To date variety of supercritical CO2 cycles were proposed by numerous authors. Multiple small-scale tests performed, and a lot of supercritical CO cycle aspects studied. Currently, 3-10 MW-scale test facilities are being built. However, there are still several pieces of SCO2 technology with the Technology Readiness Level (TRL) 3-5 and system modeling is one of them. The system modeling approach shall be sufficiently accurate and flexible, to be able to precisely predict the off-design and part-load operation of the cycle at both supercritical and condensing modes with diverse control strategies. System modeling itself implies the utilization of component models which are often idealized and may not provide a sufficient level of fidelity. Especially for prediction of off-design and part load supercritical CO2 cycle performance with near-critical compressor and transition to condensing modes with lower ambient temperatures, and other aspects of cycle operation under alternating grid demands and ambient conditions.
In this study, the concept of a digital twin to predict off-design supercritical CO2 cycle performance is utilized. In particular, with the intent to have sufficient cycle simulation accuracy and flexibility the cycle simulation system with physics-based methods/modules were created for the bottoming 15.5 MW Power Generation Unit (PGU). The heat source for PGU is GE LM6000-PH DLE gas turbine. The PGU is a composite (merged) supercritical CO2 cycle with a high heat recovery rate, its design and the overall scheme are described in detail. The calculation methods utilized at cycle level and components’ level, including loss models with an indication of prediction accuracy, are described. The flowchart of the process of off-design performance estimation and data transfer between the modules as well. The comparison of the results obtained utilizing PGU digital twin with other simplified approaches is performed. The results of the developed digital twin utilization to optimize cycle control strategies and parameters to improve off-design cycle performance are discussed in detail.