Thesis (PH.D.) -- Syracuse University, 2004. / Advisers: Elsherbeni, Atef Z. ; Arvas, Ercument. "Publication number AAT 3135873."

After more than a century of the ultimate dominance of AC in distribution systems, DC distribution is being re-considered. However, the advantages of AC systems cannot be omitted. This is mainly due to the cheap and efficient means of generation provided by the synchronous AC machines and voltage stepping up/down allowed by the AC transformers. As an intermediate solution, hybrid AC/DC distribution systems or microgrids are proposed. This hybridization of distribution systems, incorporation of heterogeneous mix of energy sources, and introducing Pulsed Power Loads (PPL) together add more complications and challenges to the design problem of distribution systems. In this dissertation, a comprehensive multi-objective optimization approach is presented to determine the optimal design of the AC/DC distribution system architecture. The mathematical formulation of a multi-objective optimal power flow problem based on the sequential power flow method and the Pareto concept is developed and discussed. The outcome of this approach is to answer the following questions: 1) the optimal size and location of energy storage (ES) in the AC/DC distribution system, 2) optimal location of the PPLs, 3) optimal point of common coupling (PCC) between the AC and DC sides of the network, and 4) optimal network connectivity. These parameters are to be optimized to design a distribution architecture that supplies the PPLs, while fulfilling the safe operation constraints and the related standard limitations. The optimization problem is NP-hard, mixed integer and combinatorial with nonlinear constraints. Four objectives are involved in the problem: minimizing the voltage deviation (ΔV), minimizing frequency deviation (Δf), minimizing the active power losses in the distribution system and minimizing the energy storage weight. The last objective is considered in the context of ship power systems, where the equipment’s weight and size are restricted. The utilization of Hybrid Energy Storage Systems (HESS) in PPL applications is investigated. The design, hardware implementation and performance evaluation of an advanced – low cost Modular Energy Storage regulator (MESR) to efficiently integrate ES to the DC bus are depicted. MESR provides a set of unique features: 1) It is capable of controlling each individual unit within a series/parallel array (i.e. each single unit can be treated, controlled and monitored separately from the others), 2) It is able to charge some units within an ES array while other units continue to serve the load, 3) Balance the SoC without the need for power electronic converters, and 4) It is able to electrically disconnect a unit and allow the operator to perform the required maintenance or replacement without affecting the performance of the whole array. A low speed flywheel Energy Storage System (FESS) is designed and implemented to be used as an energy reservoir in PPL applications. The system was based on a separately excited DC machine and a bi-directional Buck-Boost converter as the driver to control the charging/discharging of the flywheel. Stable control loops were designed to charge the FESS off the pulse and discharge on the pulse. All the developments in this dissertation were experimentally verified at the Smart Grid Testbed.

We present a video-based system designed to enhance the efficiency of emergency preparedness and recovery. Our system includes two tools, one or emergency preparedness, and the other for emergency recovery. In our emergency preparedness tool, we 1) capture videos from the physical environment (e.g., University Campus, Knowles Engineering Building, North Station Subway Station), 2) generate an immersive virtual environment using these videos, 3) provide exploration and navigation mode in the virtual environment, and 4) allow users to make annotations of the environment and review these created annotations. By using this virtual environment, emergency personnel can familiarize themselves with the environment before their actual training in the physical setting. This tool can enhance the training experience of emergency responders as well as improve the training outcomes. Our emergency recovery tool includes two parts: 1) an Android application which captures the videos in the disaster site, and 2) a server that stores all these videos and provides visual analytics support. The recorded videos (both in indoor and outdoor settings), which include location and orientation information, are uploaded to the server that supports spatial queries. We developed tools for the investigators to review these videos, e.g., review videos in specific areas of interest. To test our tools, we select both outdoor and indoor environments. The results show that our tool works well in these situations.

Swept Frequency Source Optical Coherence Tomography (OCT) requires high repetition rate and wide spectral width wavelength tunable sources at a low cost. The sampled grating distributed Bragg reflector (SG-DBR) laser provides wide wavelength tuning range while exhibiting a wavelength switching speed that is among the fastest currently available. The SG-DBR laser is used to generate linear frequency ramps with high repetition rates. Since the SG-DBR laser is currently used for the telecommunications industry in high volume, the price of the system is much lower than current OCT sources. Therefore the SG-DBR laser provides a practical solution for Swept Source OCT. Four synchronized waveforms sent to the inputs of the laser control the linear frequency ramp. Three of the waveforms control the output frequency, while the fourth waveform controls the output amplifier of the laser to keep the output power stable. Two SG-DBR lasers with overlapping wavelength coverage are also concatenated to increase the bandwidth of the swept frequency source. The linear ramp stitching points in the frequency ramps are investigated and methods are outlined to reduce them. Finally, experimental OCT tests are performed using the swept frequency sources created to analyze the linearity of our sources. From these test results, an evaluation can be made on the specifications and capabilities of the swept frequency sources and analyze their value for OCT applications.

n this thesis project, a proposed architecture for the multiple input, single output conversion stage for the DC House was designed, simulated, and tested. This architecture allows for multiple different input sources to be used to create a single higher power output source. The design uses a DC-DC boost converter with a parallelable output which has been demonstrated to allow increased total output power as a function of the number of input sources available. The parallelable output has been shown to distribute load amongst the input sources relatively closely to optimize the system. This approach is also desirable since it allows for flexibility in multiple configurations it can be used in. The design was tested using hardware and data results show the performance met and exceeded the needs of the DC House project. Data was taken for configuration with 1, 2, 3, and 4 input sources providing greater than 600W of total output power at an efficiency of greater than 92%. This architecture demonstrates the possibility of expanding the total available power for a single output in proportion to the number of available input sources.

In this research, we investigate a data processing method to capture the respiratory rate of a person by utilizing a doppler radar to monitor their body movement during respiration. We utilize a machine learning algorithm with a radar sensor to capture the chest movement of a person while breathing and determine the respiratory rate according to that movement. We are using a Random Forest classifier to distinguish between different classes of pulses. After that, the algorithm constructs a sinusoidal signal representing the breathing rate of the sample. By applying this technique, we can detect the breathing rate accurately for different subjects by analyzing the evolution of the reflected pulse while breathing. Furthermore, we can detect the change in pulse width ratio between the pulses of the classes across multiple breaths

High frequency (HF) radars have many critical applications due to the effects that physical media have on the wave's propagation. The diffraction of HF radio waves in the ionosphere allows for long range communication and radar operation. Waves travel over the horizon where they may be reflected off large scatterers such as ships, or monitor sea states over large oceanic surface areas. Furthermore, the ionosphere provides key information on solar weather. Monitoring RF reflections from the ionosphere (specifically at the polar regions) is of great importance to the scientific community. The use of the HF (3-30 Mhz) has been greatly simplified for radar transceivers in recent decades. Digital hardware can sample and process information fast enough to eliminate the need for conventional analogue down-converters. The result is an increase in sensitivity, signal to noise ratio and design simplicity. The primary advantage of digital radars is versatility. The ability to change parameters and even modes of operation means that digital radars have become more common, and have replaced or been partly integrated into most of their analogue counterparts. The SuperDARN is a network of ionosphere monitoring radars that have been in operation since the 1980s. Since its inception it has undergone multiple improvements and served the scientific community well. The 4th South African National Arctic Expedition (SANAE IV) makes use of a digital radar platform based on the third generation TIGER-3 FPGA boards. The highly adaptable nature of the transceivers provide a host of secondary applications and improvements to its analogue predecessors. The system is however not in a state that supports further development. Currently the system is programmed for a set mode of operation without access to the source software. This work details the design and implementation process followed to bring the TIGER-3 system to a state that will support further development. In this state, peripheral interfaces are designed and implemented to allow for a listening mode of operation. In this mode, the radar samples a signal from an antenna and effectively communicates the data to a personal computer via an Ethernet link. To achieve these outcomes; FPGA code (written in Verilog) was developed to implement IQ downconversion, digital filtering, and a client interface for the Ethernet link. The features were tested by recording and analysing digital outputs from the platform, and finally, by recording signal information obtained through the Ethernet interface. Supporting literature will lay the groundwork for future projects to build on the base layer implementation; with the hope of redesigning the current SuperDARN implementation in the future. Further improvements to the current system could include a range of scanning patters and multi-frequency operation.

The advent of CubeSats has provided a platform for relatively low-budget programmes to realise space missions. In South Africa, Stellenbosch University and the Cape Peninsula University of Technology have impressive space programmes and have been involved in numerous successful satellite launches. A number of CubeSat projects are currently in progress and commercial-grade Attitude Determination and Control Systems (ADCS), and communications modules, are being developed by the respective universities. The development of a CubeSat-compatible Electrical Power System remains absent, and would be beneficial to future satellite activity here in South Africa. In this thesis, some fundamental aspects of electronic design for space applications is looked at, including but not limited to radiation effects on MOSFET devices; this poses one of the greatest challenges to space-based power systems. To this extent, the different radiation-induced effects and their implications are looked at, and mitigation strategies are discussed. A review of current commercial modules is performed and their design and performance evaluated. A few shortcomings of current systems are noted and corresponding design changes are suggested; in some instances these changes add complexity, but they are shown to introduce appreciable system reliability. A single Li-Ion cell configuration is proposed that uses a 3.7 V nominal bus voltage. Individual battery charge regulation introduces minor inefficiencies, but allows isolation of cells from the pack in the case of cell failure or degradation. A further advantage is the possibility for multiple energy storage media on the same power bus, allowing for EPS-related technology demonstrations, with an assurance of minimum system capabilities. The design of each subsystem is discussed and its respective failure modes identified. A limited number of single points of failure are noted and the mitigation strategies taken are discussed. An initial hardware prototype is developed that is used to test and characterise system performance. Although a few minor modifications are needed, the overall system is shown to function as designed and the concepts used are proven.

No description available.

No description available.