# Characterization of Relative Movements between Blocks Observed in a Concrete Dam and Definition of Thresholds for Novelty Identification Based on Machine Learning Models

^{1}

^{2}

^{3}

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Methodology Proposed for the Characterization of Relative Movements between Blocks and for Threshold Definition for Novelty Identification

#### 2.1. Methodology

- The adoption of the HTT formulation for the development of the prediction model requires the selection of the main physical quantities to be used. The water height and the concrete dam body temperature variations are the selected main inputs in the proposed model.
- The separation of effects in the analysis of the relative movements between blocks may not be verified, so a model capable of mapping the interaction between the inputs is recommended. In this sense, an MLP-NN model was found to be suitable. Once the MLP-NN model is obtained, the influence of the main loads on the responses, as well as the analysis of the residuals, r, can be analysed.
- The definition of thresholds for novelty identification based on the residuals can be defined: (i) based on one record at a time, (ii) based on a time window of records, (iii) and based on the simultaneity of the responses measured, as shown below.
- A threshold definition based on one record at a time is the most used type of threshold in dam engineering and consists of the adoption of a multiplier n applied to the standard deviation of the residuals, $sd.r.$, as follows: $n\times sd.r.$ Values of n equal to 3 and 4 correspond to a confidence level of 99.73% and 99.99%, respectively, since residuals are assumed to follow a Gaussian distribution.
- A threshold definition based on a time window of records of one physical quantity is based on the moving average of the residuals, $m.a.r.$, and the moving standard deviation of the residuals, $m.sd.r.$, adopting a time window of records over time. The size of the time window may depend on the case under study.
- A threshold definition based on the simultaneity of the responses measured allows the identification of novelties related to a cluster of data (with m features) related to the m quantities considered in the analysis. $DBSCAN$ is a method suitable for this approach, in which it is necessary to adopt, as an initial step, suitable values for Epsilon distance, $\u03f5$ and MinPoints, $MinP$ (Section 2.4). Hence, the residuals related to several physical quantities $({r}^{1},\dots {r}^{m})$, with m being the number of features measured simultaneously, are assessed to define clusters of the ones that present a behaviour far from that of the remaining data.

#### 2.2. Multilayer Perceptron Neural Network Algorithm

#### 2.3. Time-Window Threshold Definition Based on the Moving Average and Moving Standard Deviation of the Residuals

#### 2.4. Multivariate Threshold Definition Based on DBSCAN Algorithm

## 3. Case Study

#### 3.1. The Feiticeiro Dam

- Horizontal movements based on the pendulum method through the use of optical telecoordinometers (4–20 mA signal);
- Vertical movements through rod strain meters with a position transducer of the vibrating wire type;
- Discharges through level gauges with ultrasonic sensors (4–20 mA signal);
- Pressures through piezometers with pressure transducers of the vibrating wire type;
- Relative movements between blocks through 3D jointmeters with a position transducer of the vibrating wire type, located in the inspection galleries;
- Relative movements between blocks through 1D jointmeters with a resistance transducer of the Carlson type, embedded in the concrete dam body;
- Concrete strain through strain meters with resistance strain meter transducers of the Carlson type, embedded in the concrete dam body;
- Concrete temperature through thermometers with resistance thermometer transducers of the Carlson type.

#### 3.2. The Analysed Data

## 4. Results and Discussion

#### 4.1. Model Formulation, Construction, and Performance

#### 4.2. Threshold Definition for a Singular Record

#### 4.3. Threshold Definition for Novelty Identification Based on a Time Period Evolution of the Residuals

#### 4.4. Threshold Definition for Novelty Identification Based on Multivariate Data

## 5. Conclusions and Final Remarks

- The definition of univariate thresholds based on the residuals, one for each quantity, is the first recommended step. This approach, through the use of multiplicative factors associated with the standard deviation of the residual, allows an easy novelty identification that, in some cases, could be related to measurement errors.
- The second type of threshold proposed in this study can be defined by the adoption of a moving window with the moving average and the standard deviation of the residuals. This kind of approach allows the specialist to have immediate information concerning the evolution over time.
- Finally, the third type of threshold proposed can be defined based on multiple factors related to the behaviour of the residuals for all the quantities considered in the model. For this approach, a strategy based on the DBSCAN algorithm, which proved to be suitable for multivariate analysis, was used.

## Author Contributions

## Funding

## Institutional Review Board Statement

## Informed Consent Statement

## Data Availability Statement

## Acknowledgments

## Conflicts of Interest

## References

- ICOLD. Surveillance: Basic Elements in a Dam Safety Process; Bulletin Number 138; International Commission on Large Dams: Paris, France, 2009. [Google Scholar]
- Lombardi, G. Advanced data interpretation for diagnosis of concrete dams. In Structural Safety Assessment of Dams; CISM: Udine, Italy, 2004. [Google Scholar]
- Swiss Committee on Dams. Methods of Analysis for the Prediction and the Verification of Dam Behaviour. In Proceedings of the 21st Congress of the International Commission on Large Dams, Montreal, QC, Canada, 16–20 June 2003. [Google Scholar]
- Leger, P.; Leclerc, M. Hydrostatic, Temperature, Time-Displacement Model for Concrete Dams. J. Eng. Mech.
**2007**, 133, 267–277. [Google Scholar] [CrossRef] - Mata, J. Interpretation of concrete dam behaviour with artificial neural network and multiple linear regression models. Eng. Struct.
**2011**, 33, 903–910. [Google Scholar] [CrossRef] - Hellgren, R.; Enzell, J.; Ansell, A.; Nordström, E.; Malm, R. Estimating the Ice Loads on Concrete Dams Based on Their Structural Response. Water
**2022**, 14, 597. [Google Scholar] [CrossRef] - Mata, J.; Tavares de Castro, A.; SÃ da Costa, J. Time frequency analysis for concrete dam safety control: Correlation between the daily variation of structural response and air temperature. Eng. Struct.
**2013**, 48, 658–665. [Google Scholar] [CrossRef] - Salazar, F.; Toledo, M.A.; Oñate, E.; Morán, R. An empirical comparison of machine learning techniques for dam behaviour modelling. Struct. Saf.
**2015**, 56, 9–17. [Google Scholar] [CrossRef] - Simon, A.; Royer, M.; Mauris, F.; Fabre, J. Analysis and Interpretation of Dam Measurements using Artificial Neural Networks. In Proceedings of the 9th ICOLD European Club Symposium, Venice, Italy, 10–12 April 2013. [Google Scholar]
- Ranković, V.; Grujović, N.; Divac, D.; Milivojević, N. Predicting piezometric water level in dams via artificial neural networks. Neural Comput. Appl.
**2014**, 24, 1115–1121. [Google Scholar] [CrossRef] - Salazar, F.; Morán, R.; Toledo, M.Á.; Oñate, E. Data-Based Models for the Prediction of Dam Behaviour: A Review and Some Methodological Considerations. Arch. Comput. Methods Eng.
**2017**, 24, 1–21. [Google Scholar] [CrossRef] - Granrut, M.; Simon, A.; Dias, D. Artificial neural networks for the interpretation of piezometric levels at the rock-concrete interface of arch dams. Eng. Struct.
**2019**, 178, 616–634. [Google Scholar] [CrossRef] - Rico, J.; Barateiro, J.; Mata, J.; Antunes, A.; Cardoso, E. Applying Advanced Data Analytics and Machine Learning to Enhance the Safety Control of Dams. In Machine Learning Paradigms: Applications of Learning and Analytics in Intelligent Systems; Tsihrintzis, G.A., Virvou, M., Sakkopoulos, E., Jain, L.C., Eds.; Springer International Publishing: Cham, Switzerland, 2019; pp. 315–350. [Google Scholar]
- Mata, J.; Salazar, F.; Barateiro, J.; Antunes, A. Validation of Machine Learning Models for Structural Dam Behaviour Interpretation and Prediction. Water
**2021**, 13, 2717. [Google Scholar] [CrossRef] - Li, M.; Wang, J. An Empirical Comparison of Multiple Linear Regression and Artificial Neural Network for Concrete Dam Deformation Modelling. Math. Probl. Eng.
**2019**, 2019, 7620948. [Google Scholar] [CrossRef] - Hariri-Ardebili, M.A.; Kianoush, M.R. Integrative seismic safety evaluation of a high concrete arch dam. Soil Dyn. Earthq. Eng.
**2014**, 67, 85–101. [Google Scholar] [CrossRef] - Ahmadi, M.T.; Izadinia, M.; Bachmann, H. A discrete crack joint model for nonlinear dynamic analysis of concrete arch dam. Comput. Struct.
**2001**, 79, 403–420. [Google Scholar] [CrossRef] - Du, X.; Tu, J. Nonlinear seismic response analysis of arch dam-foundation systems—Part II opening and closing contact joints. Bull. Earthq. Eng.
**2007**, 5, 121–131. [Google Scholar] [CrossRef] - Wang, J.; Zhang, C.; Jin, F.J. Nonlinear earthquake analysis of high arch dam–water–foundation rock systems. Earthq. Eng. Struct. Dyn.
**2012**, 41, 1157–1176. [Google Scholar] [CrossRef] - Hariri-Ardebili, M.A.; Mirzabozorg, H.; Ghaemian, M. Pulvino and peripheral joint effects on static and seismic safety of concrete arch dams. Sci. Iran.
**2014**, 20, 1579–1594. [Google Scholar] - Azmi, M.; Paultre, P. Three-dimensional analysis of concrete dams including contraction joint non-linearity. Eng. Struct.
**2002**, 24, 757–771. [Google Scholar] [CrossRef] - Hellgren, R.; Malm, R.; Ansell, A. Performance of data-based models for early detection of damage in concrete dams. Struct. Infrastruct. Eng.
**2021**, 17, 275–289. [Google Scholar] [CrossRef] - Salazar, F.; Conde, A.; Irazábal, J.; Vicente, D.J. Anomaly Detection in Dam Behaviour with Machine Learning Classification Models. Water
**2021**, 13, 2387. [Google Scholar] [CrossRef] - Ester, M.; Kriegel, H.P.; Sander, J.; Xu, X. A density-based algorithm for discovering clusters in large spatial databases with noise. KDD
**1996**, 96, 226–231. [Google Scholar] - Housner, G.W.; Bergman, L.A.; Caughey, T.K.; Chassiakos, A.G.; Claus, R.O.; Masri, S.F.; Skelton, R.E.; Soong, T.T.; Spencer, B.F.; Yao, J.T.P. Structural control: Past, present, and future. J. Eng. Mech.
**1997**, 123, 897–971. [Google Scholar] [CrossRef] - Aggarwal, C. Neural Networks and Deep Learning; Artificial Intelligence; Springer International Publishing: Cham, Switzerland, 2018; pp. 389–411. [Google Scholar] [CrossRef]
- Shokri, M.; Tavakoli, K. A review on the artificial neural network approach to analysis and prediction of seismic damage in infrastructure. Int. J. Hydromechatron.
**2019**, 2, 178. [Google Scholar] [CrossRef] - Fallahian, M.; Khoshnoudian, F.; Meruane, V. Ensemble classification method for structural damage assessment under varying temperature. Struct. Health Monit.
**2018**, 17, 747–762. [Google Scholar] [CrossRef] - Sun, L.; Shang, Z.; Xia, Y.; Bhowmick, S. Nagarajaiah, S. Review of bridge structural health monitoring aided by big data and artificial intelligence: From condition assessment to damage detection. J. Struct. Eng.
**2020**, 146, 1–22. [Google Scholar] [CrossRef] - Salehi, H.; Burgueño, R. Emerging artificial intelligence methods in structural engineering. Eng. Struct.
**2018**, 171, 170–189. [Google Scholar] [CrossRef] - Liu, R.; Yang, R.; Zio, E.; Chen, X. Artificial intelligence for fault diagnosis of rotating machinery: A review. Mech. Syst. Signal Process.
**2018**, 108, 33–47. [Google Scholar] [CrossRef] - Bishop, C.M. Neural Networks for Pattern Recognition; Oxford University Press: New York, NY, USA, 1995. [Google Scholar]
- Han, J.; Pei, J.; Kamber, M. Data Mining: Concepts and Techniques; Morgan Kaufmann Publishers: Burlington, MA, USA; Elsevier: Amsterdam, The Netherlands, 2011. [Google Scholar]
- Schubert, E.; Sander, J.; Ester, M.; Kriegel, H.P.; Xu, X. DBSCAN revisited, revisited: Why and how you should (still) use DBSCAN. ACM Trans. Database Syst. (TODS)
**2017**, 42, 1–21. [Google Scholar] [CrossRef] - Cunha, J.; Vieira Soares, P. The Emergency Action Plans of the Baixo Sabor Dams. In Proceedings of the Second International Dam World Conference 2012, Lisbon, Portugal, 21–24 April 2015. [Google Scholar]
- LNEC. Downstream Dam of Baixo Sabor Hydroelectric Scheme. In Revision of the Monitoring Plan of the Dam in Conventional Concrete; Technical Report 67/2013—DBB/NO; LNEC: Lisbon, Portugal, 2013. (In Portuguese) [Google Scholar]
- LNEC. Downstream Dam of Baixo Sabor Hydroelectric Scheme. In Revision of the Automated Data Adquisition System Plan; Technical Report 244/2013; DBB/NO; LNEC: Lisbon, Portugal, 2013. (In Portuguese) [Google Scholar]
- LNEC. Feiticeiro Dam. In Inspection of the Structure and Monitoring System Carried Out in March 2020; Technical Report 149/2020—DBB/NO; LNEC: Lisbon, Portugal, 2020. (In Portuguese) [Google Scholar]
- Cunha, J.; Mata, J.; Gonzalo, O. Structural Safety Control of the Feiticeiro Dam Based on an Automated Data Acquisition System. In Proceedings of the 85th Annual Meeting of International Commission on Large Dams, Prague, Czech Republic, 3–7 July 2017. [Google Scholar]
- R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2021. [Google Scholar]
- RStudio Team. RStudio: Integrated Development Environment for R; R Studio, Inc.: Boston, MA, USA, 2019. [Google Scholar]
- Venables, W.N.; Ripley, B.D. Modern Applied Statistics with S, 4th ed.; Springer: New York, NY, USA, 2002; ISBN 0-387-95457-0. [Google Scholar]
- Peters, A.; Hothorn, T. ipred: Improved Predictors; R Package Version 0.9-11; R Foundation for Statistical Computing: Vienna, Austria, 2021. [Google Scholar]

**Figure 3.**MLP architecture where x represents the input layer, l represents the hidden layer, L represents the output layer, b represents the bias and w represents the weights, adapted from [5].

**Figure 4.**DBSCAN illustration with MinPoints = 4 and epsilon distance represented by circles. Core points, such as A, are represented in red, B and C, in yellow, are border points, and N is a noise point, represented in blue. From [34].

**Figure 6.**Feiticeiro dam. Thermometers and 3D jointmeters included in the automated monitoring system. Legend: MWL—Maximum reservoir water level.

**Figure 7.**Time series of the relative movements between blocks measured between April 2017 and June 2022.

**Figure 8.**Time series of the relative movements between blocks measured between the 1st and the 8th of August 2018.

**Figure 11.**Opening-closing mov. measured in BT3 vs. temperature measured in T37 (

**left**). Opening-closing mov. measured in BT3 vs. water level for temperature measured in T37 equal to 10 °C (

**right**).

**Figure 12.**Time series of the fitted values and recorded values for the movements measured through 3D jointmeters.

**Figure 13.**Time series of the relative movements between block measurements and fitted values between the 1st and the 8th of August 2018.

**Figure 18.**Novelty identification based on a multivariate analysis based on the DBSCAN method (adopting Epsilon = 0.12 and MinPoints = 7).

Block | Level (m) | Position | Block | Level | Position | ||
---|---|---|---|---|---|---|---|

T10 | J7–J8 | 122.1 | u.f. | T11 | J7–J8 | 122.1 | 1 m u.f. |

T15 | J7–J8 | 122.1 | 1 m d.f. | T16 | J7–J8 | 122.1 | d.f. |

T31 | J11–J12 | 112.0 | u.f. | T32 | J11–J12 | 112.0 | 1 m u.f. |

T33 | J11–J12 | 112.0 | oth | T35 | J11–J12 | 112.0 | oth |

T62 | J15–J16 | 122.1 | u.f. | T64 | J15–J16 | 122.2 | oth |

T65 | J15–J16 | 122.1 | 1/2 th. | T66 | J15–J16 | 122.2 | oth |

Measurements | Model | |||||
---|---|---|---|---|---|---|

Min | Max | r_{Min} | r_{Max} | sd | ${\mathit{R}}^{2}$ | |

[mm] | [mm] | [mm] | [mm] | [mm] | % | |

BT3 O/C | 0.16 | 2.56 | −0.22 | 0.35 | 0.060 | 99.81 |

BT5 O/C | 0.29 | 2.85 | −0.32 | 0.39 | 0.072 | 99.79 |

BT7 O/C | 1.03 | 3.57 | −0.29 | 0.33 | 0.068 | 99.92 |

BT8 O/C | 0.70 | 2.61 | −0.21 | 0.34 | 0.063 | 99.86 |

BT10 O/C | 0.25 | 2.15 | −0.21 | 0.30 | 0.056 | 99.77 |

BT12 O/C | 0.81 | 2.72 | −0.24 | 0.37 | 0.067 | 99.85 |

BT13 O/C | 0.48 | 2.21 | −0.22 | 0.28 | 0.048 | 99.86 |

^{2}—coefficient of determination.

MinPoints | Epsilon | ||||
---|---|---|---|---|---|

0.08 | 0.10 | 0.12 | 0.14 | 0.16 | |

5 | 79 | 9 | 5 | 3 | 0 |

6 | 102 | 12 | 5 | 3 | 1 |

7 | 116 | 14 | 5 | 3 | 1 |

8 | 126 | 22 | 5 | 3 | 1 |

9 | 140 | 22 | 5 | 3 | 1 |

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |

© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Mata, J.; Miranda, F.; Antunes, A.; Romão, X.; Pedro Santos, J. Characterization of Relative Movements between Blocks Observed in a Concrete Dam and Definition of Thresholds for Novelty Identification Based on Machine Learning Models. *Water* **2023**, *15*, 297.
https://doi.org/10.3390/w15020297

**AMA Style**

Mata J, Miranda F, Antunes A, Romão X, Pedro Santos J. Characterization of Relative Movements between Blocks Observed in a Concrete Dam and Definition of Thresholds for Novelty Identification Based on Machine Learning Models. *Water*. 2023; 15(2):297.
https://doi.org/10.3390/w15020297

**Chicago/Turabian Style**

Mata, Juan, Fabiana Miranda, António Antunes, Xavier Romão, and João Pedro Santos. 2023. "Characterization of Relative Movements between Blocks Observed in a Concrete Dam and Definition of Thresholds for Novelty Identification Based on Machine Learning Models" *Water* 15, no. 2: 297.
https://doi.org/10.3390/w15020297