Research on the Application of Ductile Iron in High-Performance Automobile Parts

Abstract

Background: Ductile iron, known for its better mechanical properties, is essential in the production of high-performance automobile components. As the automotive industry strives for lightweight yet durable elements, comprehending the relationship between ductile iron material properties and part effectiveness is critical. Predicting the efficiency of these components using their properties can dramatically improve product development and manufacturing procedures.

Objectives: The primary objective of this research is to create the Ductile Iron Part Performance Predictor (DIP3), a sophisticated technique that can predict the performance rating (high, medium, or low) of ductile iron automobile parts. Using machine learning ensemble methods such as Bagging, Boosting, and Stacking, the model intends to offer precise and trustworthy performance predictions.

Methods: The DIP3 model uses a three-step ensemble technique. The first step uses Bagging with the J48 decision tree classifier to create multiple sub-models. The second step improves predictions by combining AdaBoostM1 with RandomForest as the base classifier, which improves on the initial predictions from Bagging. Finally, in the third step, the algorithm uses a SimpleLogistic meta-classifier for Stacking, which combines the results of Bagging and Boosting to make more accurate predictions. Mean/mode imputation for missing values, label encoding for categorical variables, and Min-Max normalization for numerical features are all used for data preprocessing. Information Gain is employed in feature selection to find the most important predictors of part performance.

Results and Conclusion: The DIP3 model was assessed using a variety of performance metrics, comprising accuracy, precision, recall, F1-score, and MCC. The model had a high level of predictive performance. These results prove the resilience and dependability of the DIP3 model in predicting the performance rating of ductile iron parts. The DIP3 model is a reliable and efficient solution for predicting the performance of high-performance ductile iron automobile parts. By combining Bagging, Boosting, and Stacking approaches, the algorithm offers precise, trustworthy, and interpretable predictions, which can help manufacturers improve their design and manufacturing processes.