Skip to main content

Aeroelastic design of turbine blades ii (ADTURBII)

Deliverables

Current developments in Forced Response (including the previous programme: ADTurB) focus on vibration modes excited at wake passing frequency. Low Engine Order (LEO) excitation is caused by small differences in the flow around the annulus. It is therefore much more difficult to predict because it depends upon knowledge of the sources of the variation and sensitivity of the vibration response to these small changes. For vibration at wake passing frequency it is possible to model a small number of aerofoil passages because the flow distortion is assumed to be identical for all nozzle guide vanes. An analysis method for low engine order forced response would typically need to model the whole annulus and explicitly include the vane-to-vane variation that would be representative of a real machine. Unique experiments were undertaken utilising a slightly modified existing research rotor and a continuous flow rig facility at DLR to include: - Vane to vane throat width variation in the set. - Variability in cooling flow ejection between vanes in set, - blockage simulation. Data gathered included detailed measurements of the flow field under LEO conditions using pressure transducers & Laser 2 Focus device. Each blades amplitude was recorded simultaneously via strain gauges. Supporting Aeroelastic analysis was carried out for correlation with fluid & response measurements. It is expected that by improved understanding of the excitation sources & validation of prediction tools significant reductions can be made in HCF events related to LEO.
The actual design of rotors assumes the cyclic symmetry of the part, that means, for instance, that all the blades are identical. In reality, local variations in material properties, manufacturing and ageing of the part break the cyclic symmetry. Among others, the result is that the vibration properties of the blades differ from one another. This phenomenon, called mistune, can cause an amplification of the blades response to mechanic and aerodynamic excitations. The experiments carried out during ADTURB II are very valuable because the test conditions have been carefully characterised. These results will be used to validate new design methods currently under development aiming at reducing the sensitivity of the rotor design to mistuning. These results will lead to improve the reliability of the engines and the safety of the flights. Turbomeca engines are used on airplanes as well as helicopters for both military and civil application. Moreover, improving the robustness of the design to production tolerances will reduce the production costs.
Experiences from previous projects, where many different partners were involved with different cultural and technical backgrounds, have highlighted the need to ensure common definitions and evaluations are used or carried out. This is to be enforced by creating a specific task for this purpose. A previous Framework 4 project (ADTurB, ref. BRPR-CT95-0124) established a common database including the definition of neutral data formats, and interrogation / post-processing facilities. This database contains all of the data generated during the project, securely stored electronically with password protection but is accessible to all partners via the Internet. This project built on the results and experiences within the ADTurB project and will benefit from this approach and from the investment that has already been made. It has therefore been decided to establish a data integration work package for ADTurBII. This database will be extended with the experimental and numerical results of the present project. The choice of data to be stored was based on a re-evaluation of requirements & lessons learnt in ADTurB.
The main goal in this task was to study friction damper behaviour on a simplified structure under rotation in a vacuum. The test assembly made up of 2 engine style rotor blades with a range of under platform dampers between them. This arrangement removed aerodynamic effects and allowed the friction effect on its own to be quantified. Excitation was applied directly to the blades via piezoelectric material on the blade platform, ensuring sliding of the dampers. Response was measured by means of strain gauges on the blades. The experimental set up allowed the efficiency of the friction damper element to be determined for any blade vibration mode and hence its effect on blade amplitude can be predicted. It was also possible to introduce a phase difference between the adjacent blades, simulating engine order variation and to independently vary the damper normal force by changing rotation speed.
This test series uses a modified cascade developed originally for the first ADTurB project and was used to evaluate the relative influence of aerodynamic & mechanical mistuning on the aerodynamic damping and forced response. Known displacements (torsion & bending) are applied and aerodynamic effects measured. The effects investigated were: - Frequency mistune - blade movement due to forced response, and steady/unsteady pressure measurements. - Mode shape mistuning - influence of mode shape variability (amount of torsion/bending per blade) on the aero damping. - The experimental work was carried out at EPFL