| Autor: |
Pal, Apurba, Kundu, Tamal, Datta, Aloke Kumar |
| Zdroj: |
Asian Journal of Civil Engineering; December 2023, Vol. 24 Issue: 8 p3001-3008, 8p |
| Abstrakt: |
The sensor-based Acoustic Emission (AE) technique is widely used for detecting structural damages to prevent catastrophic failure in real time. However, this approach can be further improved by incorporating soft computing approaches such as Artificial Neural Network (ANN). An ANN-based AE approach for damage detection can be a more energy-efficient alternative to existing Non-destructive tests (NDT) for huge railway systems throughout the globe. This study aims to develop an energy-efficient soft computing-based approach using surface wave propagation of AE signal simulation and ANN for the smart diagnosis of rail defects. The study uses Pencil Lead Break (PLB) as an artificial AE source to generate an AE signal equivalent to damage in microscale at various distances over the Top Flange (TF) along the rail section. A single AE sensor is mounted in the middle of the web of the rail section to obtain the simulated AE signal. The AE parameters viz. Counts, Energy, Amplitude, etc. obtained from the AE signals are used to develop an ANN model, where defect location is considered as an output of the model for localization of AE source/ defect in the rail section. The developed ANN model is tested for predicting the defect location and found to be efficient, with an error of less than 1%. This model can be a potential energy-efficient approach for sensor-based smart diagnosis of defects in the rail section in real time. |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
Full text from SpringerLink