Analysis and Comparison of the Simple and Recompression Supercritical CO2 Cycles

by John Bryant1, Henry Saari1, & Kourosh Zanganeh2
1Carleton University & 2Natural Resources Canada, CanmetENERGY

Click Here To Access The Complete Paper
Click Here To Access The Presentation

The recompression cycle has been proposed as a means of improving upon the efficiencies obtainable using a simple Brayton cycle. It differs from the simple cycle in that a portion of the mass flow is recompressed to the working pressure without first being precooled. A second compressor is required and the single recuperator of the simple cycle is split into two separate parts, so increasing system complexity, but this can be justified by the potential for increased efficiency.

This paper analyzes the efficiencies of the simple and recompression cycles over a range of operating conditions (turbine inlet temperature, compressor inlet pressure, turbine inlet pressure) and with recuperator performance defined in terms of both effectiveness and minimum attainable pinch. NIST material properties were used throughout the analyses, but pressure losses in piping and heat exchangers were neglected. Taking them into account would reduce the efficiencies of both the simple and recompression cycles, but the recompression cycle would presumably suffer more because of the greater recuperator areas required. A previously published comparison of the simple and recompression cycles is critiqued and the efficiencies of the two cycles are compared using a variety of assumptions.

It is proved that, for any given compressor and turbine conditions, the recompression cycle will always be more efficient than a simple cycle provided that the two cycles have the same precooler inlet temperature. In order to satisfy this condition, however, the recompression cycle will always require more total recuperator area It will be demonstrated that when the two cycles are compared on the basis of equal total recuperator area, the efficiency advantage of the recompression cycle is reduced and, depending upon the assumptions made, can disappear altogether.