Temperature Hotspot Detection on Printed Circuit Boards (PCBs) Using Ultrasonic Guided Waves-A Machine Learning...

Area of Science:

• Materials Science
• Electrical Engineering
• Acoustics

Background:

• Traditional temperature monitoring of printed circuit boards (PCBs) using numerous sensors can be intrusive and affect electronic component operation.
• Developing non-invasive, high-resolution temperature monitoring is critical for advanced electronics, especially in aerospace and automotive sectors.

Purpose of the Study:

• To investigate the use of ultrasonic guided waves with Piezoelectric Wafer Active Sensors (PWAS) for non-invasive PCB temperature hotspot detection.
• To evaluate the effectiveness of machine learning algorithms in identifying hotspots using a minimal sensor configuration.

Main Methods:

• Simulated wave propagation through a simplified PCB model using COMSOL multiphysics under varying temperature conditions.
• Applied machine learning algorithms to analyze sensor data for hotspot identification at component locations.
• Systematically varied the number and configuration of Piezoelectric Wafer Active Sensors (PWAS).

Main Results:

• Achieved a high accuracy of 97.6% for hotspot detection using four PWAS.
• Demonstrated a reduced accuracy of 88.1% with a single sensor in a pulse-echo configuration.
• The methodology provides sufficient spatial resolution for effective hotspot identification.

Conclusions:

• Ultrasonic guided waves offer a non-invasive and efficient solution for high-resolution PCB temperature monitoring.
• This approach enhances condition-monitoring technologies essential for the reliability and safety of electronic systems in electrified industries.
• The proposed method supports the advancement of electronics in demanding sectors like aerospace and automotive.