Design a cooling pillow to support a high-speed supercritical CO2 turbine shaft

Abstract

• High speed supercritical CO 2 turbine rotor shaft cooling zone. • Modeling a cooling pillow to cool the shaft and to provide dynamic support. • Design charts and figures for the cooling zone. • Rotor dynamic analysis through the Campbell diagram. • Trade-off analysis in designing the cooling zone. Supercritical CO 2 cycles offer a new thermal power generation paradigm but their commercial deployment has been slow due to significant turbine design challenges. Rotor bearings and seal design are identified as the critical challenges due to the high temperatures, high pressures, and high rotational speeds, a combination of extreme values not encountered in any other past turbine applications. This paper presents a new supercritical CO 2 turbine configuration that addresses these challenges by introducing a rotor shaft cooling zone to control the temperatures encountered by the seals and the bearings. This makes it possible to use proven conventional seal and bearing technology. The salient feature is referred to as a ‘cooling pillow’ that cools the rotor shaft to a tolerable seal temperature by circulating supercritical CO 2 fluid through the annulus and at the same time provides dynamic support to the rotating shaft against vibration. Turbulent heat transfer and frictional heat generation correlations for the annuli have been employed to evaluate the cooling performance and the simplified Reynolds equation for hydrodynamic fluid film lubrication has been employed to evaluate the dynamic performance. Design charts and figures are presented using non-dimensional parameters to help with the design of supercritical CO 2 turbines at different temperatures and sizes. The COMSOL rotor-bearing system simulator consisting of bearing and disks with the main shaft has been employed to produce Campbell plots and carry out a parametric study of rotor vibration frequencies varying with the cooling zone design. The cooling capacity is found to be significantly higher than the cooling load required. Moreover, the convection heat transfer coefficient and dimensionless temperature distribution are independent of radial clearance and this offers trade-offs in terms of controlling rotor vibrations. Through the pursuit of such trade-offs, it is demonstrated that dynamic performance can be improved by up to 55% and 60% for the 0.5 MW and 10 MW turbine sizes, respectively.