Many communities around the world have no access to an electricity grid. To supply power to these people, stand-alone power systems are often used, the majority of which are based on diesel generators. Rising fuel costs and environmental concerns make the use of renewable energy in stand-alone systems increasingly attractive. The research reported in this thesis was to demonstrate a stand-alone power system based exclusively on renewable energy sources. To achieve this, a DC electrical backbone is used. Power electronic converters are used to interconnect the loads and generators and hydrogen is used as an inter-seasonal energy store. The design and control of the DC based stand-alone power system forms the primary focus of this research. A demonstration system has been implemented at West Beacon Farm in the UK. Substantial data has been collected that confirms the successful operation of the system.

The penetration of renewable energy sources has increased significantly in recent years due to the ongoing depletion of conventional resources and the transition to a low carbon energy system. Renewable energy sources such as wind energy are highly intermittent and unpredictable in nature, which makes the operation of the power grid more dynamic and therefore more complex. In order to operate the power system reliably under such conditions, Phasor Measurement Units (PMUs) through the use of satellite technology can offer a state-of-the-art Wide Area Monitoring System (WAMS) for improving power system monitoring, control and protection. They can improve the operation by providing highly precise and synchronised measurements near to real-time with higher frequency and accuracy. In order to achieve such objectives, a high-speed and reliable communications infrastructure is required to transfer time-critical PMU data from remote locations to the control centre. The signals measured by PMUs are transmitted across Local and Wide Area Networks, where they may encounter excessive delays. Signal delays can have a disruptive effect and make applications at best inefficient and at worse ineffective. The main research contribution of this thesis is the performance evaluation of communication infrastructures for WAMS. The evaluation begins from inside substations and continues over wide areas from substations to control centre. Through laboratory-based investigations and simulations, the performance of communications infrastructure in a typical power system substation has been analysed. In addition, the performance evaluation of WAMS communications infrastructure has been presented. In the modelling and analysis, an existing WAMS as installed on the GB transmission system has been considered. The actual PMU packets as received at the Phasor Data Concentrator (PDC) were captured for latency analysis. A novel algorithmic procedure has been developed and implemented to automate the large-scale latency calculations. Furthermore, the internal delays of PMUs have been investigated, determined and analysed. Subsequently, the WAMS has been simulated and detailed comparisons have been performed between the simulated model results and WAMS performance data captured from the actual WAMS. The validated WAMS model has been used for analysing possible future developments as well as to test newly proposed mechanisms, protocols, etc. in order to improve the communications infrastructure performance.

Transient stability analysis is an important part of power planning and operation. For large power systems, such analysis is very time consuming and expensive. Therefore, an online transient stability assessment will be required as these large power systems are operated close to their maximum limits. In this thesis swallowtail catastrophe is used to determine the transient stability regions. The bifurcation set represents the transient stability region in terms of power system transient parameters bounded by the transient stability limits. The system modelling is generalized in such, that the analysis could handle either one or any number of critical machines. This generalized model is then tested on a three-machine as well as a seven-machine system. The results of the stability analysis done with the generalized method is compared with the time solution and the results were satisfactory. The transient stability regions determined are valid for any changes in loading conditions and fault location. This method is a good candidate for on-line assessment of transient stability of power systems. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

Contemporary power networks are gradually expanding incorporating new sources of electrical energy and power electronic based devices. The major stability issue in large interconnected power systems is the lightly damped interarea oscillations. In the light of growth of their incidents there are increased concerns about the effectiveness of current control devices and control systems in maintaining power system stability. This thesis presents a Wide Area Measurement System (WAMS) based control scheme to enhance power system stability. The control scheme has a hierarchical (two-level) structure comprising a Supplementary Wide-Area Controller (SWAC) built on top of existing Power System Stabilisers (PSSs). The SWAC's focus is on stabilising the critical interarea oscillations in the system while leaving local modes to be controlled entirely by local PSSs. Both control systems in the two levels work together to maintain system stability. The scheme relies on synchronised measurements supplied by Phasor Measurement Units (PMUs) through the WAMS and the only cost requirement is for the communication infrastructure which is already available, or it will be in the near future. A novel linear quadratic Gaussian (LQG) control design approach which targets the interarea modes directly is introduced in this thesis. Its features are demonstrated through a comparison with the conventional method commonly used in power system damping applications. The modal LQG approach offers simplicity and flexibility when targeting multiple interarea modes without affecting local modes and local controllers, thus making it highly suitable to hierarchical WAMS based control schemes. Applicability of the approach to large power systems is demonstrated using different scenarios of model order reduction. The design approach incorporates time delays experienced in the transmission of the SWAC's input/output signals. Issues regarding values of time delays and required level of detail in modelling time delays are thoroughly discussed. Three methods for selection of input/output signals for WAMS based damping controllers are presented and reviewed. The first method uses modal observability/controllability factors. The second method is based on the Sequential Orthogonalisation (SO) algorithm, a tool for the optimal placement of measurement devices. Its application is extended and generalised in this thesis to handle the problem of input/output signal selection. The third method combines clustering techniques and modal factor analysis. The clustering method uses advanced Principal Component Analysis (PCA) where its draw backs and limitations, in the context of power system dynamics' applications, are overcome. The methods for signal selection are compared using both small signal and transient stability analysis to determine the best optimal set of signals. Enhancement of power system stability is demonstrated by applying the proposed WAMS based control scheme on the New England test system. The multi-input multi-output (MIMO) WAMS based damping controller uses a reduced set of input/output signals and is designed using the modal LQG approach. Effectiveness of the control scheme is comprehensively assessed using both small signal and transient stability analysis for different case studies including small and large disturbances, changes in network topology and operating condition, variations in time delays, and failure of communication links.

Variations in the electricity consumption and production connected to the power system have to be balanced by active control. Hydropower is the most important balancing resource in the Nordic system, and will become even more important as the share of variable renewable energy sources increases. This thesis concerns balancing of active power, especially the real-time balancing called frequency control. The thesis starts in a description of the situation today, setting up models for the behaviour of hydropower units and the power system relevant to frequency control, and comparing the models with experiments on several hydropower units and on the response of the Nordic grid. It is found that backlash in the regulating mechanisms in hydropower units have a strong impact on the quality of the delivered frequency control. Then, an analysis of what can be done right now to improve frequency control and decrease its costs is made, discussing governor tuning, filters and strategies for allocation of frequency control reserves. The results show that grid frequency quality could be improved considerably by retuning of hydropower governors. However, clear technical requirements and incentives for good frequency control performance are needed. The last part of the thesis concerns the impact from increased electricity production from variable renewable energy sources. The induced balancing need in terms of energy storage volume and balancing power is quantified, and it is found that with large shares of wind power in the system, the energy storage need over the intra-week time horizon is drastically increased. Reduced system inertia due to higher shares of inverter connected production is identified as a problem for the frequency control of the system. A new, linear synthetic inertia concept is suggested to replace the lost inertia and damping. It is shown that continuously active, linear synthetic inertia can improve the frequency quality in normal operation and decrease wear and tear of hydropower units delivering frequency control.

The Mbesese Initiative for Sustainable Development (MISD) is a group aiming to help eliminate extreme poverty in Africa by creating educational opportunity. One project that the group is currently doing is to build Same Polytechnic College (SPC) in Tanzania. As part of the project, this thesis aims to study and analyze the electrical power system and distribution for the college. Based on the projected load profile of the college and high potential for solar generation in Tanzania, several different power systems utilizing local utility AC electricity and/or photovoltaic (PV) DC electricity are explored and simulated for their feasibility and performance. Analysis of each design is presented and compared to determine the most viable system based on reliability, costs, and space. Results of the study indicate that over designing the DC system may generate wasteful energy while under designing the DC system may cause the overall system to rely heavily on the AC power grid. Ultimately, this thesis demonstrates that integrating a 58.9% DC system mixed with AC system offers the highest payback while efficiently utilizing the PV system, the battery system, and provided land.

Distributed Generation (DG) is one of the technologies approved by the South African government for the country's generation expansion to meet future load demand and to support economic growth. DGs change the conventional power flow (generation, transmission to distribution) by injecting real and reactive power at distribution voltage levels. The change in the conventional power flow creates complexity in the power system grid planning due to the conversion of the power system from a passive network to an active network. Introduction of bi-directional power flow on the power system can, among other benefits reduce local power demand which opens opportunities for capital investment deferrals on the transmission and distribution sectors. Consequently, DG impact on the transmission and distribution grid planning has been studied by other researchers. However, previous studies evaluated DG integration on a regulated market and assumed a certain level of generation availability during network peaking period. None of the studies have yet evaluated the benefits on an unregulated market using real measured data. Furthermore, SA distribution network expansion is also being planned without incorporating DGs on the network because of unreliability of wind and solar energy and the network operator's inability to influence the size, location and penetration level of DGs. This planning approach forces the network operator to do more to ensure high network strength. This approach can also result in network overdesign and unnecessary capital expenditure due to the potential benefits that can be deduced from DGs. This dissertation therefore aims to investigate whether incorporating future DG integration in distribution network planning can alleviate financial ramifications of grid code compliance requirements. The data used in the simulations was obtained from the distribution network operator and comprises of both real and reactive power values with a sampling time of 60 minutes for a period of a year. Simulations were conducted for both low and high load conditions to cover the extreme ends of the network and the parameters that were assessed are thermal rating, voltage regulation and network grid losses. Results showed that thermal constraints that are expected on the network when DGs are not considered are not evident when DGs are considered. Results further revealed that there are undervoltage improvements on the network when DGs are considered, and this reduces the capital expenditure that would have otherwise been incurred without DGs to result in a grid code compliant network. Furthermore, there is evidence of reduction in losses under high load conditions and increase in losses under low load conditions in the simulation results. Reduction in losses is caused by supplementary generation from wind and solar plants while increase in losses is due to excessive generation from wind plants which necessitate transportation over long distances to the nearest load centres. In addition to location, size and penetration levels as described in the literature, technology selection for a particular load type is also of utmost important to maximise the DG benefits on the network.

Low frequency oscillations are detrimental to the goals of maximum power transfer and optimal power system security. A contemporary solution to this problem is the addition of power system stabilizers to the automatic voltage regulators on the generators in the power system. The damping provided by this additional stabilizer provides the means to reduce the inhibiting effects of the oscillations. This thesis is an investigation of the use of synchronized phasor measurements as input signals for power system stabilizers installed on the generators of a two-area, 4-machine test power system. A remote measurement feedback controller has been designed and placed in the test power system. Synchronized phasor measurements from optimally sited measurement units were shown to improve the damping of low-frequency inter-area oscillations present in the test system when the proposed controller was included in the generator feedback control loop. The benefit of the damping of these oscillations was evident through the ability to increase the tie-line power flowing in the test system once the proposed control scheme was implemented. Time-domain simulations were used to verify the robustness of the proposed control during severe events, such as a short- circuit or sudden large variations of load. / Master of Science

Energy efficiency and renewable energy become more attractive in smart grid. In order to efficiently reduce global energy usage in building energy systems and to improve local environmental sustainability, it is essential to optimize the operation and the performance of combined heat and power (CHP) systems. In addition, intermittent renewable energy and imprecisely predicted customer loads have introduced great challenges in energy-efficient buildings' optimal operation. In the deterministic optimal operation, we study the modeling of components in energy-efficient building systems, including the power grid interface, CHP and boiler units, energy storage devices, and building appliances. The mixed energy resources are applied to collaboratively supply both electric and thermal loads. The results show that CHP can effectively improve overall energy efficiency by coordinating electric and thermal power supplies. Through the coordinated operation of all power sources, the daily operation cost of building energy system for generating energy can be significantly reduced. In order to address the risk from energy consumption forecast errors and renewable energy production volatility, we utilize the approach of stochastic programming and robust optimizations to operate energy-efficient building systems under uncertainty. The multi-stage stochastic programming model is introduced so that the reliable operation of building energy systems would be probabilistically guaranteed with stochastic decisions. The simulation results show that the stochastic operation of building energy systems is a promising strategy to account for the impact of the uncertainty on power dispatch decisions of energy-efficient building systems. In order to provide absolute guarantee for the reliable operation of building energy systems, a robust energy supply to electric and thermal loads is studied by exploring the effectiveness of energy storage on energy supply against the uncertainty. The robustness can be adjusted to control the conservativeness of the proposed robust operation model. For the purpose of achieving adaptability in the robust optimal operation and attaining robustness in the stochastic optimal operation of building energy systems, we also develop an innovative robust stochastic optimization (RSO) model. The proposed RSO model not only overcomes the conservativeness in the robust operation model, but also circumvents the curse of dimensionality in the stochastic operation model.

Dynamic simulation of a large-scale electric power system involves solving a large number of differential algebraic equations (DAEs) every simulation time-step. With the ever-growing size and complexity of power grid, dynamic simulation becomes more and more time-consuming and computationally difficult using conventional sequential simulation techniques. This thesis presents a fully distributed approach intended for implementation on High Performance Computer (HPC) clusters. A novel, relaxation-based domain decomposition algorithm known as Parallel-General-Norton with Multiple-port Equivalent (PGNME) is proposed as the core technique of a two-stage decomposition approach to divide the overall dynamic simulation problem into a set of sub problems that can be solved concurrently. While the convergence property has traditionally been a concern for relaxation-based decomposition, an estimation mechanism based on multiple-port network equivalent is adopted as the preconditioner to enhance the convergence of the proposed algorithm. The algorithm is presented in detail and validated both in terms of accuracy and capability