Structural health monitoring methods.pdf

SHM Methods

1. Department of Civil Engineering
SHARAD INSTITUTE OF TECHNOLOGY COLLEGE OF ENGINEERING
YADRAV
Structural Health Monitoring Methods
Dr. Maloth Naresh

2. Flow of presentation
1 Introduction
2
Traditional Methods3
4 Vibration-based methods
References5
Structural health monitoring methods

3. Structural health monitoring (SHM)
• The failure causes, deterioration of the structure which leads to sudden failure of the structure cause, loss of
property as well as human life. In this regard, SHM gives an exact solution to avoid such failure. Fig. 1
shows the classification of the SHM.
Fig. 1: Classification of the SHM.
Introduction

4. Structural health monitoring methods
• There are many studies found on health monitoring of steel frame structures using traditional and vibration-
based techniques.
•Traditionally techniques such as ultrasonic technique, impedance-based methods, acoustic emission
techniques, modal strain energy analysis, mode shape curvature analysis, and displacement-based methods.
However, these methods often prove impractical for most structures due to their cost and time requirements.
•Consequently, researchers have shifted their focus toward vibration-based (VB) techniques.
•These techniques are built on the premise that changes in a structure's physical characteristics, particularly in
terms of stiffness loss, and changes in modal parameters provide meaningful damage indicators that reflect
the actual behavior of the structure.
•Recently, the combination of machine learning (ML) frameworks with VB responses enhanced the detection
of structural damage.
•Due to its capacity to handle uncertainty and noise effectively. ML frameworks are leveraged to automate the
process of damage pattern recognition.

5. Traditional methods
•In the past, local techniques such as acoustoelastic effect-based methods (F. Wang & Song, 2019),
ultrasonic techniques (Nadom, 2016), vision-based methods (Fukuda et al., 2010), piezoelectric
impedance methods, modal strain energy methods (Pal & Banerjee, 2015), and displacements (Park et
al., 2015) have also been utilized in SHM to identify joint damage in frame structures.

6. Fig 15. Bolted frame.
Swit (2018)
➢ Impact load applied at the joints.
➢ Collected the elastic wave response in the
form of signal.
➢ Damage developed, loosening of the
bolts and collected the responses.
➢ Changes in the signal in terms of
wavelength, identified the damage
parameter.
Acoustic emission method for SHM of bolted plane frame
20
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

7. ➢ Fatigue load applied at the joint.
➢ Collect the (signal) response.
➢ Damage developed by reducing the
cross-section of the beam.
➢ Collected the damaged responses.
➢ Based on changes in the wavelength,
developed the damage identification
parameter.
Acoustic emission method for SHM of welded plane frame
Fig 16. Welded frame.
Rucka (2015)
21
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

8. Ultrasonic method for SHM of welded tubular structures
Song (2019)
➢ More suitable for welded structures.
➢ Waves penetrates inside the structure.
➢ Response collected in the form of
speed of the ultrasonic wave.
➢ Reducing stiffness.
➢ Changes appears in the propagation
of wave (speed), developed damage
identification parameter.
Fig 17. Tubular structure.
22
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

9. Impedance method for SHM of bolted plane frame
➢ Impedance is measure of the opposition to
electrical flow.
➢ PZT patches are attached both side of the
structure.
➢ Help of PZT applied voltage at the joint,
collected response in the form of impedance.
➢ Damage developed by reducing the stiffness of
the beam.
➢ Applied voltage at the joint, collected the
response
➢ Impedance relate to the stiffness and damping,
identified the damage parameter.
Sun 2013
Fig 18. Bolted 3D-frame.
23
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

10. ➢ Bolted 3D-frame.
➢ Apiezoelectric sensorconverts physical
parameters into an electrical charge.
➢ Damage developed by loosening of the
bolts.
➢ Both strain, acceleration collected in both
cases.
➢ Changes in electrical charges, represents
the damage identification parameter.
Hasni (2017)
Fig. 20 Bolted 3D-frame.
Piezoelectric sensing method for SHM of bolted 3D-frame 24
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

11. Drawbacks of the traditional method
➢ Difficult in the implementation and more
complexity.
➢ Sensor placing is difficult at the joint.
➢ Skilled persons are required.
➢ Always supervising.
➢ High costly.
Conti… 25
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

12. Vibration methods
•The vibration-based methods are employed to assess the entire performance of the monitored structure
by converting its vibration response into a meaningful damage identification parameter that indicates
the real condition of the structure, which made the VB techniques more popular.The ultimate aim of
these techniques is to detect damage by processing data that is acquired from an acquisition system like
accelerometers and strain gauges

13. Finite Element (FE) -model updating methods
➢ The FE-method is used to find the static and dynamic behaviour of the
structure in its service periods.
➢ Update the structure based on stiffness and mass matrices.
➢ Structural damage identification, and forecast the future performance of the
structure.
26
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

14. FE-modal updating for SHM of welded plane frame
Pal (2013)
➢ Free and harmonic vibration.
➢ Collected the response in the form of
Frequency and mode shape.
➢ Reducing the cross section near the joint,
collected the responses.
➢ Modal curvature method (MCM),
identified the joint damage.
Fig. 21 (a) Plane frame.
Fig. 21 (b) Damaged frame.
27
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

15. ➢FE-model technique and the Bayesian approach.
➢Impact load is applied at the joints, collected the
acceleration.
➢Reduction of stiffness of a beam, collected the
responses.
➢Help of Bayesian approach, derive the damage
identification parameter.
Yin (2017)
FE-model updating method for SHM of bolted plane frame
Fig. 22 Bolted plane frame.
28
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

16. FE-model updating method for SHM of Welded 3D- frame
Ji (2012)
➢ Applied cyclic load at the joint.
➢ Collected the frequency and mode shapes.
➢ Removing stiffness, Collected the responses.
➢ Curve fitting method (CFM), calculated the
damage index.
Fig. 23 Welded 3D-frame.
29
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

17. ➢ The combination of the model updating
method and adaptive quadratic sum-square
error with unknown inputs (AQSSE-UN)
technique.
➢ Applied base excitation with help of shaking
table.
➢ Collected the acceleration.
➢ Reducing cross-section of the beam, collected
the responses.
➢ Apply AQSSE-UN technique, calculated the
rotational stiffness, is used as the damage
identification parameter.
Fig. 24 Bolted 3D-frame
Yang (2014)
FE-model updating method for SHM of bolted 3D- frame 30
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

18. Pal (2017)
➢ Impact hammer.
➢ Collected the frequency responses.
➢ Loosening of the bolts, collected the responses.
➢ Frequency amplitude and shape corelation is
considered as the objective function.
➢ Particle Swarm Optimisation (PSO) technique.
➢ Calculated the fixity factor of the bolt.
Fig. 25 (b) Bolted connected Frame
PSO technique for SHM of bolted plane frame
31
Fig. 25(a) PSO10-Dec-21/// 11.3O A.M (Online image)

19. Drawbacks of the FE- model updating method
➢ Involving more number of variable.
➢ Iterative method.
➢ Heavy computational work.
32
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

20. AI-based methods
33
➢Artificial intelligence (AI), that provides the ability to automatically
learn and experience from the sample data.
Fig. 27 Typical diagram of AI.
10-Dec-21/// 11.3O A.M
National Institute of Technology, Hamirpur (H.P), INDIA.

21. ➢The combination of PCA and ANN model used.
➢Applying base excitation with help of shaking
table.
➢Collected the acceleration responses.
➢Damage developed by reducing the stiffness of a
beam.
➢Using PCA, extract the features from the data.
➢Applied ANN model, classify the damage.
Bandara (2014)
ANN technique for SHM of bolted plane frame
Fig. 28 Bolted connected frame.
34
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

22. ➢ Applying fatigue at the joint.
➢ Collected the frequency responses.
➢ Reducing cross-section near the joint.
➢ Using PCA & ANN model, classify
the damage.
ANN technique for SHM of welded 3D-structure
Fig. 29 Bridge structure.
Zhang (2019)
35
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

23. ➢ Applying impact load at the joint.
➢ Collected the acceleration responses,
➢ Loosening of the bolts.
➢ Using auto-regressive (AR) modelling,
extract the feature.
➢ Based on the ANN classify the damage.
Nguyen (2017)
Fig. 30 Flange bolted tubular section.
ANN technique for SHM of bolted tubular structure
36
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

24. ➢ Convolution Neural Network (CNN) and
continuous wavelet transform (CWT).
➢ Applied Impact load at the joint.
➢Collected the acceleration.
➢Damage developed by loosening of the
bolts, collected the responses.
➢With help of CWT, responses convert into
scalogram images.
➢Using CNN model trained the data set,
classify the damage.
Paral (2020)
Fig. 31 Bolted connected Frame.
CNN technique for SHM of bolted plane frame
37
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

25. ➢ Applied the fatigue load at the joints.
➢ Collected the acceleration.
➢ Damage developed by reducing
cross-section near the joint, collected
the damaged responses.
➢ Using discrete wavelet modelling for
generating the images.
➢ Using CNN model, classify the
damage.
CNN technique for SHM of welded 3D-structure
Fig. 32 Welded gusset plate.
Dung (2019)
38
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

26. R-CNN technique for SHM of bolted structure
Fig. 33 Bolted beam
➢ Applied impact load at the joint.
➢ Collected the images of bolt (bolt angle).
➢ Damage developed by loosening of the
bolts.
➢ Collected the damaged images.
➢ R-CNN model, classify the damage.
Huynh (2019)
39
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

27. Drawbacks of the AI based-method
➢ Require large number of training data set.
➢ There is no guarantee that a specific feature/classifier set will be the
best choice for all types of structure.
➢ Computational complexity and time taking method.
➢ Continuously supervising method.
41
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

28. REFERENCES
1.Pal, J., Shah, V., & Banerjee, S. (2013). Performance of damage detection algorithms for health monitoring of joints in
steel frame structures using vibration-based techniques.International Journal of Structural Engineering,4(4), 346-360.
2.Teimoori, T., & Mahmoudi, M. (2020). Damage detection in connections of steel moment resisting frames using proper
orthogonal decomposition and wavelet transform.Measurement,166, 108188.
3.Pal, J., Banerjee, S., Chikermane, S., & Banerji, P. (2017). Estimation of fixity factors of bolted joints in a steel frame
structure using a vibration-based health monitoring technique.International Journal of Steel Structures,17(2), 593-607.
4.Paral, A., Roy, D. K. S., & Samanta, A. K. (2020). A deep learning-based approach for condition assessment of semi-rigid
joint of steel frame.Journal of Building Engineering, 101946.
5.Ji, X., Fenves, G. L., Kajiwara, K., & Nakashima, M. (2011). Seismic damage detection of a full-scale shaking table test
structure.Journal of Structural Engineering,137(1), 14-21.
6.Yang, J. N., Xia, Y., & Loh, C. H. (2014). Damage identification of bolt connections in a steel frame.Journal of Structural
Engineering,140(3), 04013064.
7.Hasni, H., Jiao, P., Alavi, A. H., Lajnef, N., & Masri, S. F. (2018). Structural health monitoring of steel frames using a
network of self-powered strain and acceleration sensors: A numerical study.Automation in Construction,85, 344-357.
8.Song, M., Xu, T., Yuan, K., Yu, H., & Sun, C. (2019). Creep failure of a steam pipe girth weld and NDT strategy on creep
damage.Engineering Failure Analysis,104, 673-681.
9.Park, J. H., Huynh, T. C., Choi, S. H., & Kim, J. T. (2015). Vision-based technique for bolt-loosening detection in wind
turbine tower.Wind Struct,21(6), 709-726.
10.Nguyen, C. U., Huynh, T. C., & Kim, J. T. (2018). Vibration-based damage detection in wind turbine towers using
artificial neural networks.Structural Monitoring and Maintenance,5(4), 507.
57
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.

29. Thankyou
58
10-Dec-21/// 11.3O A.M National Institute of Technology, Hamirpur (H.P), INDIA.