Doctoral Degrees (Physics)

Permanent URI for this collectionhttps://hdl.handle.net/10413/6603

Browse

Now showing 1 - 20 of 106

A study on the atmospheric and environmental impacts of aerosol, cloud and precipitation interaction.
(2022) Yakubu, Abdulaziz Tunde.; Chetty, Naven.
Understanding the mechanisms and processes of aerosol-cloud-precipitation interactions (ACPI) is essential in the determination of the specific role of aerosols in modulating extreme weather events and climate change in the long run. Atmospheric aerosols are mainly of various types and are emitted from differing sources. Considering they commonly exist in the heterogeneous forms in most environments, they significantly influence the incoming solar energy and the general perturbation of the clouds depending on their constituents. Thus, a systemic identification and characterisation of these particles are essential for proper representation in climate models. To better understand the process of climate change, this research explores the climate diversity of South Africa to examine aerosol sources and types concerning the atmospheric aerosol suspension over the region and their role in clouds and precipitation formation. The study further provided answers to the cause of extreme precipitation events, including drought and occasional flooding experienced over the region. Also, an insightful explanation of the process of ACPI is provided in the context of climate change. Furthermore, the research found that the effective radiative forcing (RF) over South Africa as monitored in Cape Town and Pretoria is negative (i.e., cooling effect) and provided an analysis of the cause. Similarly, the validation of some satellite datasets from MISR (Multiangle Imaging Spectroradiometer) and MODIS (Moderate Resolution Imaging Spectroradiometer) instruments against AERONET (Aerosol Robotic Network) is conducted over the region. Although a significant level of agreement is observed for the two instruments, intense improvements are needed, especially regarding measurements over water surfaces. Finally, the study demonstrated the proficiency of effective rainfall prediction from satellite instrument cloud datasets using machine learning algorithms.
Aerosol characteristics over different regions of southern Africa : using sunphotometer and satellite measurements.
(2015) Adesina, Joseph Ayodele.; Venkataraman, Sivakumar.
Aerosols and cloud play a major role in understanding and interpreting the varying earth’s energy budget. It is necessary to characterize these atmospheric particles by their sizes, chemical composition, water content etc. Aerosols can both cause heating and cooling depending on what they are made of; dust will generally tend to scatter leading to cooling effect while some species of black carbon will absorb sunlight thereby causing a heating effect. In order to assess their impact on global climate, a multiple measurement approach is necessary and specifically, we need long and short term ground-based measurements in clean and polluted environment and long term satellite measurements. In this thesis, we have used aerosol measurements from CIMEL Sunphotometer (part of the world-wide network; Aerosol Robotic Network: AERONET) over, Pretoria (25.75º S, 28.28º E) and Skukuza (24.9º S, 31.5º E) in South Africa, and satellite data from Moderate Resolution Imaging Spectroradiometer (MODIS) and Multiangle Imaging Spectroradiometer (MISR). Pretoria is situated in industrial region with adequate influence of urban/industrial aerosols while Skukuza is an agricultural based region with frequent burning of agricultural waste to clear the harvest during the late winter, spring and summer seasons. Thus, the study over industrial and agricultural regions explores more understanding about the regional radiative forcing in relation to aerosol loading and meteorology. MODIS satellite data was utilized for addressing long term trend in aerosol loading and cloud interaction studies over different locations of South Africa where no ground based sunphotometer data are available. Using six months sunphotometer data (July–December 2012), aerosol characteristics over Gorongosa were studied with particular attention to how aerosol loading evolves during the biomass burning season (spring) including pre- and post-months. The results revealed that the monthly mean aerosol optical depth (AOD₅₀₀) was at maximum in September and minimum in November. The study also investigated biomass burning and forest fire occurrences in Mozambique using MODIS active fire data. Using a year sunphotometer data (January – December 2012) obtained from Pretoria’s (CSIR_DPPS) AERONET site, aerosol was characterized by its optical, microphysical and radiative properties. The study explored meteorological effects on aerosol loading and aerosol direct radiative forcing over Pretoria. Maximum value of aerosol optical depth (AOD₅₀₀) was found during February (summer) and August (winter) while the atmospheric forcing was found to be independent of seasonal variation in AOD. Besides, AOD, Angstrom exponent (AE; α440-870), columnar water vapor (CWV), volume size distribution (VSD), single scattering albedo (SSA) and aerosol radiative forcing (ARF) were computed and their variations with their climatic implications were studied. Using the ground-based instrument of AERONET at Skukuza, we performed validation of MISR and MODIS (Terra and Aqua) level 3 AOD products using the data retrieved for the year 2010. We also carried out regression analysis on these satellite products using 10 years of dataset (2004-2013) to evaluate their performance at a hinterland and coastland stations with two distinct environments in SA. The validation showed that MISR was better correlated with sunphotometer having a coefficient of determination (R²=0.94), Aqua MODIS (R²=0.77) and Terra MODIS (R²=0.68). The long term regression analysis at the two selected locations showed MODIS products underestimating MISR. At the hinterland, MISR showed an increasing trend while MODIS products showed a decreasing trend over the study period but at the coastland MISR and Terra MODIS showed a negative trend while Aqua MODIS showed a positive trend. When the two MODIS products were compared, they were better correlated at the coastland (R²=0.66) than hinterland (R²=0.59) and when compared based on seasonal variation, they were better correlated in the winter season in both locations than any other season. The Ozone Monitoring Instrument (OMI) Ultra-Violet Aerosol Index (UVAI) which was used to monitor the absorption aerosol index showed an increasing trend over the two locations with 0.0089/yr hinterland and 0.0022/yr at coastland. In the present thesis, we also used data obtained from the Terra satellite onboard of the MODIS to investigate the spatial and temporal relationship between AOD and cloud parameters namely, water vapor (WV), cloud optical depth (COD), cloud fraction (CF), cloud top pressure (CTP) and cloud top temperature (CTT) based on 5 years (January 2008 -December 2012) of dataset over six locations in South Africa. AOD has high values during spring (September to November) but low values in winter (June to August) in all locations. In terms of temporal variation AOD was lowest at Bloemfontein 0.06±0.04 followed by Cape Town 0.08±0.02, then Potchefstroom 0.09±0.05, Pretoria and Skukuza had 0.11±0.05 each and Durban 0.13±0.05. The mean AE values for each location show a general prevalence of fine particles for most parts of the year. Our analysis of AOD and WV showed both quantities only co-vary at the beginning of the year but later in the year they tend to have opposite trend over all the locations. AOD and CF showed negative correlation for most of the locations while AOD and COD were positive over three of the locations. AOD and CTT, CTP showed similar variations in almost all the locations. The co-variation of CTT and CTP may be due to large scale meteorological variation.
The analysis and optimization of electrostatic electron optical lenses with rotational symmetry, through use of orthogonal functions.
(1978) Van der Merwe, Johannes Petrus.; Walker, Anthony David Mortimer.; Spalding, Dennis Raymond.
No abstract available.
An analysis of Pc5 pulsations observed in the SuperDARN radar data.
(2009) Magnus, Lindsay Gerald.; Rash, Jonathan Paul Stuart.; Walker, Anthony David Mortimer.
This thesis deals with the development of automatic methods for nding pulsation events in time series produced by the radars in the SuperDARN network. These methods are then applied to the detection and analysis of pulsation events illustrating the relative usefulness of radar data for the study of global pulsation dynamics. Each of the SuperDARN radars produces 1200 Doppler velocity records every hour. If backscatter is present, and there is a pulsation occurring in the same region as the scatter, the pulsation can be measured as periodic changes in the Doppler velocity of that record. There are over 85 million Doppler velocity data records for 2004. In order to identify pulsations in these data, an automated pulsation nder was developed. All records with signi cant peaks in the FFT spectra were tagged as having a pulsation present. If a record had less than 20% data missing it was termed a clean record as it was suitable for use with the automated pulsation nder. As pulsations can only be observed if there are scatter, an investigation into scatter characteristics are presented. It is shown that the occurrence of clean records is most strongly in uenced by IMF Bz, and the underlying spatial structure of the SuperDARN network. The results for the automated pulsation nder for 2004 are then presented. It was found that the average daily distribution of pulsation events, shown as a function of pulsation frequency, followed 1/f distribution with no distinct peaks. It was also found, however, that the standard deviation of the average showed peaks close to the "magic" frequencies indicating that on average there is more variation at these frequencies that any of the other pulsation frequencies measured. The occurrence of pulsations followed the clean scatter statistics both temporally and directionally telling us that the network is not suited for studying global pulsation dynamics because the variations in scatter dominated any variations in pulsation occurrence. Data from a few events identi ed by the pulsation nder are then presented to illustrate the advantages and disadvantages of using SuperDARN data for pulsation event studies. The events show a pulsation that occurs at its fundamental and third harmonics, an aliased pulsation, a pulsation, interrupted by sounding frequency changes, that shows how ionospheric scatter was tagged as ground scatter and how data from two overlapping pulsations in di erent radars can be merged to give the poloidal and toroidal characteristics of the event.
Application of machine learning in cosmology.
(2024) Nagarajappa, Chandan Ganjigere.; Ma, Yin-Zhe.
Abstract available in PDF.
Application of machine learning techniques to the description of open quantum systems.
(2023) Naicker, Kimara.; Petruccione, Francesco.; Sinayskiy, Ilya.
This work focuses on using classical machine learning (ML) models to study the quantum dynamics of excitation energy transfer (EET) within strongly coupled open quantum systems relevant to light harvesting complexes (LHCs). Direct evidence for long-lived quantum coherence has been found to play an important role in EET processes during the first step of photosynthesis in certain LHCs where excitation energy is transmitted from the antenna pigments to the reaction center in which photochemical reactions are initiated [1–3]. The numerically exact method used to simulate the dynamics in this work is the hierarchical equations of motion (HEOM) adapted by Ishizaki and Fleming to suit the quantum biological regime [4–6]. In the case of an open quantum system, such as a photosynthetic pigment-protein complex, evolving over time we can generate a set of time dependent observables that depict the coherent movement of electronic excitations through the system by solving a suitable set of quantum dynamic equations such as the HEOM. We have focused on solving two problems, the first being the inverse problem. That is, the objective is to determine whether a trained ML model can perform Hamiltonian tomography by using the time dependence of the observables as inputs. We demonstrate the capability of the convolutional neural network (CNN) to solve the inverse problem. That is, the trained CNN can accurately describe the system under study by predicting the parameters of the system Hamiltonian when given the aforementioned time dependent data. The models developed can predict Hamiltonian parameters such as excited state energies and inter-site couplings of a system up to 99.28% accuracy. The second use of the same data set of observables involves time-series analysis. Although various analytical solutions for the dynamics of open quantum systems such as the HEOM have been developed, these often require immense computational resources. We demonstrate that models such as SARIMA, CatBoost, Prophet, convolutional and recurrent neural networks can predict the long-time dynamics provided that the initial short-time dynamics is given. Our results suggest that SARIMA can serve as a computationally inexpensive yet accurate way to predict long-time dynamics.
Applications of digital holograms for the selection and detection of transverse laser modes.
(2014) Ngcobo, Sandile.; Forbes, Andrew.
The transverse mode of generally available commercial lasers in most instances is not suitable for desired applications. Shaping the laser beam either extra-cavity, that is outside the laser resonator, or intra-cavity, which is inside the laser resonator, is required to force the laser beam or cavity to oscillate on a selected desirable single laser mode. The shaped laser beam’s spatial intensity profile and propagation properties would then be suitable for the desired application. The crux of the work presented in this thesis involves intra-cavity beam shaping where we generate desirable transverse modes from inside the laser resonator and detecting such mode using digital holograms. In Chapter 1 we discuss a novel technique of modal decomposition of an arbitrary optical light field into underlying superposition of modes. We show that it can be used to extract physical properties associated with the initial light field such as the intensity, the phase and M2, etc. We show that this novel modal decomposition approach that requires no a priori knowledge of the spatial scale of the modes which lead to an optimised modal expansion. We tested the new technique by decomposing arbitrary modes of a diode-pumped solid-state laser to demonstrate its versatility. In Chapter 2 we experimentally demonstrate selective generation of Laguerre-Gaussian (LG) modes of variable radial order from 0 to 5, with zero azimuthal order. To generate these customised LG modes from within the laser resonator we show that a specialised optical element in a form of an amplitude mask is required to be inserted inside the laser resonator. The amplitude mask is designed and fabricated to contain absorbing rings which are immutably connected to the desired LG mode. The geometry of the absorbing ring radii are selected to match and coincide with the location of the selected LG mode zero intensity parts inside the cavity. We show for the first time that the generated LG modes using this method are of high mode purity and a gain mode volume similar to the desired LG mode. The results provide a possible alternative route to high brightness diode pumped solid state laser sources. In Chapter 3 we show that we can overcome the disadvantage of the specialised optical element being immutably connected to the selection of a particular mode by experimentally demonstrating a novel digital laser capable of generating arbitrary laser modes inside the laser resonator. The digital laser is realised by intra-cavity replacing an end-mirror of the resonator with a rewritable holographic mirror which is an electrically addressed reflective phase-only spatial light modulator (SLM). We show that by calculating a new computer-generated holographic gray-scale image on the SLM representing the desired customized laser mode digitally, the digital laser resonator is capable of generating the desired laser modes on demand. The results provide a new laser that can generate customized laser modes. In Chapter 4 we show that the digital laser can be used as a test bed for conceptualizing, testing, and proving ideas. We experimentally demonstrate this by using a simple laser cavity that contains an opaque ring which is digitally programmed on the SLM and an adjustable circular aperture on the output coupler mirror. We show that by manually varying the diameter of the aperture without realignment of the laser, the generated laser modes can be tuned from a Gaussian mode to a Flat-top mode. This opens up new digital methods that can be used to test laser beam shaping techniques. In Chapter 5 we outline a simple laser cavity comprising an opaque ring and a circular aperture that is capable of producing spatially tuneable laser modes, from a Gaussian beam to a Flat-top beam. The tuneability is achieved by varying the diameter of the aperture and thus requires no realignment of the cavity. We demonstrate this principle using a digital laser with an intracavity spatial light modulator, and confirm the predicted properties of the resonator experimentally. In Chapter 6 we discuss the techniques used to intra-cavity generate and detect LG beams with a non-zero azimuthal index since they are known to carry orbital angular momentum (OAM), and have been routinely created external to laser cavities. We show that the few reports of obtaining such beams from laser cavities suffer from inconclusive evidence of the real electromagnetic field. In this Chapter we revisit this question and show that an observed doughnut beam from a laser cavity may not be a pure Laguerre–Gaussian azimuthal mode but can be an incoherent sum of petal modes, which do not carry OAM. We point out the requirements for future analysis of such fields from laser resonators. In Chapter 7 we conclude and discuss future work.
Applications of light scattering and refraction by atmospheric gases.
(2002) Moorgawa, Ashokabose.; Michaelis, Max M.
LIDAR, an acronym for LIght Detection And Ranging, is a system used for studying the scattering of laser light incident on a parcel of air. This thesis investigates the atmosphere above the Durban region using two atmospheric LIDARs, referred to, in this study, as the "old LIDAR" and the "new LIDAR". The old LIDAR was used in a campaign of observation from July to October 1997 in a study of aerosol concentrations over Durban. This thesis will focus on, among other things, the local aerosol profiles for low altitude (0 to 10 km) and high altitude (10 to 35 km). In particular, the focus will shift on any long persistence in this region (it was found that the aerosol layer observed by M. Kuppen (1996) on June 1994 at 25 km may have moved to the higher altitude of 28 km in October 1997. This may be explained by stratospheric upwelling, carrying the layer to higher altitude. These aerosols are known to influence the local climate). This investigation will give some useful insight into the local atmospheric dynamics. The new LIDAR system (Rayleigh-Mie LIDAR) has been used to measure atmospheric temperatures from 20 to 60 km as well as aerosol extinction coefficients from 15 to 40 km. Height profiles of temperature have been measured by assuming that the LIDAR returns are solely due to Rayleigh scattering by molecular species and that the atmosphere obeys the perfect gas law and is in hydrostatic equilibrium (Hauchecorne and Chanin 1980). Since its installation in April 1999, the new LIDAR has been used to monitor stratospheric temperatures and aerosol concentrations from 10 to 40 km. In this study, we discuss in chapter 7 the results of a validation campaign conducted during the period of April 1999 to December 2000. Average monthly LIDAR temperatures are computed from April 1999 to December 1999 and compared with radiosonde temperatures obtained from the South African Weather Service (SAWS) at Durban. The monthly LIDAR temperature profiles over two years (1999 and 2000) were also computed and compared with the climatological model Cospar International Reference Atmosphere (CIRA)-1986 and with the average monthly European Centre for Medium Range Weather Forecast (ECMWF) temperatures . The results show that there is good agreement between LIDAR and SAWS radiosonde temperatures in the 20 and 30 km altitude range. Between 20 and 40 km, the monthly LIDAR temperatures agree closely with the CIRA-86 and ECMWF profiles. However, during winter, in the altitude range 40 to 60 km, LIDAR temperatures are warmer than CIRA-1986 and ECMWF temperatures, and they show large variability. These variations could be due to relatively fast transient phenomena like gravity waves or planetary waves propagating vertically in the stratosphere. As part of the validation process, the aerosol extinction coefficients retrieved from the LIDAR data have also been compared with the extinction coefficients measured by Stratospheric Aerosol and Gas Experiment (SAGE) II close to the LIDAR location and on coincident days. Appendix E of this thesis also investigates the concept of refraction by atmospheric gases as applied to gas lenses. A simple spinning pipe gas lens (SPGL) has been used as the objective lens of a camera to take pictures of the moon and sun spots. The SPGL is a varifocal length lens which depends on the temperature of the pipe and the angular velocity at which it spins. For our purpose a focal length of 8 m has been used. The moon pictures are compared with a lunar map so as to identify the maria.
Aurora and associated VLF phenomena.
(1978) Duthie, Desmond D.; Scourfield, Malcolm W. J.
Observations have been made at Sanae (gm. lat. -63,71°) on occurring auroral forms, (diffuse and pulsating aurora), and simultaneous occurring VLF phenomena, (whistlers and auroral hiss) . Two studies are presented in this thesis. (a) A comparison of the positions of auroral forms and the positions of field lines, along which whistlers propagate, is made and it is found that: (i) Diffuse aurora occurs on closed field lines and indirect evidence shows that this is also the case for pulsating aurora. (ii) For two periods of data the separation of diffuse aurora from the plasmapause ranges from < 0,9 L to < 0,2 L but during a third period, the diffuse aurora lies, at least partially, within the plasmasphere. (b) An investigation into the association between pulsating aurora and pulsating auroral hiss is made and it is shown that: (i) A common identical pulsation period of 0,75 s and high coherency exists between the light intensity of an auroral patch and the intensity of the associated pulsating auroral hiss. This suggests a wave-particle interaction as a common modulation mechanism. (ii) Cyclotron instability (gyroresonance) or Cerenkov radiation mechanisms occurring in the equatorial plane do not account for the time delays, typically between 0,90 s and 0,157 s, found to be present between, the two phenomena, where the incident auroral electrons, responsible for the auroral patch light intensity, are observed to arrive before the auroral hiss emissions. The results of the analysis in (a) are reported in The Journal of Atmospheric and Terrestrial Physics, 39, 1429, 1977.
Baseline demand responsiveness framework for the conventional grid through appliance scheduling by evolutionary metaheuristics.
(2020) Doho, Goncalves Justino.; Matthews, Alan Peter.; Jarvis, Alan Lawrence Leigh.
A major problem of many energy environments nowadays, is an obsolete and highly inefficient electricity supply system, the Conventional Grid (CG), characterized by a high peak to average ratio, out of an uncontrollable demand, worsened by a native lack of communications infrastructures and resources for performing a proper automated demand side management, which has resulted in blackouts, harsh user discomfort, high electricity cost, huge economic losses and a high carbon footprint. Designed to tackle this problem is the emerging Smart Grid (SG). Most research works are devoted to providing automation and efficiency to the SG (or the intermediate SG-like) environments. There is a scarcity of research devoted to providing automated demand responsiveness to the information layer deprived CG environments, although as evident, an Automated Demand Response (ADR) is badly required, since there is still a long way until we get to the SG, all the more when developing world is concerned. Such context, set our focus towards the CG. So, this research work, developed a framework for providing a "blind" baseline Demand Responsiveness (bbDR) for CG environments, wherein, a pseudo real time electricity pricing function, built from a country load profile, is used as a guiding function for the autonomous scheduling of controllable appliances, which seeks to improve electricity consumption patterns, while also preserving user satisfaction by complying to their preferences. For performance evaluation, the optimized energy consumption patterns (peak load, peak to average ratio, load and cost profiles and mean energy rate) of the controlled use of appliances, are compared to those ones produced by their uncontrolled use. The controlled usage schedules are produced by an evolutionary metaheuristics, whilst the uncontrolled usage is stochastically generated from appliances’ rate-of-use probabilities sourced from the literature. The results proved that, such framework is capable of, without DR communications, delivering meaningful, ADR-like, performances to a communications deprived CG environment. As part of the work for simulating the above bbDR framework, we developed and demonstrated a Real Parameter Blackbox Optimization Approach to Appliance Scheduling (RPBBOAS) model, which describes the household, and provides the logical interface with the optimization algorithms. This real parameter model, vis-a-vis its discrete parameter counterpart, tackles combinatorial explosion by, in a novel way, reducing the problem dimension that is traded with the external blackbox optimization algorithms, in such a way that boosts performance and widens the window of applicable algorithms. While developing the above RPBBOAS model, readily available state-of-the-art metaheuristics showed a lackluster performance, which propelled us to design a novel hybrid evolutionary metaheuristics (Hy- PERGDx) that was eventually used in the bbDR simulations. It showed a better all-around performance and robustness vs the state-of-the-art, when benchmarked on a wide range of non-linear problems. Overall, such deliveries, demonstrated the potential of the proposed bbDR framework for improving demand patterns and quality of service figures, in a communication free way, which with an appropriate follow-up development, makes it suitable for application in severely affected, communications deprived (or communications limited), energy networks such as South Africa or worse energy ecosystems.
Bulk heterojunction organic solar cells and thin film electrode buffer layers : synthesis, preparation and characterization.
(2016) Arbab, Elhadi Abdalla Adam.; Mola, Genene Tessema.
Abstract available in PDF file.
Characterisation of quantum channels in plasmonic metamaterials and bulk optical systems.
(2018) Uriri, Solomon Akpore.; Tame, Mark Simon.
Quantum channels are key to our understanding of how quantum information can be processed and transmitted. In this respect, over the past decade light has become an important carrier of quantum information. More recently, metamaterials have opened up many new exciting ways of controlling and manipulating light in the quantum regime, and in particular, controlling the polarisation and orbital angular momentum of light. In this work, we undertake an indepth characterisation of quantum channels made from plasmonic metamaterials and bulk optical systems by probing them with quantum states of light. We rst experimentally demonstrate the active control of a plasmonic metamaterial operating in the quantum regime. Using an external laser, we control the temperature of the metamaterial and carry out quantum process tomography on single-photon polarization-encoded qubits sent through, characterizing the metamaterial as a variable quantum channel. We nd that the overall polarization response can be tuned by up to 33%. Second, we experimentally realise a more complicated type of quantum channel in the form of a non-Markovian process made from the sum of two Markovian processes, and a Markovian process from two non-Markovian processes in a comparable bulk optical system. We perform quantum process tomography, and obtain high process delities. We discuss how these more complex types of quantum channel may be implemented using metamaterials.
Classical noise in quantum systems.
(2013) Akhalwaya, Ismail Yunus.; Petruccione, Francesco.
Quantum mechanics contains a fresh and mysterious view of reality. Besides the philosophical intrigue, it has also produced and continues to inspire tantalizing new technological innovations. In any technological system, the designers must contend with the problem of noise. This thesis studies classical noise in two different quantum settings. The first is the classical capacity of a quantum channel with memory. Adding forgetful-memory, attempts to push the boundaries of our understanding of how best to transmit information in the presence of correlated noise. We study the noise within two different frameworks; Algebraic Measure theory and Monte Carlo simulations. Both tools are used to calculate the capacity of the channel as correlations in the noise are increased. The second classical-quantum system investigated is atomic clocks. Using power spectral density methods we study aliasing noise induced by periodic-correction which includes the Dick Effect. We propose a novel multi-window scheme that extends the standard method of noise correction and exhibits better anti-aliasing properties. A uniting thread that emerges is that correlations can be put to good use. In the classical capacity setting, correlations occur between uses of the quantum channel. We show that stronger correlations increase the classical capacity. The benefits of correlation are even seen at a meta-level within the framework of Monte Carlo simulations. Correlations are designed into the algorithm which have nothing to do with real-world correlations, but are abstract correlations created by a Markov chain employed in the algorithm to help efficiently sample from a distribution of exponential size. Finally, in the atomic clock setting, correlations in the measured noise are used to help predict and cancel noise on a short time-scale while trying to limit aliasing. Channel capacity and precise time-keeping are distinct topics and require very different approaches to study. However, common to both topics is their application to com- munication and other tasks, the need to overcome noise and the benefits of exploiting correlations in the noise.
Clustering analysis for classification and forecasting of solar irradiance in Durban, South Africa.
(2017) Govender, Paulene.; Matthews, Alan Peter.; Brooks, Michael John.
Classification and forecasting of solar irradiance patterns has become increasingly important for operating and managing grid-connected solar power plants. A powerful approach for classification of irradiance patterns is by clustering of daily profiles, where a profile is defined as irradiance as a function of time. Classification is useful for forecasting because if the class of a day can be successfully forecast, then the irradiance profile of that day will share the general pattern of the class. In Durban, South Africa (29.871 °S; 30.977 °E), beam and diffuse irradiance profiles were recorded over a one-year period and normalized to a clear sky model to reduce the effect of seasonality, from which several variables were derived, namely minute-resolution beam, hourly-resolution beam and diffuse, and hourly-resolution beam variability. To these variables, individually and in combination, k-means clustering was applied, and beam irradiance was found to be the one that best distinguishes between sky conditions. In particular, clustering of hourly-resolution beam irradiance produced four classes with diurnal patterns characterized as sunny all day, cloudy all day, sunny morning-cloudy afternoon, and cloudy morning-sunny afternoon. These classes were then used to forecast beam and diffuse irradiance for the day ahead, in association with cloud cover forecasts from Numerical Weather Prediction (NWP) output. Two forecasting methods were investigated. The first used k-means clustering on predicted daily cloud cover percentage profiles from the NWP, which was a novel aspect of this research. The second used a rule whereby predicted cloud cover profiles were classified according to whether their averages in the morning and afternoon were above or below 50%. From both methods, four classes were obtained that had diurnal patterns associated with the irradiance classes, and these were used to forecast the irradiance class for the day ahead. The two methods had a comparable success rate of about 65%. In addition, hour-ahead forecasts of beam and diffuse irradiance were performed by using the mean profile of the forecast irradiance class to extrapolate from the current measured value to the next hour. The method showed an average improvement of about 22% for beam and diffuse irradiance over persistence forecasts. These results suggest that classification of predicted cloud cover and irradiance profiles are potentially useful for development of class-specific, multi-hour irradiance forecast models.
Computational and experimental study of thin films of polymer blends.
(2016) Megnidio-Tchoukouegno, Mireille Merlise.; Pellicane, Giuseppe.; Mola, Genene Tessema.
Abstract available in PDF file.
Computational studies of bond-site percolation.
(2007) Nduwayo, Léonard.; Chetty, Nithaya.; Lindebaum, Robert James.
Percolation theory enters in various areas of research including critical phenomena and phase transitions. Bond-site percolation is a generalization of pure percolation motivated by the fact that bond-site is close to many physical realities. This work relies on a numerical study of percolation in lattices. A lattice is a regular pattern of sites also known as nodes or vertices connected by bonds also known as links or edges. Sites may be occupied or unoccupied, where the concentration ps is the fraction of occupied sites. The quantity pb is the fraction of open bonds. A cluster is a set of occupied sites connected by opened bonds. The bond-site percolation problem is formulated as follows: we consider an infinite lattice whose sites and bonds are at random or correlated and either allowed or forbidden with probabilities ps and pb that any site and any bond are occupied and open respectively. If those probabilities are small, there appears a sprinkling of isolated clusters each consisting of occupied sites connected by open bonds surrounded by numerous unoccupied sites. As the probabilities increase, reaching critical values above which there is an infinitely large cluster, then percolation is taking place. This means that one can cross the entire lattice by going successively from one occupied site connected by a opened bond to a neighbouring occupied site. The sudden onset of a spanning cluster happens at particular values of ps and pb, called the critical concentrations. Quantities related to cluster configuration (mean cluster and correlation length) and individual cluster structure (size and gyration radius of clusters ) are determined and compared for different models. In our studies, the Monte Carlo approach is applied while some authors used series expansion and renormalization group methods. The contribution of this work is the application of models in which the probability of opening a bond depends on the occupancy of sites. Compared with models in which probabilities of opening bonds are uncorrelated with the occupancy of sites, in the suppressed bond-site percolation, the higher site occupancy is needed to reach percolation. The approach of suppressed bond-site percolation is extended by considering direction of percolation along bonds (directed suppressed bond-site percolation). Fundamental results for models of suppressed bond-site percolation and directed suppressed bond-site percolation are the numerical determination of phase boundary between the percolating and non-percolating regions. Also, it appears that the spanning cluster around critical concentration is independent on models. This is an intrinsic property of a system.
Computational study of the structure and optical properties of perovskite solar cells materials.
(2021) Kheralla, Adam Abdallah.; Chetty, Naven.
Computer simulation of nonadiabatic dynamics by means of the quantum-classical Liouville equation.
(2013) Uken, Daniel Alexander.; Sergi, Alessandro.
Simulation of quantum dynamics for many-body systems is an open area of research. For interacting many-body quantum systems, the computer memory necessary to perform calculations has an astronomical value, so that approximated models are needed to reduce the required computational resources. A useful approximation that can often be made is that of quantum-classical dynamics, where the majority of the degrees are treated classically, while a few of them must be treated quantum mechanically. When energy is exchanged very quickly between the quantum subsystem and classical environment, the dynamics is nonadiabatic. Most theories for nonadiabatic dynamics are unsatisfactory, as they fail to properly describe the quantum backreaction of the subsystem on the environment. However, an approach based on the quantum-classical Liouville equation solves this problem. Even so, nonadiabatic dynamics is di cult to implement on a computer, and longer simulation times are often inaccessible due to statistical error. There is thus a need for improved algorithms for nonadiabatic dynamics. In this thesis, two algorithms that utilise the quantum-classical Liouville equation will be qualitatively and quantitatively compared. In addition, stochastic sampling schemes for nonadiabatic transitions will be studied, and a new sampling scheme is introduced [D. A. Uken et al., Phys. Rev. E. 88, 033301 (2013)] which proves to have a dramatic advantage over existing techniques, allowing far longer simulation times to be calculated reliably.
Computer-aided diagnosis in mammography : correlation of regions in multiple standard mammographic views of the same breast.
(2006) Padayachee, Jayanethie.; Alport, Michael J.; Rae, William Ian Duncombe.
Abstract available in PDF file.
The context of problem tasks in school physical science.
(1999) Hobden, Paul Anthony.; Schuster, David.; Volmink, John.
The purpose of this study was to extend our current knowledge about what happens in physical science classrooms. The focus was the context of problem tasks. This involved the study of the situations, events and factors that relate to the solving of problem tasks at high school in order to understand their role and nature. e problem tasks that were central to this study were well defined, narrow in focus, and invariably involved the calculation of some quantity through the use of a formula and algebraic manipulation. The main questions that guided the study were as follows: What is happening in physical science classrooms? What is the nature and role of problem solving within this context? What are some of the consequences of organising teaching and learning in this manner? How do external forces influence what happens? The study aimed at describing the activities that the teachers and students were involved in and understanding how they understood their own actions. An interpretive research approach was chosen for this purpose, having as its basis a detailed descriptive foundation using classroom observation. Two high school science classrooms were studied in detail over a period of a year. The data gathered included field notes from over a hundred classroom visits, extensive video and audio records, questionnaires, classroom documents and formal an informal interviews with teachers, students and examiners. Through a process of careful and systematic analysis of the data, six assertions emerged. These assertions are supported by both particular evidence in the form of analytic narrative vignettes, quotes and extracts, and general evidence consisting of frequency data and summary tables. The analysis reveals that problem tasks occupied most of the teaching and learning time, and that the students found this experience of school science boring. Most of the problem tasks were routine in nature and of low conceptual demand. The majority of the students were unable to solve the more difficult tasks encountered in their tests and examinations. In addition, a significant number could not solve the routine problem tasks. This suggests that the predominant instructional strategies were ineffective. It was found that participants had an uncritical belief in the efficacy of teacher explanations and student practice on problem tasks. Further, the participants had different views of the role of problem tasks. A significant finding was that the examination exerted a powerful focusing influence on the classroom environment, the instructional activities and on the problem tasks used . It appeared that the ultimate goal of school physical science was to solve these types of problem task in preparation for the high stakes examination, rather than the learning of science. The study has implications both for practice and for research on the teaching and learning of school physical science. These implications are discussed in terms of instructional strategies aimed at promoting a deeper understanding of physical science. In order to improve practice it is advocated that the role of problem tasks in learning science be made explicit while at the same time new types of instructional task need to be designed to achieve our goals for school science.

Browse

Browsing Doctoral Degrees (Physics) by Title