Impact of rotor geometry optimization on the off-design ORC turbine performance

The paper describes the method of CFD based Nelder-Mead optimization of a 10 kW single-stage axial turbine operating in an ORC system working on R7100. The total-to-static isentropic efficiency is defined as an objective function. Multi-point linear regression is carried out to determine the significance of the objective function arguments and to pick up the set of particular variables and characteristic quantities (e.g. flow angles) which contribute most to improving the objective function value. This approach enables us to considerably reduce the number of optimized parameters and computational time needed for the optimization process. Variable operating parameters of ORC power systems become an everyday occurrence, which makes the analysis of ORC turbines under variable operating parameters a critical part of the design. Therefore, the paper presents the influence of a single-criterion optimization of the rotor geometry on the performance of the ORC turbine in a wide range of operating conditions. The results of optimization indicate that new geometry provides an approximately 5.5 pp. (percentage point) increase in total-to-static efficiency for the design point, and even larger improvements reaching 7 pp. For higher turbine loads, whereas for lower loads the efficiency improvements are at the level of 1.5–4 pp. •10 kW ORC axial turbine efficiency is increased during optimization.•The most significant parameters are selected by means of multi-point regression.•The effect of optimization on the off-design turbine performance is tested.•Wide range operational characteristics are made showing large efficiency gains.