Abstract Scope |
Evaluating the service conditions and the corresponding strain at high-pressure turbine blades is crucial for the safe service and maintenance of aircraft engines. However, due to the harsh environment in turbines and the complex geometry of blades, it is difficult to directly monitor the variation of their service temperatures, stresses and strains. In this work, an approach to predicting the equivalent service conditions and the local strain of directionally solidified superalloys and turbine blades was developed by integrating high-throughput creep tests and machine-learning tools. A large amount of experimental data was obtained using the flat specimens with the continuously variable cross-section and digital image correlation technique. Then, the quantitative relationship between temperature, stress, strain, time and the essential microstructure parameters was established under the help of machine learning models. The established machine learning models were then employed to predict the service conditions and the corresponding strain of a directionally solidified superalloy and a turbine blade. Finally, the applicability and limitations of this method were discussed. The development of this method provides guidance for the service evaluation of turbine blades. |