| dc.description.abstract | Wind energy is a proven energy source that does not contribute to emission of greenhouse gases, air and water pollution, or generate large quantities of waste. However, wind generation is dependent on wind speed, which is difficult to predict with high accuracy. The intermittent nature of wind generation makes its operation and planning a complex problem and there is a need for advanced analytical models to embed this uncertainty in its generation profile. This research focuses on the development of innovative mathematical modeling and analysis tools to improve our understanding of the effects of wind generation on power systems.
The overall goal of this research is to introduce novel analytical frameworks to consider the penetration of wind generation sources within the distribution and transmission networks. In particular, two main operational problems are addressed within this thesis; the Distribution Load Flow (DLF) problem and the Unit Commitment (UC) problem in the presence of wind generation.
First for the DLF problem, a novel probabilistic wind generation model is presented. The probabilistic wind generation profile, which is a function of the wind speed, is considered and an appropriate procedure is developed to classify specific levels based on wind speed, in order to reduce the number of probabilistic combinations of wind power generation. Next, a novel Probabilistic Distribution Load Flow (PDLF) approach is used to evaluate the impact of wind penetration into distribution systems. The traditional DLF program is modified to include the wind generation profiles. Three Wind Turbine (WT) models are derived and integrated within the PDLF program to examine and compare their performance. The probabilistic forward-backward sweep algorithm is developed for the first two models of WT. For the third model of WT, a probabilistic compensation-based load flow is presented. The effect of WT penetration is investigated on feeder losses, voltage profile and line flows.
Secondly, a new scenario generation and reduction technique is developed for analyzing the effects of wind generation uncertainties on short-term power system operation. A historical wind speed data set is used to obtain different wind speed clusters which are then processed through Monte Carlo Simulations (MCS), Markov-chains and a forward selection scenario reduction algorithm to obtain a reduced set of scenarios. These reduced scenarios are then incorporated into a Locational Marginal Price (LMP) based electricity market settlement and dispatch model. These UC type models incorporate system constraints and transmission constraints to examine the effects of wind generation on electricity market prices, UC decisions including generation, reserve requirement, load cleared and social welfare. Markov-chain transition matrices are developed to include the effect of the inter-hour transition correlation of wind speed from one specific hour to the following hour to improve the generation of the wind scenarios. The effect of changing wind farm capacity on system operation is also discussed. Furthermore, the impact of the wake-effect phenomena influencing off-shore wind turbines is explained.
Finally, this research examines the effect of wind generation penetration on the environmental emissions. A novel methodology is developed to evaluate the environmental impact of wind generation penetration into electrical power systems. The solution of the market dispatch UC model is studied for different cost functions with an emission cap. The relationship between changing the emission caps and the penetration level of wind energy is investigated. Furthermore, the effect on market prices is also examined when emission caps are imposed by external agencies, on the System Operator (SO). | en |
