Heriot-Watt University
ROS: The Research Output Service. Heriot-Watt University EdinburghNot a member yet
4689 research outputs found
Sort by
MOF-based materials for the photocatalytic conversion of CO2
Full text linkThe main force behind anthropogenic driven climate change is the emission of
greenhouse gases such as carbon dioxide and methane. The increase in the atmospheric
concentrations of these gases at its root is due to humanity’s over reliance on
hydrocarbons as a form of energy and a source of materials for use in industry. Alternative
energy sources and implementing Carbon Capture and Storage (CCS) technologies are
measures being taken to address the emissions from the energy and manufacturing
sectors. The reorganisation of the energy sector also presents an opportunity for CO2
utilisation (CCU)technologies to act as alternative sources of hydrocarbon products
through the conversion of carbon dioxide. Typically, CO2 utilisation processes require a
lot of energy and harsh reaction conditions, however one of the most promising methods
of CO2 utilisation comes in the form of CO2 photoreduction. Through this process CO2 is
converted into value-added products using light under mild reaction conditions making
use of a well-designed photocatalyst.
This thesis consists of six chapters outlining work carried out since 2018 focusing on the
design and synthesis of photocatalysts for CO2 photoreduction. Chapter 1 is an
introduction to the field beginning with a wider discussion of the current climate crisis,
and CCUS before moving on to describe the fundamental principles of CO2
photoreduction and the current state-of-the-art photocatalysts. Chapter 2 details the
methods of synthesis and provides a description of the photoreduction system used to
evaluate the performance of the synthesised photocatalyst. It includes a discussion of the
characterisation techniques used to analyse the samples as well as the equations used.
Chapter 3 discusses a TiO2 photocatalyst derived from the calcination of MIL-125-NH2,
it was found to be highly selective towards CO generation which is unusual in purely
TiO2 based photocatalysts. Chapter 4 expands upon the work in Chapter 3 with the aim
of improving the performance of the photocatalyst whilst maintaining the excellent
selectivity that had been observed. This is achieved through the formation of a composite
with UiO-66-NH2, and it was found that the inclusion of the MOF shifts the selectivity
towards CO production. Chapter 5 explores further the work with metal oxide/MOF
pairings through the synthesis of Cu2O/UiO-66-NH2 composites. This is the first time that
a Cu(I) oxide MOF composite has been reported without the presence of copper in the
form of Cu(0 or II). This thesis aims to show the potential of metal oxide MOF pairings
for CO2 photoreduction.Engineering and Physical Sciences Research Council (EPSRC) funding
When Primes adopt brokers in supplying complex infrastructure projects : a TCE and structural hole perspective
Full text linkPurpose:
In some complex public private infrastructure projects, the Prime contractor (the Prime) relies on
professional brokers to buy from, and coordinate with, Original Equipment Manufacturers (OEMs)
activity. Broker, the Prime and OEM triads are formed. In most purchasing studies on complex
performance (PCP) it is the Client who contracts with a Prime that is studied, and their relationship.
Extant operations management literature has very little to say on why or when these broker roles exist
and on how these professional brokers add value, which also suggests very little understanding of the
Primes’ purchasing challenges as well as their corresponding solutions. This silence has driven the
overarching aim of this thesis in exploring rationales of a Prime in adopting professional brokers in
complex infrastructure projects.
Design/methodology/approach:
Four work packages are formed from observations and 34 interviews conducted in various engineering
projects in the global power plant construction and retrofitting sector. Interview questions are informed
by theoretical insights to investigate the nature of Primes’ challenges and how they are resolved by
professional brokers. Transaction cost economics (TCE) provides this study with a foundational
explanation for why brokers exist (trilateral governance), which is then complemented by the nuances
of broker behaviour provided by structural hole theory.
Finding:
This thesis provides an empirically derived typology of a Prime’s scope of challenges, and Prime-OEM’s position in sourcing from OEMs in complex infrastructure projects. The typology confirms
Prime’s challenge in materialising clear specifications for archetypal complex product and systems, as
well as oligopolistic market patterns such as soaring and non-negotiable service price proposed by
limited suppliers. However, shifting the perspective from the Client to the Prime reveals novel
procurement challenges in sourcing non-archetypal complex product and systems. For example,
procurement at project initiation phase could see client-imposed language and geographical barriers and
asymmetric transaction interest. In addition, at project delivery phase asymmetric financial terms and
conditions, problems directing sub-system OEMs on technical specifications and on-site services could
emerge.
This thesis finds that Primes are seen to purposefully adopt different categories of professional brokers
to target specific procurement challenges. A common characteristic of these brokers is their relationship
continuity with OEMs based on historical interactions and/or perception of future transactions, which
is absent from the Primes in temporary, one-of-a-kind project settings. Through adopting different brokers in different project stages, Primes could focus on major tasks that form their competitive
advantages such as project management and on-site construction work.
Relevance/contribution:
This study examines the under investigated area of PCP by expanding the typology of procurement
challenges addressed in the literature, as well as the solutions. Such investigation refines the definition
of procurement complexity by identifying two additional factors (i.e., the extent of interdependency of
technical specifications from other product/service, and the extent of iterative process in defining
technical specifications of the product/service) that describe the level of infrastructure complexity of
product and service in complex infrastructure projects. In addition, the identification of five broker
categories extends the approaches that encourage value co-creation between buyers and suppliers on
top of formal (i.e., legal rules, standards and remedies etc.) and informal governance (i.e., workshops,
incentives and punishments etc.) mechanisms introduced in the PCP literature. The adoption of TCE
and structural hole lenses in this thesis confirms the applicability of TCE in a social network context,
and the applicability of structural hole theory in a transaction-oriented context. For practitioners, this
thesis contributes to prime contractors who undertake similar purchasing activities for complex
infrastructure projects. In particular, prime contractors could benefit from this thesis by recognising
similar procurement challenges, and the collection of approaches in resolving the corresponding
challenges through the adoption of appropriate brokers
The biomechanical changes in soft tissues due to compartment syndrome, as a route for a quantitative monitoring
No full textAcute compartment syndrome (ACS), occurring after leg trauma, involves increased pressure in muscle compartments, which disrupts blood flow and damages nerves and muscles. Timely diagnosis is crucial to prevent permanent harm.
While advances like intracompartmental pressure measurement have been made,
other imaging and biological methods lack sufficient evidence to surpass it. Developing a reliable tool for early ACS detection could significantly benefit patients
and the field. In this thesis, we proposed that assessing the mechanical properties of ACS could offer a rapid and objective method to understand its occurrence
and aid in its diagnosis. To explore this, we developed models to simulate the
physiological changes of ACS: tissue-mimicking leg phantoms derived from MRI
images, a pig model, and a human cadaver model. Then, three mechanical methods were designed and employed: a hand-held indentation device to measure how
the mechanical properties of the skin change at different pressures; Digital Image
Correlation (DIC) to visualise non-invasively the changes on the skin surface; strain
sensors to track and monitor the mechanical changes over time. The main findings
from this thesis include:
• The tissue-mimicking phantoms provided us information on how a change in
the pressure inside the leg alters its mechanical properties.
• DIC proved to be the strongest tool to monitor ACS and allowed us to
recommend a possible diagnostic threshold for ACS.
• Patient-specific mechanical FE models, developed from the CT scans of the
cadavers, proved that we need to conduct pre-clinical testing in order to
consider what happens inside the muscles
Bayesian inference procedures for dynamical systems aided by artificial neural networks
Full text linkBayesian parameter inference, which feature associated uncertainty quantifications in analyses of real-life dynamical systems, is important for informing scientific
knowledge and guiding interventive actions. In such inference of non-sequential,
complete datasets, nonlinear dynamic models admit analytically intractable trajectories, thus eliciting intractable posterior distributions and causing computational inefficiencies; for inferring sequentially arriving data, nonlinear state-space
models prohibit the identification of suitable or tractable proposal distributions,
thereby hampering sequential sampling procedures. This thesis devises methodologies for leveraging the analytical tractability and functional flexibility offered
by artificial neural networks to accurately approximate dynamical system trajectories, and then use them to facilitate posterior inference of relevant parameters in
both non-sequential and sequential instances. Specifically, the proposed network
structures are crucial in aiding (i) the uncertainty quantification in non-sequential
estimations, and (ii) the identification of suitable proposal distributions in sequential procedures. The methodologies developed are then applied on the Susceptible-Infectious-Removed model for infectious diseases to infer static parameters pertaining to contact rates and infectious durations. Major findings include significant
reductions in computational costs and enabling of sequential Bayesian inference
in settings where it was previously impossible. Overall, the incorporation of deep
learning approaches demonstrates significant impact to Bayesian inference methods
A finance lexical approach underpinned by investor’s ontology for sentiment analysis to predict stock market trends
No full textAbstract and full-text unavailable. Restricted access until 17.03.2027. Please refer to PDF
Characterisation and comparative genomics of novel aerobic methane oxidizing bacteria and their industrial potential as methane biofilters
Full text linkThe atmospheric methane (CH4) oxidizing bacteria known as upland soil
cluster α (USCα) and γ (USCγ) constitute one of the largest global biological sinks for
atmospheric CH4. USCα are widespread in soil and other environments, but despite their
widespread nature, successful characterisation has remained elusive. A major aim of this
thesis was to elucidate USCα through isolation and characterisation of its genomics and
metabolomics. As such, an enrichment culture of a novel USCα strain sampled from
Canadian permafrost was begun in 2015, along with an enrichment of a novel
Methylocella sp. sampled from German forest soil in 2017.
USCα was enriched to ~60% of its culture during this study and Methylocella sp.
enriched to ~40%. The metagenome assembled genome (MAG) of USCα denoted
PF_bin11 was 93.38% complete with 68 contigs and a 3,395,889 bp genome size.
Methylocella sp. denoted MF_bin10 was 98.07% complete with 19 contigs and a
3,364,475 bp genome size. The closest relative of PF_bin11 was identified as
Methylocapsa gorgona MG08, the only isolate of USCɑ, with an average nucleotide
identity of 96.85% and amino acid identity of 96.22%. PF_bin11 possesses a full
[Nickel/Iron] hydrogenase and CO dehydrogenase, and is likely capable of supporting
growth on low CH4 levels with H2 and CO oxidation. MF_bin10 shared genus level
similarity with Methylocella silvestris BL2 and TVC, with an average nucleotide identity
of >77% and amino acid identity of >71%. It also possessed soluble methane
monooxygenase exclusively and is likely capable of facultative methanotrophy. While
attempts were made to isolate USCα and Methylocella sp. using the streak plate method,
it is likely that USCα was successfully isolated but later contaminated by satellite
heterotrophs, while Methylocella sp. was not successfully isolated due to failure to foster
growth from single colonies.
In addition, another aim of this study was to determine whether USCα can function as
a biofilter of atmospheric CH4 after colonization of the porous surfaces of biochar. For
this, USCα and Methylocella sp. enrichments, Methylosinus sporium and M. silvestris
BL2 were tested for colonizing capabilities on a standard biochar set provided by the UK
Biochar Research Centre. CH4 was only oxidized in the presence of soft wood biochar
pyrolyzed at 550°C or 700°C and sewage sludge biochar pyrolyzed at 550°C, which was
attributed to the pH being 7.91, 8.44, and 8.17 respectively, compared to >9 pH for the
other biochar types. Subsequently, colonization of the biochar was tested through DAPI
staining with soft wood 550°C and 700°C appearing to be successfully colonized by M. sporium and the USCα enrichment culture. Sewage sludge 550°C was not successfully
colonized, which was likely ascribable to heavy metal toxicity as sewage sludge biochar
is noted to contain inordinately higher concentrations of heavy metals, especially zinc,
copper, cobalt, chromium and mercury.
The generation of a new MAG of USCα helps to expand the number of characterized
strains and further contextualize current knowledge of these lineages. As USCα was not
successfully isolated and tested for surface colonization of biochar in isolation, whether
USCα can function as a CH4 biofilter of atmospheric CH4 is still equivocal
The ambient carbonation of lime for carbon dioxide removal technologies
Full text linkIt is broadly agreed that the burning of fossil fuels since the industrial revolution is the main
driver of anthropogenic climate change. This has led to a rapid increase of the CO2
concentration in the atmosphere, trapping solar radiation and increasing average global
temperatures, thus leading to negative impacts to life on Earth. A proposed solution to mitigate
these harmful issues, is to reduce new CO2 emissions and remove legacy emissions directly
from the atmosphere, ultimately reducing atmospheric CO2 concentrations. Removal of CO2
from the atmosphere, may be achieved through carbon dioxide removal (CDR) technologies,
for example mineral carbonation. Mineral carbonation involves the reaction of alkaline cations
with CO2 to form stable carbonate minerals. Lime (CaO) can be used for mineral carbonation,
however the chemical data of the reaction of lime with CO2 in ambient conditions is sparse. In
this thesis, the ambient carbonation reaction has been explored in various conditions to aid in
the development of a CDR technology.
For lime to be used in a scaled-up CDR technology it may be important to understand
certain kinetic steps such as the relationship between layer thickness and rates & extents of
carbonation. The findings of Chapter 3 show that thinner layers of lime (2.5 mm), carbonate at
a higher rate than the thicker layers (50 mm). However, the thicker layers of lime have a higher
uptake per area, which may offer benefits for a CDR technology due to lower land requirements
and lower operational costs associated with reapplying the layers of material.
The slower carbonation rates observed in the thicker layers may be increased if mixing is
applied. This thesis tested the effect of infrequent mixing on rates and extents of ambient
carbonation of lime (Chapter 4) and found that intermittent mixing increased the rates of
ambient carbonation. Furthermore, there is no real difference when mixing weekly or monthly,
concluding that CDR technologies may benefit from low frequency mixing of layers of lime.
This PhD study tested both CaO and Ca(OH)2, and found that Ca(OH)2 carbonated at faster
rates and to higher extents, which is consistent with the literature. The production of Ca(OH)2
creates inherently small particles (2-5 µm) which, combined with this materials tendency for
dipole-dipole interactions, leads to cohesivity and poor handleability and dust issues. This
problem is commonly solved in industry by increasing the particles size via agglomeration.
The effect of agglomeration on the carbonation rate and extent has been explored (Chapter 5),
finding that agglomerated Ca(OH)2 may increase the rate and extent of reaction.
Different methods of facilitating the reaction of lime with the CO2 in the air were explored
in a series of experiments under the instruction of an industrial partner, Origen Carbon
Solutions. A ~7 m tall pilot scale re-carbonation rig (fluidised bed set up) was commissioned
which highlighted the challenges of handling a fine, cohesive powder. A similar bench scale
experiment (column experiments) showed that the availability of CO2 in a passive system may
be rate limiting step with an optimum air flow rate below 6 l/min (in a ~4.2 l column). A
different approach to air lime mixing was explored in a carbonation duct experiment in which
an attempt to carbonate the lime whilst suspended in air was made, with findings suggesting
increasing suspension time was beneficial to carbonation rates. High relative humidity systems
were also tested on both milk of lime (Ca(OH)2 suspended in water) and powdered lime. The
findings from these experiments suggest that milk of lime carbonated at higher rates and extents
in low relative humidity whereas powdered Ca(OH)2 carbonated at a higher rate and extent in the high relative humidity. These findings suggest there is an optimal moisture content needed
in processes seeking to carbonate Ca(OH)2 in ambient conditions
Investigations into labyrinthulid plant pathogens : environmental distribution, disease characterisation and potential controls
No full textLabyrinthulid protists are poorly studied members of the Phylum Labyrinthulomycetes,
although there is historical evidence from ‘wasting disease’ in eelgrass, they can be of great
ecological and pathogenic importance. The emergence of the novel terrestrial species
Labyrinthula terrestris, causative agent of ‘rapid blight’ in turf grasses, appears to be directly
linked to soil salinification. Previous studies have shown L. terrestris infects and kills
important crop species including wheat, barley, and rice. This, coupled with the increasing
impact of soil salinification globally, makes our current lack of understanding of this atypical
pathogen increasingly concerning.
The aim of the present work was to extend our understanding of the occurrence of
labyrinthulids in natural environments, to evaluate potential control methods and to explore the
relationship of labyrinthulids with their plant hosts to elucidate fundamental aspects of the host-pathogen interaction, including environmental circumstances under which they might become
pathogenic, and key aspects of host immune defence.
Broad field surveys focussed on saltmarsh habitats, as they were hypothesised to be suitable
habitats, given that all labyrinthulids so far known have a requirement for elevated sodium in
order to grow and the closest relative to L. terrestris has been previously shown an isolate from
the saltmarsh grass species (Spartina alterniflora). Labyrinthulids were successfully isolated
from saltmarshes across Scotland, on both east and west coasts. A culture collection of
labyrinthulids isolated from golf-course turf experiencing rapid blight disease was built from
samples received from golf-courses across the UK and Europe, to facilitate comparison of
diversity and pathogenicity of rapid blight pathogens to the isolates obtained in saltmarsh
habitats. Isolates obtained in pure culture were identified to closest matches in the NCBI
nucleotide database using partial 18S and 28S/ITS rRNA gene sequences. Labyrinthula
terrestris was isolated from saltmarsh turfgrass hosts and following pathogenicity assays in the
susceptible turfgrass Poa trivialis, 3 distinct pathogenic saltmarsh Labyrinthula sp. haplotypes
were identified, one of which, Laby31, has been previously documented from diseased golf-courses. Two clades revealed by Bayesian phylogenetic analysis in both rDNA sequences may
represent novel genetic diversity of L. terrestris. Nevertheless, identifiable pathogenic
haplotypes, including L. terrestris (haplotype Laby31), were found for the first time from
natural environments and were shown to produce rapid blight disease symptoms in laboratory
trials with P. trivialis. Lab trials were conducted to assess the efficacy of a range of rapid blight control treatments.
A high-throughput laboratory assay was tested in comparison to a previously field data
validated laboratory method, in the highly susceptible turfgrass host Poa trivialis. In total, 14
treatments representing various chemical classes of current and novel formulated products were
assessed for efficacy in maintaining turf vigour and reducing rapid blight disease symptoms.
Six treatments showed significant disease control, with comparable turf vigour and disease
symptoms to the untreated (uninfected) controls. Several products, applied in advance of
symptom development, showed long lasting disease suppression (up to 3 weeks), but were less
effective when applied reactively. Improved means of predicting outbreaks would improve the
sustainable use of preventative applications.
Further exploration of the labyrinthulid-host relationship revealed, for the first time, the ability
of these protists to infect and cause disease symptoms in dicotyledonous plants in the form of
the model species Arabidopsis thaliana. This finding is crucial to further studies of
labyrinthulids, since the ready availability of a wide range of molecular biology tools for A.
thaliana such as mutant lines, is unrivalled and provides a powerful means of unravelling the
molecular basis of disease between host, pathogen and environment. A rapid and robust qPCR
assay was optimised to track L. terrestris infection levels and allowed the key characteristics
of L. terrestris infection in A. thaliana, including abiotic factors, to be quantified. The work
initiated here indicated that salt stress increased infection, whilst high temperatures reduced
infection levels of L. terrestris (haplotype Laby31). Microscopic investigation over time
suggested that L. terrestris is not initially necrotrophic at point of infection and most likely
exhibits a hemi-biotrophic lifestyle under the conditions tested. Additionally, infection load in
key immune defence hormone mutants (sid2, abi-1 and ert-1) in comparison to WT Col
revealed that Salicylic Acid (SA) was highly significant in defence against L. terrestris in A.
thaliana. Abscisic acid (ABA) was also revealed to be significant, which as a result of this
hormones important role in salinity stress regulation reveals the utility of the model host in
unravelling the infection dynamics for this pathogen
Training sets for algorithm tuning : exploring the impact of structural distribution
Full text linkAlgorithm tuning is a growing area of research and practice. In particular, this area
of work is developing in the context of metaheuristic algorithms, which involve configuring the parameters of an algorithm on the basis of a "training set" of problem
occurrences. There has been significant interest (but to date, limited research) in
the relationship between the training set and the generalisation performance, often assuming that if an algorithm is targeted at problems of a certain type, then
the training set of problem instances should represent a diverse collection of problems of that type. The current research investigates this assumption, focusing on
how certain structural aspects of training set problem instances affect generalisation
performance.
In our first study, using travelling salesmen problem (TSP) instances across both
uniform and Gaussian distributions, we sought evidence regarding the veracity of
a "widely held assumption" about training sets and generalisations that training
sets should be pulled from the same distribution as the problem instances. We
found evidence against this assumption, showing that the algorithms on problem instances from the same distribution as the training sets did not perform better than
those that were trained on distinct training sets; algorithms did, however, appear
to perform consistently better when TSP distributions were Gaussian, as opposed
to uniform. In our second study, we replicated our findings from the first, showing again that the performance of algorithms on problem instances from the same
distribution as the training set did not perform better. Extending these findings,
we explored whether problem difficulty played a role in generalisation, finding that
greater difficulty by using fitness-distance correlation in training appears to be linked
to better performance, but only in two-region problems. Finally, the third study explored these effects in the context of vehicle routing
problems (VRPs). In our final study, we again found evidence against the ‘widely
held assumption’, and even found that algorithms trained on different training sets
seemed to perform better than those trained on the same distributions. We further
showed that the difficulty of distribution which we measured by using fitness-distance
correlation was tied to higher performance, confirming what was found in the second
study. Overall, the primary finding across both TSP and VRP contexts is that many
researchers have a prior assumption that they must train it on instances from a
distribution similar to the problem instances. One significant limitation is that these
data were generated using a synthetic data approach, and simplified to an extent that
creates a differentiation between these and real-world data. However, the findings
do have implications for the development of a-world optimisers for specific problem
classes, such as VRPs with particular depot and customer distribution setups
Analysis of multibeam reflector antennas for satellite applications
Full text linkThe demand for high-capacity satellite communication systems has grown exponentially in recent years, driven by the need for high-speed connectivity. As the demand
rises, the efficient utilization of satellite resources becomes imperative. This has
underscored the importance of developing advanced antenna technologies to meet
user requirements. Multibeam Reflector Antennas (MBRA) offer a promising solution by enabling multiple beams, which can simultaneously serve multiple users
or regions from a single satellite platform. This technology is one of the preferred
architectures for Very High Throughput Satellite (VHTS) systems to achieve the
requirements from geostationary orbit (GEO) due to the high gain offered by the
system optics. They combine a reflector with an array of feed elements which can
generate multiple spot beams. To fully exploit the capabilities of the satellite, multi-disciplinary optimisation becomes a very important topic in the design process of
the payload system. In this context, efficient estimation of the multi-beam coverage
characteristics becomes a step towards the desired system performance evaluation.
This thesis presents the research work carried out towards an efficient way of estimating the beam patterns of multibeam reflector antennas when the number of
beams employed is very large. To that end, an in-depth study of reflector antennas
for space applications and their analysis is described. HERAS (HEriot-watt Reflector Antenna Solver) tool is presented, an in-house tool developed for the analysis of
reflector antennas using Matlab. This tool is used as part of this research as the core
element for the rapid estimation of hundreds of beams. Two different configurations
of MBRA are considered: Single Feed Per Beam (SFPB) and Array Fed Reflector
(AFR) antennas. Different acceleration techniques are developed to efficiently estimate the antenna patterns for each of the configurations. These include a threshold
method, interpolation methods or computational techniques. They are essentially
based on the reduction of the number of farfield points to be calculated or the size of
the numerical integration to be performed for each beam calculation. In the case of
AFR antennas, beamforming is one of the key steps in the calculation of the beam
patterns and hence, a study of different beamforming techniques is also presented.
To support the results presented in this thesis, a benchmark with respect to available
commercial software (GRASP from TICRA) is provided. Last, parametric studies
at the system level are presented, which shows the potential of the rapid estimation
of coverage characteristics in the design process of the satellite payload
Access Repository Dashboard
Do you manage ROS: The Research Output Service. Heriot-Watt University Edinburgh ? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!