Research on an efficient prediction for deformations of thread connections based on deep transfer learning

Area of Science:

• Mechanical Engineering
• Computational Science
• Artificial Intelligence

Background:

• Threaded connections are critical in mechanical assemblies and experience complex deformations under various loads.
• Accurate prediction of these deformations is vital for ensuring structural integrity and safety.
• Existing methods like Finite Element Analysis (FEA) can be computationally intensive for complex scenarios.

Purpose of the Study:

• To develop an efficient method for rapid multi-dimensional deformation prediction in threaded connections under combined loading.
• To leverage deep transfer learning (TL) to improve prediction accuracy and generalization across different bolt types and loading conditions.
• To provide a computationally cost-effective alternative to traditional FEA for engineering design and structural assessment.

Main Methods:

• Development of a simplified static analysis model for deformation prediction under combined loads, validated against FEA.
• Pre-training a deep neural network (DNN) on a large deformation dataset to learn complex load-deformation relationships.
• Application of transfer learning (TL) to adapt the DNN for various bolt types and loading scenarios, using Latin Hypercube Sampling (LHS) for dataset generation.

Main Results:

• The proposed TL method significantly reduces computational costs, requiring only 0.14% of the computation time compared to FEA for complex loading conditions.
• Prediction accuracy reaches up to 96.6% for different bolt types after applying transfer learning.
• The method accurately simulates nonlinear deformations and demonstrates strong generalization capabilities.

Conclusions:

• The developed deep transfer learning approach offers an efficient and accurate solution for predicting threaded connection deformations under combined loading.
• This method provides a substantial computational advantage over FEA, enabling real-time applications in engineering design and structural analysis.
• The study highlights the potential of AI-driven methods for enhancing the safety and efficiency of mechanical component design.