Heinrich Stüer

Optimization Strategy for a Coupled Design of the Last Stage and the Successive Diffuser in a Low Pressure Steam Turbine

In this study, an effective yet numerically simple approach for a coupled design of the last stage running blade and diffuser is presented. The method applied uses a two-dimensional streamline curvature code combined with a boundary layer solver for the prediction of flow separation within the diffuser. An accurate representation of the diffuser flow is vital for the assessment of the overall performance. Thus, the major influences from the turbine stage on the diffuser flow, i.e., the tip leakage jet and the swirl of the flow, are taken into account. Secondary effects like blade wakes are neglected. The basic capability of the method to correctly represent the flow is demonstrated by a comparison with three-dimensional CFD simulations of a sample configuration. Solid correlation can be found between both cases. For the optimization process, a genetic algorithm is used. Optimization parameters include the blade exit angle and the diffuser contour. The results of the optimization are again scrutinized with the assistance of three-dimensional CFD simulations.

Aerodynamic Mistuning of Structurally Coupled Blades

This paper discusses the impact of aerodynamic mistuning in the context of a structurally coupled blade system. The aerodynamic mistuning assumes a difference in blade loading (between adjacent blades) but no change in eigenfrequency or modeshape. Thus, emphasis is placed on the profile shape impact of transonic rotor blades on aeroelasic damping. An aeroelastic calculation example of the proposed mistuning strategy is present for a coupled blade arrangement. The impact of mistuning compared to the tuned datum profile is discussed. It is shown that there exist significant differences for the two configurations but that the net impact of mistuning is smaller for the coupled system investigated compared to a freestanding blade design.Copyright

Transient Forced Response Analysis of Mistuned Steam Turbine Blades During Start-Up and Coast-Down

It is well-known that the vibrational behavior of a mistuned bladed disk differs strongly from that of a tuned bladed disk. A large number of publications dealing with the dynamics of mistuned bladed disks is available in the literature. The vibrational phenomena analyzed in these publications are either forced vibrations or self-excited flutter vibrations. Nearly all published literature on the forced vibrations of mistuned blades disks considers harmonic, i. e. steady-state, vibrations, whereas the self-excited flutter vibrations are analyzed by the evaluation of the margin against instabilities by means of a modal, or rather than eigenvalue, analysis. The transient forced response of mistuned bladed disk is not analyzed in detail so far. In this paper, a computationally efficient mechanical model of a mistuned bladed disk to compute the transient forced response is presented. This model is based on the well-known Fundamental Model of Mistuning. With this model, the statistics of the transient forced response of a mistuned bladed disk is analyzed and compared to the results of harmonic forced response analysis.Copyright

Validation of a Linearized Navier-Stokes Based Flutter Prediction Tool: Part 1—Numerical Methods

This is the first of two papers describing the validation o a tool chain for flutter prediction. This first paper provides an overview of the numerical methods and their verification. The second paper presents the detailed validation of the tool chain on the basis of experimental data obtained from measurements of an annular cascade sector comprising slightly twisted turbine blades. The tool chain consists of commercial programs for grid generation, structural analysis and the computation of the stead flow, and software developed at DLR for data conversion, aeroe- lastic preprocessing, and solving the time-linearized RANS equations. The time-linearized solver computes the unsteady flow in the frequency domain based on the linearization of the RANS equations about a steady solution which, in the process chain presented here, is provided by the commercial solver. We inves- tigate the issues that arise from using different spatial discretizatiom schemes and turbulence models for the computation of the steady and the time-linearized solutions. Results are presented for Standard Configurations 4, 10 and 11, and a freestanding tur- bine blade from a rear stage of a stationary gas turbine. It is shown that with an appropriate choice of the spatial discretiza- tion parameters good agreement with reference data can be ob- tained.