| About this Abstract |
| Meeting |
MS&T24: Materials Science & Technology
|
| Symposium
|
Materials Processing and Fundamental Understanding Based on Machine Learning and Data Informatics
|
| Presentation Title |
Multi-Layer Graded Thermal Barrier Coating Design via Deep Reinforcement Learning |
| Author(s) |
Ningxuan Wen, Hai Xiao, Dongsheng Li, Fei Peng |
| On-Site Speaker (Planned) |
Ningxuan Wen |
| Abstract Scope |
Thermal barrier coatings (TBCs) are commonly used in power generation systems or aerospace for extreme temperature environments protection. Recently, multi-layered, graded TBCs have been proposed that integrate the bonding layer, environmental barrier layer, and thermal barrier layer. The performance of TBCs is highly dependent on matching the properties and thickness of the different layers. The goal of this study is to find the optimal designs of multi-layer, graded TBC with minimum thermal stress and maximum thermal insulation using the reinforcement learning (RL) method. In this study, the design parameters for the coating are material composites and thickness of each layer. The thermal stress and heat transport were modeled using the FEM to feed the data for RL training. After trained, the RL algorithm is capable of automatically generating, evaluating, and improving the TBC design. The effectiveness and validity of the output TBC design is examined using FEM simulation. |