Thermofluid design and performance analysis of an air-cooled heat rejection system for a 50 MWe sCO2 concentrated solar power plant

Thesis / Dissertation

2024

Permanent link to this Item
Authors
Supervisors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher

University of Cape Town

License
Series
Abstract
There is increasing interest in renewable energy technologies, both globally and within South Africa. An emerging renewable energy technology is concentrated solar power (CSP) plants utilising supercritical carbon dioxide (sCO2) as a working fluid. This technology promises a more compact and efficient alternative to the current Rankine cycle based CSP plants. However, no such plant currently exists at the utility scale, and little is known about the heat rejection systems of such plants. If these plants are to be realised at a utility-scale, further research is required with regards to the process design, physical layouts, and operational philosophy of these heat rejection systems. The present work includes the development of the conceptual layout and steady-state onedimensional (1D) thermofluid network models of forced draft direct air-cooled heat rejection systems for a proposed 50 MWe sCO2 CSP plant. The models were developed using opensource software libraries in Python 3.8.5. The heat rejection systems consist of cooling cells, each with horizontal multi-pass staggered finned tube banks employing axial forced draft fans. In contrast to existing studies, the designs consider the mechanical integrity of the components e.g., tube wall thickness and the availability of standard components e.g., commercial fans. The number of cooling cells, fan selection, and tube geometry is determined by coupling the heat rejection system model to a sCO2 power cycle model developed by others, and then developing precooler (PC) and intercooler (IC) designs which maximise cycle efficiency. The findings indicate that cycle efficiency increases with increased heat transfer area in the air-cooled heat rejection systems. However, any optimisation of the PC and IC needs to be performed together, and with the cycle model for the best results. Furthermore, an increase in the thermal efficiency of the power cycle was observed when including the detailed cooler models, as opposed to the simplified cooler models used in the original cycle model, demonstrating the value of including more detailed models in cycle-level studies. The methodology applied in this study may be used in future works to size heat rejection systems for sCO2-CSP plants. The PC and IC designs were then modelled in the Flownex ® SE software to perform quasi-steady state studies, and to assist with the development of complete dynamic plant models by others. The Python and Flownex models agree well, with the heat rejection rates within 3%. The Flownex ® SE model features improved (additional) discretisation of the heat exchangers on the sCO2-side and includes distribution headers which consider mechanical integrity and simulate realistic flow distribution. Since air-cooled heat exchangers are known to be sensitive to ambient conditions, the performance of the systems under varying ambient air temperatures were investigated. This was done using data for a typical year for the Upington area in South Africa. Additionally, the performance of the systems under cycle part load conditions were also investigated. The results indicate that the performance of the systems is especially sensitive to air temperatures and cycle part-load operation. For temperatures lower than the nominal design temperature, and at cycle part-load operation, over-cooling of the working fluid occurs. This can be controlled by bypassing areas of the heat rejection system and by switching off fans. However, to ensure sufficient cooling at higher temperatures, the heat-rejection systems need to be oversized
Description

Reference:

Collections