Extracting Lamb wave vibrating modes with convolutional neural network.

Area of Science:

• Micro-acoustics
• Wave physics
• Machine learning applications

Background:

• Micro-acoustic devices like SAW and BAW are crucial for IoT and mobile communications.
• Increasing data speeds necessitate higher operating frequencies, driving the development of Lamb wave devices.
• Lamb wave devices exhibit complex vibration modes, posing challenges for traditional analysis methods.

Purpose of the Study:

• To develop an automated method for identifying and extracting complex Lamb wave modes.
• To overcome the limitations of existing methods used for SAW and BAW devices.
• To enhance the applicability and performance of Lamb wave devices in high-frequency applications.

Main Methods:

• A machine learning approach utilizing convolutional neural networks (CNNs) was employed.
• The CNN model was trained to recognize and extract specific Lamb wave modes.
• A pre-trained model was utilized for efficient identification of vibration modes.

Main Results:

• The proposed method successfully identified and extracted the first two anti-symmetric and symmetric modes of Lamb waves.
• The technique demonstrated effectiveness in varisized plate structures.
• The machine learning approach proved adept at handling the complex vibrational patterns of Lamb waves.

Conclusions:

• The developed machine learning method provides an effective solution for automatic Lamb wave mode extraction.
• This approach can be extended to other micro-acoustic devices and wave types.
• The method is expected to significantly advance the practical application of Lamb wave devices.