Investigations into rapid aeroelastic stability estimation in turbomachines for preliminary design

Abstract

Turbomachines are a vital part of today's energy and transportation industry. They need to perform reliably in order to maximize profit and minimize costs. Fuel efficiency and exhaust gas property demands need to be considered in order to meet ecological concerns. To meet all these criteria, rotor and stator row geometries of modern turbomachines are more and more often designed with automated optimization tools. To date, these preliminary design tools include aerodynamic efficiency and structural durability considerations. However, the aeroelastic stability of turbomachinery stages is traditionally treated very late in the design process after the end of the automated preliminary design cycle. This practice causes very large costs if a redesign is necessary due to aeroelastic problems. To include aeroelastic stability considerations into an automated preliminary design tool chain, methods need to be available to assess the flutter susceptibility of a given geometry configuration. There exists a wide range of possible aeroelastic procedures with varying numerical effort and different degrees of accuracy to judge the aeroelastic stability of a given geometry. In order to find a method to estimate the aeroelastic stability with affordable numerical effort for industrial purposes yet with sufficient accuracy, the development of computational performance and numerical aeroelastic methods over the past years is reviewed. It is proposed that scientific methods for aeroelastic stability estimation from ~20-25 years ago might be suitable to be used as preliminary design tools, today. To investigate the feasibility and accuracy of this proposed approach, three traditional methods for aeroelastic stability assessment covering the affordable range of numerical complexity were considered: • As a baseline approach which resembles the used approach at DLR ~7 years ago, the possibility of automating and using the energy method in combination with a 3D time-linear RANS solver was investigated. Various approaches of reducing the numerical cost of this approach were evaluated. • To gauge the absolute minimum of numerical effort possible to generate an aeroelastic assessment, the validity of a one-equation design rule (reduced angular frequency at aerodynamic design point) used in the middle of the 20 th century was investigated. • Finally, the possibility of using a simpler unsteady aerodynamic method (unsteady potential flow) in combination with the energy method was investigated. To do so, a vortex lattice method was equipped with turbomachinery features to judge its potential to deliver fast and reliable or correctable unsteady aerodynamic results with harmonically deforming blades. It was concluded that the most promising approach to rapid aeroelastic stability analysis in preliminary design was the use of a 3D time-linear RANS solver in combination with suitable reduction techniques. If the computed cases are selected with a suitable algorithm, this approach might be used in preliminary design of turbomachinery stages, starting today.