Modeling of a turbine bladerow with stagger angle variation using the multi-fidelity influence superposition method

Journal article

Manufacturing variabilities can significantly affect a turbine bladerow's performance. With the push for turbine's efficiency and reliability more than ever, the understanding of manufacturing variability impacts is sought after. It has been shown that a multi-passage simulation domain is required to capture the interaction among varied blades. This is a challenge for conventional single-passage methods, largely due to the requirement of prohibitive computational resources. The present work introduces an attempt to solve this challenge by combining the multi-fidelity method and the influence superposition approach. The proposed methodology combines the accuracy of a high-fidelity simulation and the speed of a low-fidelity simulation. Therefore, the multi-fidelity predictions tend to be both accurate and fast. The key enabler is that only a small set of configurations is needed to pre-compute the source term. Two representative geometries of a subsonic low-pressure turbine and a transonic high-pressure turbine have been chosen as the test cases. Each test case is subject to two further stagger angle variation patterns, namely alternating and sinusoidal pattern. The low-fidelity method has been shown to be inadequate for the transonic high-pressure turbine test case, owning to its incapability to capture correctly the shockwave formation and the shockwave/wake interaction. On the other hand, the proposed multi-fidelity method has been successful to match qualitatively and quantitatively compared to the direct high-fidelity solution. More interestingly, the multi-fidelity method has a reduced computational overhead of one order of magnitude compared to the direct high-fidelity simulation. With an ability to accurately and efficiently predict the manufacturing variability effects, this method provides a tool for engineers to explore and optimize their blading designs.

Authors: Phan, HM

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: Elsevier
• Journal: Aerospace Science and Technology
• Volume: 121
• Article number: 107318
• Publication date: 2022-01-05
• Acceptance date: 2021-12-30
• DOI: 10.1016/j.ast.2021.107318
• ISSN: 1270-9638
• Language: English
• Keywords: mis-staggering; turbomachinery; multi-fidelity method; FFR; manufacturing variations