Preliminary Design and Model Assessment of a Supercritical CO2 Compressor

• Main Authors: Zhiyuan Liu, Weiwei Luo, Qingjun Zhao, Wei Zhao, Jianzhong Xu
• Format: Article
• Language: English
• Published: MDPI AG 2018-04-01
• Series: Applied Sciences
• Subjects: SCO2; compressor design; loss model; numerical simulation
• Online Access: http://www.mdpi.com/2076-3417/8/4/595

The compressor is a key component in the supercritical carbon dioxide (SCO2) Brayton cycle. In this paper, the authors designed a series of supercritical CO2 compressors with different parameters. These compressors are designed for 100 MWe, 10 MWe and 1 MWe scale power systems, respectively. For the 100 MWe SCO2 Brayton cycle, an axial compressor has been designed by the Smith chart to test whether an axial compressor is suitable for the SCO2 Brayton cycle. Using a specific speed and a specific diameter, the remaining two compressors were designed as centrifugal compressors with different pressure ratios to examine whether models used for air in the past are applicable to SCO2. All compressors were generated and analyzed with internal MATLAB programs coupled with the NIST REFPROP database. Finally, the design results are all checked by numerical simulations due to the lack of reliable experimental data. Research has found that in order to meet the de Haller stall criterion, axial compressors require a considerable number of stages, which introduces many additional problems. Thus, a centrifugal compressor is more suitable for the SCO2 Brayton cycle, even for a 100 MWe scale system. For the performance prediction model of a centrifugal compressor, the stall predictions are compared with steady numerical calculation, which indicates that past stall criteria may also be suitable for SCO2 compressors, but more validations are needed. However, the accuracy of original loss models is found to be inadequate, particularly for lower flow and higher pressure ratio cases. Deviations may be attributed to the underestimation of clearance loss according to the result of steady simulation. A modified model is adopted which can improve the precision to a certain extent, but more general and reasonable loss models are needed to improve design accuracy in the future.