27716 research outputs found
Sort by
The Air Quality Co-benefits of a Low Carbon Economy in Europe
No full textAir quality in Europe has improved since the Industrial Revolution, however, Europe still experiences greater than average mortality due to air pollutant exposure. Many of the sources of air pollutants are the same as those of greenhouse gases and thus, in theory, societal developments designed to mitigate climate change could also improve air quality.
This is however, more complex than it seems on the face of it, as the dominant sources of air pollutants can modify the efficacy of climate mitigation for achieving air quality co-benefits. Secondary pollution can also pose an issue, particularly regarding ozone, which is projected to increase as the climate changes. Nevertheless, if air quality co-benefits of climate change mitigation can be achieved in Europe, this could be very impactful for public health, counteract any mitigation costs and help position European policymakers as climate leaders, potentially encouraging more ambitious mitigation in countries that emit more greenhouse gases.
There are many studies that consider co-benefits, however, this is challenging to model. Modelling air quality tends to require very detailed and computationally expensive models that run at a fine enough resolution to capture elevated urban concentrations and have a detailed representation of chemistry. Modelling climate change however, is best undertaken using global climate models, which trade resolution and chemistry for the ability to capture physical changes in the climate system over very long periods. Up to now, most of the research uses the latter. Additionally, recently, new scenarios for driving these models were developed. These Shared Socioeconomic Pathways provide a greater representation of the societal changes, many of which would be extremely meaningful for air quality, than the prior Representative Concentration Pathways upon which much of the existing research is based.
This thesis uses the more recent SSPs and a regional atmospheric chemistry model to model air quality co-benefits of a low-carbon economy in Europe compared to a “business as usual” pathway and a “low mitigation pathway”. This better represents the impacts at a scale relevant to European policymakers, and allows for more detailed analysis of the drivers of future air quality local to Europe. It does however, have a limitation in that the physical effects of climate change on air quality are not considered
Transfer learning for population-based SHM
Full text linkStructural health monitoring (SHM) systems aim to proactively identify damage and provide diagnostic information to support maintenance decisions in mechanical, aerospace, and civil infrastructure. A critical challenge for the application of SHM systems -- particularly those that provide contextual information -- is the feasibility and cost of acquiring comprehensive data. Population-based SHM (PBSHM) presents a potential solution by leveraging data from related structures. However, differences between structures often prevent conventional machine learning models from generalising across domains. This issue motivates the use of transfer learning, which seeks to improve predictive performance in a target domain using data from a related source domain.
In PBSHM, target structures will often only have data for a limited range of health states. Therefore, to enable transfer when target labels are sparse, this thesis presents novel statistic alignment (SA) methods that require only undamaged target data. These methods are shown to facilitate the generalisation of models learnt using only labelled source data.
Quantifying similarity between structures and their features is essential to ensure that transfer learning will yield positive results. This thesis investigates using physics knowledge to address limitations with data-based similarity measures in sparse-data scenarios. This approach is incorporated into a feature-selection criterion to identify transferable, damage-sensitive features. Subsequently, it is used within a regression framework to predict the quality of predictions when transferring between a specific source/target pair, supporting decisions about when transfer is appropriate.
Previous work has not considered how to incorporate transfer learning into an online framework that updates as labels are collected during a monitoring campaign. Thus, a Bayesian model is proposed that uses the SA methods to define mappings early in the monitoring campaign and updates sequentially as labels are obtained. This model is integrated into an active-sampling strategy that guides inspections by selecting the most informative observations to label
Time-Series Clustering and Visualization for Insights into Multimorbidity Progression
Full text linkThe increasing availability of vast amounts of data in electronic health records (EHR) offers immense opportunities to extract valuable insights, particularly through the application of machine learning techniques like clustering. This thesis focuses on clustering time-series data extracted from medical records, with the aim of identifying meaningful clusters of patient sequences. While clustering methods are well-established for static datasets, clustering time series data presents unique challenges, especially when it comes to selecting the most relevant solution from many valid clustering outputs.
In this work, we develop a two-stage methodology for clustering time-series data. The first stage simplifies high-dimensional sequence data, while the second stage focuses on identifying clusters within these sequences. We also address the issue of comparing multiple clustering solutions by introducing a novel approach that combines a graphical user interface (GUI) with a graph-based representation of the relationships between different clustering solutions. This framework allows for intuitive, simultaneous exploration and comparison of multiple valid solutions, helping to reduce the space of possible results and aiding in the interpretation of alternative outcomes.
Our methodology is applied to the domain of multimorbidity, a significant healthcare
challenge characterised by the coexistence of multiple chronic conditions. By applying our tools to multimorbidity datasets, we gain insights into the progression of chronic illnesses and their interactions
Nanoscale Zinc Substituted Hydroxyapatites: Potential Bone Grafting Biomaterials with Antibacterial Properties
No full textQuantifying the impacts of land-use change and management on tropical biodiversity across scales
No full textTropical forests harbour exceptional biodiversity and are vital for sustaining future human needs, yet they face severe threats, particularly from agricultural land-use change. In this thesis, I examine the impacts of livestock agriculture on tropical biodiversity using a comprehensive field-based dataset across local, regional, and near-national scales. Birds are used as a model taxon due to their ease of sampling, extensive data availability, and their representation of broader ecological impacts. I assess phylogenetic and functional diversity to reflect biodiversity’s multidimensionality and identify land-use practices that best conserve it.
Chapter 1 introduces the broader context and key aspects of the research problem. Chapter 2 examines the local scale in lowland tropical forests, finding that deforestation reduces both phylogenetic and functional diversity, with land-sparing practices offering better conservation outcomes. Chapter 3 assesses land sharing and sparing strategies across a topographically diverse region spanning elevational gradients, showing that land sparing remains the more effective approach despite elevational differences. Chapter 4 explores forest conversion impacts at regional spatial scales, revealing phylogenetic diversity loss driven by impoverishment within clades, though entire lineages are not systematically lost, a pattern that is generally scale independent. Chapter 5 explores functional diversity change, finding that although functional richness declines, the overall functional structure remains resilient. However, forest loss is especially detrimental to dispersal-limited forest birds. Chapter 6 synthesises the main findings of the thesis, discusses their implications, and highlights potential applications along with directions for future research.
Overall, losses of phylogenetic and functional diversity appear less severe than those typically reported using taxonomic metrics at local scales. Nonetheless, the findings highlight the need for more sustainable agricultural practices that prioritise forest conservation. Approaches that share land with wildlife appear incompatible with species dependent on intact habitats, suggesting that strategies to spare and restore native forest may be more effective
Novel hybrid permanent magnet interior PM synchronous machines
No full textThe rising cost of rare-earth permanent magnets (REPMs) in recent years has provided a challenge in developing high performance PM machines at a competitive price. Therefore, this thesis focuses on novel hybrid PM (HPM) machines with the synergies of high energy product REPM and low-cost ferrite PM (FEPM), with a particular emphasis on enhancement of the ratio of torque to REPM volume.
Six HPM interior PM synchronous machines (IPMSMs) with different topologies using symmetrical and asymmetric rotor structures have been developed and investigated in this thesis. The electromagnetic performance, including open circuit characteristics, output torque, torque and dq-axis inductances with current advancing angle, and efficiency, as well as mechanical robustness, demagnetisation withstand capability, and PM cost, are analysed and compared with those of a REPM-based V-shape baseline using the same stator and specifications of the commercialised Nissan Leaf 2012 IPMSM. The comparison confirms that the proposed HPM machines can produce the same torque at a lower volume of REPM consumption than that of the baseline. It is worth mentioning that the proposed HPM machines have been built and tested in a small size as the proof of concept.
Meanwhile, using the frozen permeability method, the output torques of the proposed HPM machines are divided into three components, including the reluctance torque, and the FEPM and REPM torques. Any improvement in either reluctance torque, FEPM torque or both, along with the torque enhancement caused by the magnetic field shifting effect in asymmetric topologies will result in a lower required REPM torque while maintaining the desired level of output torque. Consequently, a lower volume of REPM would be required which leads to the reduction of total PM cost. As a result, at the same torque and size, the contributions of torque components and volumes of both PM types are compared. It is shown that the HPM utilisation in topologies with a combined spoke arrangement of PM with V- or delta-shape structures can effectively improve the ratio of torque to REPM volume. Meanwhile, in the topologies where spoke FEPM is not used, a combination of asymmetric rotor topology with V- and U-shape structures of PMs can also improve the ratio of torque to REPM volume. Therefore, the proposed HPM machines in this thesis are potential candidates for electric vehicle (EV) applications at a reduced cost
Energy-efficient Tracking of Mobile Audio Sources via Emergent Distributed Systems
Full text linkTracking multiple mobile audio sources in acoustic scenes where the layout, targets, and requirements change rapidly is a fundamental problem in the research field of tracking only through listening with computational means. Recent developments in the field of robotics and artificial intelligence have enabled researchers to further the capabilities of such systems – interconnected or not – towards solving the localisation and tracking problem. Nonetheless, such research focuses primarily on managing the accuracy of such systems with little care for energy (i.e. battery) efficiency, especially in applications where movement is required. Meanwhile, highly dynamic acoustic scenes are not always accounted for in the designs using mobile listeners.
This thesis attempts to bridge these gaps by attempting to solve this problem with a focus on energy efficiency: reaching the targets in a timely manner conserving as much energy as possible. To achieve this goal a suitable system has been designed and implemented, while bio-inspired computing has provided the key inspiration towards developing a listening and tracking strategy that can expertly adapt to such scenarios. Established machine-learning techniques have been employed to further optimise this strategy, ultimately achieving even higher efficiency through adaptation of psychological research towards improved collaborative problem solving via emergence engineering.
The key contributions of this thesis are thus: a distributed system framework based on microservices tailored for modern devices capable of listening and tracking with both simulation and real-world deployment capabilities, an adaptive strategy that can be utilised for standalone system solutions, and an even more efficient approach for cooperative solutions. An example application could be the deployment of several small robots in disaster scenarios for reaching and aiding trapped individuals (e.g. building on fire with heavy smoke). Finally, the interdisciplinary research process followed throughout this undertaking aspires to offer an incentive for other researchers to pursue similar avenues for innovative applications or efficient solutions in the pertinent domains
Deep generative model for synthesising and analysing cardiac magnetic resonance images
Full text linkCardiovascular disease (CVDs) is still the main disease causing many deaths around the world. According to the World Heart Federation's 2023 World Heart Report, approximately 20.5 million deaths in 2021 were attributed to CVDs, accounting for nearly one-third of global fatalities. Over the past few decades, deep learning algorithms have increasingly been applied in magnetic resonance imaging (MRI) in the medical field, and in particular, have become central to the diagnosis and prediction of CVDs. However, the dynamic motion of the heart and its complex and changeable anatomy pose many challenges to the interpretation of cardiac magnetic resonance (CMR) data. Traditional manual analysis methods are time-consuming and provide variable results. At the same time, generative models have advanced medical image analysis, especially for downstream cardiac image analysis tasks. The aim is to use these synthetic images as viable alternatives to real data in deep learning model training, providing cutting-edge solutions in data segmentation, registration, and strain analysis.
This thesis systematically investigated several probabilistic generative models applied specifically to cardiac image analysis, including multi-channel variational autoencoders (VAEs), generative adversarial networks (GANs), and latent diffusion models (LDMs), using cine CMR and tagging CMR images as primary subjects. Cine CMR provides high-resolution dynamic sequences to assess cardiac morphology and myocardial function throughout the cardiac cycle. Tagging CMR enables the quantification of myocardial deformation by encoding spatial modulation patterns into the myocardium. The efficacy of these models is validated through multiple metrics and downstream tasks such as cardiac segmentation and myocardial strain analysis. Initially, we comprehensively reviewed existing deep learning-based image generation techniques in medical image synthesis. Subsequently, we introduced a sparse multi-channel VAE to learn the joint latent representation of cine and tagging CMR images. The proposed model can generate tagging CMR from cine CMR alone, thereby
enabling myocardial strain estimation straight from cine CMR images. This represents a novel approach within cardiac imaging research and could potentially replace the conventional clinical use of tagging image sequences as a basis for myocardial motion and strain analysis. Furthermore, we introduced an innovative framework employing latent denoising diffusion implicit models (DDIM) to synthesise full-spatial cine CMR images. We investigated whether these synthetic images can serve as viable substitutes for real data in downstream cardiac image analysis tasks. Building upon this, we present a novel spatial-temporal generative model that leverages latent DDIM conditioned on demographic and clinical factors, capable of synthesising realistic 4D cardiac cine CMR image sequences.
Overall, the methodologies presented in this research demonstrate potential for innovation and practical applications. The method introduced here may potentially revolutionize traditional clinical diagnosis and intervention methods, and introduce new perspectives on applying deep learning models in medical imaging. These models show promising performance in the generative field, not only promising insights into cardiac conditions, but also advancing the development of personalized medical diagnosis and prediction solutions in the field of cardiology
Landscape Planning and Characterisation For Ecotourism within Protected Areas in Saudi Arabia: A case Study in Crown Prince Mohammed Bin Salman Royal Reserve
No full textAbstract
This study delivers the first National Landscape Character Assessment (LCA) for Saudi Arabia, adapting a methodology traditionally developed for temperate, Western landscapes to an arid, culturally distinct context. While in the UK and Europe LCA frameworks are primarily based on vegetation, topography, and settlement patterns, this research modifies the approach to reflect the unique ecological and geographical characteristics of Saudi Arabia.
A mixed-methods approach was employed comprising:
1. Desk Study: Review of international LCA frameworks and regional studies to develop a Saudi-specific classification system.
2. Site Investigation: Field investigations were conducted to support both the national-scale LCA of Saudi Arabia (Chapter 4) and the local assessment within PMBSRR, including Wadi Al Disah, presented in Chapter 5.
3. Questionnaire Survey (153 participants): Assessment of public perceptions of ecotourism potential in Wadi Al Disah and PMBSRR.
4. Semi-Structured Interviews (21 participants): Engagement with key stakeholders and government representatives to explore the potential role of LCA in planning policy and ecotourism development.
Findings demonstrate that LCA can be effectively adapted for arid environments and that the resulting landscape classifications provide an evidence base for conservation and sustainable tourism planning. At the national level, the research highlights inconsistencies between existing protected areas and landscape character types, while the local-scale analysis identifies zones suitable for ecotourism development based on landscape sensitivity and distinctiveness.
This work provides a framework for landscape classification in Saudi Arabia and offers a methodological foundation for future planning, aligning with national goals under Vision 2030. The study underscores the broader potential for applying LCA in Gulf countries with similar climatic and geographic contexts
The Impact of Physicochemical Properties of Formulation Ingredients on Drug Product Flow and Compaction
Full text linkSuccessful tablet manufacturing is greatly dependent on good flow and compaction behaviour of the active pharmaceutical ingredient (API) and excipients. These are governed by the physicochemical and mechanical properties of the materials. Like most organic crystalline materials, APIs are highly anisotropic, with weak van der Waals interactions often yielding needle-like, brittle crystals with poor flow and compaction. Additionally, the API-excipient ratio further impacts the processability and quality of the final tablet. Studying multi-component systems can prove challenging as API-excipient interactions are complex, and decoupling interparticle cohesion-adhesion and their impact on flow and compaction can be difficult to determine. This is especially true for modern formulations with low-dose excipients. Robust predictive models for formulation optimisation and processability remain limited. Thus, the need to utilise tools which enhance the understanding of how composition affects tabletability, ultimately optimising the manufacturing process.
This PhD combines computational and experimental methods to investigate the interplay between physicochemical and mechanical properties of the drug mefenamic acid (MA) and excipient d-mannitol (DM), and their impact on flow and compaction.
Molecular modelling successfully predicted the thin platy morphology of MA, driven by its anisotropic hydrogen bonding at its capping faces, and the columnar prismatic particles of DM, due to its homogenous distribution of -OH interactions. Predictions of their surface interactions revealed MA to have stronger cohesivity driven by strong dispersive interactions (93.53 %) and DM stronger adhesivity, displaying a better balance of dispersive to polar interactions (75.29 % and 24.71 % respectively). Prediction of their mechanical properties revealed both compounds to be brittle in nature; however, DM's extensive h-bond network allows for the occurrence of plastic deformation.
The predictions correlated well with experimental results. MA displayed poor flow (HR: 1.37, AOR: 44.55°), no plasticity (Py: 357.14 MPa), high porosity (E: 0.22) and low tensile strength (0.095 MPa). On the contrary, DM displayed good flow (HR:1.13, AOR 39.20), good compressibility (E: 0.03) and plasticity (Py: 65.79 MPa) and high tensile strength (1.07 MPa). Binary mixtures of 50-50, 65-35 and 75-25 MA to DM were examined in the same manner. The 50-50 blend provided the best balance between flow, compressibility and tensile strength. Whilst increasing the amount of MA, the blend's properties increasingly resembled those of pure MA. The findings were further validated through X-ray computed tomography (XCT), where powder flow, consolidation patterns, tablet compressibility and particle orientation were examined
Data sources:
Access Repository Dashboard
Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇