Carleton University Institutional Repository
Not a member yet
20816 research outputs found
Sort by
Development of a Residential-Scale PV/T Driven Combined Cooling, Heating, and Power System
Get PDFPhotovoltaic/Thermal (PV/T) driven combined cooling, heating, and power (CCHP) systems are a promising low-carbon alternative for meeting building energy demands in the residential sector. This thesis focuses on the development of a novel PV/T driven CCHP topology and corresponding control strategy for use in residential buildings. To achieve the objectives, a research plan comprising three parts is devised. In the first part of the research, a computational fluid dynamic (CFD) - based numerical modelling approach is developed to simulate the operation of a PV/T driven domestic hot water (DHW) system under dynamic conditions. Using the case study location of Ottawa, Canada, 48 different scenarios are developed which involve two different working fluids (i.e., air and a water-ethylene glycol solution) and two PV/T collector design alternatives (i.e., with and without fins) for each month of the year. Results show that the solar fraction is greater in the air-based system than in the water-based system for all months, and values as high as 90.1% and 84.3% are obtained, respectively. In the second part, a numerical model of a PV/T driven CCHP system, controlled via a novel dual-tank latent heat thermal energy storage (LHTES) strategy, is developed in TRNSYS and C++. The system performance is investigated via a case study that considers a three-bedroom home located in Ottawa, Canada. The phase change material (PCM) melting temperature, load loop supply temperature, and LHTES tank height are varied in the analysis resulting in 27 annual simulation scenarios. For the considered simulations, the annual solar and electricity fractions are shown to vary between 10.3% and 39.7%, and 23.2% and 33.2%, respectively. In the third part of the thesis, the performance of the above-mentioned PV/T driven CCHP system is further assessed for use in a low-energy multi-unit residential building. Two North American case study locations representing a heating dominated (Ottawa, CA) and a cooling dominated (Albuquerque, USA) climate are examined by considering two solar collector array area sizes. Results show that operating the system in Albuquerque relative to Ottawa results in an increase in solar fraction and electricity fraction of up to 62% and 40%, respectively
Delay Optimization in Digital FinFET Circuits Using a Modified Method of Logical Effort
Get PDFDelay optimization in digital circuits is crucial, especially at advanced nodes like 12nm, which use FinFET transistors. This thesis extends the classical logical effort method, originally designed for continuous-width transistors (e.g., 65nm CMOS), to optimize delay in FinFET-based circuits. Unlike traditional transistors, FinFETs have discrete width parameters—fins, fingers, and multiplicities—requiring modifications to the standard approach. Two key adaptations are introduced. First, transistor sizing is mapped to discrete FinFET dimensions. Second, a correction factor (Cg/Cd = 1.33) is incorporated to account for the gate-to-diffusion capacitance ratio. These adjustments ensure compatibility with FinFET technology while maintaining accuracy in delay prediction. The optimized method achieves less than 8% error in delay estimation. Simulations show over 20% performance improvement when FinFET sizes are optimized, demonstrating its effectiveness. This framework enables efficient delay optimization, balancing speed and power in FinFET digital circuits
Methodologies for Using Neural Networks to Facilitate an Aerial Vehicle’s Fully Automated and Uncrewed Landing on a Moving Ship
Get PDFAn autonomous Uncrewed Aerial Vehicle (UAV) attempting to perform a vertical landing on a moving ship’s deck must be capable of predicting the ship’s motion in order determine the most opportune landing time. If the UAV is acting independent of the ship additional constraints are introduced; computation resources are limited to what can be mounted on the drone and the UAV must predict motion from noisy UAV-mounted sensors. The work in this thesis presents a methodology for creating datasets which reflect ship motion that would be measured by a UAV attempting to land. Eight Neural Network (NN) models are evaluated for their performance in order to determine which architectures are suited for the predictions used in the landing system. Each model’s ability to minimize error, handle input noise, and their computational efficiency is studied. Simulations are used to evaluate each model’s in-situ capabilities and it is shown that the methodology presented creates feasible NN models
Speed Control of Permanent Magnet Synchronous Motor Employing Direct Voltage Maximum Torque Per Ampere
Get PDFPermanent magnet synchronous motors (PMSMs) have become more attractive and competitive for various industrial applications. In the constant torque region, a commonly adopted technique for maximum torque utilization of PMSMs is known as maximum torque per ampere control (MTPA). This technique aims to achieve the desired torque with minimal current consumption. Conventional MTPA control approaches typically utilize PI-based cascaded control loops, which lead to significant time costs in practical. In contrast, direct voltage control methods (DVC) are alternative techniques to track the MTPA points by directly manipulating the voltage vector amplitude and angle, eliminating the need for cascaded regulation. However, the few developed direct voltage control methods still rely on complex torque estimation, numerical approximations of a given motor, lookup tables, and lengthy iterative calculations to determine its control laws. In view of these limitations, this thesis aims to introduce various analytical simplified direct voltage maximum torque per ampere speed control approaches. Initially, a simple current sensing-based direct voltage MTPA control is introduced for IPMSM propelling electric vehicles. The design strategy’s voltage amplitude and angle control laws are analytically derived from the motor’s electrical model based on measured stator current, eliminating the need for torque estimation, control law approximation, or iterative solution. Direct Voltage MTPA techniques are model-based approaches that depend heavily on time-varying electrical parameters. Furthermore, shaft speed and position are of great importance in these strategies. To overcome the adverse effects of parameter variation and mechanical sensor drawbacks, a multiparameter estimation-based sensorless adaptive direct voltage MTPA control is introduced for IPMSM. A fuzzy logic control-based cascaded model reference adaptive system (FLC-MRAS) is suggested for online parameter tracking and provides precise rotor speed and position. To further reduce the control system’s complexity and improve its reliability, a simple current sensorless MTPA control is proposed to control the speed of the IPMSM and SPMSM. The proposed strategy eliminates the need for current sensing, transformations, and regulation loops. The validity of the designed strategies is confirmed experimentally, and the results obtained are compared to the conventional field-oriented MTPA (FOC). Additionally, the performance of the developed strategies is quantitatively assessed
