Abstract <p>Ultrasonic testing of thick-walled polyethylene pipes is challenged by energy loss, resulting in weak echo signals from deep defects. To enhance the detection of these weak signals, a time-frequency energy concentration method is presented. The fractional adaptive superlet transform combines multiple wavelet transform results with distinct bandwidths through geometric averaging, providing superior time-frequency analysis capabilities than single wavelet transforms. However, its time-frequency representation exhibits the issue of instantaneous frequency deviation. The proposed method addresses the issue via instantaneous frequency-embedding, leading to improved accuracy in instantaneous frequency estimation. Numerical signal analysis reveals higher accuracy in instantaneous frequency estimation using this method, compared to other time-frequency processing methods. When applied to detecting deep defects in thick-walled polyethylene pipes, the method shows an 18.9% increase in weak signal enhancement capability compared to the continuous wavelet transform. Finally, the results demonstrate the method’s accuracy in clarifying instantaneous frequency changes and enhancing instantaneous amplitudes of weak signals, offering a promising approach for the detection of deep defects in thick-walled polyethylene pipes.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Enhancing Deep Hole Defects Visibility in Ultrasonic Detection for Thick-Walled Polyethylene Pipes via Time-Frequency Energy Concentration

  • Chaolei Chen,
  • Huaishu Hou,
  • Shiwei Zhang,
  • Mingxu Su,
  • Zhifan Zhao,
  • Chaofei Jiao

摘要

Abstract

Ultrasonic testing of thick-walled polyethylene pipes is challenged by energy loss, resulting in weak echo signals from deep defects. To enhance the detection of these weak signals, a time-frequency energy concentration method is presented. The fractional adaptive superlet transform combines multiple wavelet transform results with distinct bandwidths through geometric averaging, providing superior time-frequency analysis capabilities than single wavelet transforms. However, its time-frequency representation exhibits the issue of instantaneous frequency deviation. The proposed method addresses the issue via instantaneous frequency-embedding, leading to improved accuracy in instantaneous frequency estimation. Numerical signal analysis reveals higher accuracy in instantaneous frequency estimation using this method, compared to other time-frequency processing methods. When applied to detecting deep defects in thick-walled polyethylene pipes, the method shows an 18.9% increase in weak signal enhancement capability compared to the continuous wavelet transform. Finally, the results demonstrate the method’s accuracy in clarifying instantaneous frequency changes and enhancing instantaneous amplitudes of weak signals, offering a promising approach for the detection of deep defects in thick-walled polyethylene pipes.