Modeling water use at thermoelectric power plants

Abstract

The withdrawal and consumption of water at thermoelectric power plants affects regional ecology and supply security of both water and electricity. The existing field data on US power plant water use, however, is of limited granularity and poor quality, hampering efforts to track industry trends and project future scenarios. Furthermore, there is a need for a common quantitative framework on which to evaluate the effects of various technologies on water use at power plants. To address these deficiencies, Part 1 of this thesis develops an analytical system-level generic model (SGEM) of water use at power plants. The S-GEM applies to fossil, nuclear, geothermal and solar thermal plants, using either steam or combined cycles, and outputs water withdrawal and consumption intensity, in liters per megawatt-hour. Two validations of the S-GEM are presented, one against data from the literature for a variety of generation types, the other against field data from coal plants in South Africa. Part 2 of the thesis then focuses on cooling systems, by far the largest consumers of water in most power plants. The water consumption of different cooling systems is placed on a common quantitative basis, enabling direct comparison of water consumption between cooling system types, and examination of the factors that affect water consumption within each cooling system type. The various cost, performance, and environmental impact tradeoffs associated with once-through, pond, wet tower, dry, and hybrid cooling technologies are qualitatively reviewed. Part 3 examines cooling of concentrating solar power (CSP) plants, which presents particular problems: the plants generate high waste heat loads, are usually located in water-scarce areas, and are typically on the margin of economic viability. A case study is conducted to explore the use of indirect dry cooling with cold-side thermal energy storage, in which cooling water is chilled and stored at night, when ambient temperatures are lower and the plant is inactive, and then used the following day. This approach is shown to hold promise for reducing the capital, operational, and performance costs of dry cooling for CSP.