Advanced Turbine Aerodynamic Design Utilizing a Full Stage CFD

Gas turbines for power generation are required to operate more efficiently than ever before for both economic and environmental reasons. Because of this situation, an advanced multistage turbine design and optimization system is required to improve upon existing turbine designs where viscous CFD codes had already been applied on a single row or single stages basis. An advanced CFD code for multistage design applications has been developed at Mitsubishi Heavy Industries (MHI) and has been applied to the redesign of a four stage single shaft turbine. The front 3 stages of the turbine are highly cooled using about 20% cooling air. The outstanding performance of this redesigned turbine has been demonstrated at MHI’s engine test facility. This paper focuses on the customization of the Denton code [5] for industrial usage, the validation of the customized code employing experimental data, and finally the use of the code in executing a successful redesign. Code development and validation are discussed in terms of prediction accuracy for the basic aerodynamic design parameters such as exit flow angle and cascade losses. Through-flow design parameters such as pressure ratio and reaction of each stage are also addressed. Especially important in modern high temperature turbines is the location and distribution of cooling and leakage air being introduced into the main gas-path. The proper treatment of these flows is very important because of the mixing losses and the temperature migration downstream. These important considerations in any analysis approach are discussed and it is shown how they are treated in the customized CFD code. Consistency between the customized CFD code and other parts of the existing aerodynamic design procedure are carefully examined. This is important because aerodynamic parameters have different modeling fidelities in the different parts of the design system. Computer execution times are a very important consideration when utilizing advanced CFD codes. This issue is addressed from the perspective of an industrial design organization. In validating the customized code, special attention was placed on tip clearance leakage flow behavior and seal air migration from the hub wall. Local changes of total pressure and temperature distributions affect the local velocity triangles and local static pressure distributions on the airfoil and end-wall surfaces. Airfoil section geometry and three-dimensional stacking to maximize the turbine efficiency are also considered and discussed. The validated code was subsequently used to execute a redesign of a large frame industrial turbine. This is discussed in some detail. The redesigned turbine has completed full scale engine testing and has been shown to have met all design goals. The CFD predictions are compared with special measurements taken in the engine such as the inter-stage span-wise total pressure and temperature distributions as well as the efficiency trend versus engine load. These comparisons prove the capability of the advanced multistage CFD code.