Skip to main content

Data-driven crack behavior anomaly identification method for concrete dams in long-term service using offline and online change point detection


Cracking is a common threat to dam structural safety. It is desirable to establish models that can accurately assess the influence of cracks on the structural safety of concrete dams in time. The structural condition assessment for dams can be categorized into offline structural state review based on historical monitoring data and online real-time detection based on updated data. Moreover, the offline review can be further divided into two scenarios, depending on whether the number of change points is known in advance. To solve the above practical problems, three different offline and online changepoint detection (CPD) methods, including dynamic programming segmentation, bottom–up segmentation, and online Bayesian CPD methods are introduced. A concrete gravity-arch dam with 300 m length and 5 m depth horizontal cracks stretched across the downstream of various blocks in long-term service is used as the case study. Crack opening displacement collected by resistance joint meters is used to demonstrate the feasibility of the proposed identification methods. The experimental results show that the underlying change points of crack behavior can be accurately and timely detected by the proposed model, and the exact dates of change points can also be obtained. The calculated results are roughly consistent with the observations of visual inspections and are consistent with the recorded historical engineering management report. The proposed model does not require prior physical knowledge about concrete cracks. It is practical and flexible to be embedded in dam automated structural health monitoring systems to deal with large-scale monitoring data related to structural changes.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13

Data availability

The data used to support the findings of this study are included within the supplementary information file(s).


  1. 1.

    Li Y, Bao T, Shu X et al (2020) A hybrid model integrating principal component analysis, fuzzy C-means, and Gaussian process regression for dam deformation prediction. Arab J Sci Eng.

    Article  Google Scholar 

  2. 2.

    Wu Z, Li J, Gu C, Su H (2007) Review on hidden trouble detection and health diagnosis of hydraulic concrete structures. Sci China Ser E Technol Sci 50:34–50.

    Article  Google Scholar 

  3. 3.

    Chen B, Wu Z, Liang J, Dou Y (2017) Time-varying identification model for crack monitoring data from concrete dams based on support vector regression and the bayesian framework. Math Probl Eng.

    Article  Google Scholar 

  4. 4.

    Li Y, Bao T, Gao Z et al (2021) A new dam structural response estimation paradigm powered by deep learning and transfer learning techniques. Struct Health Monit.

    Article  Google Scholar 

  5. 5.

    Li Y, Bao T, Chen H et al (2021) A large-scale sensor missing data imputation. Measurement.

    Article  Google Scholar 

  6. 6.

    Li X, Zhang X, Gu C, Hongzhong Xu (2008) Abnormality diagnosis of crack based on fracture mechanics. J Hohai Univ Sci Ed 2008(02):209–212.

  7. 7.

    Plizzari GA (1997) LEFM applications to concrete gravity dams. J Eng Mech 123:808–815.

    Article  Google Scholar 

  8. 8.

    Wan HP, Ni YQ (2018) Bayesian modeling approach for forecast of structural stress response using structural health monitoring data. J Struct Eng 144:1–12.

    Article  Google Scholar 

  9. 9.

    Bao T, Li J, Zhao J (2019) Study of quantitative crack monitoring and POF layout of concrete dam based on POF-OTDR. Sci Sin Technol 49:343–350.

    Article  Google Scholar 

  10. 10.

    Glisic B, Inaudi D (2012) Development of method for in-service crack detection based on distributed fiber optic sensors. Struct Health Monit 11:161–171.

    Article  Google Scholar 

  11. 11.

    Li H, Bao T, Gu C, Chen B (2019) Vibration feature extraction based on the improved variational mode decomposition and singular spectrum analysis combination algorithm. Adv Struct Eng 22:1519–1530.

    Article  Google Scholar 

  12. 12.

    Ohno K, Ohtsu M (2010) Crack classification in concrete based on acoustic emission. Constr Build Mater 24:2339–2346.

    Article  Google Scholar 

  13. 13.

    Jeong S, Ferguson M, Hou R et al (2019) Sensor data reconstruction using bidirectional recurrent neural network with application to bridge monitoring. Adv Eng Inform 42:100991.

    Article  Google Scholar 

  14. 14.

    Li ZC, Gu CS, Wang ZZ, Wu ZR (2015) On-line diagnosis method of crack behavior abnormality in concrete dams based on fluctuation of sequential parameter estimates. Sci China Technol Sci 58:415–424.

    Article  Google Scholar 

  15. 15.

    Dorcas Wambui G (2015) The power of the pruned exact linear time (PELT) test in multiple changepoint detection. Am J Theor Appl Stat 4:581.

    Article  Google Scholar 

  16. 16.

    Cho H, Fryzlewicz P (2015) Multiple-change-point detection for high dimensional time series via sparsified binary segmentation. J R Stat Soc Ser B Stat Methodol 77:475–507.

    MathSciNet  Article  MATH  Google Scholar 

  17. 17.

    Ni YQ, Zhang QH (2019) A Bayesian machine learning approach for online wheel condition detection using track-side monitoring. In: 2018 Int Conf Intell Rail Transp ICIRT 2018.

  18. 18.

    Aminikhanghahi S, Cook DJ (2017) A survey of methods for time series change point detection. Knowl Inf Syst 51:339–367.

    Article  Google Scholar 

  19. 19.

    Lu G, Zhou Y, Lu C, Li X (2017) A novel framework of change-point detection for machine monitoring. Mech Syst Signal Process 83:533–548.

    Article  Google Scholar 

  20. 20.

    Wan HP, Ni YQ (2019) Binary segmentation for structural condition classification using structural health monitoring data. J Aerosp Eng 32:1–9.

    Article  Google Scholar 

  21. 21.

    Killick R, Fearnhead P, Eckley IA (2012) Optimal detection of changepoints with a linear computational cost. J Am Stat Assoc 107:1590–1598.

    MathSciNet  Article  MATH  Google Scholar 

  22. 22.

    Li Z, Gu C, Wu Z (2013) Nonparametric change point diagnosis method of concrete dam crack behavior abnormality. Math Probl Eng.

    MathSciNet  Article  MATH  Google Scholar 

  23. 23.

    Hu J, Wu S (2019) Statistical modeling for deformation analysis of concrete arch dams with influential horizontal cracks. Struct Health Monit 18:546–562.

    Article  Google Scholar 

Download references


This research has been supported by the National Key Research and Development Program, Grant/Award Number: China2018YFC1508603, the National Natural Science Foundation of China, Grant/Award Number: 51579086, 51739003, Postgraduate Research & Practice Innovation Program of Jiangsu Province, Grant/Award Number: KYCX21_0515. The data preparation work from Associate Professor Bo Chen and the support from Anhui Reservoir Management Office are grateful.

Author information



Corresponding author

Correspondence to Tengfei Bao.

Ethics declarations

Conflict of interest

The authors declare no conflict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (RAR 663 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Li, Y., Bao, T., Shu, X. et al. Data-driven crack behavior anomaly identification method for concrete dams in long-term service using offline and online change point detection. J Civil Struct Health Monit 11, 1449–1460 (2021).

Download citation


  • Structural health monitoring
  • Crack behavior change
  • Condition classification
  • Gravity-arch dam
  • Bottom–up segmentation
  • Bayesian online detection