Investigating supercritical carbon dioxide power cycles and the potential of improvement of turbine leakage characteristics via a barrier gas

The supercritical carbon dioxide power cycle has recently received tremendous interest owing to its compactness and enhanced efficiency. However, the inherent characteristics of supercritical carbon dioxide and high rotational speeds pose unique challenges in designing cycle turbomachinery sealing. Conventional sealing technology can result in an efficiency penalty as high as 0.55–0.65%. This drives the need for designing a more efficient sealing system to enable a supercritical carbon dioxide power cycle. In this study, leakage characteristics of a 500 °C class supercritical carbon dioxide turbine with labyrinth seals are modeled and investigated. The labyrinth seal model was validated against experimental results. In contrast to previous studies, a new leakage management scheme using a barrier gas was proposed and simulated. The carbon dioxide from the main cycle is utilized as the barrier gas; hence, no additional cost is incurred to produce the high-pressure barrier gas. The comprehensive parametric study revealed that the performance of the labyrinth seal with barrier gas is mainly influenced by leakage inlet pressure, barrier gas injection pressure, the differential pressure between the barrier gas and exit pressure, and the length of the seals. Subsequently, the performance of the proposed design is optimized using a pattern search algorithm in the MATLAB environment. The electric power required to reinject the leaked flow into the main cycle is taken as the objective function. The leakage flowrates for the benchmark and proposed cases are 0.0071 kg/s and 0.0108 kg/s, respectively. The proposed design reduced the reinjection power by collecting the leakage at a higher pressure of 832.28 kPa, compared to 101.325 kPa for the benchmark case. The optimization results show that, via the proposed scheme, the leakage recovery power demand is 2.897 kW whereas it is 3.725 kW for the benchmark case. Thus, the proposed scheme results in a 22.24% saving in electric power, compared to the benchmark case. Furthermore, it is reported for the proposed case that the amount of leakage avoiding turbine expansion is 0.0086 kg/s while for the benchmark case it was 0.0071. For the simulation conditions considered in this study, the extra leakage results in a 0.15% power penalty from the turbine, which is insignificant compared to the power saving in the leakage compression power.