Dissertation/Thesis Abstract

Stochastic Drought Risk Analysis and Projection Methods For Thermoelectric Power Systems
by Bekera, Behailu Belamo, Ph.D., The George Washington University, 2016, 166; 3725243
Abstract (Summary)

Combined effects of socio-economic, environmental, technological and political factors impact fresh cooling water availability, which is among the most important elements of thermoelectric power plant site selection and evaluation criteria. With increased variability and changes in hydrologic statistical stationarity, one concern is the increased occurrence of extreme drought events that may be attributable to climatic changes. As hydrological systems are altered, operators of thermoelectric power plants need to ensure a reliable supply of water for cooling and generation requirements. The effects of climate change are expected to influence hydrological systems at multiple scales, possibly leading to reduced efficiency of thermoelectric power plants. This study models and analyzes drought characteristics from a thermoelectric systems operational and regulation perspective. A systematic approach to characterize a stream environment in relation to extreme drought occurrence, duration and deficit-volume is proposed and demonstrated. More specifically, the objective of this research is to propose a stochastic water supply risk analysis and projection methods from thermoelectric power systems operation and management perspectives. The study defines thermoelectric drought as a shortage of cooling water due to stressed supply or beyond operable water temperature limits for an extended period of time requiring power plants to reduce production or completely shut down. It presents a thermoelectric drought risk characterization framework that considers heat content and water quantity facets of adequate water availability for uninterrupted operation of such plants and safety of its surroundings. In addition, it outlines mechanisms to identify rate of occurrences of the said droughts and stochastically quantify subsequent potential losses to the sector. This mechanism is enabled through a model based on compound Nonhomogeneous Poisson Process. This study also demonstrates how the systematic approach can be used for better understanding of pertinent vulnerabilities by providing risk-based information to stakeholders in the power sector.

Vulnerabilities as well as our understanding of their extent and likelihood change over time. Keeping up with the changes and making informed decisions demands a time-dependent method that incorporates new evidence into risk assessment framework. This study presents a statistical time-dependent risk analysis approach, which allows for life cycle drought risk assessment of thermoelectric power systems. Also, a Bayesian Belief Network (BBN) extension to the proposed framework is developed. The BBN allows for incorporating new evidence, such as observing power curtailments due to extreme heat or lowflow situations, and updating our knowledge and understanding of the pertinent risk. In sum, the proposed approach can help improve adaptive capacity of the electric power infrastructure, thereby enhancing its resilience to events potentially threatening grid reliability and economic stability.

The proposed drought characterization methodology is applied on a daily streamflow series obtained from three United States Geological Survey (USGS) water gauges on the Tennessee River basin. The stochastic water supply risk assessment and projection methods are demonstrated for two power plants on the White River, Indiana: Frank E. Ratts and Petersburg, using water temperature and streamflow time series data obtained from a nearby USGS gauge.

Indexing (document details)
Advisor: Francis, Royce A.
Commitee: Mazzuchi, Thomas A., Omitaomu, Olufemi A., Santos, Joost R., Shittu, Ekundayo
School: The George Washington University
Department: Engineering Management and Systems Engineering
School Location: United States -- District of Columbia
Source: DAI-B 77/02(E), Dissertation Abstracts International
Subjects: Climate Change, Engineering, Energy
Keywords: Bayesian belief nets, Climare risk modeling, Compound nhpp, Energy loss modeling, Probabilistic risk assessment, Thermoelectric power plants
Publication Number: 3725243
ISBN: 9781339093994