A Comparison Study for Off-Design Performance Prediction of a Supercritical CO2 Compressor With Similitude Analysis

Abstract

Abstract Most of the power plants operating nowadays mainly have adopted a steam Rankine cycle or a gas Brayton cycle. To devise a better power conversion cycle, various approaches were taken by researchers and one of the examples is an S-CO2 (supercritical CO2) power cycle. Over the past decades, the S-CO2 power cycle was invented and studied. Eventually the cycle was successful for attracting attentions from a wide range of applications. Basically, an S-CO2 power cycle is a variation of a gas Brayton cycle. In contrast to the fact that an ordinary Brayton cycle operates with a gas phase fluid, the S-CO2 power cycle operates with a supercritical phase fluid, where temperatures and pressures of working fluid are above the critical point. Many advantages of S-CO2 power cycle are rooted from its novel characteristics. Particularly, a compressor in an S-CO2 power cycle operates near the critical point, where the compressibility is greatly reduced. Since the S-CO2 power cycle greatly benefits from the reduced compression work, an S-CO2 compressor prediction under off-design condition has a huge impact on overall cycle performance. When off-design operations of a power cycle are considered, the compressor performance needs to be specified. One of the approaches for a compressor off-design performance evaluation is to use the correction methods based on similitude analysis. However, there are several approaches for deriving the equivalent conditions but none of the approaches has been thoroughly examined for S-CO2 conditions based on data. The purpose of this paper is comparing these correction models to identify the best fitted approach, in order to predict a compressor off-design operation performance more accurately from limited amount of information. Each correction method was applied to two sets of data, SCEIL experiment data and 1D turbomachinery code off-design prediction code generated data, and evaluated in this paper.