Exploring Machine Learning-Based Fault Monitoring for Polymer-Based Additive Manufacturing: Challenges and...

Area of Science:

• Additive Manufacturing
• Machine Learning
• Mechanical Engineering

Background:

• Three-dimensional printing, or additive manufacturing (AM), is vital for rapid prototyping.
• Monitoring AM processes enhances print quality, minimizing material waste and cost.
• Machine learning (ML) is increasingly used for automating fault detection in AM.

Purpose of the Study:

• To survey recent research on ML-based mechanical fault monitoring systems in FDM 3D printing.
• To identify common ML algorithms and sensor data used for fault diagnosis in FDM.
• To highlight underexplored research areas, such as SLM-based mechanical fault monitoring.

Main Methods:

• Reviewing studies that apply ML algorithms to sensor data from 3D printers.
• Analyzing data from sensors measuring attitude, acoustic emission, acceleration, and vibration.
• Focusing on mechanical fault analysis in fused deposition modeling (FDM).

Main Results:

• ML algorithms effectively diagnose and identify mechanical faults in FDM printers.
• Sensor data, including vibration and acoustic emission, are crucial for fault detection.
• Existing research primarily focuses on FDM, with limited exploration of other AM methods.

Conclusions:

• ML shows significant promise for enhancing the reliability and efficiency of FDM 3D printing.
• Further research is needed in areas like SLM-based mechanical fault monitoring.
• Integrating ML for real-time fault detection can optimize AM processes and reduce operational costs.