A spatial–temporal analysis approach for flutter predictions using decoupled and fully-coupled methods

Abstract

Accurate and efficient numerical predictions of flutter is essential for the turbomachinery industry. The decoupled method has been the main workhorse of the industry for decades thanks to its low computational resources required. However, there arises the need for an innovative propulsion architecture and a lightweight material, which hinders the accuracy and effectiveness of the decoupled method. Hence, the use of fully-coupled methods for flutter predictions has been steadily increasing. Despite more demanding computational resources needed to perform the fully-coupled simulations, the large amount of data generated in the time-domain is not often used efficiently. Typically the aeroelastic stability is judged in a straightforward manner using the logarithmic decrement. Although this global parameter is useful to compare the aeroelastic stability among configurations, it yields almost no meaningful understanding of the physical vibration mechanisms. In the present paper, a spatial–temporal analysis approach based on the Hilbert transform is proposed. It can work well with the non-linear non-stationary data typically generated from the time-domain fully-coupled solutions. In addition, the Hilbert transform-based analysis approach is designed to be consistent with the conventional energy method when the newly developed method is applied to the decoupled simulation data. The Hilbert transform-based approach is applied to three computational examples, each with a particular aeroelastic flow mechanisms of interest (i.e. tip clearance, laminar separation bubble, stall). The new approach has been shown to effectively elucidate the associated vibration mechanisms using the decomposition method in space and time.