Speaker
Description
Abradable coatings such as AlSi-PES and AlSi-hBN enable small tip clearances in aeroengine axial compressors by reducing the severity of blade-casing interactions. However, as these abradables are typically plasma sprayed onto the internal surfaces of the casing, the precise makeup and properties of every spray batch are not well defined. This poses significant challenges regarding the simulation of blade-casing interactions and design of new systems, therefore a method for determining abradable properties is needed to better guide these processes.
An inverse analysis method utilizing both particle swarm optimisation and an Artificial Neural Network (ANN) has been created to predict the homogenised mechanical properties of an AlSi-PES abradable based on its Rockwell Hardness. Following this, the homogenised properties are set as target values in a second optimization process where the abradable constituent material properties and microstructural information such as PES content and void volume fraction are determined via a tensile test of an AlSi-PES RVE. The newly determined homogenised properties are to be used during blade-casing interaction simulations, allowing for the system behaviour following an interaction with a specific abradable to be investigated. Furthermore, the microstructural information can be used to better predict the specific abradable failure mechanisms.
This top-down-approach enables a target Rockwell Hardness to be used to determine the homogenised and constituent material properties as well as basic microstructural information of an AlSi-PES abradable. This information can then be used to better inform simulations of blade-casing interactions and the design process.
| Speaker Country | United Kingdom |
|---|