In the present work, Proper Orthogonal Decomposition (POD) has been applied to a large dataset obtained in two Low Pressure Turbine (LPT) cascades in order to identify the most influencing parameters that affect the loss generation. To this purpose, the dataset is constituted by the total pressure loss coefficient distribution in the pitch-wise direction for different inflow conditions. The experiments have been conducted under realistic unsteady inflow varying the Reynolds number, the reduced frequency and the flow coefficient in order to cover a database that is large enough to provide qualitative information to the designers. A moving bar system has been used for the wake generation, allowing the variation of the bar speed and bar count independently. The two cascades in analysis are characterized by the same axial chord but different pitch-to-chord ratio and different flow angles, hence two different values of the Zweifel number. The total pressure loss coefficients have been measured using the recently developed procedure that requires the simultaneous acquisition of the signal of two Kiel probes located upstream and downstream of the cascades. This method provides an accurate evaluation of the losses due to the wake distortion and dilation processes into the cascade passage and to the wake-boundary layer interaction process.

The novel POD-based procedure is applied to the total pressure loss coefficients that have been measured for the different inflow conditions. The POD mode distribution shows the spatial region where losses occur, the POD eigenvectors provide how such losses vary for the different design conditions and the POD eigenvalues provide the rank of the approximation. In particular, the first POD mode alone provides a satisfactory approximation of the losses generated in the blade wake region for all cases under analysis. Conversely, in order to accurately reproduce the losses that occur in the potential flow region, the number of necessary POD modes is about a quarter of the test conditions. Thus, the low-rank representation of the loss distribution may predict the losses in a wider design space and it allows a rapid view of the spatial region where losses are produced, as well as a clear identification of the parameters that mostly influence them. This procedure may offer a new tool for designers to optimize the unsteady operation of low pressure turbine blades when a large number of parameters are considered in the definition of the design space. Furthermore, since this analysis shows that the rank of the problem is much lower than the test conditions, the paper proposes a further method to reduce the number of measurement tests a priori.


Low Rank Education of Cascade Loss Sensitivity to Unsteady Parameters by Proper Orthogonal Decomposition

Division / Sponsor:
International Gas Turbine Institute (IGTI)

Paper Proceedings: https://asmedigitalcollection.asme.org/GT/GT2020/volume/84058

Turbo Expo - Virtual Conference

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.