Sub-surface thermal measurement in additive manufacturing via machine learning-enabled high-resolution fiber optic...

Area of Science:

• Materials Science
• Mechanical Engineering
• Optical Sensing

Background:

• Microstructure evolution in additively manufactured parts dictates mechanical properties.
• Layer-wise printing involves complex thermal cycles (re-melting, reheating) impacting microstructure.
• Current analysis relies on time-consuming numerical models due to limited sub-surface temperature measurement.

Purpose of the Study:

• To develop an effective sub-surface temperature measurement technique for additive manufacturing.
• To overcome the spatial resolution limitations of traditional fiber optic sensors.
• To enable detailed study of thermal phenomena during metal 3D printing.

Main Methods:

• Implementation of a machine learning-assisted approach for demodulating chirped-fiber Bragg grating (CFBG) sensor signals.
• Development of a robust sensor embedding technique for laser additive manufacturing.
• Utilizing CFBG sensors for high-resolution thermal distribution mapping.

Main Results:

• Achieved a significant improvement in spatial resolution from millimeter to 28.8 µm.
• Successfully demonstrated the sensor's capability to measure sharp thermal gradients and rapid cooling rates.
• Developed a method to embed fragile sensors with minimal damage and ensured close contact.

Conclusions:

• The machine learning-assisted CFBG sensor offers a powerful tool for real-time sub-surface temperature monitoring in additive manufacturing.
• This technique overcomes previous limitations in spatial resolution and sensor fragility.
• The developed sensor system holds significant potential for advancing the fundamental understanding of metal additive manufacturing physics.