Low-Cycle Fatigue Life Prediction of Titanium-Based Intermetallic Alloys Using Machine Learning and Finite Element...

Area of Science:

• Materials Science
• Mechanical Engineering
• Computational Materials Science

Background:

• Ti2AlNb-based alloys are critical for high-temperature structural applications.
• Understanding low-cycle fatigue (LCF) is essential for component reliability.
• Existing predictive models require validation and improvement for these advanced alloys.

Purpose of the Study:

• To explore the LCF behavior of Ti2AlNb alloy components at high temperatures.
• To validate fatigue life simulations using Seeger's theory and Lemaitre's model against experimental data.
• To evaluate machine learning algorithms (LSTM, ELM, PLS) for predicting LCF life.

Main Methods:

• Application of Seeger's fatigue life theory and an improved Lemaitre damage evolution model.
• High-temperature experimental fatigue life testing for validation.
• Comparative analysis of Long Short-Term Memory (LSTM), Extreme Learning Machine (ELM), and Partial Least Squares (PLS) algorithms.

Main Results:

• Experimental fatigue tests confirmed the accuracy of the simulation models.
• ELM showed superior performance in predicting high-temperature LCF life compared to LSTM and PLS.
• The study provides a validated and efficient predictive framework for Ti2AlNb alloys.

Conclusions:

• Seeger's theory and Lemaitre's model, when validated experimentally, accurately predict LCF life in Ti2AlNb alloys.
• Extreme Learning Machine (ELM) offers a robust and practical approach for predicting the LCF life of these components.
• This research establishes an efficient methodology for assessing the fatigue durability of Ti2AlNb-based structural materials.