Search

Energy storage is a field which governs almost all aspects of daily life. Batteries and capacitors are utilised in devices at every scale, from mobile phones all the way up to the national grid. Efficient energy storage requires extensive research into chemistries and configurations, utilising novel materials to assemble devices. These often have a significant effect on features including the energy and power density of a cell. One such property is the pore structure of an electrode. Electrochemical Double-Layer Capacitors, also known as supercapacitors, utilise meso/microporous electrodes with a wide distribution of pore sizes. Larger pores typically facilitate ion diffusion and a high device power density, yet smaller micropores provide a large specific surface area for the electric double layer which facilitates a high device energy density. Recent advancements in post-lithium ion battery technology (such as lithium-sulphur and lithium air batteries) utilise these micropore surface areas for redox reactions. Developing accurate models of energy storage devices allows for the simulation of the processes within the cells without the need for expensive and time-consuming experimental testing. Many models have been developed in the past, however sweeping assumptions are often made with respect to the pore structure. These generally assume a uniform pore size and structure, and do not take into account the effects of different pore sizes present in a true electrode material. Pore network models are able to do this, but require complex or expensive pieces of software and a long simulation time. In this project, a novel continuum model of mass and charge transport has been developed catering for the many pore sizes in an electrode. This model allowed for the simple implementation of a pore size distribution into the simulation of the processes in several supercapacitor configurations, a lithium-sulphur battery, and two lithium-oxygen battery configurations. This novel approach involved characterising an electrode material through analysis of the pore structure and pore size distribution. A novel mathematical transient volume averaged model was developed to solve for the mass transport of species within pores of different sizes. The use of this model in each case demonstrated good agreement with experimental data and allowed for analysis of the effects of different pore structures on the activity of cells. It was found that larger pores (in the macropore region) facilitate mass transport of species, and smaller pores (in the micropore region) have a reduced rate of mass transport and a higher rate of reaction (in batteries) and stern layer formation (in supercapacitors). This novel model demonstrated the importance of modelling ion transport through multiple pore sizes of an electrode material. This model also addresses the need for more complex energy storage device simulation in a reasonable solving time without the need for expensive pieces of software.

The plasma treatment of few-layer graphene (FLG) was investigated for the effect on the performance in supercapacitors and microsupercapacitors. Modifications to the FLG surfaces were proven by comprehensive studies using characterisation techniques including elemental microanalysis, X-ray photoelectron spectroscopy, potentiometric titration, zeta-potential measurements and dispersion stability analysis. A thermal pre-treatment to yield dried FLG was shown to increase the FLG surface charge and density due to the removal of adsorbed water and incorporation of carboxyl and phenolic functional groups. The thermal treatment was used before all characterisation methods were applied. An Ar gas plasma treatment on dried FLG was shown to introduce carboxyl and phenolic surface functional groups and reduce material variability. With increasing treatment time of Ar plasma, the FLG oxygen content increased by 1 at% due to the presence of a larger number of carboxyl functional groups. The introduction of H2 gas at 3 wt% in Ar gas plasma produced a different functionalised FLG with a smaller quantity of carboxyl groups. Compared to the Ar gas plasma treated material, the H2/Ar gas treated material had a lower surface charge and density. NH3 gas plasma increased the nitrogen content of the FLG starting material two-fold according to XPS and elemental microanalysis. A multi-step treatment consisting of H2/Ar mixed gas plasma followed by NH3 gas plasma gave a further surface increase in nitrogen content by five times relative to the starting material. Electrode films were manufactured using polytetrafluoroethylene (PTFE) and styrene-butadiene rubber (SBR) as non-conductive binders and poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT/PSS) as a conductive binder. The electrode films were constructed into supercapacitors and microsupercapacitors. External testing with PTFE binder showed promise for the H2/Ar mixed gas plasma treated material with a specific capacitance of 195 F g-1 at 1 mA cm-2 in sulfuric acid electrolyte, a 50% increase relative to untreated FLG devices. The other treated materials did not match this performance because they did not contain low concentrations of oxygen surface functional groups and had large quantities of sp3 hybridised carbon atoms. The supercapacitor devices were studied with sulfuric acid and potassium hydroxide aqueous electrolytes. The devices with potassium hydroxide electrolyte did not match the performance of the materials in sulfuric acid electrolyte due to materials’ incompatibilities. Supercapacitor testing was additionally carried out with SBR and carboxymethyl cellulose (CMC) in sandwiched devices alongside potassium hydroxide and sulfuric acid electrolytes. The manufacture of the electrode films required 20 wt% of SBR-CMC binder. The electrochemical results were indistinguishable and had large resistances. An extensive investigation into the manufacture of flexible electrode films with FLG, multi-walled carbon nanotubes (MWCNT) and PEDOT/PSS in composite electrode materials was carried out. The electrode films were laser-etched into interdigitated patterns for planar micro-supercapacitor devices with the application of polyvinylalcohol-phosphoric acid gel electrolyte. These devices performed best with a mass ratio of 1:3:1 (PEDOT/PSS:FLG:MWCNT), and with NH3 functionalised FLG and acid functionalised MWCNT. Gravimetric capacitances of 120 F g-1 at 5 mV s-1 and volumetric capacitances of 20 F cm-3 at 5 mV s-1 were obtained for the NH3:MWCNT(Acid) combination during long cycling tests (10,000 cycles) and showed capacity retentions > 80%. In-situ Raman microscopy analysis suggested that the PEDOT/PSS component underwent pseudo-capacitive, reversible changes during cycling tests but the dominant electric-double layer capacitive-like response was due to the FLG and MWCNT materials, which were highly stable.

This thesis sets out to answer the following question: To what extent can inverter reactive power control, combined with voltage regulators, increase network effciency and hosting capacity of distributed generation? The first part of this thesis is concerned with analytic methods for studying this question; therefore, a two-bus network model is used. Two aspects of the problem are considered using this model. Firstly, the maximum power transfer capabilities of distributed generation are considered. This bounds the maximum generation of a network, even if the marginal costs of both real and reactive power are zero. It is shown that losses caused by the reactive power and generation result in maximum power transfer before stability limits, in networks with high impedance (R=X) ratios. Secondly, analytic bounds on a proposed 'relative energy loss fraction' are derived in closed form, using the LinDistFlow approximate solution. The bounds are validated by comparison with both the exact solution of the two-bus equations, and on a set of unbalanced distribution networks. The relative energy loss fraction was found to be as high as 30%, with the approximate bounds able to estimate the true bounds to within 5%. The next part of this thesis studies how reactive power and taps can be used to increase hosting capacity under uncertainty in the locations and sizes of domestic generation. A linear method is proposed that is shown to reduce the computing runtime by as much as 1000 times. A proposed hosting capacity sensitivity is shown empirically to correlate well with the error caused by linearization. Centralised control of taps and regulators are shown to increase hosting capacity by as much as 70%. Finally, an optimization is proposed to study the impact of taps and domestic inverter control on the feeder total power draw. The use of regulator control in isolation results in benefits of up to a 5% reduction in feeder power; the use of inverter reactive power control further reduces the feeder power by up to 0.5% of the load. A control scheme is proposed with a reduced communications overhead, which can obtain up to 97% of the potential benefits of full inverter control.

This thesis investigates the impact that home charging of a large private fleet of electric vehicles would have on the power system. A large multi-regional travel survey dataset is used to model vehicle use and charging spatially heterogeneously, and a selection of representative network models are used to assess the impact of charging on system operation. A stochastic data-driven model is proposed to model uncontrolled charging of vehicles, and convex optimisation is used to calculate the optimal smart charging strategy. The power system is commonly broken down into the generation, transmission, and distribution systems. The operation of each of these systems will be impacted by the addition of EV charging to residential networks. A variety of objectives have been proposed for smart charging, each of which would protect the system in a different way. Existing research tends to focus on a single part of the system, and considers only the smart charging objective that most benefits that part of the system. Here, the three systems are modelled simultaneously, and a large range of smart charging objectives are investigated. The value of explicit loss minimising smart charging is quantified, compared to a simpler and more standard load flattening algorithm. These results are used to propose a novel optimisation formulation which reduces losses without requiring extensive network information. The value of bi-directional smart charging is also quantified compared to uni-directional smart charging, in order to investigate the viability of residential vehicle-to-grid. It is demonstrated that it is not possible to optimise the transmission level and distribution level systems simultaneously, and the penalty of only optimising for one is quantified. A method for finding a compromising solution between both system levels is proposed, which exploits the sections of the distribution where components are over-specified. Two specific case studies are investigated. The majority of the analysis in the thesis is based on the GB power system, however the Texas system is also presented as a comparative case study.

The conventional grid system is facing great challenges due to the fast growing electricity demand throughout the world. The smart grid has emerged as the next generation of grid power systems, aimed at providing secure, reliable and low cost power generation, distribution and consumption intelligently. The smart grid communication system within the smart grid network is of fundamental importance to support data transfer and information exchange within the smart grid system. The National Institute of Standards and Technology has identified wireless communications as an important networking technology to be employed in power systems. The reliability of the data transmission is essential for the smart grid system to achieve high accuracy for the power generation, distribution and consumption. In this thesis, we investigate cooperative communications to improve transmission reliability in smart grid networks. Although many issues within cooperative communication have already been addressed, there is a lack of research efforts on cooperative communication for the wireless smart grid communication system which has its own network features and different transmission requirements. In our research, the smart grid communication networks were studied, and cooperative communications in smart grid networks were analysed. The research work mainly focuses on three problems: the application of cooperative relay communications to modern smart grid communication networks, the cooperative relay-based network development strategy, and the optimization of cooperative relay communication for smart grids. For the first problem, the application of cooperative relay communication to a home area network (HAN) of smart grid system is presented. The wireless transmission reliability is identified as the issue of most concern in wireless smart grid networks. We model the smart grid HAN as a wireless mesh network that deploys cooperative relay communication to enhance the transmission reliability. We apply cooperative relay communication to provide a user equipment selection scheme to effectively improve the transmission quality between the electricity equipment and the smart meter. For the second problem, we address the network design and planning problem in the smart grid HAN. The outage performance of direct transmission and cooperative transmission was analysed. Based on the reliability performance metric that we have defined, we propose a HAN deployment strategy to improve the reliability of the transmission links. The proposed HAN deployment strategy is tested in a home environment. The smart meter location optimization problem has also been identified and solved. The simulation results show that our proposed network deployment strategy can guarantee high reliability for smart grid communications in home area networks. For the third problem, the research focuses on the optimization of the cooperative relay transmission regarding the power allocation and relay selection in the neighbourhood area network (NAN) of the smart grid system. Owing to the complexity of the joint optimization problem, reduced-complexity algorithms have been proposed to minimize the transmission power, at the same time, guarantee the link reliability of the cooperative communications. The optimization problem of power allocation and relay selection is formulated and treated as a combinatorial optimization problem. Two sub-optimal solutions that simplify the optimization process are devised. Based on the solutions, two different algorithms are proposed to solve the optimization problem with reduced complexity. The simulation results demonstrate that both two algorithms have good performance on minimizing the total transmission power while guaranteeing the transmission reliability for the wireless smart grid communication system. In this thesis, we consider cooperative communications in a smart grid scenario. We minimize the outage probability and thus improve the reliability of the communications taking place in the smart grid by considering the optimization problem of power control, relay selection and the network deployment problem. Although similar problems might have been well investigated in conventional wireless networks, such as the cellular network, little research has been conducted in smart grid communications. We apply new optimization techniques and propose solutions for these optimization problems specifically tailored for smart grid communications. We demonstrate that, compared to naively applying the algorithms suitable for conventional communications to the smart gird scenario, our proposed algorithm significantly improves the performance of smart grid communications. Finally, we note that, in future work, it will be possible to consider more complex smart grid communications system models. For example, it is worthwhile considering hetregeneous smart communications by combining HAN and wide area networks (WAN). In addition, instead of assuming that all communications have the equal priority, as in this thesis, more comprehensive analysis of the priority of the smart grid communication can be applied to the research.

The technology of photovoltaics provides a renewable and ecofriendly way of conversion of photons to electricity. Low cost photovoltaics with high efficiency and extended life time are some of the main goals during their fabrication. Perovskite solar cells (PSC) have been promoted to potential candidates in the field of photovoltaics. In this thesis, chapter 1 provides a general background theory about the field of perovskites and their use to fabricate PSC. In chapter 2, experimental details are provided, supporting the work that was made and the relevant theory of characterization techniques used. Chapter 3, mainly focusing on a low-cost, energy efficient and water-free synthesis of lead iodide (PbI2) a material widely used for the fabrication of perovskites, by planetary ball milling (mechanochemical), controlling both stoichiometry and number of polytypic phases. Chapter 4, focusing on the stability of PSC fabricated with an optimum stoichiometry between different ions present in the photoactive layer. Synthesis of controllably doped or undoped methylammonium lead triiodide (CH3NH3PbI3) micro-crystals with various amounts of dopants of either methylammonium bromide (CH3NH3Br) or formamidinium iodide (CH(NH2)2I) or CH(NH2)2I and bromine (Br2), was achieved. Characterization techniques used, included X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), Rutherford backscattering spectrometry (RBS), atomic force microscopy (AFM), scanning electron microscopy (SEM), UV-Vis spectroscopy, controllable chemical synthesis of micro-crystals and testing operational PSC devices. The results presented in chapter 3, indicated that the use of the sub-stoichiometric PbI2 (when iodine content is below the optimum 1 : 2 ratio of Pb : I atoms) during the fabrication of the photoactive layer in PSC devices, results in unreacted PbI2. The presence of unreacted PbI2 reduces the average power conversion efficiency (PCE) of the PSC from ~ 15.5 % to ~ 4.6 %. Whereas, the results presented in chapter 4, indicated that controllable doping of CH3NH3PbI3 micro-crystals with different incorporated ions, reveals a minimum of distortion in the resulted perovskite crystal structure. That might lead to an optimum composition in perovskites with enhanced chemical stability as function of the exposed conditions. Un-encapsulated PSC devices fabricated using controllably doped CH3NH3PbI3 micro-crystals, showed a reduction in the average PCE from ~14 (±1.5)% to ~13.7 (±0.5)%, after storage in dark and under ambient conditions for 240 hours.

This thesis improves charge selective contacts for perovskite solar cells. First, mesoporous modification of n-type metal oxides is discovered to shift the work function of the perovskite, which enables n-type doping at the interface. A chapter on ZnO contacts elucidates chemical degradation mechanisms at the perovskite-ZnO interface, and a methylammonium-free perovskite is employed to suppress these reactions. Next, a Cs salt-treatment is developed to improve both the thermal stability and the Voc of cells on SnO2. In the following chapter, a novel organic p-type is one of the first alternatives to Spiro-OMeTAD that enables 20% power conversion efficiency. Finally, a dopant-free photo-cross-linkable polymer is presented that offers exceptional thermal stability for p-i-n cells, with device lifetimes exceeding 3,000 hours at 85°C.

Given the large energy resource available in ocean waves, wave energy converters have been developed over the last decades for power extraction. Various concepts exist, and research efforts are now focussed on reducing their levelised cost of energy. The device structure has been identified to have the highest cost reduction potential. For this reason, a number of hull geometry optimisation studies have been performed in recent years. In these studies, costs have been mostly represented through the device size or weight, and devices have been optimised for specific sea conditions, based generally on simple shapes such as spheres or cylinders. However, there is no consensus in the employed methodology and resulting shapes might be difficult to manu facture or unable to survive in high energetic seas. The goal of this thesis is, therefore, to develop a device-agnostic methodology for geometry optimisation of wave energy converters, which enables the generation of improved hull shapes that reduce the levelised cost of electricity. An existing approach for single body floating point-absorbers, exhibiting some of the best practices found in this field, is re-implemented and extended to improve its robustness for its application to different case studies. Each of the elements composing this approach (how the geometry is defined, the choice of objective function and the choice of optimisation algorithm and set-up) are then evaluated and their suitability is assessed through comparison to other strategies. The method is then applied to a range of study cases, such as to study the effect of location and of the choice of modes-of-motion for power extraction on the optimal hull shape. Further extensions of the method to include manufacturability and reliability considerations, as well as to include the effect of mass distribution are investigated. As a result, recommendations are formulated for the set-up of an early stage WEC design geometry optimisation process. Additionally, trends for the hull shape design are identified for the considered cases - depending on, location, the choice of the modes of motion for power extraction, and how costs are accounted for.

The main focus of my PhD is to investigate and explore more efficient, reduced switching devices count four-level π-type converter topologies for low voltage applications (Vdc < 600V). This thesis presents the work on four-level π-type converters during my PhD period. Average analytical mathematical models for analyzing the device power loss and the efficiency of the four-level π-type converter topology have been established. It has been found out that with the same input power level and the same dc input voltage, the four-level π-type converter can provide higher efficiency when switching frequencies are above 5 kHz compared to two-level and three-level converters due to lower switching losses. In order to resolve dc-link neutral points (NP) voltages unbalancing issues of neutral points clamped (NPC) multilevel converters, a carrier-based modulation (CB-PWM) NPs’ voltages balancing control with optimum zero-sequence signals injection for the four-level π-type converter has been investigated. Simulations and 300V dc input voltage experimental results proved that with a back-to-back configuration, the proposed control method is able to balance the three dc-link capacitors’ voltages even at high modulation indices and high-power-factor conditions. For the purpose to make the four-level π-type converter topology work as a single-end converter (inverter or rectifier) with balanced dc-link capacitors’ voltages, a modified topology based on the original four-level π-type converter has been developed and analyzed. This new topology is named as the hybrid-clamped four-level π-type converter with one more flying capacitor (FC) as well as two additional switching devices. With such layout modification, more redundant switching states can be generated to regulate neutral path currents. Therefore, the hybrid clamped four-level π-type converter does not have modulation index or power factor limitations when operating as a single-end converter.

The smart grid incorporates an integrated system of smart meters and communication networks that enable two-way communication between utilities and consumers. The granular information from smart meters can be used to improve the load forecast and influence consumer's energy consumption patterns through demand side management (DSM). However, for localized studies of power system, using a large quantity of smart meter data having high level of noise preclude the use of computationally intensive techniques. Reduction of smart meter data to extract the load profiles and smoother load profiles at lower aggregation level (individual consumer or small groups of consumers) are highly desirable for use in linear techniques for power system studies. Therefore, this thesis addresses the challenges of smart meter data size, complexity, variability and volatility for efficient use in load forecasting and DSM. This thesis presents a novel clustering-based approach for analysis of smart meter data, aimed at more accurate and detailed load profiling, reduced profile complexity and improved load forecast accuracy and DSM solutions. The approach uses an innovative clustering algorithm to reduce the data size by proposing new cluster validity indices. The extremely volatile profiles having high levels of noise and complexity are linearized using Taylor series linearization process to alleviate the non-linearity and complexity of profiles. Finally, particle swarm optimization is applied for energy optimization in linearized profiles. The approach is demonstrated on Irish smart meter dataset and simulated PV data, to achieve improved load forecast accuracy using artificial neural network and improved DSM solutions using linear optimization with reduced computational burden. Investigations suggest that proposed clustering algorithm can produce clusters with high intra-cluster pattern similarity as a result of the introduction of new stopping criteria specifically tailored for load forecasting applications. A comparison between the proposed alternative profiles and raw profiles further suggests that the alternative profiles guide the underlying energy consumption with reduced complexity making them computationally efficient. Use of the alternative profiles suggests that the load forecasting accuracy can potentially be higher compared to raw profiles. The alternative profiles in combination with the novel cluster selection approach provide higher peak reduction by shifting the load from peak hours to off-peak hours and higher monetary benefits for the participating consumers. The proposed clustering algorithm and the alternative profiles represent an advancement of the conventional load profiling approach, benefiting system operators through more accurate forecasting and efficient DSM. The novel mathematical framework suggested in this thesis provides an advancement to the new knowledge in the area of smart metering and smart power grids.

The wide range of applicability of the Pelton turbines with their robustness has made them popular turbines for renewable energy generation. Being impulse type turbine, the Pelton buckets are subject to complex turbulent multiphase flows with free surfaces. This turbine has thus been studied with great interest by Computational Fluid Dynamics (CFD) researchers to visualise and evaluate the flow patterns. However, the investigations for design optimisation have been conducted independently for each feature or by ‘trial and hit’ method due to the complexity in incorporating substantial number of design parameters. Analysing the complex fluid flow on each stochastic pre-optimised design was rather challenging due to expensive computational needs of CFD and restricted manufacturing possibilities limiting the usage of artificial intelligence for optimisation. Today, development of accurate and inexpensive CFD models and innovative manufacturing technology such as rapid prototyping has made complex freeform shape possible to simulate and manufacture. Such optimisation will require the combination of CFD and surface generation with optimisation in a novel analysis approach. In this work, Adapted Delaunay Triangulation (ADT) method is proposed as a lightweight alternative for surface generation from the random points cloud. The surface generated from ADT method is used with genetic algorithm optimisation engine developed inhouse. DualSPHysics is used as the CFD application to search for the optimal results with minimum computational costs. It was observed while comparing ADT with NURBS based geometry configuration that the computation time can be reduced by 3 folds and the storage space required is reduced by half. The DualSPHysics method completes a CFD simulation with acceptable accuracy within 3,876 seconds as opposed to 172,800 seconds (48 hours) with Fluent. Each simulation with DualSPHysics requires only 354 MB as compared to the 250 GB storage space required for Fluent. The optimised Pelton bucket showed a simulated improvement of 9.4% and experimental improvement of 1.9% on power output efficiency. The conducted research has provided a successful result and can be reused for higher accuracy or even for other CFD responsive surface optimisation applications.

The Achilles heel of many wearable and electronic textile (E-textile) devices is their power requirement, which has been a major hurdle in the adoption of E-textiles. To keep these devices continuously powered without frequent recharging or bulky energy storage devices, many have proposed integrating energy harvesting capability into clothing. Solar energy harvesting has been one of the most investigated avenues for this due to the abundance of solar energy and maturity of photovoltaic technologies. This research investigated a novel approach for realising solar energy harvesting with textiles by embedding miniature solar cells (SCs) within the fibres of a yarn, thus delivering a robust and consumer-friendly solution for powering wearable and mobile devices. SCs were first soldered onto fine copper wires and encapsulated inside of resin micro-pods, before being covered by a fibrous sheath, to realise solar cell embedded yarns (solar-E-yarns) that can be readily converted into fabrics with conventional fabric manufacturing processes such as weaving and knitting. Preliminary investigations conducted using miniature photodiode embedded E-yarns laid the foundation for embedding photovoltaic devices within yarns. A mathematical model was also formulated to characterise the performance of photovoltaic devices embedded in yarns and was experimentally validated using photodiodes to evaluate the effects of the resin micro-pod on photovoltaic response. Subsequently solar-E-yarns were fabricated using silicon SCs. The photovoltaic response of these solar-E-yarns were studied at each stage of the E-yarn fabrication process and under a range of test conditions including different light intensities, incident light angles, ambient temperatures and humidity levels. Solar-E-yarn performance could be further enhanced by impregnating the photoactive sides of the yarns with an optically clear resin, as well as by using bifacial SCs. A series of fit-for-purpose tests including wash durability tests were conducted on the solar-E-yarns which revealed that the solar-E-yarn embedded fabrics could undergo domestic laundering and maintained ~90% of the original power output after 15 machine wash cycles, which was vastly superior to other solutions proposed in the literature. To demonstrate the energy harvesting capability, prototype demonstrators were created by weaving solar-E-yarns. A solar fabric demonstrator with ~25cm2 active area generated up to ~2.15 mW/cm2 under one sun illumination and maintained both the feel and aesthetics of a normal textile. The fabric demonstrator was capable of charging various electronic storage and powering low power mobile devices. The research has generated a wealth of knowledge on the fabrication, performance and the utility of the solution for regular clothing applications. These attributes will enable these solar fabrics to feature in future wearable electronics and electronic textiles to provide a continuous supply of power, without having to compromise on comfort, aesthetics or wash durability.

Power distribution systems, specifically where they intersect with the built environment, are highly underemphasised versus generation in power planning. In a time of technology advances and cost declines in distribution automation and related technologies, this is an area of high potential for improving energy efficiency. This is particularly of impact in developing countries where urbanisation is rapidly increasing. Evidence shows that the same missed opportunities and sub-optimal distribution planning techniques are repeatedly found across multiple geographies. In this research, tools were developed to rank these problems and create solutions. These tools were endorsed by power industry executives from three countries. Following this, the tools were applied in a developing corridor near the Thailand-Cambodia border where power density is increasing, in order to develop power system solutions for live infrastructure projects. The solutions include technologies such as distributed generation, microgrids, digital monitoring systems, CCHP units, and power storage. The solutions from the live example were then honed and endorsed in an interview with Thai power sector experts. The final research and tools developed were confirmed capable of producing actionable solutions for planners across the public and private sectors, who focus on power distribution in urbanising, developing counties.

Many government policies have been implemented worldwide aiming at accelerating the development of renewable energy systems and therefore, reducing carbon emissions. This has led to increased wind energy penetration. However, the uncontrollable and inflexible nature of wind power sources causes various problems regarding the operation of power systems, such as frequency problems and increased balancing costs to back up the volatile wind power production. The revenue of wind farms is also affected by (high) imbalance costs in case of direct participation in electricity markets. Battery storage units, and especially lithium-based batteries, can mitigate some of these problems by stabilising the wind power output. However, lithium-based batteries have high capital costs and lifetime problems associated with the way they are operated. In this thesis, we examine how to optimally dispatch wind battery storage units to manage the imbalance costs of the wind farm and increase its revenue while also extending the lifetime of the battery. We begin by proposing and testing with empirical data a stochastic wind speed model that captures both the stochastic short-term variations and the daily cycle that wind speed follows. Given the daily cycle and the stochastic parameters of wind speed, the expected wind power output is computed for the whole day and it follows the trend of the empirical wind power output. This is important since it provides valuable information to those making investment and operational decisions linked to wind farms combined with storage units regarding the availability of the wind farm and the time of the day at which lower and higher wind power generation occurs. Then, compared to previous literature that considered time-shifting of wind supply to higher price periods and managing the imbalance costs of the wind farm in isolation, we examine these two services combined for a battery storage unit co-located with a Spanish wind farm. It is shown that under the presence of the battery storage unit the revenue of the wind farm is increased. Then, when the capacity of the battery storage unit is restricted, the revenue of the wind farm is reduced on daily and annual basis but the lifetime of the battery is extended. However, for the investment in wind battery storage to become profitable overall, the capital costs of the battery need to decrease significantly and further revenue streams need to be stacked. Lastly, we examine a bigger wind farm located in the UK. In this case, two different business cases are examined for the wind operator along with subsidised Contracts for Difference. It is proven that with the current capital costs of the wind farms and of the battery storage units, it is more profitable for the wind operator to hold a Power Purchase Agreement with another party than owning a battery storage unit and trading wind power directly in the intraday market. Then, for both business cases examined, the capital cost of the wind farm and the strike price of the Contracts for Difference are the most critical factors affecting the profitability of the investment and also, if subsidies are totally removed from both business cases, they prove not to be economically justifiable. Lastly, we prove that considering only the calendar life of the battery (and not the cycle life) leads to significant overestimation of the investment.

As advanced metering infrastructure (AMI) and wide area monitoring systems (WAMSs) are being deployed rapidly and widely, the conventional power grid is transitioning towards the smart grid at an increasing speed. A number of smart metering devices and real-time monitoring systems are capable to generate a huge volume of data on a daily basis. However, a variety of generated data can be made full use of to advance the development of the smart grid through big data analytics, especially, deep learning. Thus, the thesis is focused on data analysis for smart grids from three different aspects. Firstly, a real-time data driven event detection method is presented, which is quite robust when dealing with corrupted and significantly noisy data of phase measurement units (PMUs). To be specific, the presented event detection method is based on a novel combination of random matrix theory (RMT) and Kalman filtering. Furthermore, a dynamic Kalman filtering technique is proposed through the adjustment of the measurement noise covariance matrix as the data conditioner of the presented method in order to condition PMU data. The experimental results show that the presented method is indeed quite robust in such practical situations that include significant levels of noisy or missing PMU data. Secondly, a short-term residential load forecasting method is proposed on the basis of deep learning and k-means clustering, which is capable to extract similarity of residential load effectively and perform prediction accurately at the individual residential level. Specifically, it makes full use of k-means clustering to extract similarity among residential load and deep learning to extract complex patterns of residential load. In addition, in order to improve the forecasting accuracy, a comprehensive feature expression strategy is utilised to describe load characteristics of each time step in detail. The experimental results suggest that the proposed method can achieve a high forecasting accuracy in terms of both root mean square error (RMSE) and mean absolute error (MAE). Thirdly, an online individual residential load forecasting method is developed based on a combination of deep learning and dynamic mirror descent (DMD), which is able to predict residential load in real time and adjust the prediction error over time in order to improve the prediction performance. More specifically, it firstly employs a long short term memory (LSTM) network to build a prediction model offline, and then applies it online with DMD correcting the prediction error. In order to increase the prediction accuracy, a comprehensive feature expression strategy is used to describe load characteristics at each time step in detail. The experimental results indicate that the developed method can obtain a high prediction accuracy in terms of both RMSE and MAE. To sum up, the proposed real-time event detection method contributes to the monitoring and operation of smart grids, while the proposed residential load forecasting methods contribute to the demand side response in smart grids.

The investigation of power generation from large-scale tidal stream turbine arrays has become of increased interest following the progressive developments in the tidal stream energy sector. This research develops a multi-scale modelling approach, spanning from device scale through array scale up to channel scale, to assess the impact of alternative turbine operating points on the power output, energy yield and the resource. Initiating the multi-scale modelling approach, Reynolds-averaged Navier-Stokes (RANS) based device-scale computational methods are suitable to model operating points in the form of the local thrust coefficient and tip speed ratio of a turbine. Rotor and blade loading shows a good agreement within 13% against Linear Momentum Actuator Disc Theory, numerical and scaled experimental rotor studies. Both methods were found to give accurate wake predictions within 2% to 7% in the intermediate and far-wake of the rotor. The Constant Actuator Disc (AD) method is chosen to be extended to an array model because of the similarly accurate single turbine wake prediction, the capacity to apply pre-defined operating points, and the direct analogy to porous disc type models which are commonly used in array experiments at small geometric scale. The extended RANS-AD array model shows that the net array power coefficient does not vary by more than 15% between 5.26 and 6.15 as it is largely insensitive to a change of operating point or flow direction. The net thrust coefficient varies significantly with flow direction and operating point, ranging from 6.77 to 12.83. In-array turbine loading ranges between +30% and -70% for constant array operating points, indicating unfavourable operation of turbines far from design conditions which must be accounted for in order to avoid a detrimental impact on device performance and life expectancy. Mitigating measures which halve the rotor load variation can be achieved by employing row-specific operating points. This is confirmed by an experimental porous disc study, showing a discrepancy of in-array and net thrust predictions within 10%. The range of expected basin efficiency values as ratio of net array power to thrust of arrays with three rows is between 0.79 and 0.45, reaching favourable values above 0.7 for low thrust coefficients. The time variation of the array performance is assessed in a channel-scale model. The definition of the channel velocity change as a function of the removed power, incorporating the pre-determined basin efficiency, captures the interaction between array and resource more realistically. The impact of row-specific operating strategies on energy yield as well as power and drag force profile depends on the local flow regime, favouring peak flows above rated speed. Simplistic operating strategies over-predict the annual energy yield by almost 5% and are unable to capture the power output profile and extent of hourly interval changes.

This thesis provides a novel dimension for DR in optimising network OHLs asset's life considers operator's preferences regarding OHL ageing criticality and network reliability requirements. A probabilistic emergency DR model is provided defining the operators' requirements of demand response power and duration participation considering network topology and components' failures. It also proposes an advanced realistic DR restoration model that accounts for operators' financial penalties for partial DR restoration. The outcomes of this study will assist the operators in structuring more credible and realistic DR contracts activated at network contingencies. It also informs the operators with a reduced emergency loading limit applied on critically aged lines having DR availability at their receiving ends. This model could assist the operators in enhancing their asset management platforms under the smart grid paradigm. When applying the method on the IEEE RTS 24-bus system, the results show improvements in expected energy not supplied and expected interruption costs reaches 30% with reduction in network ageing reaching 55% such that the ageing of long-term emergency rating of most critically aged lines is reduced with EDR by almost 24% and 38% depending on the ageing criticality. This work also explores an extra flexibility option generated from the BESS through developing a battery degradation model integrated with the standard reliability framework through a multi-year network analysis accounting for BESS degradation risks. This study informs operators with the techno-economic worth of accelerating battery degradation as well as its impact on the long-term network planning analysis in terms of altering the net present value and affecting the operators' profits. The results show that the NPV is almost 18M$ higher when accelerated BESS degradation is considered. The work also examines the worth of investments in OHL time-varying thermal ratings (TVTR) to minimise the required storage investments. Hence, the cost constraints for grid-scale battery storage are handled economically allowing more grid-scale storage deployment. The results of a specific case study on the IEEE RTS 24-bus show that investing in TVTR with costs representing 1.8 % of the BESS capital cost displaces a BESS size worth 12.2% of BESS capital costs with preserving acceptable network reliability levels. In summary, this work is one step towards having a pool of smart grid technologies mix operated in economic way with optimised reliability levels and minimised smart grid technology risks.

Crystalline silicon is by far the most prevalent material in the manufacture of solar cells, principally due to their low cost of fabrication. However, depending on the growth method used to produce silicon crystals, unwanted impurities, and defects, can remain. These defects can be electrically active, acting as traps for charge carriers, reducing carrier lifetime, and ultimately leading to reduced efficiency in finished solar cells. In this thesis, two sets of defects in crystalline silicon are studied using electrical techniques - capacitance voltage (C-V) measurements, deep level transient spectroscopy (DLTS), Laplace DLTS, and minority carrier transient spectroscopy (MCTS), with the aim of understanding how their impact on cell efficiency might be reduced. In the first part of this thesis, the electrical activity of transition metal impurities, which can be common contaminants of cheaper silicon that has not been through costly chemical purification processes, is investigated. Although certain transition metals can be removed by gettering, slow diffusing metal atoms can remain, and alternative means of reducing their electrical activity, and hence their effect on cell efficiency, are required. One potential means to achieve this is the use of hydrogen as a passivation tool, as it can bond to metallic impurities, rendering them electrically inactive. The effect of hydrogen on electrical activity has been investigated for two of the slowest diffusing metals, vanadium and molybdenum, in intentionally contaminated silicon treated with a hydrogen plasma. We find that hydrogen passivation occurs in n-type silicon, with an observed reduction in the electrical activity of vanadium atoms, but that no such bonding occurs in p-type silicon for either vanadium or molybdenum. We suggest that this difference is due to the fact that metal and hydrogen atoms are both positively charged in p-type silicon, and experience Coulombic repulsion. The possibility of forcing metal-hydrogen bonding in p-type silicon using optical excitation is then investigated and found to be successful for both vanadium and molybdenum following illuminated annealing at temperatures 800 °C, after which they are not seen to form upon subsequent annealing between 450 and 700 °C. Finally, the influence of nitrogen doping during the float zone growth process on the formation of these thermally activated defects is investigated. We show that the concentration of defects formed during annealing is significantly greater in nitrogen doped than nitrogen lean silicon, and suggest that this effect is due to the differences in the concentration of vacancies between the two materials.

The standard calculation methods of IEEE, CIGRE, and IEC working groups are widely employed by the industry to determine current-carrying capability of bare overhead (OHL) conductors. However, these methods include several simplifications within the heat transfer mechanisms between the conductor and its surroundings. At present, they are different particularly at low wind conditions and are not well formulated for the use on High-Temperature Low-Sag (HTLS) conductors at 'high' temperatures. Therefore, the source of these simplifications are discussed in the first part of the thesis in a comparative study of considering the fundamental physical concepts of the involved heat transfer mechanisms. This project is aimed at enhancing the appropriateness of the existing calculation methods and subsequently developing a unified current-temperature calculation approach for OHL conductors. For this, the impact of mixed convective cooling mechanism, conductor structural design (e.g. aerodynamic), and bundle configurations which are simplified within the standard methods are studied during the low wind and 'high' operating temperatures. The Finite Element Modelling (FEM) of COMSOL Multi-physics software package, acknowledged as an appropriate tool to capture these simplified elements is used to perform various heat transfer phenomena analyses on both conventional and HTLS conductors. Based on these results, new equations that advance the existing standard methods to calculate convective cooling mechanism PC, radiative cooling mechanism PR, and radial temperature gradient Tc-Ts of overhead line (OHL) conductors are subsequently formulated. These proposed methods could be utilised for round and trapezoidal conductors that operate at less than 1 m/s and temperatures up to 250 °C. The thesis also discusses possible thermal benefits from applying grease on conductors and provides recommendations to advance the existing current-temperature calculation practices of bundled conductors. The main outcome of this evaluation highlights the importance of using high accuracy calculation approach that captures the wire shapes, the combined effect of natural and forced convection mechanisms, and position of conductor with respect to wind in bundle configuration particularly at low wind and 'high' operating temperatures as presented by the formulated equations within the thesis. Due to these improvements, the errors of the standard methods for single conductors that usually ranging from 2% - 6% within the low-wind and medium-wind zones are reduced by at least 50%. These findings are valuable for designing OHL power networks operational limit especially when actual weather conditions are available in order to maximise their security clearance and power transfer capability.

Distributed generators (DGs) are proposed as a possible solution to supply economic and reliable electricity to customers. It is adapted to overcome the challenges that are characterized by centralized generation such as transmission and distribution losses, high cost of fossil fuels and environmental damage. This work presents the basic principles of integrating renewable DGs in low voltage distribution networks and particularly focuses on the operation of DG installations and their impacts on active and reactive power. In this thesis, a novel technique that applies the stochastic approach for the operation of distribution networks with considering active network management (ANM) schemes and demand response (DR) within a joint active and reactive distribution market environment is proposed. The projected model is maximized based on social welfare (SW) using market-based joint active and reactive optimal power flow (OPF). The intermittent behaviour of renewable sources (such as solar irradiance and wind speed) and the load demands are modelled through Scenario-Tree technique. The distributed network frame is recast using mixed-integer linear programming (MILP) that is solved by using the GAMS software and then the obtained results are being analysed and discussed. In addition, the impact of wind and solar power penetration on the active and reactive distribution locational prices (D-LMPs) within the distribution market environment is explored in terms of the maximization of SW considering the uncertainty related to solar irradiance, wind speed and load demands. Finally, a realistic case study (16-bus UK generic medium voltage distribution system) is used to demonstrate the effectiveness of the proposed method. Results show that ANM schemes and DR integration lead to an increase in the social welfare and total dispatched active and reactive power and consequently decrease in active and reactive D-LMPs.

Dye-sensitized solar cells are a strong contender for next-generation photovoltaic technology with niche applications as solar-powered windows. The performance of a DSSC is particularly susceptible to the dye sensitizer, which is adsorbed onto the surface of a wide-band-gap semiconductor such as TiO2, to form the working electrode. The nature by which such surfaces are sensitized stands to influence the resulting dye···TiO2 interfacial structure and thence the operational performance of the DSSC working electrode. To this end, atomic force microscopy is first used to image the nanoscopic formation of dye···TiO2 interfacial structures. Four well-known DSSC dyes (N3, N749, MK-2, SQ-2) adsorbed onto amorphous TiO2 substrates were chosen for this study. The results show that all of these dyes present some form of aggregation upon the sensitization on TiO2, whose spatial distributions reveal distinct nanoaggregate particle characteristics. This nanoparticle structural assembly persists even when these dye···TiO2 interfaces are fabricated using hundredfold diluted dye sensitization concentrations. Time-tracking scanning of the N749 dye reveals the forming progress of the nanoaggregates. Simulations and experiments were subsequently combined to conduct systematic studies on the optical and structural properties of two commercial dyes, RK-1 and SQ-2. The results reveal that the deprotonation of the cyanoacrylic acid group in RK-1 occurs upon dye adsorption onto TiO2, which induces a hypsochromic shift of its absorption band. DFT modeling also indicates the involvement of cyano groups in anchoring to TiO2 which makes the dye···TiO2 binding structure more stable. Fourier transform infra-red spectroscopy data prove that COO/CN (A2) and bridging bidentate (BB) binding modes co-exist in the RK-1···TiO2 interfacial structure. As for SQ-2, its simulation explains the distinct optical properties of this squaraine dye. Additionally, desorption experiments reveal a low dye loading of SQ-2 on TiO2. Based on the aforementioned study, DSSCs co-sensitized with RK-1 and SQ-2 were studied using spectral and electrical characterization, with supporting DFT simulations. Results show that the dye absorption band can be effectively extended with this co-sensitization strategy to realize panchromatic DSSCs. The co-sensitized DSSCs can gain higher short-circuit current density and power conversion efficiency with specific molar ratios of the co-sensitizing solution, although open-circuit voltage remains lower than that of the cell based on pure RK-1 dye sensitizers. The analysis also suggests that the decrease in the open-circuit voltage may result from the forming of aggregation when SQ-2 is co-sensitized with RK-1. The thesis ends with suggesting future work corresponding to the present study.

Silicon-based materials present high theoretical capacities (4200 mAh/g), making them among the most promising materials for Li-ion battery anodes. However, the major problem with them is massive volume change (∼ 300%) that occurs during cycling, resulting in pulverisation and significant capacity drop of these anode materials. Besides, the use of additives and metallic current collectors to make the anodes reduces the overall battery energy density as these components could not contribute any capacities to the battery. In this work, free-standing silicon-coated multi-wall carbon nanotube (MWCNT) anodes were developed without any polymeric binders and carbonaceous additives being added. These materials were achieved through exploiting atmospheric pressure chemical vapour deposition as well as advanced nanomaterials fabrication and processing methods. Electrodes were prepared utilising MWCNT papers (so-called buckypapers) as robust, electrically conductive flexible (22.2 S/cm) and free-standing scaffolds since the capacity of these buckypapers is too low to use as a new active material for the battery (0.3 mAh/cm² in the 300th cycle). Hence, these buckypapers were functionalised with silicon nanoparticles attached through a strong silicon carbide interface and coated with a protective amorphous carbon layer. The SiC interface was partially developed between Si NPs and MWCNTs as excessive SiC could reduce electrochemical capability on the material due to its low electroactivity. The coin cell battery test demonstrated a high areal capacity of up to 2.01 mAh/cm2. These free-standing Si-CNT-SiC-C materials showed good rate capability and stability performance (0.55 mAh/cm² after 300 cycles with coulombic efficiency of 99.7%). As a result, the material interface design, which is the key of this work, enhances the connection of Si NPs to the MWCNT walls and addresses the issue of Si detachment. A post-mortem result also revealed the structural stabilisation of the electrode after the repeated cycling test. In addition, the pouch-cell battery was fabricated to confirm the use of this free-standing Si-CNT-SiC-C electrode without the influence of the battery component. This new anode architecture performs better than the buckypaper, thereby increasing the areal capacity by 83.3%, allowing better functionality for this material to be implemented in a full-cell battery system.

There is significant interest in aprotic lithium-air batteries due to their high theoretical specific energy. During discharge, O2 is reduced at the positive electrode and forms Li2O2. Cells are typically discharged in O2, not air, because CO2 and H2O can interfere with the discharge reaction. However, lithium-air batteries will need to use atmospheric O2 if they are to supplant state-of-the-art lithium-ion cells. Therefore, it is important to establish how detrimental CO2 and H2O are to the battery. The former is well-known to induce the formation of Li2CO3 in Li-O2 batteries, but the recent work has suggested that H2O appears to be beneficial at low concentrations in some solvents (e.g. glyme ethers), increasing discharge capacities and still forming Li2O2, while in other solvents, such as acetonitrile (CH3CN), LiOH is the discharge product. Several mechanisms have been proposed to rationalise these findings, but as yet, there is no consensus on the role of H2O on O2 reduction. The purpose of this work was to understand how H2O affects O2 reduction in electrolytes using acetonitrile (CH3CN), dimethyl sulfoxide (DMSO) and tetraethylene glycol dimethyl ether (TEGDME) solvents, and, why the 4e- reduction appears to be unfavourable at low H2O concentrations. Electrochemical and spectroscopic analysis found that, in CH3CN, 4e- reduction occurred at lower H2O concentrations than in DMSO and TEGDME. A mechanism based on the ability of the H2O/solvent mixture to stabilise OH- was proposed, with mixtures that stabilise OH- promoting 4e- O2 reduction. The mechanism was confirmed by using pressure cells to identify the electrochemical reaction occurring during discharge. Cells using DMSO and TEGDME solvents underwent 2e- O2 reduction, even with 1 M H2O concentrations. Finally, TEGDME cells were discharged in a 13% RH at 25 °C O2 atmosphere, corresponding to 1 M H2O in solution, and Li2O2 was confirmed as the discharge product, demonstrating that it is possible for electrolytes to withstand a near-atmospheric humidity.

The number and size of renewable projects, such as Offshore Wind Farms (OWFs), has been rapidly growing during the last years, mainly due to the increasingly great environmental concerns. The submarine cables used to transmit the power generated offshore to the mainland are crucial for the entire project’s economic viability. Although cables are manufactured in rather cost-efficient ways and delivered in reasonable timelines thanks to the progress made in insulating material technology, they are presently subjected to a hard compromise: fixed costs are pushed to go down as much as possible, but at the same time the chance of failures is required to be minimised. The golden ratio in this difficult problem can certainly be sought to optimising the cable design. Three-core (3C), HVAC cables are presently the most cost-effective technical solution for offshore power transmission. They are also expected to be so in the future, at least regarding the interconnection of OWFs located in reasonable distance from shore. To optimise the cable design, the current carrying capacity of the cable, often called as “ampacity”, needs to be determined as accurately as possible. Due to electromagnetic induction, additional induced losses are generated inside the cable, which are dissipated in the form of heat from the cable to its surroundings. In order to investigate any likely optimisation margins, the way these losses are generated needs to be clearly understood. In parallel, the heat paths that enable the dissipation of heat inside the cable must be in depth considered. The existing calculation methods allow for such an analysis and cover, in theory, the larger cable sizes required in modern OWFs. However, empirically derived approximations are often used in these methods instead of rigorously extracted, mathematical solutions and sometimes refer to cable types different from the modern submarine cables. Furthermore, the physical models implied usually rely on simplifying assumptions that are expected to work sufficiently for smaller cables sizes, but need to be benchmarked in larger sizes. Thus, the existing calculation methods have to be reviewed and improved, where necessary. In order to allow for a quantitative analysis around the accuracy of the presently used methods, models representing more realistically the physical phenomena involved are developed. In 3C cables, the 2-D nature of heat transfer cannot be omitted, due to the physical proximity between the power cores. Traditional methods imply 1-D, radial analysis, which is in principle incapable of capturing the heat transfer occurring in the angular direction. Comparisons between the existing, traditional methods and the models developed demonstrate that this effect can be significant in larger cables. A submarine cable often encounters various conditions, which in some cases may be thermally adverse, forming the so-called “hotspots”. Cables armoured with non-magnetic steel wires are preferred in these points, due to lower induced losses. To avoid any unnecessary increase in conductor size and, thus, any economic impact such an increase would have, an optimum design is sought for. For this purpose, numerical models capable of representing the AC phenomena involved are developed. These are benchmarked against the existing analytical methods and the thermal gain obtained from the more realistic loss generation is assessed. Cables being armoured with magnetic steel wires are typically preferred in the main subsea section, due to techno-economic reasons. The cable geometry in this case influences the physical model to a great extent. Unfortunately, this is not considered by the traditional methods of calculating the losses, due to its inherent complexity. By applying 3-D electromagnetic analysis, it is possible to study the effect of the cable geometry on the induced losses. Hence, it becomes feasible to evaluate the accuracy level afforded by the traditional methods and, thus, anticipate the potential for design optimisation and further cost reduction.

The recent development of small scale electronics working within an integrated wireless sensor network has led to the massive potential monitoring of human and structure health. Such devices however are often limited by their battery life, thus there is a great need for energy harvesting to increase the lifespan of these devices. This thesis presents a novel vibration energy harvester based upon dielectric elastomers. A number of numerical models and device setups are investigated, where the device was subjected to a wide variety of harmonic excitation conditions as well as random vibrations. The numerical model was developed through experimentation to determine the nonlinear material properties of a commonly used material, V HBTM 4910. This allowed the device to be compared favourably against other energy harvesters of similar volume. The proposed device is capable of producing a maximum energy density of 9.15J/kg at an excitation frequency of 35Hz. Although observations made regarding the influence of the material nonlinearity have predicted that with a slight increase in the material nonlinearity, the device could significantly increase its energy density to 4359J/kg which would occur at the extremely useful frequency of 3Hz. The creation of this numerical model to simulate an energy harvester also allowed the direct comparison between a promising new electrical scheme and a well developed conventional scheme. Specific investigations were carried out on the device size and orientation, which highlighted an extremely effective setup which can harvest energy from a wide range of excitation conditions and orientations.

Recent advances in semiconductor technology have facilitated placing many cores on a single chip. This has led to increases in system architecture complexity with diverse application workloads, with single or multiple applications running concurrently. Determining the most energy-efficient system configuration, i.e. the number of parallel threads, their core allocations and operating frequencies, tailored for each kind of workload and application concurrency scenario is extremely challenging because of the multifaceted relationships between these configuration knobs. Modelling and classifying the workloads can greatly simplify the runtime formulation of these relationships, delivering on energy efficiency, which is the key aim of this thesis. This thesis is focused on the development of new models for classifying single- and multi-application workloads in relation to how these workloads depend on the aforementioned system configurations. Underpinning these models, we implement and practically validate low-cost runtime methodologies for energy-efficient many-core processors. This thesis makes four major contributions. Firstly, a comprehensive study is presented that profiles the power consumption and performance characteristics of a multi-threaded many-core system workload, associating power consumption and performance with multiple concurrent applications. These applications are exercised on a heterogeneous platform generating varying system workloads, viz. CPU-intensive or memory-intensive or a combination of both. Fundamental to this study is an investigation of the tradeoffs between inter-application concurrency with performance and power consumption under different system configurations. The second is a novel model-based runtime optimization approach with the aim of achieving maximized power normalized performance considering dynamic variations of workload and application scenarios. Using real experimental measurements on a heterogeneous platform with a number of PARSEC benchmark applications, we study power normalized performance (in terms of IPS/Watt) underpinned with analytical power and performance models, derived through multivariate linear regression (MLR). Using these models we show that CPU intensive applications behave differently in IPS/Watt compared to memory intensive applications in both sequential and concurrent application scenarios. Furthermore, this approach demonstrate that it is possible to continuously adapt system configuration through a per-application runtime optimization algorithm, which can improve the IPS/Watt compared to the existing approach. Runtime overheads vii are at least three cycles for each frequency to determine the control action. To reduce overheads and complexity, a novel model-free runtime optimization approach with the aim of maximizing power-normalized performance considering dynamic workload variations has been proposed. This approach is the third contribution. This approach is based on workload classification. This classification is supported by analysis of data collected from a comprehensive study investigating the tradeoffsbetweeninter-applicationconcurrencywithperformanceand power under different system configurations. Extensive experiments have been carried out on heterogeneous and homogeneous platforms with synthetic and standard benchmark applications to develop the control policies and validate our approach. These experiments show that workload classification into CPU-intensive and memory-intensive types provides the foundation for scalable energy minimization with low complexity. Thefourthcontributioncombinesworkloadclassificationwithmodel based multivariate linear regression. The first approach has been used to reduce the problem complexity, and the second approach has been used for optimization in a reduced decision space using linearregression. This approach further improves IPS/Watt significantly compared to existing approaches. This thesis presents a new runtime governor framework which interfaces runtime management algorithms with system monitors and actuators. This tool is not tied down to the specific control algorithms presented in this thesis and therefore has much wider applications.

The proliferation of stochastic renewable energy sources (RES) such as photovoltaic and wind power in the power system has made the balancing of generation and demand challenging for the grid operators. This is further compounded with the liberalization of electricity market and the introduction of real-time electricity pricing (RTP) to reflect the dynamics in generation and demand. Energy storage sources (ESS) are widely seen as one of the keys enabling technology to mitigate this problem. Since ESS is a costly and energy-limited resource, it is economical to provide multiple services using a single ESS. This thesis aims to investigate the operation of a single ESS in a grid-connected microgrid with RES under RTP to provide multiple services. First, artificial neural network is proposed for day-ahead forecasting of the RES, demand and RTP. After the day-ahead forecast is obtained, the day-ahead schedule of energy storage is formulated into a mixed-integer linear programming and implemented in AMPL and solved using CPLEX. This method considers the impact of forecasting errors in the day-ahead scheduling. Empirical evidence shows that the proposed nearoptimal day-ahead scheduling of ESS can achieve a lower operating cost and peak demand. Second, a fuzzy logic-based energy management system (FEMS) for a grid-connected microgrid with RES and ESS is proposed. The objectives of the FEMS are energy arbitrage and peak shaving for the microgrid. These objectives are achieved by controlling the charge and discharge rate of the ESS based on the state-of-charge (SoC) of ESS, the power difference between RES and demand, and RTP. Instead of using a forecasting-based approach, the proposed FEMS is designed with the historical data of the microgrid. It determines the charge and discharge rate of the ESS in a rolling horizon. A comparison with other controllers with the same objectives shows that the proposed controller can operate at a lower cost and reduce the peak demand of the microgrid. Finally, the effectiveness of the FEMS greatly depends on the membership functions. The fuzzy membership functions of the FEMS are optimized offline using a Pareto based multi-objective evolutionary algorithm, nondominated sorting genetic algorithm- II (NSGA-II). The best compromise solution is selected as the final solution and implemented in the fuzzy logic controller. A comparison was made against other control strategies with similar objectives are carried out at a simulation level. Empirical evidence shows that the proposed methodology can find more solutions on the Pareto front in a single run. The proposed FEMS is experimentally validated on a real microgrid in the energy storage test bed at Newcastle University, UK. Furthermore, reserve service is added on top of energy arbitrage and peak shaving to the energy management system (EMS). As such multi-service of a single ESS which benefit the grid operator and consumer is achieved.

With the development of power electronics, the replacement of silicon by a promising candidate becomes necessary in the field of high power applications. The GaN-based devices are attractive for the high power switching applications, owing to their superior advantages of high breakdown electrical field, high carrier mobility, and fast switching speed. However, the realization of normally-off GaN-based devices for high voltage and low resistance applications is not fully accomplished. In this thesis, the simulation, fabrication, and characterization of the AlGaN/GaN MIS-HEMTs for improving high-power properties are carried out. The TCAD simulation was first implemented to understand the effect of gate dielectric parameters and Al2O3/GaN interface states on the C–V behavior of AlGaN/GaN MIS-capacitors. After that, an economical and effective method of the 1- Octadecanethiol treatment on the GaN surface prior to the Al2O3 gate dielectric deposition proposed to improve the Al2O3/GaN interface quality. The GaN-based Metal-Insulator-Semiconductor devices treated by HCl, O2 plasma and ODT have been demonstrated. The ODT treatment is found capable of suppressing native oxide and also passivating the GaN surface effectively, hence the interface quality of the device is considerably improved. The interface traps density of Al2O3/GaN has been calculated to be around 3.0x1012 cm-2 eV-1 for devices with the ODT treatment, which is a relatively low value reported using Al2O3 for the gate dielectric in GaN-based MIS devices. Moreover, there is also an improvement in the gate control characteristics of MIS-HEMTs fabricated with the ODT treatment. In addition, a simulation of off-state breakdown voltage and electric field profiles in the MIS-HEMTs as functions of the device structures was carried out. In order to improve the high voltage performances of the devices, the AlGaN/GaN MIS-HEMTs with SiNx single-layer passivation, Al2O3/SiNx bilayer passivation, and ZrO2/SiNx bilayer passivation are investigated. High-k dielectrics are adopted as the passivation layer on MIS-HEMTs to suppress the shallow traps on the GaN surface. Besides, highk dielectrics passivated MIS-HEMTs also show improved breakdown characteristics, and that is explained by the 2-D simulation analysis. The fabricated devices with highk dielectrics/SiNx bilayer passivation exhibit improved power performance than the devices with plasma enhanced chemical vapor deposition-SiNx single layer passivation, including lower leakage currents, smaller current collapse, and higher breakdown voltage. The Al2O3/SiNx passivated MIS-HEMTs exhibit a breakdown voltage of 1092 V, and the dynamic Ron is only 1.14 times the static Ron after off-state VDS stress of 150 V. On the other hand, the ZrO2/SiNx passivated MIS-HEMTs exhibit a higher breakdown voltage of 1203 V, and the dynamic Ron is 1.25 times the static Ron after offstate VDS stress of 150 V. Furthermore, in order to realize the GaN-based devices with normally-off operations, the AlGaN/GaN MIS-FET with a fully-recessed gate structure was firstly investigated. The devices exhibited a large on-state resistance, which is not desirable for high power applications. After that, a novel normally-off AlGaN/GaN MIS-HEMTs structure with a ZrOx trap charging layer is proposed. The deposition of the ZrOx charge trapping layer on the partially recessed AlGaN in conjunction with the Al2O3 gate dielectric was developed. The fabricated MIS-HEMTs presented a threshold voltage of +1.51 V and a maximum drain current density of 779 mA/mm, which accompanied a low on-resistance of 7 O·mm. Moreover, switching after an off-state VDS,Q stress of 200 V, the degradation of dynamic on-resistance was a low value of 1.5, indicating of a satisfactory interface between ZrOx and GaN. Furthermore, the devices exhibit a high breakdown voltage of 1447 V. Though further improvement is needed on the charges storage stability, the results indicate a significant potential of employing the ALD-ZrOx charge trapping layer to realize normally-off the GaN-based devices for high power applications.

Thermoelectric materials have the unusual but highly desirable property of converting between heat and electricity. While their poor efficiencies have so far limited them to niche applications such as space exploration, they have great potential for waste heat recovery. Recent years have seen the demonstration of large performance improvements through the use of nanostructured materials. These materials have shown efficiency gains predominantly through dramatic decreases in the thermal conductivity, however it is desirable to achieve these thermal conductivity reductions without also impacting on the material’s electronic properties. For this, a high level of understanding of the electronic transport through such structures is needed. This thesis uses a variety of simulation methods—ranging from the classical to the quantum mechanical and including a Monte Carlo simulator constructed as part of the work—to explore the electronic transport through nanostructured systems. We identify key optimization guidelines to maintain and even enhance the electronic transport in the presence of nanostructures. Most significantly we outline a new concept we term “clean filtering” which shows the potential to provide substantial increases in thermoelectric performance.

In this thesis we consider the chemo-mechanical response of a multi-material lithium-ion battery anode nano-particle at equilibrium. The stresses are induced by lithium being intercalated into the nano-particle, causing the anode materials to expand. The presence of several anode materials expanding to different extents induces stress even at equilibrium. The model we primarily study is a linear elasticity model derived under the assumption that the stress-free expansion of each anode material is small. We couple the mechanical model to a diffusion-based model governing the transport of lithium in each material through stress-assisted diffusion. At equilibrium, the chemical potential of the lithium, which depends both on the lithium concentration and the hydrostatic stress, is uniform throughout the nano-particle, governing the distribution of lithium between the materials. Silicon is a promising anode material with a high capacity for lithium, but it expands to around 380% its original size when fully lithiated, leading to several issues during battery usage. We apply our equilibrium model to a nano-particle consisting of a silicon core and a surrounding graphite shell. The aim of using both materials is to mitigate the expansion of silicon without sacrificing too much capacity. We analyse the lithium concentration, displacement and stresses within each material, and use the results to optimise the volume of the silicon core based on several measures of success. Void spaces and porous silicon are introduced into the nano-particle design, using the method of multiple scales, to improve on the simple core–shell design, to moderate success. We extend the model to a geometrically nonlinear elastic model, and analyse this after finding that solutions cease to exist above a certain state of charge. Finally, we incorporate yielding and fracture of both materials into the linear model, based on perfect plasticity.

Recently, the fast technological development of organic solar cells (OSCs) proves that it has a promising future in the sustainable energy market. Interlayers play a significant role in OSCs, which can enhance device performance and stability. However, many of the fundamental mechanism and design principles of the interlayers have not been adequately studied. The theoretical chapters of this thesis deliver a summary of the state-of-the-art photovoltaic (PV) technologies and the fundamentals of organic photovoltaics (OPVs). The basic knowledge of interlayers within OSCs is introduced, too. The methodology chapter introduces the materials used in this thesis. Besides, fundamentals and technical parameters of the methodologies are summarised in this chapter, e.g. X-ray diffraction (XRD), grazing-incidence wide-angle X-ray scattering (GIWAXS), photoluminescence (PL), spectroscopic ellipsometry, external quantum efficiency (EQE), and internal quantum efficiency (IQE). In the experimental chapters, a new molecule (hexapropyltruxene) is introduced as an interlayer material in small molecule organic solar cells (SMOSCs), and its functions within OSCs are investigated. Firstly, hexapropyltruxene is used as an interlayer in the SubPc/C60 solar cell stack. The experimental results show that hexapropyltruxene works well as an exciton blocking layer (ExBL) between the MoOx and SubPc layers. By adding 3.8 nm of hexapropyltruxene between MoOₓ and SubPc layers, the power conversion efficiency (PCE) of the SubPc/C60 solar cells is enhanced from 2.6 % to 3.0 %, achieving a ca. 15 % improvement. Besides, it is found that hexapropyltruxene has a weak molecular template growth effect on the growth of SubPc thin film. Secondly, the application of hexapropyltruxene as an interlayer in the solar cell stack first demonstrated by interuniversity microelectronics centre (IMEC) is also investigated. The results reveal that hexapropyltruxene has a very weak exciton blocking effect between MoOₓ layer and alpha-6T donor layer. The structural characterization also suggests that hexapropyltruxene can induce the growth of different alpha-6T polymorphs. In the final experimental chapter, the polymorphism of alpha-6T induced by different substrates with hole contact layers (HCLs) is studied. The competing growth of low temperature (LT)-, high temperature (HT)-, and beta-phases of alpha-6T is observed in the thin film growth. The results of the experiments suggest that widely-used HCLs such as MoOx, PEDOT:PSS, and MeO-TPD have the template growth effect on alpha-6T molecules. In addition, the substrate temperature in the deposition process is significant in the competing growth of alpha-6T polymorphs. Corresponding solar cell performances show a correlation with the polymorph composition within alpha-6T layers and the performances of OSCs. To conclude, less "standing-up" but more "lying-down" molecules lead to better light absorption, which results in better solar cell performance. PEDOT:PSS is the most appropriate HCL for our alpha-6T/SubPc solar cell stack, and low substrate temperature (20 C) is more suitable than high temperature (120 C) in the growth of alpha-6T thin films for solar cell application.

The evolution of microgrids represents a significant step towards the transition to more sustainable power systems. Recent trends in microgrids include the integration of renewable energy resources (RERs), alternative energy resources (AERs) and energy storage systems (ESSs). However, the integration of these systems creates new challenges on microgrid operation because of their stochastic and intermittent nature. To mitigate these challenges, determining the appropriate size together with the best energy management strategy (EMS) systems are essential to ensure economic and optimal performance. This thesis presents an investigation into sizing and energy management of microgrids. In the first part of the thesis, an analytical and economic sizing (AES) approach is developed to find the optimal size of a grid-connected photovoltaicbattery energy storage system (PV-BESS). The proposed approach determines the optimal size based on the minimum levelised cost of energy (LCOE). Fundamental to this approach obtains an improved formula of LCOE which includes new parameters for reflecting the impact of surplus PV energy and the energy purchased from the grid. In the second part of this thesis, an integrated framework is proposed for finding the best size-EMS combination of a stand-alone hybrid energy system (HES). The HES consists of PV, BESS, diesel generator, fuel cell, electrolyser, and hydrogen tank. The proposed framework includes three consecutive steps; first, performing the AES to obtain the initial size of the HES, second, implementing the initial EMS using finite automata (FA) and instantiating multiple EMSs; and third, developing an evaluation model to assess the instantiated EMSs and extract the featured conditions to produce an improved EMS. Then the AES approach is re-exercised using the improved EMS to obtain the best size-EMS combination. The core of this framework is utilising FA to implement various EMSs and capturing the impact of selecting the best EMS on the sizing of the HES. Furthermore, a sensitivity analysis is performed to address the uncertainty in demand and solar radiation data showing their effect on the HES performance. The analysis is carried out by assuming variations in solar radiation and demand annual data. Several scenarios are generated from the sensitivity analysis, and a number of performance indices are computed for each scenario. Following that, a vii fuzzy logic controller is designed using the performance indices as fuzzy input sets. The objective of this controller is to modify the EMS obtained from the integrated framework. This can be accomplished by detecting any changes in the demand and solar radiation and accordingly modify the operating conditions of the diesel generator, fuel cell, and electrolyser. The performance of the proposed approaches is validated using real datasets for both demand and solar radiation. The results show the optimal size and EMS for both grid-connected and stand-alone microgrids. Moreover, the designed fuzzy logic controller enables the microgrid to mitigate the uncertainty in the demand and generation data. The proposed approaches can be used with various scales of microgrids to extract manifold benefits where reliability, environmental and cost requirements can not be tolerated.

Photovoltaic PV systems have shown a significant growth in recent years driven by the increased efficiency and reductions in the cost of PV modules. Today, distribution generation based PV systems have a major contribution to the total electricity production worldwide. However, in areas with high penetration of PV system connected to the grid, interaction may arise between the grid and PV system. This research focuses on the effect of grid operating conditions on the performance of grid-connected PV inverter systems. A simulation model of the system under investigation has been developed to evaluate the impact of frequency deviation, grid voltage distortion and grid impedance variation on the harmonic performance of the injected current as well as the voltage at the point of the common coupler. Proportional resonance PR controller is employed to regulate the current produced from the PV inverter due to its reputation in tracking sinusoidal signals. The obtained simulation results demonstrate that the harmonic performance of the grid current and PCC voltage can be significantly influenced by the change of grid operating conditions. In particular, grid impedance variation can result in a shift in the system resonance frequency leading to distortion in network current and voltage. To adapt the PR controller to the grid impedance variation, a novel adaptive PR controller which takes into account the change in grid impedance is proposed. The adaptive consists of a high-order digital band-pass filter and chain of statistical signal processing technique. Simulation results show that the harmonic performance of grid current and PCC voltage can be enhanced and the proposed APR controller is robust against impedance variation. Finally, the proposed control method is experimentally implemented and the obtained results validate the effectiveness of the proposed control structure.

In many real-world engineering systems, the rate of variation in the system state depends on the past states, which characteristic is called delay or a time delay. As time delay will degrade the system dynamic performance and even destroy the system stability, the stability analysis of a system with a time delay has been investigated in many real control systems such as load frequency control scheme of power systems, genetic regulatory network, digital filter, electricity market and economic dispatch of the power system. Several improved stability criteria of time delay systems are proposed and applied in power system control with time delay such as load frequency control and electricity market. In this thesis, the Lyapunov-Krasovskii functional and the Wirtinger inequality have been investigated to establish several new stability criteria at first. The Wirtinger-based inequality used in a linear system with two additive time-varying delays is investigated and applied in load frequency control system in chapter 2. In chapter 3, an improved stability criterion based on Wirtinger-type double integral inequality was developed and applied in genetic regulatory networks. Then, in chapter4, the proposed method has been extended to a discrete system and applied in digital filters. Chapter 5 presents a further improve a stability criterion using Wirtinger-type three integral inequality and applied in the power system, including load frequency control system and economical dispatch system. In chapter 6, the stability analysis of an electricity market with a time-varying delay is studied. The dynamic model of an electricity market is presented to minimize the operation cost. The impact of a time delay on this dynamic model is analysed via a stability criterion developed in Chapter 5. The main contribution of the thesis is not only proposes several stability criteria based on the Lyapunov-Krasovskii functional and Wirtinger inequality but also applying the methods in power systems, genetic regulatory network and digital filter. Effectiveness of those new stability criteria has been verified via simulation studies based on numerical examples and those industry time-delay systems.

Wind energy is an interesting prospect due to its availability and sustainability along the coast and offshore locations. Development of offshore wind farms is projected to increase very rapidly within the next decades, which is expected to efficiently improve the electricity challenges in areas located along the sea. While the offshore wind energy is collected by turbine structures, which are supported by towers and foundations, the challenge remains on the provision of cost-effective foundations to resist the lateral loads from wind, waves and dynamic actions of the wind turbine nacelle-rotor. The popular foundations used are monopiles, having a competitive advantage of stability, ease of installation and low cost of materials compared to other types. The loads acting on monopiles are cyclic in nature and can be resisted by earth pressure mobilised in the soil surrounding the pile. The cyclic loads can affect the strength and stiffness of both the soil and pile, leading to accumulated rotation and change of overall stiffness. Studies have been carried out regarding these effects, however, there has not been a clear understanding of the response of a stiff pile when subjected to many cycles. The literature has revealed that the current method (p-y curves method) for analysing and designing the offshore monopiles is insufficient, tend to overestimate the stiffness of the rigid piles, and leading to interference between resonance and natural frequency of the wind turbines. The method usually regards the soil as a series of non-linear wrinkle spring and derives its base on the empirical relationships developed from full-scale tests on slender piles. The deficiency of the prevailing design approach, therefore, justifies a need for further research to develop a model which will monitor how the monopiles foundations respond to both monotonic and cyclic loading. This thesis, therefore, presents an experimental and theoretical research approach that will improve the understanding of the response of monopile foundations in sands when subjected to both monotonic and cyclic loading. The experimental work involves a comprehensive design and development of a new mechanical loading system in a geotechnical centrifuge, with model tests scaled to represent full-scale wind-turbine monopiles. The test programme is designed to identify the key mechanisms driving pile response, including investigating the monotonic loading behaviour as well as the response of a pile to long term cyclic loads. The methodology developed and data collected from this thesis will provide a potential contribution to the establishment of a better understanding of monopile responses within the field. The experiment was carried out by initially, scaling down the prototype monopile using a 1:100 geometric scale. Three monotonic tests were carried out at 100g to identify the responses under monotonic loading, estimate the ultimate capacity and determine the initial (tangent) stiffness of the pile-soil system. One monotonic test was conducted at 30g as a reference to the cyclic test results. The parameters extracted from the monotonic tests are used as the basis for the design of cyclic loading system and analysis of the cyclic test results. Available models from the literature were modified to capture the ground global response of the pile under monotonic loading. The models were employed in a kinematic approach, with soil being modelled as a series of spring elements and the pile as an elastic beam element. The model pile in the centrifuge was not instrumented, hence assumptions were made to match the global response of the monotonic centrifuge tests at the ground surface. With the absence of data along with the embedded depth of the pile, the fitting constants were used to estimate the ultimate capacity and the concept of the modulus of subgrade reaction. The model was also used to the published centrifuge test results of the past monopile research. The methodology developed for cyclic load tests incorporated the effects of cyclic loading on the response of a monopile. The validity of the model is supported by centrifuge tests in which a stiff model pile, installed in cohesionless soil, was subjected to a series of load cycles with load amplitude and frequency. The tests were carried out to investigate the responses of the newly developed model. Due to technical challenges during the model testing, all cyclic tests were achieved at centrifuge acceleration of 30g instead of the 100g for the monotonic tests. Selected tests were used to examine the total cyclic pile-head response to gain an insight into the accumulation of pile-head displacement and change in cyclic secant stiffness. Power and logarithmic functions were used to predict the accumulated displacement and cyclic stiffness variation using the data from the centrifuge, respectively. The key experimental findings of the cyclic tests were then used to develop a theoretical model that captures the unload-reload hysteresis behaviour. The model function, called the modified Romberg Osgood (MR-O), is rigorous yet simple and is framed where the pile-head cyclic response is modelled as the hysteresis loops for the backbone, unloading and reloading curves. The model was calibrated against the cyclic centrifuge tests and successfully reproduce the main elements of the pile response with high accuracy. However, it does not predict precisely the cyclic accumulation and change in cyclic stiffness as shown in the experiment, thus further improvement will still be required. Nevertheless, the newly developed model and suggested methodologies in this thesis can be used as a primary stage for research developers to understand the behaviour of foundations supporting offshore wind turbines, with scientific justification based on the centrifuge model scale tests.

In this work, organic materials have been studied for their application in photoelectrochemical water oxidation and the electrochemical reduction of CO2 in water, using porous and linear conjugated polymers and hydrogels. The use of such organic materials for solar fuel generation is highly attractive as they are cheap, relatively easy to synthesise and their structure-properties can be fine-tuned for selective gas uptake, changes in optical and physical properties such as band gap, pore size, surface area and solubility. The use of a conjugated microporous polymer containing a bipyridine (CMP-(bpy)20) unit allowed for the incorporation of a well-known CO2 molecular electrocatalyst, [Mn(bpy)(CO)3Br], in order to prevent the undesirable process of dimerisation of Mn centres prior to the formation of the catalytically active reduced species. Engineering of the electrode was explored via drop casting a suspension of CMP-(bpy)20-Mn in a polymer/acetonitrile solution, such as Nafion or PANI to aid adhesion to the electrode surface and exploit secondary functionalities such as proton and electrical conduction respectively. Although dimerisation of Mn catalyst was prevented, CO Faradaic Efficiencies and the electroactive content were low. This is likely due to only a small fraction of catalytic centres becoming electroactive due to structural distortions across the polymer backbone resulting in insulated regions, and only Mn that is in direct contact with the electrode surface is electroactive. We also explored growing CMP-(bpy)20 onto the surface of a GCE to produce thin film electrodes in an attempt to increase the planarity of CMP-(bpy)20 and diffusion through the porous structure. Two soluble linear polymers, P8s and P56, known to be good photocatalysts for hydrogen evolution were spectroscopically studied via transient absorption spectroscopy to gain insight into their photocatalytic mechanism and the role of solvents present. Due to the ease of processability and superior hydrogen evolution rate of P8s, we then went on to explore P8s as a photoanode for water oxidation purposes. XPS measurements showed P8s to be a suitable material for overall water splitting as it straddles the band gap for both hydrogen and oxygen evolution. Improved photocurrents were achieved by the fabrication of a heterojunction using TiO2 by a factor of 30 reaching photocurrents up to 20.97 µA.cm-2 . Photoactivity was further enhanced by increasing the amount of absorbed photons by using high surface area mesoporous TiO2. Transient absorption spectroelectrochemistry gave insight into the excited state dynamics of the photoanode and their lifetimes, which showed a longlived hole in the material. Utilisation of the hole was demonstrated by the addition of a hole scavenger into the electrolyte, followed by the incorporation of a co-catalyst for water oxidation. An amino acid functionalised perylene bisimide (PBI-A) hydrogel was studied in various hydrated states using UV-Visible absorption spectroscopy and neutron scattering to explore how drying affects changes in structure. Although the overall networks were similar, drying of a PBI-A hydrogel to form a xerogel resulted in an irreversible change in packing on the molecular scale between PBI-A units within a fibre. Spectroscopic studies showed a longlived charge separated state under visible light illumination unlike previous reports of dominant UV activity for this class of materials. Therefore we tested the materials as photoanodes for water oxidation and explored how the difference in water content within a gel affects photoactivity and their excited state dynamics. Utilisation of the photogenerated hole was determined from oxygen evolution experiments in the addition of an IrOx water oxidation co-catalyst.

Lithium-ion (Li-ion) batteries are the present state-of-the-art energy storage technology used in mobile phones and electric vehicles. This thesis focuses on different lithium intercalation compounds that can be used as Li-ion cathode materials in order to improve the electrochemical performance of lithium-ion batteries, with particular focus on the study of Li-rich mixed metal oxides, where oxygen is a fundamental part of the redox activity of these compounds. A general background about Li-ion batteries is presented in Chapter 1 and a detailed explanation of the experimental methods, specially focused on X-ray photoelectron spectroscopy (XPS), is explained within Chapter 2. Chapter 3 of this thesis investigates how the effect of synthesis route of lithium manganese mixed metal oxides, with the formula LiNi1-x-yCoxMnyO2, can affect the electrochemical performance of the material. Four LiNi1/3Co1/3Mn1/3O2 samples were synthesised using two synthetic methods: resorcinol-formaldehyde sol-gel polymerisation and hydroxide co-precipitation synthesis; these four samples were compared with a commercial sample. Using XPS, Raman and scanning electron microscopy, a direct relationship between the particle size and the electrochemical performance, as well as the link between the used synthetic method and the surface reactivity was shown. Chapter 4 studies the electrochemical behaviour of two Li-rich transition metal oxides: Li1.2Ni0.13Mn0.54Co0.13O2 and Li1.2Ni0.32Mn0.4Co0.08O2. The anionic redox activity and the formation of Onspecies is shown via XPS and Raman spectroscopy at different voltages, and show the importance of the oxygen redox activity in order to obtain a high electrochemical performance. In Chapter 5, Li-rich intercalation compounds with rock-salt type structures with the general formula of Li4+xNi1-xWO6 is reported, where a third-row transition metal is used in order to stabilise the structure. In this chapter, the studied material is the nonstoichiometric Li4.15Ni0.85WO6 and a detailed study in XPS and Raman at different voltages shows the formation of stable peroxo-type species that can explain the high capacity obtained, as well as the observed large voltage hysteresis (> 2.0 V) that is characteristic of this material.

A sustainability assessment framework combines life cycle and triple bottom line approach is proposed in this study; and a life cycle sustainability assessment model that is generic and suitable to examine and compare sustainability performance of decentralised electricity technologies on a regional scale is designed under the proposed framework. The assessment model is designed based on the context of Northeast region of England; the framework is generic, and the model can be tailored to be suitable to assess different technologies in different regions. In the proposed model, sustainability performance is evaluated using three sets of nineteen indicators in total, with five examining the techno-economic impact, twelve measuring the environmental impact and two assess the social impact of selected energy technologies. Three decentralized energy technologies were assessed in this thesis, they are solar photovoltaic (PV), onshore wind and biomass. Three types of most commonly deployed solar photovoltaic electricity generation systems are considered to represent the current technology, they are: monocrystalline (s-Si), polycrystalline (p-Si) and Cadmium telluride (CdTe) thin film. Three wind turbines with highest installation capacity are considered to be representative for present day onshore wind technology, they are: Vesta V80, Vesta V90, and Repower MM82; For biomass technology, the largest biomass combined heat and power plant both within the region and the UK –Wilton 10 is considered to be representative of state of art for the technology. Results obtained from the assessment is then ranked and compare against each other to conclude the sustainability performance of each assessed technology. ReCiPe method is also applied as part of sensitivity analysis; and finally data quality assessment is carried out using criteria produced by Stamford and Azapagic (2012, p. 415). The study reveals that no technology is superior to another; the sustainability performance needs to be expressed in relation to the resource availability and regional development strategy. The common belief that renewable energy is totally emission free is because the significant environmental impacts associated with upstream manufacturing and end-of-life process are not accounted for. For example solar PV is almost emission free during electricity generation but production of the system components do pose significant environmental impact; its merit resides being an effective tool to alleviate fuel poverty due to its ability to reduce energy bills for the system host and its low capital cost. Since sustainability is a dynamic process, the choice for the most sustainability electricity options will also progress over the time; depending on the need of society and resource availability. Planning for a sustainable energy future requires holistic review of suitable energy options and strategic energy planning.

The modern power system has gone through a lot of changes over the past few years. It is no longer about providing one-way power from sources to various loads. Power monitoring and management have become an increasingly essential task with the growing trend to provide users more information about the status of the loads within their energy consumption so that they can make an informed decision to reduce usage and cost or request desired maintenance. Computational intelligence has been successfully implemented in the electrical power systems to aid the user, but these research studies about this are generally conducted on the conventional alternative current (AC) macro-grids. Until now, little work has been done on direct current (DC) and the focus on smaller DC grids has been even less. In recent years, the evolution of electrical power system has seen the proliferation of direct current (DC) appliances and equipment such as buildings, households and office loads. This number keeps increasing with the advancement in technology and consumer lifestyles changes. Given that DC power supplies are getting more popular in the form of photovoltaic panels and batteries, it is possible for Extra Low Voltage (ELV) DC households or office pico-grids to come into use soon. This research recognises and addresses this research gap in the monitoring and managing of the DC picogrids. It recommends and applies the bottom-up monitoring and management approach in smaller scale grids and in larger scale grids. It innovatively categorises the loads in the grids into dumb loads that do not have intelligence and communication features and smart loads that have these features. While targeting at these ELV DC pico-grids, this research presents solutions that provide users useful information on load classification, load disaggregation, anomaly warning and early fault detection. It provides local and remote sensing with the alternative use of hardware to lessen the computational burden from the main computer. The inclusion of remote monitoring has opened a window of opportunities for Internet of Things (IoT) implementation. These solutions involve the blending of computational intelligence techniques with enhanced algorithms, such as K-Means algorithm, k-Nearest Neighbours (kNN) classification, Naïve Bayes Classification (NBC) Theorem, Statistical Process Control (SPC) and Long Short-Term Memory Recurrent Neural Network (LSTM RNN). As demonstrated in this research, these solutions produce high accuracy results in load classification and early anomaly detection in both AC and DC pico-grids. In addition to the load side, this research features a short-term PV energy forecasting technique that is easily comprehensible to users. This research contributes to the implementation of the Smart Grid with possible IoT features in DC pico-grids.

In this thesis, a possible solution for improving the utilization of wind power energy without causing impact to the power grid by using microgrid concept, which promises the function of making a cluster of generators and loads behave like a controllable unit, is proposed. This microgrid concept is achieved by a proposed hierarchical control method which contains three levels: decentralized level 1(primary or bottom level)is of fast response and can satisfy the essential power system operation requirements; centralized level 2 (secondary or middle level) eliminates control target deviation and rely on low band width communication channels to achieve accurate control objectives within several seconds; level 3(tertiary or top level) is the system organization level in which on-site functions such as seamless states transition operation and off-site functions such as setting points optimization are executed. A 4-bus microgrid model built in MATLAB/Simulink is used for individual validation tests of the different control level. The following experiments are executed in the wind power generators integrated CERTS 20-bus microgrid system. The simulation results show the system variations caused by wind power can be digested within the microgrid system with cooperation of the local loads, backup generators and energy storage devices which are operated by the hierarchical control system. In which case, the impact of wind power fluctuation can be absorbed by using microgrid concept with the proposed hierarchical control method.

The connected and automated vehicle (CAV) is a promising technology that will reshape the transport. It has potentials in reducing fuel consumption as well as improving capacity and safety. The traffic control of CAVs on urban roads is investigated in the thesis, which consists of two aspects: intersection control and trajectory planning. For the intersection control with CAVs, a bilevel programming model integrating intersection control with trajectory planning is proposed to improve the efficiency of the intersection when all vehicles are CAVs. The feedback structure adapts a wide range of conditions and improves capacity. The cooperation between the two levels and their linear properties ensure reasonable solving time. A platoon-based method is also proposed to improve the calculation speed. For the trajectory planning with CAVs, a platoon-based eco-driving model is proposed using model predictive control. The model is used in mixed autonomy traffic and considers traffic efficiency, fuel consumption, and driving comfort. The cooperation between automated vehicles and human-driven vehicles reduces the negative impact of eco-driving on the following vehicles and reduces fuel consumption even further. The performance under different platoon sizes and penetrations of automated vehicles are also tested in the simulation. At an intersection controlled by the adaptive traffic control system, the vehicle may not be able to get accurate information on the future signal timing. A multi-phase model predictive control model is proposed to reduce fuel consumption by considering the stochastic signal information. Two driving regimes are considered based on the state of the traffic signal at each time step: accelerating to pass the intersection when the light is green or keeping waiting for the possible green light in the next time step. This model is further extended to the case that the vehicle can receive the information on the future signal timing some seconds in advance. An additional eco-driving strategy is considered in this case.

In this work a step away from spin coating deposition was taken in order to make a more realistic approach with a view towards future of production of perovskite large scale. Instead of traditional spin casting, a bar coating deposition was used. This method provides a simple way to spread large areas with no wasted of precursor solution. At the same time it offers the advantage of producing different types of perovskites processed at low and high temperature, allowing a broad variety of options of tunability of the film properties. Over the course of this work, an attempt has been made to optimize deposition by a roll–to–roll technique, bar coating, and to approach the phenomena of crystallization and processes involving in degradation and stabilization of mixed halide perovskite.

Silicon has been highlighted as a promising anode material in lithium-ion batteries due to its step change in capacity verses conventional graphite. The cycling of silicon anodes within a lithium-ion battery (LIB) leads to degradation and capacity fade due to the 280% volume change of silicon. Many avenues of silicon synthesis have been explored to produce nanostructures which can withstand this change in volume. Magnesiothermic Reduction (MgTR) of silica to silicon shows significant promise over other syntheses in scalability, economic and environmental aspects for producing porous silicon nanostructures. The problem with MgTR is a lack of understanding regarding the pore evolution of porous silicon based on reduction parameters and precursor material, which in turn limits predictive design for desired applications. Here we show for the first time that the pore structure of porous silicon is strongly related to the interconnectivity of silicon crystallites. We show that the MgTR is a thermodynamically driven equilibria which determines the purity of the silicon product. Higher temperatures also cause sintering of silicon nanocrystallites. We show that it is the interconnectivity of these crystallites determine the pore size and distribution within porous silicon. These findings apply to a wide variety of porous silica precursors and we show this mechanism is true for the introduction of pores into nonporous quartz after MgTR. Further, we show that by exploiting this mechanism, mesoporous silicon can be produced which has excellent promise for LIB applications with a capacity of 2170 mAh/g after 100 cycles. As a second section of the thesis, we focused on the use of silica directly in LIBs. The use of silica has potential advantages over other silicon based active materials, upon reduction with lithium silicon is produced and contained within a supporting structure of inactive material. However, there is no detailed understanding of how the lithium-silica reduction reaction progresses and the chemical nature of the products. Here we develop a new method to effectively monitor the rate of electrochemical reduction of silica and propose a mechanistic understanding of this process. In addition, and for the first time, we characterise the existence of elemental silicon in the reduced structures. Our proposed mechanism is based upon the initial insulating nature of the silica active material and how electronic conduction pathways are formed in the reduced material. We apply the principles of this mechanism to reduce the length of the electrochemical reduction reaction to 13 hour compared with 400 hour reduction times reported in the literature. The findings herein provide a significant step change and hence can be taken forward to design optimal materials for LIB applications. These results strongly support the potential for reduction in silicon costs for LIB in both economic and environmental terms as well as for a reverse engineering approach to design specific porous silicon and silica for desired applications.

Introduction: South Asians (SAs) have a low prevalence of atrial fibrillation (AF) compared with Caucasians despite a higher prevalence of hypertension, diabetes mellitus and coronary artery disease. The aim of this thesis was to determine whether this was related to an under-detection of the arrhythmia and if not, whether differences in left atrial (LA) size, electrophysiological properties or autonomic function in SAs might help to explain this disparity. Methods: Retrospective and prospective cohort studies were performed on SA and Caucasian participants using data from implantable cardiac devices, cardiac magnetic resonance imaging scans, invasive electrophysiology studies and a range of non-invasive cardiac investigations. Results: The cumulative incidence of subclinical AF was significantly lower in SAs compared with Caucasians (log rank p=0.002) with an annual event rate of 6.9% versus 13.9%. In comparison with Caucasians, SAs were of a smaller height with lower lean body mass and higher waist:hip ratio; had lower minimum (27.7±11.1 ml vs 34.9±12.3 ml, p=0.002) and maximum LA volumes (64.7±21.1 ml vs 80.9±22.5 ml, p<0.001) even after matching for body surface area; lower P wave dispersion (males 28.0(12)ms vs 25.0(12)ms, p=0.039; females 24.0(12)ms vs 22.0(12)ms, p=0.004) and P wave terminal force in lead V1 (males 0.031(0.04)mm•s vs 0.021(0.03) mm•s, p=0.023; females 0.036(0.04)mm•s vs 0.034(0.04)mm•s, p=0.030), electrophysiological variations related to the inhomogeneity of LA conduction and LA size respectively; increased heart rate (82.5(18)bpm vs 78.0(18)bpm, p=0.024), lower atrioventricular (280(50)ms vs 300(60)ms, p=0.001) and ventriculoatrial (300(60)ms vs 320(93)ms, p=0.013) effective refractory periods and lower heart rate variability (in SA males), suggestive of sympathetic predominance. Conclusions: SAs have reduced LA size and evidence of increased sympathetic tone and reduced inhomogeneity in LA conduction. The morphological, electrophysiological and autonomic differences identified in SAs may help to explain why this ethnic group has a lower prevalence of AF.

The connectivity of the power grid is increasing with the internet of things, and low carbon technologies being deployed to help enhance smart grid performance and reliability. Meanwhile, they also increase the digital complexity and dependency of cyber assets, which might be vulnerable to cyber-physical threats, and hence may impact the reliability of power systems. Due to cyber-threats’ unpredictable nature, the interactive operation of low carbon technologies with cyber-physical systems is becoming a challenging task for smart grids. This thesis proposes novel mathematical frameworks to estimate the availability of photovoltaics and heat pumps with cyber-physical components. These frameworks are developed to quantify the level of risk posed by cyber-threats to the interactive operation of photovoltaics and heat pumps, using Markov-Chains. The availability framework considers the severity of random cyber-attacks on photovoltaics and the probability of cyber-threats with mean time to detection-time on heat pump operation. Sensitivities of the repair times of cyber-physical component for photovoltaics and sensitivities of cyber-attack-detection time for heat pumps are also evaluated. The impact of cyber threats on the interactive operation of photovoltaics and heat pumps are considerable and inconsistent, however the propagation of cyber-threats can be restricted by appropriate means of photovoltaics. For heat pumps, operational reliability substantially decreases due to the unavailability of their control panel. Contributions of this thesis include an availability model for photovoltaic configurations, an innovative approach to assess the reliability of a photovoltaic integrated power system with cyber-physical interactions, the availability estimation of heat pump with variable detection time, and an enhanced cyber-intrusion process model for reliability analysis of heat pumps. The findings offer insight into the impact of cyber-physical system availability and its importance on power system reliability.

Converter transformers are a key component in High-Voltage Direct Current (HVDC) systems. Differential protection is used as the main protection scheme for converter transformers to mitigate damage caused by short-circuits within transformers. However, in addition to conventional challenges, e.g. magnetizing inrush and current transformer (CT) saturation, threats posed by DC bias, the implementation of IEC 61850 and cyber security have become new challenges, where the security and dependability of differential protection of converter transformers could be jeopardized. Therefore, this thesis concentrates on the investigation into the effect of DC bias on converter transformer differential protection as well as determining if existing converter transformer protection schemes are adequate or could be improved. This thesis also conducts performance testing and cyber security assessment of an IEC 61850 compatible transformer protection relay. This thesis starts with an investigation into the effect of DC bias on converter transformer differential protection. DC bias phenomenon and origins are first discussed to provide a full insight into the categories and characteristics of DC bias. A transformer model based on unified magnetic equivalent circuit (UMEC) is developed to simulate converter transformers. Its capability of representing hysteresis is confirmed by comparisons between hand calculations and simulation. Appropriate models, including converter transformers, CTs and DC bias signatures, are properly built to aid in the investigation. The results demonstrate the impact of DC bias on converter transformer differential protection is negligible when protection CTs are properly specified. Harmonic blocking is used for converter transformer differential protection to identify inrush currents, but it may incorrectly block the protection during internal faults. By means of the CIGRE HVDC benchmark test system, the operating behaviour of differential protection with harmonic blocking when applied to converter transformer internal faults is evaluated. The main factor that accounts for differential protection failing to trip is converter transformer half-cycle saturation, resulting from the heavy DC component in the fault current. To properly identify inrush currents, a method based on mathematical morphology (MM) is proposed. The effectiveness of the proposed method is validated by extensive simulation, and the results indicate this MM-based method has an advanced capability of discriminating between inrush and internal faults. The proposed method can also properly cope with extreme cases, such as sympathetic inrush, inrush with CT saturation, high through-fault with CT saturation and inrush accompanied by a turn-to-turn fault. Differential protection of converter transformers is achieved by collaborations among differential relays, CTs and circuit breakers. Latest transformer protection implements IEC 61850 protocol to convey protection measurements and trip signals, but this implementation should not degrade the overall performance of differential protection. To assess the response of an IEC 61850 compatible transformer protection relay, a fully digitalised hardware testbed is built in the laboratory. The assessments are comprised of underlying protection performance tests and relay responses under different network contingencies. Test results indicate the differential protection function, achieved by IEC 61850 based data packets, satisfies the protection requirements of transformer protection. Meanwhile, the relay accepts high-volume network traffic before its performance deteriorates. The implementation of IEC 61850 also encounters cyber security threats. In the end, the laboratory testbed is updated, and cyber security assessment is carried out. The response of an IEC 61850 compatible transformer protection relay under various cyber attack scenarios is assessed and observed. The tests demonstrate once the relay is exposed in a public network and discovered by a malicious actor, it is vulnerable to cyber attacks. It is easy for a malicious actor to disrupt protection functions/services and to cause catastrophic consequences.

The suitability of the numerical weather prediction model, weather research and forecasting (WRF), is assessed for evaluation of large-scale offshore wind farm sites with a particular focus on conditions in North West India. Prediction accuracy of the wind resource is established for typical operating conditions, intervals of peak wind speed, and over short time-scales during the occurrence of sea breeze. The analysed conditions are employed to evaluate alternative large-scale offshore wind deployment scenarios. Energy yield and economic viability of wind farm sites of up to 2 GW installed capacity are considered with fixed or floating infrastructure. Wind turbine energy yield is dependent on the accurate prediction of wind speed distribution occurrence and this is obtained with a root mean square error (RMSE) within 1.3% for representative time periods within a year. For typical operating intervals, the wind speed occurrence distribution is also shown to be accurate with RMSE of 0.3 to 1% for multiple sites located on flat terrain, at up to 300 km spacing. The mean wind speeds are predicted within 1.3 to 5.1% of error with improvement in prediction accuracy for wind speeds greater than wind turbine rated speed of 11ms-1. The wind speed is predicted within 5.1%, including during days on which sea-breeze events cause rapid change in wind speed and direction over short time intervals. Wind turbine design selection also requires information on the wind profile and turbulence intensity. The wind profile RMSE is within 0.7 to 7.8% over different wind conditions; turbulence intensity is generally underpredicted for a nearshore mast, however across all wind speeds and sites this is within -2 to +1.6%, relative to average measured values in the range 11.4 to 13.4%. The high prediction accuracy of wind speed distribution over multiple locations and of the prediction of wind speed time series, including during short-duration sea-breeze events, provides confidence in the use of WRF for evaluation of offshore wind farm power supply, potentially reducing reliance on extensive field measurements. Deployments of up to 2 GW of wind turbine capacity located along the Gujarat coastline could provide energy supply of between 7.6 to 8.7TWh for sites located at distances between 7 km and 70 km from shore. The site further from shore is in water depths of up to 80 m, greater than the depths suited to typical bed-fixed structures such as monopiles. A techno-economic study of alternative sites and infrastructure indicates that floating systems are expected to have a net project benefit of 409£m. However, sea breeze is found to have a stronger influence on the energy yield of offshore wind farms at the nearshore sites, with a gain in annual energy yield by 5.8% compared to 2% at far offshore locations. Whilst this improves the viability of nearshore locations, the floating offshore wind farm sites remain more profitable and on this basis are recommended over nearshore locations with bottom fixed platforms.