Simulation and loss investigation of large-scale axial turbines operating with supercritical carbon dioxide mixtures

Supercritical carbon dioxide (sCO2) mixtures have been found promising in improving the power generation efficiency for concentrated solar power (CSP) applications. The utilisation of these novel working fluids poses various design challenges. For the turbine design, the available mean line loss models were developed for conventional working fluids while their applicability in designing sCO2 mixture turbines has not yet been assessed. This research aims to verify the turbine design, previously developed using the mean line models, utilising numerical simulations due to the lack of experimental data. The study is extended to assess the applicability of the mean line loss models at off-design operating conditions. The aerodynamic performance is simulated using 3D viscous steady-state Reynolds averaged Navier Stokes computational fluid dynamics (CFD) simulations. The blade stresses have been evaluated using finite element analysis (FEA) to assess the safety of the design. The preliminary design has been further improved through blade shape optimisation in which CFD and FEA simulations have been utilised to improve the performance, satisfy the cycle requirements, and maintain acceptable stress limits. Aerodynamic losses have been investigated utilising CFD simulations for different mixtures, power scales, and pressure ratios to improve the understanding of aerodynamic losses in large-scale axial turbines operating with sCO2 mixtures. An improved loss breakdown estimation approach has been developed to address the shortcomings of the previously published approaches by considering the interaction between loss sources and the boundary layer thickness variation for each model. This is considered important for dense working fluids such as pure CO2 and CO2 mixtures. Subsequently, CFD simulations have been utilised to evaluate the off-design performance of various flow path geometries. The CFD simulation results showed the suitability of the mean line model in predicting the performance of large-scale axial turbines operating with sCO2 mixtures, with a total-to-total efficiency deviation of less than 2.2%. However, a large deviation of 6.7% was observed in the mass flow rate obtained from the 3D blades generated based on the mean line results. This elevates the importance of blade shape optimisation to constraint the mass flow rate within 2% of the design value, while also considering stress limits. Off-design simulations revealed the mean line models' limited accuracy away from the design point, showing a 17.5% efficiency deviation at 84% of the design mass flow coefficient