Related Experiment Video
Updated: Jul 10, 2025
09:12
A Method for Studying the Temperature Dependence of Dynamic Fracture and Fragmentation
Published on: June 28, 2015
8.6K
Data-driven based fracture prediction of notched components
Hossein Talebi1, Bahador Bahrami1, Mohammad Daneshfar1
1Fatigue and Fracture Research Laboratory, Center of Excellence in Experimental Solid Mechanics and Dynamics, School of Mechanical Engineering, Iran University of Science and Technology,Narmak 16846, Tehran, Iran.
Summary
A data-driven approach accurately predicts notched component fracture load using machine learning. Gaussian process regression achieved 92% accuracy, outperforming other models for structural integrity assessment.
Area of Science:
- Materials Science and Engineering
- Computational Mechanics
- Data Science
Background:
- Predicting fracture load in notched components is crucial for structural integrity.
- Existing methods often struggle with complex geometries and mixed-mode loading conditions.
- A data-driven approach offers a promising alternative for accurate fracture load prediction.
Purpose of the Study:
- To develop and evaluate machine learning models for predicting the fracture load of notched brittle components.
- To identify key features influencing fracture behavior under mixed-mode I/II loading.
- To compare the performance of Gaussian process regression, decision tree ensemble, and artificial neural networks.
Main Methods:
- Collected and pre-processed over 1500 fracture test data points from literature.
- Selected six critical features using Neighbourhood Component Analysis (NCA).
- Trained and optimized Gaussian process regression (GPR), decision tree ensemble, and artificial neural network (ANN) models using Bayesian optimization.
Main Results:
- GPR achieved 92% accuracy, decision tree ensemble 89%, and ANN 88% in predicting fracture load on unseen data.
- GPR demonstrated superior performance due to its ability to model nonlinear relationships and provide uncertainty estimates.
- Models successfully predicted fracture load for VO-shaped notches not included in training.
Conclusions:
- Data-driven machine learning models show high potential for accurate fracture load prediction in notched components.
- Gaussian process regression is particularly effective for this task, offering robust predictions and uncertainty quantification.
- The developed approach can enhance structural integrity assessments and design optimization.
Related Concept Videos
01:25Yield Criteria for Ductile Materials under Plane Stress
165
In designing structural elements and machine parts using ductile materials, it is crucial to ensure that these components withstand applied stresses without yielding. Yielding is initially determined through a tensile test, which evaluates the material's response to uniaxial stress. However, tensile stress is insufficient when components face biaxial or plane stress conditions This condition requires advanced criteria to predict failure.
The Maximum Shearing Stress Criterion, also known as...
The Maximum Shearing Stress Criterion, also known as...
165
01:21Fatigue
186
Fatigue occurs when materials rupture under repeated or fluctuating loads, even at stress levels far below their static breaking strength. It typically results in brittle failure, even for ductile materials. It is a critical consideration in designing machines and structural components subjected to repetitive or varying loads. The nature of these loadings can range from fluctuating loads like unbalanced pump impellers causing vibrations to repeatedly bending a thin steel rod wire back and forth...
186
01:24Stress-Strain Diagram - Brittle Materials
2.5K
Brittle materials, including glass, cast iron, and stone, exhibit unique characteristics. They fracture without considerable change in their elongation rate, indicating that their breaking and ultimate strength are equivalent. Such materials also show lower strain levels at the point of rupture. The failure in brittle materials predominantly results from normal stresses, as evidenced by the rupture created along a surface perpendicular to the applied load. These materials do not display...
2.5K
01:18Mechanical Characteristics of Steel
580
The mechanical characteristics of steel are assessed through various tests that evaluate its strength, toughness, and flexibility. These tests include tension, torsion, impact, bending, and hardness assessments, each providing crucial information about steel's suitability for specific applications.
The tension test is fundamental for determining tensile strength. In this test, a steel specimen is stretched using a gripping device until it breaks. The data collected during this test are used...
The tension test is fundamental for determining tensile strength. In this test, a steel specimen is stretched using a gripping device until it breaks. The data collected during this test are used...
580
01:19Members Made of Elastoplastic Material
99
The behavior of elastoplastic materials under bending stresses, particularly in structural members with rectangular cross-sections, is crucial for predicting material responses and understanding failure modes. Initially, when a bending moment is applied, the stress distribution across the section follows Hooke's Law and is linear and elastic. This distribution means the stress increases from the neutral axis to the maximum at the outer fibers, up to the elastic limit.
As the bending moment...
As the bending moment...
99
01:14Plastic Deformations
89
It is essential to understand how structural members behave under plastic deformation when the bending stress exceeds the material's yield strength. This state of deformation permanently alters the shape of the member, in contrast to the linear elastic behavior observed before yielding. The strain at any point in the member is expressed in terms of maximum strain. Notably, the neutral axis, which coincides with the centroid during elastic bending, shifts away from the centroid under plastic...
89