7103 research outputs found
Sort by
End-to-end optimization of lipid nanoparticle manufacturing for mRNA delivery
Full text linkWith the emergence of SARS-CoV-2 in 2020, mRNA vaccines have gained global attention. Currently,
lipid nanoparticles (LNPs) are the most clinically advanced drug delivery system for the delivery of
nucleic acids. Despite the extensive literature on LNPs in recent years, challenges persist regarding
their development and production. In fact, most research papers focus on the therapeutic targets of
LNPs, while less attention is given to understanding the challenges associated with their
manufacturing, especially on an industrial-production scale, including scalability, reproducibility,
encapsulation efficiency and long-term storage.
This thesis focused on the end-to-end workflow of LNPs manufacturing, covering production,
purification, and freeze-drying, while also addressing storage conditions. Beginning with LNP
production, the effects of microfluidic parameters on LNP manufacturing were investigated while the
preclinical scalable production of LNPs using various microfluidic devices was also evaluated. Moving
on to purification, the second step of LNP manufacturing, the typical bottlenecks associated with this
stage were assessed, with a focus on tangential flow filtration (TFF) as this method is commonly used
on an industrial level. The effect of TFF speed and diafiltration volumes on LNPs characteristics were
evaluated, along with the challenges related to scaling up the purification process. mRNA LNPs storage
also represents a challenge due to the fragile nature of mRNA. With the aim of exploring lyophilisation
as a technique for preserving mRNA LNPs, a series of freeze-drying cycles were conducted to identify
the optimal parameters for producing mRNA LNPs with acceptable critical quality attributes (CQAs)
and the in vitro and in vivo activity of the lyophilised product was evaluated to determine the
effectiveness of the method. This thesis also explored the role of lipid selection in shaping the quality,
stability, and performance of the final product. In particular, the contribution of PEGylated lipids having
different alkyl chain lengths (DMG-PEG 2000 versus DSG-PEG 2000) to the physicochemical
characteristics and performance of mRNA LNPs was investigated, as well as the impact in vitro and in
vivo of the ionisable lipid (ALC-0315, DLin-MC3, and SM-102).
The results presented demonstrate that all steps of LNP manufacturing influence the CQAs of the
particles, from the choice of lipids, which can either limit or enhance their efficiency, to the selection
of microfluidic parameters, buffers, purification methods, and lyophilisation conditions, highlighting
the importance of carefully considering each individual step.With the emergence of SARS-CoV-2 in 2020, mRNA vaccines have gained global attention. Currently,
lipid nanoparticles (LNPs) are the most clinically advanced drug delivery system for the delivery of
nucleic acids. Despite the extensive literature on LNPs in recent years, challenges persist regarding
their development and production. In fact, most research papers focus on the therapeutic targets of
LNPs, while less attention is given to understanding the challenges associated with their
manufacturing, especially on an industrial-production scale, including scalability, reproducibility,
encapsulation efficiency and long-term storage.
This thesis focused on the end-to-end workflow of LNPs manufacturing, covering production,
purification, and freeze-drying, while also addressing storage conditions. Beginning with LNP
production, the effects of microfluidic parameters on LNP manufacturing were investigated while the
preclinical scalable production of LNPs using various microfluidic devices was also evaluated. Moving
on to purification, the second step of LNP manufacturing, the typical bottlenecks associated with this
stage were assessed, with a focus on tangential flow filtration (TFF) as this method is commonly used
on an industrial level. The effect of TFF speed and diafiltration volumes on LNPs characteristics were
evaluated, along with the challenges related to scaling up the purification process. mRNA LNPs storage
also represents a challenge due to the fragile nature of mRNA. With the aim of exploring lyophilisation
as a technique for preserving mRNA LNPs, a series of freeze-drying cycles were conducted to identify
the optimal parameters for producing mRNA LNPs with acceptable critical quality attributes (CQAs)
and the in vitro and in vivo activity of the lyophilised product was evaluated to determine the
effectiveness of the method. This thesis also explored the role of lipid selection in shaping the quality,
stability, and performance of the final product. In particular, the contribution of PEGylated lipids having
different alkyl chain lengths (DMG-PEG 2000 versus DSG-PEG 2000) to the physicochemical
characteristics and performance of mRNA LNPs was investigated, as well as the impact in vitro and in
vivo of the ionisable lipid (ALC-0315, DLin-MC3, and SM-102).
The results presented demonstrate that all steps of LNP manufacturing influence the CQAs of the
particles, from the choice of lipids, which can either limit or enhance their efficiency, to the selection
of microfluidic parameters, buffers, purification methods, and lyophilisation conditions, highlighting
the importance of carefully considering each individual step
Contextual comparing and constraining : a grounded theory study of undergraduate engineering group design projects
Full text linkEngineering design projects are significant components of engineering undergraduate students’ programmes and can be transformative in developing important professional competencies such as communication and teamwork skills.
This study contributes to the limited knowledge in qualitative research into undergraduate engineering design projects by exploring the questions: what are the prominent concerns of engineering students undertaking such design projects and how do they manage these? The study involved an exploration of undergraduate Chemical Engineering students undertaking a substantial design project at the University of Strathclyde.
Student groups were observed in supervisory meetings to sensitise the researcher and accompany the primary data collection method of intensive interviews. Informing the approach to the study was Grounded Theory (GT) methods for analysis, these involved techniques such as initial (line-by-line) coding; memo-writing; focused coding and theoretical sampling. Following extensive coding and categorisation of concepts, data from other studies in engineering design projects conducted by Goncher (2012) and Morgan (2017) were introduced and integrated into the emerging understanding.
This research demonstrated that students’ beliefs around the significance of the design projects; their socialisation with others and their expectations of the educational setting impacted their approach to design significantly. Students engaged in social processes related to comparing and constraining to manage a range of contextual factors. Comparing involved evaluating various design ideas and solutions to make sense of open-ended problems; while constraining refers to the methods used to limit the design space and context to reduce ambiguity. These processes also relate to both social dynamics (such as collaborative efforts and information sharing) and technical factors (like minimising risks and relying on models). Students use such strategies to refine their designs iteratively and align with project requirements to constrain their emergent designs. Supervisory feedback also plays a crucial role in guiding these processes, with effective feedback helping to constrain creative exploration within feasible boundaries.Engineering design projects are significant components of engineering undergraduate students’ programmes and can be transformative in developing important professional competencies such as communication and teamwork skills.
This study contributes to the limited knowledge in qualitative research into undergraduate engineering design projects by exploring the questions: what are the prominent concerns of engineering students undertaking such design projects and how do they manage these? The study involved an exploration of undergraduate Chemical Engineering students undertaking a substantial design project at the University of Strathclyde.
Student groups were observed in supervisory meetings to sensitise the researcher and accompany the primary data collection method of intensive interviews. Informing the approach to the study was Grounded Theory (GT) methods for analysis, these involved techniques such as initial (line-by-line) coding; memo-writing; focused coding and theoretical sampling. Following extensive coding and categorisation of concepts, data from other studies in engineering design projects conducted by Goncher (2012) and Morgan (2017) were introduced and integrated into the emerging understanding.
This research demonstrated that students’ beliefs around the significance of the design projects; their socialisation with others and their expectations of the educational setting impacted their approach to design significantly. Students engaged in social processes related to comparing and constraining to manage a range of contextual factors. Comparing involved evaluating various design ideas and solutions to make sense of open-ended problems; while constraining refers to the methods used to limit the design space and context to reduce ambiguity. These processes also relate to both social dynamics (such as collaborative efforts and information sharing) and technical factors (like minimising risks and relying on models). Students use such strategies to refine their designs iteratively and align with project requirements to constrain their emergent designs. Supervisory feedback also plays a crucial role in guiding these processes, with effective feedback helping to constrain creative exploration within feasible boundaries
The political determinants of public infrastructure investment in Middle Income Countries : the visibility and targetability of white elephants
Full text linkMiddle Income Countries are often stuck in a trap, due to the political distortions that
drive their allocation of investment. This dissertation explores these political
determinants of public infrastructure investment in Middle-Income Countries [MICs],
focusing on how political competition drives the allocation of investment. I argue that
insecure incumbents in MICs, prioritize investments in economic infrastructure over
social infrastructure due to their higher visibility and targetability, which make them
politically valuable. Theoretically, when accountability is low, this misallocation can lead
to inefficient, wasteful investments and "White Elephants" that hinder development. By
adopting a political economy approach, the dissertation disaggregates investment
categories by function to examine the visibility and targetability mechanisms and their
distortionary effects on resource allocation. Using empirical analysis of MICs and a
detailed case study, this dissertation shows how low horizontal and vertical
accountability mean that politics can distort the allocation of investment, resulting in
overinvestment in potentially wasteful-albeit politically valuable- economic infrastructure
at the expense of -often-more necessary social investments. The findings contribute to
understanding, not only why MICs overinvest in economic infrastructure, but also why
investment in MICs often fails to translate into growth, providing critical insights into the
political dynamics behind resource misallocation and its implications for escaping the
MIT.Middle Income Countries are often stuck in a trap, due to the political distortions that
drive their allocation of investment. This dissertation explores these political
determinants of public infrastructure investment in Middle-Income Countries [MICs],
focusing on how political competition drives the allocation of investment. I argue that
insecure incumbents in MICs, prioritize investments in economic infrastructure over
social infrastructure due to their higher visibility and targetability, which make them
politically valuable. Theoretically, when accountability is low, this misallocation can lead
to inefficient, wasteful investments and "White Elephants" that hinder development. By
adopting a political economy approach, the dissertation disaggregates investment
categories by function to examine the visibility and targetability mechanisms and their
distortionary effects on resource allocation. Using empirical analysis of MICs and a
detailed case study, this dissertation shows how low horizontal and vertical
accountability mean that politics can distort the allocation of investment, resulting in
overinvestment in potentially wasteful-albeit politically valuable- economic infrastructure
at the expense of -often-more necessary social investments. The findings contribute to
understanding, not only why MICs overinvest in economic infrastructure, but also why
investment in MICs often fails to translate into growth, providing critical insights into the
political dynamics behind resource misallocation and its implications for escaping the
MIT
Study of long-distance high-temperature superconductor cables for HVDC power transmission
Full text linkThe carbon neutrality goal of achieving net zero by 2050 has sparked significant interest
in offshore wind farms. With offshore wind farms being installed from short distances to
long distances away from the seashore, a submarine power transmission infrastructure is
a necessity. Considering power cable length and low net effective losses, a high-voltage
direct-current (HVDC) system has been chosen over HVAC. HVDC cables are considered
to be made out of high-temperature superconductor (HTS) material. HTS power cables
hold immense potential for efficient, low-loss, high-current-density, and compact power
transmission. But, unlike copper/aluminium, the HTS material has a sharp transient
behaviour as a function of the operating current, temperature, and magnetic field.
However, the susceptibility of HTS cables to faults in the grid, resulting in quenching or
permanent damage due to joule heating, poses a critical challenge to their real-world
resilience.
The main aim of this thesis is to study and evaluate the long-distance HTS subsea power
cable for offshore to onshore power transmission along the radial and axial direction of
the 2G HTS power cable in terms of electrical, thermal, hydraulic and quench
characteristics during the normal and grid fault conditions. Prior to the installation of longdistance cable experiments and lab prototypes, modelling and simulation work is needed
to predict cable behaviour. Furthermore, modelling the HTS power cable was necessary
as part of enhancing cable performance and design. To tackle the problem described
above, a computationally efficient high-fidelity model over the whole length of the cable
was needed to understand HTS cable characteristics. The complex design of the
superconductor power cable was modelled using Finite Element Method (FEM).
However, for long-distance superconductor electrical-thermal-hydraulic studies, the FEM simulation involves a computational burden and for the electrical network, it is not
suitable.
Novel discretised electrical-thermal-hydraulic SIMSCAPE components with non-linear
resistive behaviour dependent on current and temperature in MATLAB/SIMSCAPE
software were modelled, partitioning the cable into discrete blocks to understand the
temperature along the axial length and to determine the impact of transient conditions on
a long-distance superconducting power cable in transmission. The model has the
advantage of having the flexibility to change the fault location and cryocooler spacing
length along the length of the cable, including the HTS and copper former and LN2 layers.
The experiment measured the joint resistance of the tape incorporated into the model. This
study is unique in that it is the first to look at 100 km long-distance HVDC subsea HTS
for offshore to onshore grid use capable of transmitting more than 1 GW of power at 100
kV voltage and an ampacity of 10 kA. The primary technical challenges in modelling this
long-distance HVDC subsea HTS cable are due to its non-linear resistive properties,
which vary with temperature, current, and magnetic field. These variations result in the
cable's electrical and thermal behaviour changes along its length, complicating accurate
modelling. Distributed RLC network models are implemented using Simscape cable
blocks to address these complexities. Each Simscape block includes thermal, electrical,
and hydraulic parameters of respective cable layers, considering conduction, convection,
radiation heat losses, non-linear resistance, critical current, and pressure drop and
temperature. It allows the integration of cryocoolers at specific points along the cable. For this study, an HVDC HTS power cable was considered and modelled both as a
lumped element and as a distributed element model with 100 elements to compare and
evaluate the cable parameters along the length. The parameters include temperature
distribution, resistance, critical current, current, and losses at different spots throughout
the length of the cable. To simulate the transient condition, a pole-to-ground (PG) fault is
considered and the current distribution between the copper former and HTS tapes is
studied. Using this cable model, the maximum temperature of the HTS and coolant both
in the superconducting state and transient state has been evaluated and presented.
This research is the first to develop a 2D T-A of a subsea HTS cable with a 10-kA
transport current to investigate the electromagnetic behaviour of the superconductor and
the impact of the harmonics generated in the AC/DC converters on the cable. The crosssectional area of the HTS cable was modelled in 2D using COMSOL Multiphysics to
examine harmonic ripple losses in the cable. These findings were further validated through
experimental investigations on the superconductor tapes.
This study underscores the benefits of integrating Superconducting Fault Current
Limiters (SFCL) with HTS cables in the network, showcasing load sharing between the
superconductor and copper former during steady and transient state operations, HTS
quench and recovery time.The carbon neutrality goal of achieving net zero by 2050 has sparked significant interest
in offshore wind farms. With offshore wind farms being installed from short distances to
long distances away from the seashore, a submarine power transmission infrastructure is
a necessity. Considering power cable length and low net effective losses, a high-voltage
direct-current (HVDC) system has been chosen over HVAC. HVDC cables are considered
to be made out of high-temperature superconductor (HTS) material. HTS power cables
hold immense potential for efficient, low-loss, high-current-density, and compact power
transmission. But, unlike copper/aluminium, the HTS material has a sharp transient
behaviour as a function of the operating current, temperature, and magnetic field.
However, the susceptibility of HTS cables to faults in the grid, resulting in quenching or
permanent damage due to joule heating, poses a critical challenge to their real-world
resilience.
The main aim of this thesis is to study and evaluate the long-distance HTS subsea power
cable for offshore to onshore power transmission along the radial and axial direction of
the 2G HTS power cable in terms of electrical, thermal, hydraulic and quench
characteristics during the normal and grid fault conditions. Prior to the installation of longdistance cable experiments and lab prototypes, modelling and simulation work is needed
to predict cable behaviour. Furthermore, modelling the HTS power cable was necessary
as part of enhancing cable performance and design. To tackle the problem described
above, a computationally efficient high-fidelity model over the whole length of the cable
was needed to understand HTS cable characteristics. The complex design of the
superconductor power cable was modelled using Finite Element Method (FEM).
However, for long-distance superconductor electrical-thermal-hydraulic studies, the FEM simulation involves a computational burden and for the electrical network, it is not
suitable.
Novel discretised electrical-thermal-hydraulic SIMSCAPE components with non-linear
resistive behaviour dependent on current and temperature in MATLAB/SIMSCAPE
software were modelled, partitioning the cable into discrete blocks to understand the
temperature along the axial length and to determine the impact of transient conditions on
a long-distance superconducting power cable in transmission. The model has the
advantage of having the flexibility to change the fault location and cryocooler spacing
length along the length of the cable, including the HTS and copper former and LN2 layers.
The experiment measured the joint resistance of the tape incorporated into the model. This
study is unique in that it is the first to look at 100 km long-distance HVDC subsea HTS
for offshore to onshore grid use capable of transmitting more than 1 GW of power at 100
kV voltage and an ampacity of 10 kA. The primary technical challenges in modelling this
long-distance HVDC subsea HTS cable are due to its non-linear resistive properties,
which vary with temperature, current, and magnetic field. These variations result in the
cable's electrical and thermal behaviour changes along its length, complicating accurate
modelling. Distributed RLC network models are implemented using Simscape cable
blocks to address these complexities. Each Simscape block includes thermal, electrical,
and hydraulic parameters of respective cable layers, considering conduction, convection,
radiation heat losses, non-linear resistance, critical current, and pressure drop and
temperature. It allows the integration of cryocoolers at specific points along the cable. For this study, an HVDC HTS power cable was considered and modelled both as a
lumped element and as a distributed element model with 100 elements to compare and
evaluate the cable parameters along the length. The parameters include temperature
distribution, resistance, critical current, current, and losses at different spots throughout
the length of the cable. To simulate the transient condition, a pole-to-ground (PG) fault is
considered and the current distribution between the copper former and HTS tapes is
studied. Using this cable model, the maximum temperature of the HTS and coolant both
in the superconducting state and transient state has been evaluated and presented.
This research is the first to develop a 2D T-A of a subsea HTS cable with a 10-kA
transport current to investigate the electromagnetic behaviour of the superconductor and
the impact of the harmonics generated in the AC/DC converters on the cable. The crosssectional area of the HTS cable was modelled in 2D using COMSOL Multiphysics to
examine harmonic ripple losses in the cable. These findings were further validated through
experimental investigations on the superconductor tapes.
This study underscores the benefits of integrating Superconducting Fault Current
Limiters (SFCL) with HTS cables in the network, showcasing load sharing between the
superconductor and copper former during steady and transient state operations, HTS
quench and recovery time
Isotope hydrology and hydrogeochemistry of surface water and groundwater in the Lake Malawi basin to support sustainable water resource management
No full textThis thesis presents groundbreaking body of research aimed at examining meteoric water inputs and their interactions with surface water and groundwater systems to enhance understanding
of Lake Malawi basin’s hydrological processes to provide transformative insights into the basin's water cycle. This was achieved through an innovative multi-method approach that integrated stable isotopic and hydrogeochemical tracers. The thesis adopted an integrated-holistic methodological framework that considered precipitation, groundwater and surface water as interconnected systems in the basin’s water cycle. The aim and objectives of the research were achieved through a series of
scientific publications, forming a significant body of research with critical-practical and policy implications. The thesis successfully established a baseline for stable isotopes in precipitation, a
fundamental benchmark for future isotopic studies in Malawi and the wider region. This foundational isotopic baseline provided groundbreaking isotopic insights into Malawi’s water cycle
by establishing a pioneering Local Meteoric Water Line, consistent with the Global Meteoric Water Line. The thesis revealed predominant oceanic and continental influences marked by moisture recycling likely from the African Great Lakes’ large scale evaporation effect and seasonal shifts in
the Intertropical Convergence Zone. The thesis provides increased and unique insights into origin of groundwater and geochemical evolution dynamics within the Lake Malawi basin’s aquifer
systems. The thesis demonstrates primary influence of non-evaporated modern precipitation on groundwater recharge, initiating critical geochemical reactions with host rock minerals upon
infiltration into the subsurface zone. The rock mineral-water reactions release essential ionic species and form secondary minerals, with sodium and calcium ions being predominant cations, thereby influencing groundwater chemistry. Through geochemical modelling insights, the thesis clarified that groundwater evolution is dominated by silicate dissolution reactions, highlighting critical relationships between meteoric water inputs, evaporative processes, and spatial-seasonal variations in groundwater hydrochemistry. The thesis further revealed distinct hydrogeochemical signatures and widespread nitrate occurrence, the latter underscoring potential human impacts on groundwater
quality. The thesis established a basin-scale isotopic and hydrochemical baseline characterization of surface water using isotopic and chemical tracers. It showed that surface water flows are primarily influenced by non-evaporated modern precipitation and groundwater baseflow inputs, with distinct seasonal variability. Riverine flows exhibited greater isotopic variability than lakes
and reservoirs, with predominant wet season depleted signals at higher elevations and dry season enriched signals in lowlands. Convergence of highland-plateau and rift margin river systems
resulted in isotopic depletion within rift valley escarpment zones and progressive enrichment through lakeshore flood plains. The thesis highlights significant contribution of local precipitation to surface water flows and retention of isotopic enrichment signals in lakes likely due to high residence time. The thesis aligns with Malawi's Water Policy, Sustainable Development Goal 6, Water Resources Act-2013, and 2063 Development Agenda, advocating for expanded monitoring networks, cross-disciplinary collaborations, and community engagement. It also supports the objectives of GloWAL Network, IWAVE, GNIP, and GNIR, emphasizing regional collaboration and comprehensive data collection and sharing. Overall, this thesis provides new and unique knowledge
critical for advancing understanding of Lake Malawi basin’s hydrological cycle. It serves as a call to action for evidence-based sustainable water management, underscoring the importance of
innovative, interdisciplinary, and participatory approaches. The thesis outcomes are expected to inform and guide development of sustainable water resource management strategies that aligns with
objectives of SDG 6 and IWRM practices. This thesis delivers a new and unique body of research with critical policy, practical and scholarly implications.This thesis presents groundbreaking body of research aimed at examining meteoric water inputs and their interactions with surface water and groundwater systems to enhance understanding
of Lake Malawi basin’s hydrological processes to provide transformative insights into the basin's water cycle. This was achieved through an innovative multi-method approach that integrated stable isotopic and hydrogeochemical tracers. The thesis adopted an integrated-holistic methodological framework that considered precipitation, groundwater and surface water as interconnected systems in the basin’s water cycle. The aim and objectives of the research were achieved through a series of
scientific publications, forming a significant body of research with critical-practical and policy implications. The thesis successfully established a baseline for stable isotopes in precipitation, a
fundamental benchmark for future isotopic studies in Malawi and the wider region. This foundational isotopic baseline provided groundbreaking isotopic insights into Malawi’s water cycle
by establishing a pioneering Local Meteoric Water Line, consistent with the Global Meteoric Water Line. The thesis revealed predominant oceanic and continental influences marked by moisture recycling likely from the African Great Lakes’ large scale evaporation effect and seasonal shifts in
the Intertropical Convergence Zone. The thesis provides increased and unique insights into origin of groundwater and geochemical evolution dynamics within the Lake Malawi basin’s aquifer
systems. The thesis demonstrates primary influence of non-evaporated modern precipitation on groundwater recharge, initiating critical geochemical reactions with host rock minerals upon
infiltration into the subsurface zone. The rock mineral-water reactions release essential ionic species and form secondary minerals, with sodium and calcium ions being predominant cations, thereby influencing groundwater chemistry. Through geochemical modelling insights, the thesis clarified that groundwater evolution is dominated by silicate dissolution reactions, highlighting critical relationships between meteoric water inputs, evaporative processes, and spatial-seasonal variations in groundwater hydrochemistry. The thesis further revealed distinct hydrogeochemical signatures and widespread nitrate occurrence, the latter underscoring potential human impacts on groundwater
quality. The thesis established a basin-scale isotopic and hydrochemical baseline characterization of surface water using isotopic and chemical tracers. It showed that surface water flows are primarily influenced by non-evaporated modern precipitation and groundwater baseflow inputs, with distinct seasonal variability. Riverine flows exhibited greater isotopic variability than lakes
and reservoirs, with predominant wet season depleted signals at higher elevations and dry season enriched signals in lowlands. Convergence of highland-plateau and rift margin river systems
resulted in isotopic depletion within rift valley escarpment zones and progressive enrichment through lakeshore flood plains. The thesis highlights significant contribution of local precipitation to surface water flows and retention of isotopic enrichment signals in lakes likely due to high residence time. The thesis aligns with Malawi's Water Policy, Sustainable Development Goal 6, Water Resources Act-2013, and 2063 Development Agenda, advocating for expanded monitoring networks, cross-disciplinary collaborations, and community engagement. It also supports the objectives of GloWAL Network, IWAVE, GNIP, and GNIR, emphasizing regional collaboration and comprehensive data collection and sharing. Overall, this thesis provides new and unique knowledge
critical for advancing understanding of Lake Malawi basin’s hydrological cycle. It serves as a call to action for evidence-based sustainable water management, underscoring the importance of
innovative, interdisciplinary, and participatory approaches. The thesis outcomes are expected to inform and guide development of sustainable water resource management strategies that aligns with
objectives of SDG 6 and IWRM practices. This thesis delivers a new and unique body of research with critical policy, practical and scholarly implications
Development of targeted dendrimer-based nanomedicines for gene delivery to the brain following intravenous administration
No full textLeveraging machine learning for financial forecasting : a dual approach for meme stock price and GDP prediction
Full text linkThis dissertation explores the adoption of machine learning in financial analysis through a systematic investigation of stock price prediction and GDP forecasting. The first study examines the relationship between online sentiment and meme stock price movements. This is done by conducting sentiment analysis on Reddit’s WallStreetBets discussion thread and identifying daily popular stock tickers and their sentiment scores. The study then uses several recurrent neural networks, including variations of Long Short-Term Memory (LSTM) models such as single-layered LSTM, regular stacked LSTM, bidirectional LSTM model and Gated Recurrent Unit (GRU) model. These models were trained on five years of historical and technical data, highlighting the impact of online sentiment on Meme stock price fluctuations and the potential of AI-driven models to capture these dynamics. Models were tested for real-time applications for three consecutive days. Results demonstrated that the single-layered LSTM model outperformed other models with low error rates. For example, NVDA with average RMSE: 4.64, MAE: 3.38, MAPE: 0.035, and similar performance observed for ASTS with average RMSE 2.03, MAE: 0.92, MAPE 0.091 and LUNR (RMSE: 0.41, MAE 0.28, MAPE 0.066). However, by the third day regular stacked LSTM model slightly outperformed for NVDA, while single-layered LSTM dominated with better predictive power for other stocks (SMCI – RMSE 112.726, MAE: 86.946, MAPE: 0.111; AI – RMSE: 1.6, MAE: 1.197, MAPE: 0.036). The second study extends the use of AI in macroeconomic forecasting, focusing on the prediction of the GDP of the United Kingdom (UK) using vital macroeconomic variables, including energy prices, unemployment rate, inflation, net migration and Real Effective Exchange Rate (REER) from 1990-2018. For the prediction, several machine learning models, such as Support Vector Regression (SVR), Random Forest (RF) and Gradient Boosting Machines (GBM), were implemented and compared together with Shapley Additive exPlanantions (SHAP). The models were assessed using evaluation metrics such as Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and R2 score. The findings underscored the significant role of macroeconomic variables in economic forecasting and illustrated the potential of AI-driven models to provide valuable insight into financial markets and economic indicators. Among them scaled SVR model achieved best performance with RMSE: 83,492. 048, MAE: 77, 219.274, MAPE: 4.2% and R2 score of 0.042. Together, these studies demonstrate the adoptability and potential of machine learning in addressing complex financial and economic prediction tasks and underline practical implications. The integration of sentiment analysis for stock price prediction and macroeconomic modelling for GDP forecasting showcases machine learning’s ability to handle diverse data types, from unstructured textual data in online platforms to structured economic indicators. By combining these approaches, the research highlights how AI can uncover hidden patterns and relationships that traditional financial models might overlook, providing a more nuanced understanding of market behaviors and economic trends.This dissertation explores the adoption of machine learning in financial analysis through a systematic investigation of stock price prediction and GDP forecasting. The first study examines the relationship between online sentiment and meme stock price movements. This is done by conducting sentiment analysis on Reddit’s WallStreetBets discussion thread and identifying daily popular stock tickers and their sentiment scores. The study then uses several recurrent neural networks, including variations of Long Short-Term Memory (LSTM) models such as single-layered LSTM, regular stacked LSTM, bidirectional LSTM model and Gated Recurrent Unit (GRU) model. These models were trained on five years of historical and technical data, highlighting the impact of online sentiment on Meme stock price fluctuations and the potential of AI-driven models to capture these dynamics. Models were tested for real-time applications for three consecutive days. Results demonstrated that the single-layered LSTM model outperformed other models with low error rates. For example, NVDA with average RMSE: 4.64, MAE: 3.38, MAPE: 0.035, and similar performance observed for ASTS with average RMSE 2.03, MAE: 0.92, MAPE 0.091 and LUNR (RMSE: 0.41, MAE 0.28, MAPE 0.066). However, by the third day regular stacked LSTM model slightly outperformed for NVDA, while single-layered LSTM dominated with better predictive power for other stocks (SMCI – RMSE 112.726, MAE: 86.946, MAPE: 0.111; AI – RMSE: 1.6, MAE: 1.197, MAPE: 0.036). The second study extends the use of AI in macroeconomic forecasting, focusing on the prediction of the GDP of the United Kingdom (UK) using vital macroeconomic variables, including energy prices, unemployment rate, inflation, net migration and Real Effective Exchange Rate (REER) from 1990-2018. For the prediction, several machine learning models, such as Support Vector Regression (SVR), Random Forest (RF) and Gradient Boosting Machines (GBM), were implemented and compared together with Shapley Additive exPlanantions (SHAP). The models were assessed using evaluation metrics such as Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and R2 score. The findings underscored the significant role of macroeconomic variables in economic forecasting and illustrated the potential of AI-driven models to provide valuable insight into financial markets and economic indicators. Among them scaled SVR model achieved best performance with RMSE: 83,492. 048, MAE: 77, 219.274, MAPE: 4.2% and R2 score of 0.042. Together, these studies demonstrate the adoptability and potential of machine learning in addressing complex financial and economic prediction tasks and underline practical implications. The integration of sentiment analysis for stock price prediction and macroeconomic modelling for GDP forecasting showcases machine learning’s ability to handle diverse data types, from unstructured textual data in online platforms to structured economic indicators. By combining these approaches, the research highlights how AI can uncover hidden patterns and relationships that traditional financial models might overlook, providing a more nuanced understanding of market behaviors and economic trends
Optimisation of large-scale offshore wind farms considering turbine layout, cable layout, and co-located energy storage systems
Full text linkThis thesis aims to explore two ways in which the electrical infrastructure of a gigawattscale offshore wind farm may be optimised to reduce costs and aid the increasing deployment of renewable energy capacity. The two areas considered are (1) the integration of
a cable layout optimisation with a turbine layout optimisation forming a novel concurrent optimisation framework, and (2) the integration of energy storage systems into a
cable layout optimisation for the purpose of peak-shaving power in the cables enabling
alternative cable sizes to be used and reducing peak electrical losses.
While there are a wide variety of approaches to turbine layout optimisation, the
particle swarm optimisation (PSO) method offers a good balance between accuracy and
computational expense, enabling large-scale problems to be handled by standard desktop PCs. The turbine layout problem is formulated as a grid-based layout, fully defined
by eight variables, to comply with maritime navigation and search and rescue regulations, while allowing some deviation to maximise energy capture. The eight variables
defining the grid are optimised by means of PSO, followed by a novel micro-siting function to move individual turbines and increase energy yield. The method was compared
to SSEs in-house method, matching the energy capture of a case study of their Berwick
Bank site to within 0.3%.
Two cable layout optimisation methods from the literature are selected, which are
the widely used mixed-integer linear programming (MILP) method, and ant colony optimisation (ACO) representing the increasing use of heuristic approaches. These are
compared to a novel ACO-based method, ACOsp, that employs a decomposition strategy. The ACOsp method is shown to maintain the good quality solutions of the MILP
approach, with solutions 0.0-1.4% more expensive than optimal, while also demonstratii
ing the computational efficiency of heuristic approaches, useful for large-scale problems.
An optimisation framework considering the concurrent optimisation of turbine and
cable layouts is proposed, with comparison made to a sequentially optimised solutions,
isolating the impact of this integration. Solutions of the integrated, concurrent, approach show improved objective values where the increase is statistically significant.
For a case study site with 164-165 turbines, the approach increases the objective value
(for this maximisation problem) by 0.45%, which is slightly less than the increase found
by the addition of one further turbine at 0.55-0.57%.
Considering the limitations of the investigated cable layout optimisation approaches,
a following study proposed a MILP-based optimisation in combination with a decomposition strategy, MILPsp. The MILPsp method maintained the accuracy of the MILP
method and reduced computational expense, improving on the earlier ACOsp method.
Variables describing ESS are integrated into the MILPsp algorithm to determine the
impact of using co-located ESS on the array cable network. It was found that very few
charging strategies were able to deliver meaningful peak shaving to the power in the array cables and those that did required a very large ESS capacity to do so (3MW/64MWh
for a site using 8MW turbines). Further, the required cost of the co-located ESS was
prohibitively low, at <£1,800, compared to real ESS prices at the time of writing.
Ignoring cost restrictions, using ESS within the cable layout optimisation, for a site
containing 122 8MW turbines, was able to reduce cable network costs by 0.22-1.85%.This thesis aims to explore two ways in which the electrical infrastructure of a gigawattscale offshore wind farm may be optimised to reduce costs and aid the increasing deployment of renewable energy capacity. The two areas considered are (1) the integration of
a cable layout optimisation with a turbine layout optimisation forming a novel concurrent optimisation framework, and (2) the integration of energy storage systems into a
cable layout optimisation for the purpose of peak-shaving power in the cables enabling
alternative cable sizes to be used and reducing peak electrical losses.
While there are a wide variety of approaches to turbine layout optimisation, the
particle swarm optimisation (PSO) method offers a good balance between accuracy and
computational expense, enabling large-scale problems to be handled by standard desktop PCs. The turbine layout problem is formulated as a grid-based layout, fully defined
by eight variables, to comply with maritime navigation and search and rescue regulations, while allowing some deviation to maximise energy capture. The eight variables
defining the grid are optimised by means of PSO, followed by a novel micro-siting function to move individual turbines and increase energy yield. The method was compared
to SSEs in-house method, matching the energy capture of a case study of their Berwick
Bank site to within 0.3%.
Two cable layout optimisation methods from the literature are selected, which are
the widely used mixed-integer linear programming (MILP) method, and ant colony optimisation (ACO) representing the increasing use of heuristic approaches. These are
compared to a novel ACO-based method, ACOsp, that employs a decomposition strategy. The ACOsp method is shown to maintain the good quality solutions of the MILP
approach, with solutions 0.0-1.4% more expensive than optimal, while also demonstratii
ing the computational efficiency of heuristic approaches, useful for large-scale problems.
An optimisation framework considering the concurrent optimisation of turbine and
cable layouts is proposed, with comparison made to a sequentially optimised solutions,
isolating the impact of this integration. Solutions of the integrated, concurrent, approach show improved objective values where the increase is statistically significant.
For a case study site with 164-165 turbines, the approach increases the objective value
(for this maximisation problem) by 0.45%, which is slightly less than the increase found
by the addition of one further turbine at 0.55-0.57%.
Considering the limitations of the investigated cable layout optimisation approaches,
a following study proposed a MILP-based optimisation in combination with a decomposition strategy, MILPsp. The MILPsp method maintained the accuracy of the MILP
method and reduced computational expense, improving on the earlier ACOsp method.
Variables describing ESS are integrated into the MILPsp algorithm to determine the
impact of using co-located ESS on the array cable network. It was found that very few
charging strategies were able to deliver meaningful peak shaving to the power in the array cables and those that did required a very large ESS capacity to do so (3MW/64MWh
for a site using 8MW turbines). Further, the required cost of the co-located ESS was
prohibitively low, at <£1,800, compared to real ESS prices at the time of writing.
Ignoring cost restrictions, using ESS within the cable layout optimisation, for a site
containing 122 8MW turbines, was able to reduce cable network costs by 0.22-1.85%
Employee wellness matters : exploring employee physical activity and related wellbeing initiatives in the workplace
No full textBackground: In recent years the focus on workplace wellbeing initiatives including physical activity has gained considerable attention. The movement has been driven by the increased awareness of the association between employee wellbeing and key organisation outcomes such as improved productivity, increased job satisfaction, lower attrition, and improved overall health. However, despite this growing area in the literature, little research has focused on how wellbeing leads who are tasked with implementing such initiatives understand their role, the challenges they face, and the factors that shape the success or failure of wellbeing initiatives. To address this gap, the present study aimed to explore the effectiveness of physical activity and related wellbeing initiatives and identify ways to support employee wellbeing in the future. Methods: Ethics approval was granted by the University of Strathclyde Department of Psychological Sciences and Health ethics committee. Participants were recruited by sending an email to a database of workplace wellbeing leads of an organisation that supports both organisational wellbeing in the United Kingdom. Participants were eligible if they were over 18 years old and self-identified as being responsible for supporting their organisation wellbeing initiatives and work for an organisation based in the United Kingdom. After obtaining consent, participants provided demographic information and were scheduled for 30-minute online interviews via the online platform Zoom. These interviews followed a semi-structured guide and were transcribed verbatim. A thematic analysis of the transcripts was then conducted, resulting in the identification of key themes and sub-themes. Results: From the thematic analysis of 12 participant transcripts, three main themes and six subthemes were identified. The main themes were organisational culture, support for the wellbeing leadership role, and health and wellbeing programme components. Within organisational culture, the sub-themes included the influence of managers and senior leadership on employee wellbeing, and barriers to employee participation in wellbeing initiatives. Under support for the wellbeing leadership role, the sub-themes were the wellbeing role appointment pathway, and the training provided for wellbeing leads. Lastly, the health and wellbeing programme components theme included physical activity interventions and future considerations to support employee wellbeing. Conclusions: Senior leadership and manager support is pivotal to ensuring employee wellbeing initiatives impact at employee level. Barriers to participation were linked to hybrid work patterns and remote staff. Health and wellbeing initiatives, particularly physical activity initiatives, varied, with few linked to organisation objectives. Wellbeing leads were often appointed the role based on tenure, current role or personal interest, but formal training in how to deliver effective wellbeing programmes and initiatives were limited. Recommendations for future research are to explore how wellbeing initiatives can be better embedded into organisations and how tailored interventions can address the needs of a diverse workforce. Recommendations for practice from this research call for organisations to provide formal training for wellbeing
leads to equip them with the necessary skills to implement effective wellbeing
programmes with a particular focus on behaviour change.Background: In recent years the focus on workplace wellbeing initiatives including physical activity has gained considerable attention. The movement has been driven by the increased awareness of the association between employee wellbeing and key organisation outcomes such as improved productivity, increased job satisfaction, lower attrition, and improved overall health. However, despite this growing area in the literature, little research has focused on how wellbeing leads who are tasked with implementing such initiatives understand their role, the challenges they face, and the factors that shape the success or failure of wellbeing initiatives. To address this gap, the present study aimed to explore the effectiveness of physical activity and related wellbeing initiatives and identify ways to support employee wellbeing in the future. Methods: Ethics approval was granted by the University of Strathclyde Department of Psychological Sciences and Health ethics committee. Participants were recruited by sending an email to a database of workplace wellbeing leads of an organisation that supports both organisational wellbeing in the United Kingdom. Participants were eligible if they were over 18 years old and self-identified as being responsible for supporting their organisation wellbeing initiatives and work for an organisation based in the United Kingdom. After obtaining consent, participants provided demographic information and were scheduled for 30-minute online interviews via the online platform Zoom. These interviews followed a semi-structured guide and were transcribed verbatim. A thematic analysis of the transcripts was then conducted, resulting in the identification of key themes and sub-themes. Results: From the thematic analysis of 12 participant transcripts, three main themes and six subthemes were identified. The main themes were organisational culture, support for the wellbeing leadership role, and health and wellbeing programme components. Within organisational culture, the sub-themes included the influence of managers and senior leadership on employee wellbeing, and barriers to employee participation in wellbeing initiatives. Under support for the wellbeing leadership role, the sub-themes were the wellbeing role appointment pathway, and the training provided for wellbeing leads. Lastly, the health and wellbeing programme components theme included physical activity interventions and future considerations to support employee wellbeing. Conclusions: Senior leadership and manager support is pivotal to ensuring employee wellbeing initiatives impact at employee level. Barriers to participation were linked to hybrid work patterns and remote staff. Health and wellbeing initiatives, particularly physical activity initiatives, varied, with few linked to organisation objectives. Wellbeing leads were often appointed the role based on tenure, current role or personal interest, but formal training in how to deliver effective wellbeing programmes and initiatives were limited. Recommendations for future research are to explore how wellbeing initiatives can be better embedded into organisations and how tailored interventions can address the needs of a diverse workforce. Recommendations for practice from this research call for organisations to provide formal training for wellbeing
leads to equip them with the necessary skills to implement effective wellbeing
programmes with a particular focus on behaviour change
Access Repository Dashboard
Do you manage STAX (Strathclyde Repository)? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!