Research on Internal Damage Identification of Wire Rope Based on Improved VGG Network.

Area of Science:

• Engineering
• Computer Science
• Materials Science

Background:

• Detecting internal damage in wire ropes is challenging.
• Existing methods may lack accuracy and efficiency.

Purpose of the Study:

• To develop an improved VGG model for accurate internal wire rope damage identification.
• To enhance the VGG model's performance through specific modifications.

Main Methods:

• Wire rope damage signals transformed into time-frequency spectrograms using Short-Time Fourier Transform.
• VGG model enhanced with an attention mechanism, batch normalization, and optimized pooling/fully connected layers.
• Joint-sample uniformly distributed cross-entropy used as the loss function.

Main Results:

• The improved VGG model achieved an identification accuracy of up to 98.84% for internal wire rope damage.
• The model demonstrated reduced computational cost and shorter training times.
• Enhanced generalization ability compared to traditional VGG models.

Conclusions:

• The proposed enhanced VGG model effectively identifies internal wire rope damage.
• The modifications significantly improve recognition accuracy, efficiency, and generalization.
• This method offers a robust solution for wire rope integrity monitoring.