Edge computing-based proactive control method for industrial product manufacturing quality prediction.

Area of Science:

• Manufacturing Engineering
• Artificial Intelligence
• Data Science

Background:

• Intelligent manufacturing integrates advanced technologies like big data and AI for enhanced industrial processes.
• Product quality prediction is a key application, but traditional models struggle with imbalanced datasets common in manufacturing.
• Existing cloud computing models face bandwidth and resource limitations in industrial settings.

Purpose of the Study:

• To propose an effective method for product quality prediction using imbalanced data.
• To address the limitations of traditional cloud computing in industrial manufacturing environments.
• To validate the proposed method's practicality and effectiveness in a real-world production line.

Main Methods:

• A Synthetic Minority Over-sampling Technique (SMOTE) combined with XGBoost (SMOTE-XGboost) for imbalanced data classification.
• Joint optimization of hyperparameters for the SMOTE-XGboost model.
• Integration of edge computing technology to overcome cloud computing limitations.

Main Results:

• The proposed SMOTE-XGboost method demonstrates superior performance in predicting brake disc quality compared to other classification methods.
• Edge computing effectively addresses bandwidth and resource constraints in industrial applications.
• The method's practicality and effectiveness were validated through a case study.

Conclusions:

• The SMOTE-XGboost method offers a robust solution for imbalanced data classification in product quality prediction.
• Edge computing integration enhances the feasibility of advanced predictive models in intelligent manufacturing.
• This approach significantly improves quality prediction accuracy in industrial settings like brake disc production.