Author Affiliations
School of Control Science and Engineering, Shandong University, Jinan, Shandong 250061, Chinashow less
Fig. 1. New FBG vibration sensor. (a) I-typed section; (b) structure diagram; (c) physical drawing
Fig. 2. Modal simulation of the sensor. (a) Displacement nephogram of the first order modal; (b) strain nephogram of the first order modal
Fig. 3. Strain distribution. (a) Strain distribution of black curve position with length of 20.0 mm; (b) strain distribution at the grating paste position
Fig. 4. Sensor performance simulation. (a) Frequency response at acceleration of 2.0g; (b) accelerated response at vibration frequency of 300 Hz
Fig. 5. Testing system of the sensor
Fig. 6. Time-frequency response curves of the senor. (a) Time domain at 90 Hz; (b) frequency domain at 90 Hz; (c) time domain at 350 Hz; (d) frequency domain at 350 Hz
Fig. 7. Amplitude frequency response curve of sensor
Fig. 8. Spectra of short grating at different frequencies. (a) 90 Hz; (b) 350 Hz
Fig. 9. Acceleration response repeatability curve
Fig. 10. Anti-interference characteristic curve
Fig. 11. Structural damage simulation of I-beam
Fig. 12. Schematic of I-beam working condition
Fig. 13. Time-frequency domain response characteristic of the senor. (a)Time domain response of health status; (b) frequency domain response of health status; (c) time domain response of one crack; (d) frequency domain response of one crack ; (e) time domain response of two cracks; (f) frequency domain response of two cracks
Fig. 14. Energy ratio of three operating frequency bands. (a) Health status; (b) one crack; (c) two cracks
Fig. 15. Wavelet packet energy ratio characteristic
Fig. 16. Identification and comparison results between ELM and BP
Damage mode | Output form |
---|
Health status | 1 | One crack | 2 | Two cracks | 3 |
|
Table 1. Damage mode and corresponding output form
Recognition model | Number of tests /groups | Number of errors /groups | Accuracy /% |
---|
ELM | 90 | 3 | 96.7 | BP | 90 | 6 | 93.3 |
|
Table 2. Verification result of two damage identification models