Abstract:The preparation process and composition design of heavy-section ductile iron are the key factors affecting its fracture toughness. These factors are challenging to address due to the long casting cycle, high cost and complex influencing factors of this type of iron. In this paper, 18 cubic physical simulation test blocks with 400 mm wall thickness were prepared by adjusting the C, Si and Mn contents in heavy-section ductile iron using a homemade physical simulation casting system. Four locations with different cooling rates were selected for each specimen, and 72 specimens with different compositions and cooling times of the heavy-section ductile iron were prepared. Six machine learning-based heavy-section ductile iron fracture toughness predictive models were constructed based on measured data with the C content, Si content, Mn content and cooling rate as input data and the fracture toughness as the output data. The experimental results showed that the constructed bagging model has high accuracy in predicting the fracture toughness of heavy-section ductile iron, with a coefficient of coefficient ( R 2 ) of 0.9990 and a root mean square error (RMSE) of 0.2373.

What problem does this paper attempt to address?

The paper primarily aims to address the issue of predicting the fracture toughness of heavy-section ductile iron. Specifically, the research focuses on the following aspects: 1. **Background**: Heavy-section ductile iron is widely used in fields such as nuclear power, wind energy, nuclear waste storage and transportation, and high-speed railways due to its excellent fracture toughness, tensile strength, ductility, and radiation resistance. However, due to its complex preparation process and composition design, the fracture toughness of this material is influenced by many factors, including a long casting cycle, high cost, and the complexity of influencing factors. 2. **Research Objective**: To overcome the above challenges, this paper aims to study the effects of carbon (C), silicon (Si), and manganese (Mn) content on the fracture toughness of heavy-section ductile iron through physical simulation methods and a self-made physical simulation casting system. Based on these data, machine learning models are constructed to predict the fracture toughness of heavy-section ductile iron. 3. **Experimental Design**: Researchers prepared 18 cubic physical simulation test blocks with different C, Si, and Mn contents, each with a wall thickness of 400 mm. Samples were taken from different positions of each test block to simulate different cooling rates, and 72 heavy-section ductile iron samples with different compositions and cooling times were further prepared. The data collected from the fracture toughness tests of these samples were used to construct the prediction model. 4. **Model Construction**: Based on the collected data, the research constructed six machine learning models (including XGBoost, Support Vector Regression, Multilayer Perceptron Regression, Gaussian Process Regression, Bagging, and Random Forest) to predict the fracture toughness of heavy-section ductile iron. Among them, the XGBoost model showed high accuracy in predicting the fracture toughness of heavy-section ductile iron (coefficient of determination \(R^2\) of 0.9990 and root mean square error RMSE of 0.2373). In summary, this study effectively addresses the issue of predicting the fracture toughness of heavy-section ductile iron through physical simulation experiments combined with machine learning methods, providing strong support for optimizing the composition design and preparation process of heavy-section ductile iron in actual production.

Prediction of heavy-section ductile iron fracture toughness based on machine learning

A Novel Machine-Learning Based Framework for Calibrating Micromechanical Fracture Model of Ultra-Low Cycle Fatigue in Steel Structures

Determination of ductile fracture properties of 16MND5 steels under varying constraint levels using machine learning methods

Machine Learning Models for Predicting Rock Fracture Toughness at Different Temperature Conditions

Prediction of fracture toughness of concrete using the machine learning approach

Predicting Solidification Cracking Susceptibility of Stainless Steels Using Machine Learning

A Self-Training-based Approach for Aluminum Alloy Casting Quality Prediction

Prediction of Fatigue Crack Damage Using In-Situ Scanning Electron Microscopy and Machine Learning

Machine learning (ML) based models for predicting the ultimate bending moment resistance of high strength steel welded I-section beam under bending

Numerical Simulation and Machine Learning Prediction of the Direct Chill Casting Process of Large-Scale Aluminum Ingots

Composition design and optimization of Fe–C–Mn–Al steel based on machine learning

Bending Strength Prediction of the Cu-Sn Alloy Through a Visual Quantization Model Integrated With Microstructure Characterization and Machine Learning

Fracture prediction in transverse fillet welded joints of high strength structural steel

A multi-barrier model assisted CAFE method for predicting ductile-to-brittle transition with application to a low-carbon ultrahigh-strength steel

Hot Ductility Prediction Model of Cast Steel with Low-Temperature Transformed Structure during Continuous Casting

Fracture prediction of Fe-SMA under monotonic and low cycle fatigue loading

Fracture prediction for metal sheet deformation under different stress states with uncoupled ductile fracture criteria

Coupling of Solidification and Heat Transfer Simulations with Interpretable Machine Learning Algorithms to Predict Transverse Cracks in Continuous Casting of Steel

Machine learning techniques for prediction of failure loads and fracture characteristics of high and ultra-high strength concrete beams

Study on Thermal Simulation of Solidification in Heavy Section Ductile Iron

Machine Learning Algorithms for Prediction and Characterization of Cohesive Zone Parameters for Mixed-Mode Fracture