A Method for Predicting the Creep Rupture Life of Small-Sample Materials Based on Parametric Models and Machine...

Area of Science:

• Materials Science
• Mechanical Engineering
• Computational Science

Background:

• Creep rupture life prediction models vary in applicability and accuracy.
• Small-sample materials pose challenges for traditional prediction methods.
• Accurate creep life prediction is crucial for material design and safety.

Purpose of the Study:

• To develop an improved method for predicting creep rupture life.
• To enhance the accuracy and applicability of predictions for small-sample materials.
• To quantify the influence of input variables on creep rupture life.

Main Methods:

• Compared time-temperature parametric models and machine learning models.
• Developed a hybrid model combining parametric and machine learning approaches.
• Expanded training data using parametric equations to improve machine learning accuracy.

Main Results:

• The proposed hybrid method demonstrated superior prediction accuracy.
• Quantitative comparison using RMSE, MAPE, and R² confirmed model performance.
• The machine learning component effectively quantified the influence of input variables.

Conclusions:

• The novel hybrid method significantly improves creep rupture life prediction accuracy and applicability.
• Data expansion via parametric models overcomes limitations of small-sample datasets.
• This approach offers a robust solution for predicting material behavior under creep conditions.