system as well as steady-state errors are subsequently investigated to demonstrate the PEMEL stack's effective contribution to frequency regulation services. Furthermore, the potential resilience benefits of frequency control services and frequency sensitivity analysis for a modified IEEE-13-bus distribution system have been investigated. The results show that the system with PEMEL control loop responds to frequency changes faster than standard synchronous generators, implying the proposed PEMEL stack has a significant potential for enhancing frequency stability, resilience, and robustness.The thesis also includes a system modelling and performance analysis of a renewable hydrogen energy hub (RHEH) connected to an AC/DC hybrid microgrid. When power is in excess, a coordinated power flow control strategy minimises generation and demand mismatches while also producing green hydrogen. A modified hybrid approach involving perturb and observe, and particle swarm optimization techniques is suggested for tracking maximal power points. The thesis also presents a novel compensation technique for reducing intermittencies in a PV-powered microgrid that employs a hybrid multilevel energy storage system. This method solves memory and temporal delays through improved storage capacity. Simulations using a 24-hour solar irradiance profile are used to evaluate the technique on a 100-kW grid-interactive PV-dominated microgrid system.The thesis presents a grid-integrated analytical model for the co-production of hydrogen from an offshore hybrid energy system. The system supplies a percentage of its capacity to the onshore grid facility while producing hydrogen. The electricity is quantified based on a market price and total offshore generation. A case study is presented for a hypothetical 10 MW hybrid offshore energy system in NSW, Australia. The thesis proposes a probabilistic approach to size the HESS and a two-layer distributed energy management strategy for the HESS. Simulation results show that the supercapacitor bank can handle high-frequency fluctuations and avoid round-trip losses associated with HESS. The two-layer distributed energy management strategy extends operating lifetime and reduces operation costs of HESS, and also, maximizes the utilization HESS by avoiding excessive switching between FCs and Els, and maintaining an equal number of ON/OFF operations by ELs and FCs.</p
Anode Optimization Strategies for High-performance Aqueous Zinc-ion Batteries
No full textRecent years, various alkali metal ion batteries have been widely studied and gradually applied in modern society.[1–3] However, rising raw material and organic components prices, and the inherent safety hazards of the unstable system have seriously hindered their further application.[4,5] Therefore, it is becoming increasingly important to develop a promising battery to complement or even replace those organ alkali metal-ion batteries. Aqueous zinc-ion batteries are considered as a promising alternative due to their high theoretical capacity, abundant and cheap resources, and the feasibility of aqueous electrolyte systems due to the aqueous stability of zinc metal.[6–8] In recent years, cathode material like manganese oxide,[9,10] vanadium oxide,[11,12] organic molecules,[13–15] etc. and corresponding reaction mechanisms that are well matched to aqueous zinc ion battery systems have been intensively studied. However, the low Coulombic efficiency (CE), zinc dendrites, and unstable zinc deposition problems in zinc anode hinder the further development of zinc ion batteries. At the same time, due to the inability to form a stable protective solid electrolyte interface (SEI) in situ during zinc plating and stripping processes, the complex relationship among zinc anolyte, interface, and zinc host also makes the research on the reaction mechanism of zinc anode more diverse and complicated.[16–18] The current anode optimization methods could be divided into three categories: electrolyte modification,[19,20] zinc anode surface coating,[21,22] and zinc anode structure redesign.[23,24] Herein, we have developed multiple zinc anode optimization strategies to improve the stability of zinc plating/stripping. These strategies have been effectively applied to tune the electrochemical performance of aqueous zinc ion battery for establishing a highly reversibility and long-lived zinc anode.</p
In-situ Engineering Carbon-based Tribofilm Lubrication with Catalytically Active Layered Double Hydroxide
No full textThe abstract for this item has not been populated.</p
Dietetic Practice With Patients Undergoing Alcohol Withdrawal: Towards Recommendations for Practice
No full textAlcohol remains a leading risk factor for death and disability significantly contributing to global disease burden. Social and environmental factors can influence alcohol consumption and lead to hazardous patterns of use contributing to the development of other health conditions such as Alcohol Use Disorder (AUD), cardiovascular disease and nutritional abnormalities. Nutritional abnormalities may include electrolyte disturbances, micronutrient deficiency and/or malnutrition. These may be a result of the impact alcohol has on nutrient metabolism and digestion leading to malabsorption or may be through insufficient dietary intake secondary to food insecurity. In some instances, the negative effects of alcohol consumption may be experienced by others, such as those who are victims of domestic violence. Admission to hospital for detoxification and alcohol withdrawal may be required to interrupt heavy and sustained periods of alcohol use. This allows for comprehensive assessment by a multidisciplinary team of healthcare professionals which may include a dietitian. Despite the known presence of nutritional abnormalities, there is a lack of clearly defined nutritional management guidelines to support healthcare professionals and dietitians.The aim of this thesis was to comprehensively examine and synthesise the evidence on nutrition interventions for patients hospitalised for alcohol withdrawal; the various approaches to nutrition care for this cohort; the perspectives of healthcare professionals on the role of the dietitian; and the perspective of patients regarding nutrition. This analysis aims to identify key implications for clinical practice, by enhancing the understanding and application of nutrition and the role of the dietitian in this setting. This was achieved through an explanatory sequential design approach encompassing six studies examining current nutritional practice drawing on case studies in a metropolitan health service in Sydney, Australia to highlight implications for practice.</p
Deciphering Agitation in Elderly Care: An Analytical Approach Using Large Language Models Llama 2
No full textAgitation is a prevalent and challenging symptom in residents with dementia within Residential Aged Care (RAC) facilities. Effective management of agitation is crucial for enhancing the quality of care and improving the working conditions of caregivers. This research aims to uncover the prevalence and patterns of agitation and evaluate the effectiveness of various management strategies for these agitation symptoms.The dataset contains 3,164 progress records from 957 residents across 40 RAC facilities.Generative AI and large language models (LLM), such as Meta’s Llama 2 and Open AI’s GPT-3.5, have shown remarkable capabilities in processing and analysing vast amounts of unstructured data. Therefore, these models were utilised in this research to extract and categorise agitation symptoms, observed locations, contributing factors, interventions, and outcomes from clinical notes. Afterwards, statistical methods such as descriptive statistics, Chi-Squire Test and correlation analysis were employed to analyse the extracted agitation symptom, location of symptom onset, caregiver interventions and outcomes.The analysis uncovered the prevalence and patterns of eight distinct types of agitation symptoms, with the most frequently occurring symptoms being refusal of care, physical restlessness, verbal aggression and physical aggression. Dining areas, living rooms, and hallways were identified as the most common locations for agitation incidents. Each note is assessed by the LLM to determine if the symptom was improved, if the interventions were ineffective or if the outcome was not documented. A significant gap in documentation practices was noted, with a high proportion of notes lacking documented outcomes, highlighting the need for improved recording to better assess intervention effectiveness.The major intervention strategies for managing agitation in dementia within RAC facilities include psychosocial and behavioural interventions, pharmacological management, physical health and comfort measures, and activity-based interventions. The outcomes indicated that psychosocial and behavioural interventions are the most effective, especially for managing verbal aggression and refusal of care. An earlier study highlighted resisting, wandering, restlessness, complaining, arguing, outbursts, pacing, speaking in excessively loud voices, using profane language, and threatening as prevalent symptoms of agitation [1], largely coinciding with this study at a categorical level and emphasising the dynamic nature of agitation symptoms among dementia care recipients.This research contributes to the understanding of agitation management in dementia care and underscores the potential of LLMs to extract key clinical information from clinical notes, inform care strategies, and ultimately improve the quality of life for residents and caregivers in RAC facilities. The dynamic nature of agitation symptoms among dementia care recipients emphasises the need for tailored management strategies in RAC facilities to address these behavioural changes. Future research should explore the introduction of underutilised interventions and conduct longitudinal studies to assess the long-term outcomes of agitation interventions.</p
Predictive Analysis For Lung Cancer Using Artificial Intelligence (AI)
No full textLung cancer remains a significant challenge due to its complexity, rapid progression, and variety of clinical presentations. Although advances in diagnosis and treatment of the disease continue, millions of individuals are still affected by the disease every year. The use of machine learning in healthcare informatics has emerged as a transformative tool that offers significant potential for improving the prediction of outcomes for common diseases, such as lung cancer. There are several challenges in developing machine learning models for lung cancer prognosis, particularly due to the imbalance in datasets regarding the time to death. The majority of lung cancer patients survive between 1- or 2-years post-diagnosis, leading to a disproportionate amount of data reflecting shorter times to death. Thus, models trained on such datasets tend to be biased towards predicting these more common outcomes and struggle to accurately predict longer times to death of 5 years or more, which are less represented. Moreover, the complexity of lung cancer progression adds another layer of difficulty. Lung cancer frequently metastasizes to multiple organs, such as the bones, brain, and liver, simultaneously. This multifaceted spread introduces heterogeneity in the data, requiring models to account for a wide range of variables and interactions to make accurate predictions. On the other hand, the rarity of data on patients with multiple organ involvement exacerbates the challenge, as the model has limited examples from which to learn these complex patterns. Furthermore, while efforts continue to refine these predictive models, the issue of data scarcity for long-term survivors and those with multi-organ metastases remains a significant hurdle. Accordingly, there is a pressing need to enhance data collection and model training strategies to better represent and predict times to death for all patient groups, including those with rarer, long-term survival rates and multiple metastases. By addressing these challenges, we can improve the accuracy of predictive analytics in lung cancer care, thus significantly impacting patient management and treatment outcomes.Furthermore, in this thesis, the issues of imbalanced data related to time to death and predicting bone, brain, and liver metastases in lung cancer patients are addressed. For the imbalance data issue, several models have been developed to predict the time to death. These include a standard approach, under sampling using random under sampling, and oversampling using Synthetic Minority Over-sampling Technique. Meanwhile, the challenges of multilabel classification for predicting bone, brain, and liver metastases simultaneously in lung cancer patients are tackled by evaluating label dependencies, feature importance, and algorithm performance across three categories: problem transformation, algorithm adaptation, and ensemble methods. Relationships between the labels for bone metastasis, brain metastasis, and liver metastasis are investigated to better understand their interactions. Additionally, computational efficiency regarding training and testing times is analyzed to ensure the practicality of these models for clinical use. Additionally, a chapter in this thesis is dedicated to evaluating the best models among those developed for imbalanced data and multilabel classification. This evaluation involves using external data sets to determine how well these models perform with data not seen during training, thereby assessing their generalizability and effectiveness in broader clinical settings.The findings of predicting time to death suggest that the use of Synthetic Minority Over-sampling Technique enhances the performance of classifier models when compared to the standard approaches. In most cases, oversampling leads to higher accuracy, AUC, and F1-score, indicating an improvement in predictive capabilities. This enhancement is likely attributed to the increased sample size of the minority class, which allows the models to better learn the characteristics of both classes. On the other hand, under sampling generally yields lower performance relative to both standard data and oversampling. This reduction in performance may be due to the loss of critical information from the majority class, potentially impairing the model's ability to generalize and make accurate predictions. On the other hand, the multilabel classification method presented in Chapter 4 revealed that important factors such as AJCC staging and age are critical, as shown by feature importance analysis. While the label powerset random forest showed better sub-accuracy and lower Hamming loss, the RakEL random forest excelled in terms of F-1 score, micro-precision, and micro-recall. The effectiveness of the models varied, with some excelling in understanding the connections between different labels and others being more effective in specific clinical situations where high precision and recall are essential. Overall, these findings suggest that multilabel classification models are not only useful for predicting metastasis but also for identifying significant label dependencies and highlighting key features in lung cancer patients. After evaluating the models for predicting time to death and multilabel classification (bone, brain, and liver metastases) in lung cancer patients using external data, it was observed that while the models generally perform well, their performance declines with the external dataset. An exception was noted with the random forest model used to predict time to death; its performance on the external dataset was almost as good as on the training dataset. The Random Forest model is less likely to overfit compared to more complex models. This advantage comes from its method of using multiple decision trees to make predictions. Each tree in the Random Forest is trained on a different set of data and features. This method helps balance out any errors or biases that might appear if only one tree was used. However, the overall results of this evaluation suggest that there is overfitting during the model training process. This indicates that the training data may not sufficiently represent the diverse characteristics of demographic and clinical data. Consequently, the models primarily learn from the data in the training set and do not perform well when applied to external data.</p
Exploring the Experiences of Mental Health Consumers and Nursing Students at a Therapeutic Recreation Camp: A Novel Arts-Based Methodological Study
No full textBackground: In group settings people can develop bonds which are strong and contribute greatly to their feelings of wellbeing and satisfaction in life. Groups reduce isolation, bring hope, and connect people to something bigger than their individual selves. When positive contact takes place between two different groups, diversity between these groups can be a means to increase social capital. However, there is little research about contact through combined immersive group experiences, from the perspectives of both groups.Aim: The aim of this study was to explore mental health consumers’ and nursing students’ experiences of being in a group together at a therapeutic recreation camp.Research Design: A qualitative paradigm and arts-based methodology were used for this study. The research setting was a therapeutic recreation (TR) camp, where all participants, who were mental health consumers and nursing students, were immersed together in groups. Data were gathered through written contributions of participants to canvas artboards collectively. Contributions were guided by a question specific to their cohort. Consumers were asked: What is your experience of being immersed with nursing students? Students were asked: What is your experience of being immersed with consumers? Data from each cohort were thematically analysed, using Braun and Clarke’s (2006) method.Findings: There were ten consumer elements and twelve student elements which informed six themes.The consumer themes were Safe, Supportive, Interactions; Teaching and Learning with Students, and An Affirming Environment. The student themes were Evolving Perspectives; Positive View of Consumers, and Rich Experiences with Consumers. The essence of meaning reflecting the themes of both cohorts was Psychosocial Expansion.Discussion: Psychosocial Expansion is a concept which aligns with and extends sociological theories of group dynamics. Group self-expansion, relational self-expansion and positive regard toward outgroups were examined. Immersive experiences where two groups combined at a TR camp, led to Psychosocial Expansion for both mental health consumers and nursing students, fostering growth and benefits within their collective groups.Conclusion and Implications: This study generates new qualitative understandings of the impact of contact and immersion between two groups. A novel arts-based data collection method using canvas artboards was used to effectively explore the combining of mental health consumers and nursing students together in an immersion experience. The outcomes of the study have strong implications for community integration for mental health consumers, immersive education, and prejudice reduction.</p
Advancing <i>in vitro</i> models of human dorsal root ganglia sensory neurons using pluripotent stem cells and 3D bioprinting technologies
No full textDuring everyday interactions with the world, the sensory nervous system plays a crucial role in detecting, relaying, and processing information from the internal and external environments of the body. First-order peripheral sensory neurons (SNs) are central to these functions, transducing physical stimuli into chemical and electrical signals to provide sensations like touch, pain, and proprioception. Our knowledge of the sensory nervous system has been largely founded by animal studies, however recent works have highlighted fundamental differences between human and animal tissues. Due to limited access to primary human tissue, human pluripotent stem cells (hPSC) present a promising alternative by allowing the in vitro differentiation of human SNs. Current methods for the study of hPSC-derived SNs largely rely on two-dimensional (2D) culture platforms, which do not fully recapitulate the three-dimensional (3D) dynamics of native sensory tissues. To address these limitations, this thesis explores novel tools for more advanced modelling of human SNs, from approaches for improved hPSC differentiation to the 3D bioprinting of cell-laden scaffolds.We first describe the characterisation of a genetically modified hPSC line, H9NGN2, engineered with an inducible Neurogenin-2 (NGN2) gene cassette for the rapid and efficient differentiation of hPSCs to induced SNs (iSNs). Validation of this cell line confirmed its pluripotency and capacity for tri-lineage differentiation, establishing suitability for producing neural crest cells (NCCs), the developmental progenitors to iSNs. NCCs were enriched for CD271+ expression and differentiated into functional iSNs over 21 days in vitro (DIV), expressing key sensory neuronal markers and demonstrating essential electrophysiological activity. This work establishes H9NGN2 as a robust tool for generating iSNs, providing the foundation for subsequent 3D bioprinting studies.Building on this advance, we describe the development of an extrusion bioprinting method for generating 3D iSN scaffolds. A GelMA-based bioink was optimised for printability and cytocompatibility with NCCs, allowing for subsequent iSN differentiation. Bioprinted constructs facilitated the differentiation of NCCs into functional iSNs, expressing sensory-specific markers and exhibiting membrane excitability properties by 21 DIV. These results highlight the potential of bioprinting technologies to generate more physiologically relevant in vitro models of the sensory nervous system, offering new possibilities to study neuronal development, function, and disease in 3D.We then introduce a proof-of-concept 3D bioprinted model of human skin cocultured with iSNs to expand upon this work. Using the same GelMA-based bioink, primary human dermal fibroblasts, immortalised human Schwann cells, and H9NGN2-derived NCCs were bioprinted to form 3D dermal scaffolds, with immortalised human keratinocytes seeded on the surface of scaffolds to enable epidermal formation. After 28 DIV, discrete dermal and epidermal structures were observed, with neurite outgrowth from iSNs marking a key step towards the generation of an innervated skin model. While further optimisation is needed to enhance innervation and guide neurites towards the epidermis, these findings present an important advancement towards the fabrication of complex bioprinted cocultures with a nervous system component.Overall, this thesis presents significant strides in the development of advanced in vitro models of human SNs. By establishing a robust platform for the differentiation of hPSCs into functional iSNs and demonstrating the feasibility of 3D bioprinting for generating complex tissue constructs, this work addresses key limitations of current 2D culture systems. The creation of an innervated human skin model further highlights the potential for bioprinting to generate physiologically relevant coculture models for studying sensory biology. These developments pave the way for more sophisticated investigations into SN function and offer promising applications in regenerative medicine, disease modelling, and drug discovery. While optimisation challenges remain, this research establishes a strong foundation for advancing innovative new technologies to model human sensory neurons that can be applied to further study and gain knowledge about the human sensory nervous system.</p