7103 research outputs found
Sort by
Automatic CNN channel selection and effective detection on face and rotated aerial objects
Get PDFBalancing accuracy and computational cost is a challenging task in computer vision. This is especially true for convolutional neural networks (CNNs), which required far larger scale of processing power than traditional learning algorithms. This thesis is aimed at the development of new CNN structures and loss functions to tackle the unbalanced accuracy-effciency issue in image classification and object detection, which are two fundamental yet challenging tasks of computer vision. For a CNN based object detector, the main computational cost is caused by the feature extractor (backbone), which has been originally applied to image classification.;Optimising the structure of CNN applied to image classification will bring benefits when it is applied to object detection. Although the outputs of detectors may vary across detection tasks, the challenges and the design principles among detectors are similar. Therefore, this thesis will start with face detection (i.e. a single object detection task), which is a significant branch of objection detection and has been widely used in real life. After that, object detection on aerial image will be investigated, which is a more challenging detection task.;Specifically, the objectives of this thesis are: 1. Optimising the CNN structures for image classification; 2. Developing a face detector which enables a trade-off between computational cost and accuracy; and 3. Proposing an object detector for aerial images, which suppresses the background noise without damaging the inference efficiency.;For the first target, this thesis aims to automatically optimise the topology of CNNs to generate the structure of fixed-length models, in which unnecessary convolutional kernels are removed. Experimental results have demonstrated that the optimised model can achieve comparable accuracy to the state-of-the-art models, across a broad range of datasets, whilst significantly reducing the number of parameters.;To tackle the unbalanced accuracy-effciency challenge in face detection, a novel context enhanced approach is proposed which improves the performance of the face detector in terms of both loss function and structure. For loss function optimisation, a hierarchical loss, referred to as 'triple loss' in this thesis, is introduced to optimise the feature pyramid network (FPN) based face detector. For structural optimisation, this thesis proposes a context-sensitive structure to increase the capacity of the network prediction. Experimental results indicate that the proposed method achieves a good balance between the accuracy and computational cost of face detection.;To suppress the background noise in aerial image object detection, this thesis presents a two-stage detector, named as 'SAFDet'. To be more specific, a rotation anchor-free-branch (RAFB) is proposed to regress the precise rectangle boundary. Asthe RAFB is anchor free, the computational cost is negligible during training. Meanwhile,a centre prediction module (CPM) is introduced to enhance the capabilities oftarget localisation and noise suppression from the background. As the CPM is only deployed during training, it does not increase the computational cost of inference. Experimental results indicate that the proposed method achieves a good balance between the accuracy and computational cost, and it effectively suppresses the background noise at the same time.Balancing accuracy and computational cost is a challenging task in computer vision. This is especially true for convolutional neural networks (CNNs), which required far larger scale of processing power than traditional learning algorithms. This thesis is aimed at the development of new CNN structures and loss functions to tackle the unbalanced accuracy-effciency issue in image classification and object detection, which are two fundamental yet challenging tasks of computer vision. For a CNN based object detector, the main computational cost is caused by the feature extractor (backbone), which has been originally applied to image classification.;Optimising the structure of CNN applied to image classification will bring benefits when it is applied to object detection. Although the outputs of detectors may vary across detection tasks, the challenges and the design principles among detectors are similar. Therefore, this thesis will start with face detection (i.e. a single object detection task), which is a significant branch of objection detection and has been widely used in real life. After that, object detection on aerial image will be investigated, which is a more challenging detection task.;Specifically, the objectives of this thesis are: 1. Optimising the CNN structures for image classification; 2. Developing a face detector which enables a trade-off between computational cost and accuracy; and 3. Proposing an object detector for aerial images, which suppresses the background noise without damaging the inference efficiency.;For the first target, this thesis aims to automatically optimise the topology of CNNs to generate the structure of fixed-length models, in which unnecessary convolutional kernels are removed. Experimental results have demonstrated that the optimised model can achieve comparable accuracy to the state-of-the-art models, across a broad range of datasets, whilst significantly reducing the number of parameters.;To tackle the unbalanced accuracy-effciency challenge in face detection, a novel context enhanced approach is proposed which improves the performance of the face detector in terms of both loss function and structure. For loss function optimisation, a hierarchical loss, referred to as 'triple loss' in this thesis, is introduced to optimise the feature pyramid network (FPN) based face detector. For structural optimisation, this thesis proposes a context-sensitive structure to increase the capacity of the network prediction. Experimental results indicate that the proposed method achieves a good balance between the accuracy and computational cost of face detection.;To suppress the background noise in aerial image object detection, this thesis presents a two-stage detector, named as 'SAFDet'. To be more specific, a rotation anchor-free-branch (RAFB) is proposed to regress the precise rectangle boundary. Asthe RAFB is anchor free, the computational cost is negligible during training. Meanwhile,a centre prediction module (CPM) is introduced to enhance the capabilities oftarget localisation and noise suppression from the background. As the CPM is only deployed during training, it does not increase the computational cost of inference. Experimental results indicate that the proposed method achieves a good balance between the accuracy and computational cost, and it effectively suppresses the background noise at the same time
Steady and unsteady waves generated by a moving body and multiple bodies
No full textWhen a single ship moves through calm water, it generates a steady surface wave. Although vessels do spend the majority of their operational time traveling at a constant speed in the waterways of uniform depth, there are circumstances when the unsteady effects can be significant. Among these examples are:;In practical operations, when a ship maneuvers in a port/harbour/lock environment. Under this circumstance, a ship is likely to travel in close proximity to waterway boundaries that have an abrupt change, i.e., a step-change in bank dimension or bottom depth;;When conducting ship model tests in a towing tank, the ship model is accelerated from the rest to the target speed. The measured resistance was found to experience persistent periodic oscillations after the target speed was achieved (Doctors et al., 2008);;When a ship is moving during overtaking (or being overtaken) or passing other ships in a dense shipping traffic environment. The interaction between ships will initiate the unsteady forces/moments.;The aforementioned unsteady effects are associated with the unsteady waves on the free surface. The two main objectives of this thesis are 1) to develop a linear method and implement a numerical programme to simulate the steady waves generated by a single body and to predict the wave interference between multiple bodies; 2) to develop an unsteady and nonlinear methodology to predict the unsteady waves generated by the aforementioned unsteady circumstances.;Chapters 2 & 3 are presented in this thesis to achieve objective 1). Chapter 2 deals with the steady waves generated by a single body. The numerical demonstration case in this chapter is a ship passing a false bottom in towing tank tests. Chapter 3 extends the methodology developed in Chapter 2 to investigate the steady wave interference phenomenon by multiple bodies. The numerical demonstration case in this chapter is the steady hydrodynamic interaction between human swimmers.;Chapter 4 & 5 are presented in this thesis to achieve objective 2). Chapter 4 will introduce the unsteady boundary condition to the mathematical model developed in Chapters 2 & 3. In particular, a nonlinear and unsteady free surface boundary condition will be implemented to account for the unsteady effects initiated by the acceleration or the changing water depth. The numerical demonstration case in Chapter 4 is the unsteady waves generated by an accelerating ship. The methodology developed in Chapter 4 will be extended in Chapter 5 to account for the unsteady interaction between multiple bodies. A superposition method will be deployed to investigate the waves generated by two ships at different speeds. The numerical demonstration case in Chapter 5 is the hydrodynamic interaction between two ships during the overtaking operation.When a single ship moves through calm water, it generates a steady surface wave. Although vessels do spend the majority of their operational time traveling at a constant speed in the waterways of uniform depth, there are circumstances when the unsteady effects can be significant. Among these examples are:;In practical operations, when a ship maneuvers in a port/harbour/lock environment. Under this circumstance, a ship is likely to travel in close proximity to waterway boundaries that have an abrupt change, i.e., a step-change in bank dimension or bottom depth;;When conducting ship model tests in a towing tank, the ship model is accelerated from the rest to the target speed. The measured resistance was found to experience persistent periodic oscillations after the target speed was achieved (Doctors et al., 2008);;When a ship is moving during overtaking (or being overtaken) or passing other ships in a dense shipping traffic environment. The interaction between ships will initiate the unsteady forces/moments.;The aforementioned unsteady effects are associated with the unsteady waves on the free surface. The two main objectives of this thesis are 1) to develop a linear method and implement a numerical programme to simulate the steady waves generated by a single body and to predict the wave interference between multiple bodies; 2) to develop an unsteady and nonlinear methodology to predict the unsteady waves generated by the aforementioned unsteady circumstances.;Chapters 2 & 3 are presented in this thesis to achieve objective 1). Chapter 2 deals with the steady waves generated by a single body. The numerical demonstration case in this chapter is a ship passing a false bottom in towing tank tests. Chapter 3 extends the methodology developed in Chapter 2 to investigate the steady wave interference phenomenon by multiple bodies. The numerical demonstration case in this chapter is the steady hydrodynamic interaction between human swimmers.;Chapter 4 & 5 are presented in this thesis to achieve objective 2). Chapter 4 will introduce the unsteady boundary condition to the mathematical model developed in Chapters 2 & 3. In particular, a nonlinear and unsteady free surface boundary condition will be implemented to account for the unsteady effects initiated by the acceleration or the changing water depth. The numerical demonstration case in Chapter 4 is the unsteady waves generated by an accelerating ship. The methodology developed in Chapter 4 will be extended in Chapter 5 to account for the unsteady interaction between multiple bodies. A superposition method will be deployed to investigate the waves generated by two ships at different speeds. The numerical demonstration case in Chapter 5 is the hydrodynamic interaction between two ships during the overtaking operation
Exploring the material discursive production of organisational learning within the storytelling performances of merger and acquisition practitioners
Get PDFHow do entangled reconfigurings of spacetimematter-ing co-constitute organisational learning within the storytelling performances of M&A practitioners? This thesis explores how we might consider this question theoretically while engaging with it empirically. In engaging with this theoretically I read insights from Boje's storytelling theory and Barads agential realism paradigm through one another. Framed within an agential realism paradigm, this study employs a qualitative, post-humanist approach that considers storytelling as a material-discursive practice co-constituted by humans and non-humans, and explores how learning is enacted within storytelling practices.;The empirical material with which to think from and with this question was mainly drawn from 58 conversations enacted over a three-year engagement with 9 M&A practitioners located across 5 different countries and several tools used in their day to day M&A practices. These practitioners were undergoing changes to their core tasks and work processes. The conversations with the practitioners were enacted mainly through Skype and Zoom. In addition to the more formal conversations, the study draws on conversations with practitioners via WhatsApp chat and email exchanges.;This thesis tells a number of entangled stories, and explores how dis/continuous learning is enacted within the storytelling practices of M&A managers and how this learning enables changes to their existing organisational M&A practices. In particular it examines how embodied storytelling informs and guides how actors approach learning.;The literature review examines prior publications in the field of M&As from both functionalist and interpretivist paradigm. The research account tells stories about production of learning within conventional narrative apparatus, living story apparatus and intraview apparatus. The study explores how different apparatuses allow different opportunities for learning, therefore learning is about response-ability, becoming attuned to agential cutting as they occur within storytelling practices. In the discussion chapter, insights from the storytelling of M&A scholars, M&A practitioners and the researcher are then read through one another.;The primary contribution of the thesis is methodological, as it provides an in-depth longitudinal exploration of the process of dis/continuous learning within M&As, thus addressing the limited presence of storytelling perspectives of learning in M&As. Additionally, the thesis advances an ontoepistemological approach to investigating storytelling and learning. Drawing on insights from quantum physics this study illustrates that storytelling occurs through diffraction and learning is about bodymind attunement to agential cutting as they occur within day to day practices.;The study argues for the significance of space, time and matter-ing to learning, and the need for practitioners to become more attuned to the affective flows between human and non-humans within their day to day practices. More generally, the thesis has argued for the value of material-discursive storytelling for understanding how managers learn during complex events, while identifying limitations to conventional narrative, living story and intraview apparatuses for the production of learning.How do entangled reconfigurings of spacetimematter-ing co-constitute organisational learning within the storytelling performances of M&A practitioners? This thesis explores how we might consider this question theoretically while engaging with it empirically. In engaging with this theoretically I read insights from Boje's storytelling theory and Barads agential realism paradigm through one another. Framed within an agential realism paradigm, this study employs a qualitative, post-humanist approach that considers storytelling as a material-discursive practice co-constituted by humans and non-humans, and explores how learning is enacted within storytelling practices.;The empirical material with which to think from and with this question was mainly drawn from 58 conversations enacted over a three-year engagement with 9 M&A practitioners located across 5 different countries and several tools used in their day to day M&A practices. These practitioners were undergoing changes to their core tasks and work processes. The conversations with the practitioners were enacted mainly through Skype and Zoom. In addition to the more formal conversations, the study draws on conversations with practitioners via WhatsApp chat and email exchanges.;This thesis tells a number of entangled stories, and explores how dis/continuous learning is enacted within the storytelling practices of M&A managers and how this learning enables changes to their existing organisational M&A practices. In particular it examines how embodied storytelling informs and guides how actors approach learning.;The literature review examines prior publications in the field of M&As from both functionalist and interpretivist paradigm. The research account tells stories about production of learning within conventional narrative apparatus, living story apparatus and intraview apparatus. The study explores how different apparatuses allow different opportunities for learning, therefore learning is about response-ability, becoming attuned to agential cutting as they occur within storytelling practices. In the discussion chapter, insights from the storytelling of M&A scholars, M&A practitioners and the researcher are then read through one another.;The primary contribution of the thesis is methodological, as it provides an in-depth longitudinal exploration of the process of dis/continuous learning within M&As, thus addressing the limited presence of storytelling perspectives of learning in M&As. Additionally, the thesis advances an ontoepistemological approach to investigating storytelling and learning. Drawing on insights from quantum physics this study illustrates that storytelling occurs through diffraction and learning is about bodymind attunement to agential cutting as they occur within day to day practices.;The study argues for the significance of space, time and matter-ing to learning, and the need for practitioners to become more attuned to the affective flows between human and non-humans within their day to day practices. More generally, the thesis has argued for the value of material-discursive storytelling for understanding how managers learn during complex events, while identifying limitations to conventional narrative, living story and intraview apparatuses for the production of learning
Beyond opium, a history of refined drugs and government regulation in modern China, c. 1871-1945
Get PDFThis thesis examines the history of refined drugs in China from around 1871 to 1945. It traces the process that how refined drugs including morphine and cocaine were introduced and established in China and what drove the Chinese government to regulate them. Through an analysis of a wide range of primary sources, it explores how the Chinese state sought to draw and to police a line between licit and illicit use of refined drugs. It contributes to the historiography by shifting perspective from opium to refined drugs and by making use of some archival sources which have not been consulted in connection with the history of drugs in modern China before.;This research demonstrates that the authority of deciding what were licit and illicit use of refined drugs were in different hands in different periods. Until the twentieth century, the power belonged to the medical missionaries and later the pharmaceutical companies. Between around 1900 and 1910 the Qing administration realised that it should have this power, and it drew on a regulatory model borrowed from abroad to declare that only medical and scientific use is legitimate.;This then became complicated by international discussions of the same issue and a struggle in China by Chinese doctors and drug stores that used Western medicines to have themselves included in the licit/legitimate/legal category. The Bureau of Drug Inspection and Standardisation in 1920s Shanghai and the National Narcotics Bureau established in the 1930s indicate the efforts by the Chinese state to make its control of the issue permanent, with varying degrees of success.;Throughout the period which this thesis covers, the authority on refined drugs was gradually decolonised and centralised into the hand of the central government of the Chinese state. Drawing and policing a line between licit and illicit use of refined drugs was a process extending the authorities of the Chinese state over the society and the economy. It was also a process which was constitutive of state building and modernisation of China.This thesis examines the history of refined drugs in China from around 1871 to 1945. It traces the process that how refined drugs including morphine and cocaine were introduced and established in China and what drove the Chinese government to regulate them. Through an analysis of a wide range of primary sources, it explores how the Chinese state sought to draw and to police a line between licit and illicit use of refined drugs. It contributes to the historiography by shifting perspective from opium to refined drugs and by making use of some archival sources which have not been consulted in connection with the history of drugs in modern China before.;This research demonstrates that the authority of deciding what were licit and illicit use of refined drugs were in different hands in different periods. Until the twentieth century, the power belonged to the medical missionaries and later the pharmaceutical companies. Between around 1900 and 1910 the Qing administration realised that it should have this power, and it drew on a regulatory model borrowed from abroad to declare that only medical and scientific use is legitimate.;This then became complicated by international discussions of the same issue and a struggle in China by Chinese doctors and drug stores that used Western medicines to have themselves included in the licit/legitimate/legal category. The Bureau of Drug Inspection and Standardisation in 1920s Shanghai and the National Narcotics Bureau established in the 1930s indicate the efforts by the Chinese state to make its control of the issue permanent, with varying degrees of success.;Throughout the period which this thesis covers, the authority on refined drugs was gradually decolonised and centralised into the hand of the central government of the Chinese state. Drawing and policing a line between licit and illicit use of refined drugs was a process extending the authorities of the Chinese state over the society and the economy. It was also a process which was constitutive of state building and modernisation of China
An isogeometric boundary element method for three-dimensional lifting flows
No full textIn this PhD thesis an Isogeometric Boundary Element Method (IGA-BEM) for three dimensional steady lifting flows based on Morino's [50] formulation is presented. A potential flow assumption is used and the unknown perturbation potential satisfies Laplace's equation. Application of Green's identities leads to a Boundary Integral Equation (BIE) that is enhanced with kinematic and dynamic boundary conditions.;Analysis suitable T-splines are used for the representation of all boundary surfaces and the unknown perturbation potential is approximated by the same T-spline basis as the one used for the geometry. The BIE is discretised by enforcing it on the generalised version of Greville points for unstructured T-meshes. A novel numerical application of the so-called Kutta condition is introduced that utilises the advantages of IGA with regard to the smoothness of the trailing edge curve basis functions.;This leads to a quadratic system that is solved by a Newton-Raphson iterative scheme. The method is applied for three different test cases and shows good agreement with existing experimental results and superior behaviour when compared to a low order panel method.The effect of the tip singularity on Kutta condition is also investigated for different levels of refinement and positions of the trailing edge collocation points.In this PhD thesis an Isogeometric Boundary Element Method (IGA-BEM) for three dimensional steady lifting flows based on Morino's [50] formulation is presented. A potential flow assumption is used and the unknown perturbation potential satisfies Laplace's equation. Application of Green's identities leads to a Boundary Integral Equation (BIE) that is enhanced with kinematic and dynamic boundary conditions.;Analysis suitable T-splines are used for the representation of all boundary surfaces and the unknown perturbation potential is approximated by the same T-spline basis as the one used for the geometry. The BIE is discretised by enforcing it on the generalised version of Greville points for unstructured T-meshes. A novel numerical application of the so-called Kutta condition is introduced that utilises the advantages of IGA with regard to the smoothness of the trailing edge curve basis functions.;This leads to a quadratic system that is solved by a Newton-Raphson iterative scheme. The method is applied for three different test cases and shows good agreement with existing experimental results and superior behaviour when compared to a low order panel method.The effect of the tip singularity on Kutta condition is also investigated for different levels of refinement and positions of the trailing edge collocation points
Applications of ToF-SIMS imaging in pharmaceutical manufacturing, quantitative characterisation through image analysis strategies and multimodal imaging approaches
No full textDC fault protection in HVDC system and the impact on AC frequency response
Get PDFThe integration of renewable energy generations requires the transmission of bulky power over long distance and in many applications, HVDC transmission systems become a more preferable choice compared to conventional HVAC systems. Thus, the reduction of conventional synchronous generators in the system coupled with the increase of DC connections reduces system inertia and weakens the effect of primary frequency regulation leading to larger frequency derivation in the event of large transients. This thesis investigates frequency support strategies utilizing the flexible power flow controllability of MMC-HVDC. Different control strategies for providing frequency support in a 2-terminal and 3-terminal HVDC systems are studied in detail.;For HVDC systems, DC protection strategy can significantly impact on the duration of power outage in the event of a DC fault. The current definition of maximum loss-of-infeed for an AC network does not consider the duration of the power outage, and the impacts of DC fault protection arrangements which result in different speed of power restoration, on the system frequency of connected AC network, have not been properly understood. Different DC protection arrangements using DC switches, fast and slow DC circuit breakers on frequency response of the connected AC networks are investigated. A 3-terminal meshed HVDC system is studied to demonstrate system behaviour during DC faults. It is found that both the amount of power loss and outage duration affect system frequency response, and thus the two need to be considered simultaneously when determining the maximum 'loss-of-infeed' for future AC-DC hybrid power grids.;The full-bridge submodule based MMCs have DC fault blocking and active fault current control capability. An energy based virtual damping control for FB-MMC is proposed to rapidly de-energise large meshed DC network in the event of a DC fault by quickly absorbing energy from the DC network through the FB-MMCs. The proposed method regulates the DC terminal current of the FB-MMC to follow the DC voltage to behave as a virtual damping resistor to quickly suppress the potential circulating DC fault currents. This enables fast fault isolation using DC switches and thereby fast fault recovery after fault isolation. The fault isolation time is significantly reduced from around 50ms with the existing FB-MMC fault control method to around 15 ms. The validity of the proposed control is verified in a three-terminal meshed DC network.The integration of renewable energy generations requires the transmission of bulky power over long distance and in many applications, HVDC transmission systems become a more preferable choice compared to conventional HVAC systems. Thus, the reduction of conventional synchronous generators in the system coupled with the increase of DC connections reduces system inertia and weakens the effect of primary frequency regulation leading to larger frequency derivation in the event of large transients. This thesis investigates frequency support strategies utilizing the flexible power flow controllability of MMC-HVDC. Different control strategies for providing frequency support in a 2-terminal and 3-terminal HVDC systems are studied in detail.;For HVDC systems, DC protection strategy can significantly impact on the duration of power outage in the event of a DC fault. The current definition of maximum loss-of-infeed for an AC network does not consider the duration of the power outage, and the impacts of DC fault protection arrangements which result in different speed of power restoration, on the system frequency of connected AC network, have not been properly understood. Different DC protection arrangements using DC switches, fast and slow DC circuit breakers on frequency response of the connected AC networks are investigated. A 3-terminal meshed HVDC system is studied to demonstrate system behaviour during DC faults. It is found that both the amount of power loss and outage duration affect system frequency response, and thus the two need to be considered simultaneously when determining the maximum 'loss-of-infeed' for future AC-DC hybrid power grids.;The full-bridge submodule based MMCs have DC fault blocking and active fault current control capability. An energy based virtual damping control for FB-MMC is proposed to rapidly de-energise large meshed DC network in the event of a DC fault by quickly absorbing energy from the DC network through the FB-MMCs. The proposed method regulates the DC terminal current of the FB-MMC to follow the DC voltage to behave as a virtual damping resistor to quickly suppress the potential circulating DC fault currents. This enables fast fault isolation using DC switches and thereby fast fault recovery after fault isolation. The fault isolation time is significantly reduced from around 50ms with the existing FB-MMC fault control method to around 15 ms. The validity of the proposed control is verified in a three-terminal meshed DC network
Creation of an all-optical Bose-Einstein condensate
No full textThis thesis reports on the experimental setup of the initial stages of the two-species, 87Rb and 85Rb, bosonic quantum gas microscope. In particular on the initial cooling and trapping of 87Rb atoms, with the creation of a large all-optically cooled Bose-Einstein condensate as a proof of the temperatures and densities achieved. In most cold atoms experiments, the Bose-Einstein condensate is achieved by evaporating the atoms in a magnetic trap.;Due to the swallowness of magnetic traps, the evaporation sequence takes longer than in all-optical traps. One of the reasons for using all-optical traps is to have a high repetition rate.;After a historical introduction to the topics of atomic physics, ultracold atoms in optical lattices and quantum simulation, we provide a basic summary on atomic physics and laser cooling. It is followed by a description of the experimental setup. The following chapters detail the cooling and trapping process, culminating in the creation of an all-optical Bose-Einstein condensate. The final chapter offers an overview on the future stages of the two-species bosonic quantum gas microscope.;In this thesis we detail our cooling techniques, such as grey molasses and the evaporative cooling process. We achieve temperatures as low as 4μK in clouds of 3*109 atoms after grey molasses and a BEC of 1.8*106 atoms with a total sequence time under 5s. There is still room for improvement in our BEC creation process and we are confident that higher numbers can be achieved.This thesis reports on the experimental setup of the initial stages of the two-species, 87Rb and 85Rb, bosonic quantum gas microscope. In particular on the initial cooling and trapping of 87Rb atoms, with the creation of a large all-optically cooled Bose-Einstein condensate as a proof of the temperatures and densities achieved. In most cold atoms experiments, the Bose-Einstein condensate is achieved by evaporating the atoms in a magnetic trap.;Due to the swallowness of magnetic traps, the evaporation sequence takes longer than in all-optical traps. One of the reasons for using all-optical traps is to have a high repetition rate.;After a historical introduction to the topics of atomic physics, ultracold atoms in optical lattices and quantum simulation, we provide a basic summary on atomic physics and laser cooling. It is followed by a description of the experimental setup. The following chapters detail the cooling and trapping process, culminating in the creation of an all-optical Bose-Einstein condensate. The final chapter offers an overview on the future stages of the two-species bosonic quantum gas microscope.;In this thesis we detail our cooling techniques, such as grey molasses and the evaporative cooling process. We achieve temperatures as low as 4μK in clouds of 3*109 atoms after grey molasses and a BEC of 1.8*106 atoms with a total sequence time under 5s. There is still room for improvement in our BEC creation process and we are confident that higher numbers can be achieved
Detection of pentaerythritol tetranitrate (PETN) by raman and surface enhanced raman scattering (SERS)
No full textMilitary grade explosives such as pentaerythritol tetranitrate (PETN) remain both security and environmental threats. One common method of detecting explosives involves colourimetric testing which is rapid and portable, however, this method lacks specificity as it reacts to specific functional groups in a molecule, therefore resulting in false positives for non-explosive molecules containing the same functional group. Bulk samples of explosives can also be detected using X-ray diffraction and ion-mobility spectroscopy, however, these techniques require large, expensive instrumentation and highly trained staff to operate them. The techniques commonly used today for explosive detection work best on nitro-aromatic explosives. The aim of this research was to develop a method of detecting PETN in a quick and sensitive assay format, which could be used for in field analysis. Initially, this research focused on the detection of PETN using a previously commercially available substrate called KlariteTM. The enhancement observed was investigated to find out if the PETN molecule concentrated into the structural pits of the substrate, or the roughened gold top layer was responsible for the enhancement. However, evidence was found to support both theories and no conclusions could be drawn. As it was difficult to detect low levels of PETN using surface enhanced Raman scattering (SERS), it was decided to fragment the analyte and use the product in an azo reaction. This resulted in the development of four assays which could detect PETN using absorption and/or resonance Raman spectroscopy. However, the assays required a lengthy reaction process (75 minutes), making the assays unlikely to be useful for in field detection. Finally, the resonance Raman assays were further developed into a single SERS assay, which resulted in more sensitive detection of PETN. The timings of the reaction were optimised and reduced so completion was reached within 10 minutes. The assay was found to be translatable for use in field detection as the Raman instrumentation used was portable, and the reaction was fast. Overall, SERS substrates and Raman spectroscopic techniques (resonance Raman and SERS) can be applied to detect PETN to 1 mg L-1 (3.2 μM), leading to comparable limits of detection and quicker reaction times than those found in recent literature regarding the detection of PETN with similar methods.Military grade explosives such as pentaerythritol tetranitrate (PETN) remain both security and environmental threats. One common method of detecting explosives involves colourimetric testing which is rapid and portable, however, this method lacks specificity as it reacts to specific functional groups in a molecule, therefore resulting in false positives for non-explosive molecules containing the same functional group. Bulk samples of explosives can also be detected using X-ray diffraction and ion-mobility spectroscopy, however, these techniques require large, expensive instrumentation and highly trained staff to operate them. The techniques commonly used today for explosive detection work best on nitro-aromatic explosives. The aim of this research was to develop a method of detecting PETN in a quick and sensitive assay format, which could be used for in field analysis. Initially, this research focused on the detection of PETN using a previously commercially available substrate called KlariteTM. The enhancement observed was investigated to find out if the PETN molecule concentrated into the structural pits of the substrate, or the roughened gold top layer was responsible for the enhancement. However, evidence was found to support both theories and no conclusions could be drawn. As it was difficult to detect low levels of PETN using surface enhanced Raman scattering (SERS), it was decided to fragment the analyte and use the product in an azo reaction. This resulted in the development of four assays which could detect PETN using absorption and/or resonance Raman spectroscopy. However, the assays required a lengthy reaction process (75 minutes), making the assays unlikely to be useful for in field detection. Finally, the resonance Raman assays were further developed into a single SERS assay, which resulted in more sensitive detection of PETN. The timings of the reaction were optimised and reduced so completion was reached within 10 minutes. The assay was found to be translatable for use in field detection as the Raman instrumentation used was portable, and the reaction was fast. Overall, SERS substrates and Raman spectroscopic techniques (resonance Raman and SERS) can be applied to detect PETN to 1 mg L-1 (3.2 μM), leading to comparable limits of detection and quicker reaction times than those found in recent literature regarding the detection of PETN with similar methods
Big Data Tools and Techniques CS988 exam papers
No full textAccess restricted to staff and registered students of the University of Strathclyde.PAST EXAM PAPERS ARE NO LONGER BEING ADDED TO STAX. PLEASE VISIT SUPRIMO TO ACCESS AN UP-TO-DATE COLLECTION OF PAST EXAM PAPERS: https://suprimo.lib.strath.ac.uk
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!