Hemerson Lizarbe-Alarcón*
| David Palomino-Pariona | José Estrada-Cárdenas | Jhon Tacas-Evanan | Alex Ircañaupa-Huamani | Amílcar Tacuri-Gamboa | Rocky Ayala-Bizarro | Rualth Bravo-Anaya
© 2025 The authors. This article is published by IIETA and is licensed under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).
OPEN ACCESS
This study proposes a model based on artificial neural networks (ANNs) for predicting preliminary design parameters of cyclopean concrete dams in high-Andean micro-watersheds, used in water harvesting and storage projects. A database was developed using digital elevation models (DEMs) and technical records from projects executed by the Sierra Azul Fund Executing Unit (UEFSA) between 2019 and 2024. Geospatial processing of the DEMs yielded morphometric variables and storage volumes, which were used as model inputs, while the dam axis location, length, and height were the outputs. The selected architecture was a multilayer perceptron (MLP) with two hidden layers, trained via backpropagation with the Adam optimizer. Results show R² values greater than 0.90 and acceptable average errors for preliminary design use. It is concluded that the model reduces variability inherent in empirical methods and enhances standardization in formulating water harvesting projects in high-Andean regions. This approach improves the efficiency, accuracy, and sustainability of hydraulic infrastructure development in rural contexts with high water vulnerability.
Adam optimization algorithm, cyclopean concrete dams, hydraulic infrastructure, predictive modeling, DEMs, design optimization, ANNs, Sierra Azul Fund Executing Unit
The design of hydraulic structures, such as the cyclopean concrete dams used in water harvesting and storage projects in Peru, is a fundamental component of sustainable water resource management in high Andean regions. These infrastructures are essential for the storage, regulation, and utilization of runoff water, helping to mitigate the effects of climate variability and drought that affect rural communities with limited access to water [1]. In this context, the construction of cyclopean concrete dams is presented as a viable solution due to their durability, strength, and lower cost compared to other, more complex hydraulic structures [2]. However, despite their importance, the design process for these dams continues to be carried out primarily using empirical methods and professional criteria based on experience, which can lead to significant variations in the projected dimensions and thus affect hydraulic efficiency, structural stability, and construction costs [3].
Advances in science and technology have led to the integration of advanced computational tools in the field of civil engineering, including artificial intelligence (AI). Artificial neural networks (ANNs), as one of the most representative AI methods, have proven effective in identifying complex patterns within large volumes of data and in generating highly accurate predictive models [4, 5]. In concrete engineering, these techniques have been successfully applied for material characterization, mix optimization, strength estimation, and real-time structural monitoring [6, 7].
Similarly, in the field of hydraulics, neural networks have been used for analyzing dam behavior, predicting flow rates, and simulating hydrological processes in basins [8, 9]. However, most of these studies focus on post-construction analysis and monitoring or on evaluating structural performance in operational scenarios [7, 10].
Despite the progressive advancements in the application of AI in civil engineering, a significant gap has been identified in the technical literature: the absence of predictive models based on ANNs for the preliminary design of cyclopean concrete dams, particularly in high-altitude Andean micro-basins where terrain morphology, potential storage volume, and water availability depend on topographic parameters derived from digital elevation models (DEMs). This absence is mainly due to the limited systematization of geospatial databases in rural areas, the geomorphological heterogeneity of the Peruvian Andes, and the historical reliance on empirical methodologies in the formulation of water catchment projects [11].
In this sense, the need arises to develop a predictive model that integrates topographic, hydrological and geometric information to support the preliminary design of cyclopean dikes in water harvesting and storage projects, allowing for obtaining initial dimensions with greater precision and consistency, reducing formulation times, minimizing cost overruns and improving the technical sustainability of projects in the national territory.
Therefore, this study proposes the development of an ANN model trained on 242 technical records and DEMs from projects executed by the Sierra Azul Fund Executing Unit (UEFSA). The model allows for the prediction of key dam design parameters, such as axis location, dam length, and height, based on geomorphological variables of the micro-basin.
The fundamental purpose of this research is to demonstrate that ANNs represent an efficient, accurate, and applicable tool in high Andean rural contexts for the preliminary design of hydraulic infrastructure. In this way, it seeks to contribute to the transition from empirical methods to computer-aided design processes, aligned with modern engineering standards.
In Peru, the design and construction of cyclopean concrete dams is a fundamental activity within the hydraulic infrastructure projects carried out by the Ministry of Agrarian Development and Irrigation (MIDAGRI) through the UEFSA. These structures are intended to store and regulate water resources in high Andean regions, where seasonal rainfall and climate variability cause periods of water scarcity that affect food security and agricultural irrigation [2, 12]. Therefore, the construction of cyclopean dams represents an efficient and low-cost technical alternative compared to more complex structures.
However, the design of these types of dams faces several limitations when traditional methodologies are used alone. Conventional procedures rely primarily on empirical criteria and the experience of the responsible professional, which can lead to discrepancies in estimating the dimensions and the appropriate location of the structure [3]. Furthermore, these methods do not always consider, in an integrated manner, the interaction between topographic, geotechnical, and hydrological factors that influence the dam's performance, which can result in over- or under-dimensioning that affects safety and construction costs [13].
In this regard, the use of advanced computational technologies contributes to improving the design process. ANNs have become established tools capable of analyzing large datasets and recognizing complex patterns that are not easily identifiable using traditional statistical methods [14]. These networks allow for the establishment of nonlinear relationships between topographic, hydraulic, and structural variables, making them a suitable alternative to support the prediction of the design of cyclopean concrete dams.
Several studies have demonstrated that AI-based methodologies offer greater adaptability and accuracy compared to conventional techniques, particularly in the analysis of concrete structures and the monitoring of their behavior under various external factors [15]. Furthermore, the application of ANNs in structural engineering has enabled significant advances in structural health monitoring, deformation prediction, and improved decision-making in hydraulic infrastructure [10].
However, in the Peruvian context, the adoption of AI-based algorithms for hydraulic infrastructure design remains limited. Factors explaining this situation include the need for systematized geospatial databases, a lack of specialized training, and a historical reliance on empirical methods [11]. This underscores the need to promote research that integrates topographic, hydrological, and structural information with advanced predictive models to improve design accuracy and efficiency.
This research aims to develop a model based on ANNs to predict the design parameters of cyclopean concrete dams in water harvesting and storage projects. The goal is to strengthen the preliminary design process, reduce uncertainties, optimize resources, and contribute to the sustainable development of hydraulic solutions in vulnerable areas of the country.
Research objectives: Design of the appropriate model to predict the design of cyclopean concrete dams using ANNs in water catchment and planting projects, Ayacucho, 2024.
3.1 Hydraulic infrastructure in high Andean micro-basins
In the high Andean regions of Peru, seasonal rainfall variability and limited natural soil moisture retention affect water availability for productive activities and human consumption. Faced with this scenario, water harvesting and storage projects have become established as intervention strategies to improve water regulation in micro-watersheds. The construction of cyclopean concrete dams is a low-cost, durable, and morphologically adaptable infrastructure alternative [2, 12]. These dams allow for the storage of significant volumes of runoff, creating reservoirs that supply water for irrigation and multiple uses. However, site selection and dam dimensions are highly dependent on local topography and the hydrological behavior of the micro-watershed, requiring more precise and replicable design processes.
3.2 Geospatial modeling applied to dam design
The following models were used:
3.2.1 DEMs
DEMs allow for the representation of the terrain surface with altimetric detail, enabling the acquisition of morphometric variables such as slope, drainage direction, channel depth, and contributing areas [16, 17]. In dam design, DEMs are fundamental for identifying the dam section, estimating storage capacity, and modeling geometric reservoir scenarios. Recent studies demonstrate that the accuracy of these variables depends on the DEM resolution and the geospatial processing method [18, 19].
3.2.2 Delineation of micro-basins and closing section
The delineation of micro-watersheds using flow accumulation algorithms has allowed for the automated identification of strategic points for dam construction [20]. However, the literature indicates that the shape of the watershed and the longitudinal slope of the channel have a non-linear influence on the length and height of the dam [21, 22]. This complex relationship justifies the use of data-driven predictive models.
3.3 Materials and structural behavior of cyclopean concrete dams
Cyclopean concrete dams combine concrete with large rock inclusions, reducing costs and increasing mass stability [23]. Their design requires evaluating stresses, hydrostatic pressures, and foundation conditions. However, in small and medium-sized dams, the literature indicates that the final dimensions are particularly sensitive to geomorphological parameters of the site, rather than to complex structural processes [7, 24]. This reinforces the importance of having reliable geometric prediction methods.
3.4 Applications of AI in hydraulic engineering
3.4.1 AI and machine learning in hydrology
AI has been used for flow estimation, runoff prediction, and water resource modeling [8, 9, 15]. The ability of ANNs to recognize nonlinear patterns allows them to overcome the limitations of deterministic hydrological models in basins with limited information or high spatial variability.
3.4.2 RNA in concrete structures
In structural engineering, neural networks have been applied to estimating the mechanical properties of concrete, evaluating deformations, and monitoring gravity dams, demonstrating efficiency compared to parametric correlation methods [10, 25]. However, most studies focus on post-construction evaluation, not preliminary design.
3.5 Limitations of the literature and identified gap
While AI-based models have been developed for hydrology and structural assessment, no studies have been identified that use ANNs to predict design parameters for cyclopean dams (axis location, length, and height) from morphometric variables derived from DEMs, particularly in high-Andean micro-basins. The absence of this approach limits the systematization of design and perpetuates reliance on empirical methods.
3.6 Contribution of the study
This work addresses this gap by developing a multi-layer ANN model capable of predicting the preliminary design parameters of cyclopean concrete dams using DEM data and technical files, contributing to improving the accuracy, consistency and efficiency in the formulation of water harvesting and storage projects.
The research was developed in four main methodological stages: (i) collection and structuring of the geospatial and technical database, (ii) processing of the DEM and delimitation of micro-basins, (iii) extraction of geometric parameters and definition of the input and output variables of the model, and (iv) construction, training and validation of the ANN for the prediction of design parameters of cyclopean concrete dams.
4.1 Database compilation and organization
Technical files approved by the UEFSA, corresponding to projects executed between 2019 and 2024, were used, identifying 242 cyclopean concrete dams built in high Andean micro-basins. From each file, UTM coordinates of location, micro-basin area, storage volume, and structural design dimensions (length and height of the dam) were extracted.
DEMs from the ALOS PALSAR and SRTM services were also used, with resolutions of 12.5 m and 30 m, respectively, the former being selected for its better performance in areas of abrupt relief. The main characteristics of the technical files and DEMs used are summarized in Table 1.
Table 1. Characteristics of the files and MDE used
|
No. |
Name of the Qocha |
Easting Coordinate (m) |
North Coordinate (m) |
Elevation (masl) |
Zone |
|
1 |
Challhuacocha |
547583 |
8531043 |
4471.00 |
18 L |
|
2 |
Hatunqocha |
547606 |
8526998 |
4196.00 |
18 L |
|
3 |
Huayllupatata |
543693 |
8530843 |
4607.90 |
18 L |
|
… |
… |
… |
… |
… |
… |
|
241 |
Aculla 01 |
703077 |
8324770 |
4797.17 |
18 L |
|
242 |
Aculla 02 |
703460 |
8325515 |
4777.99 |
18 L |
4.2 DEM processing
The DEM was corrected to remove artifacts and spurious depressions using the Fill Sinks algorithm. Subsequently, slope, flow direction, and flow accumulation maps were calculated using raster hydrological analysis techniques.
The micro-watershed delineation was carried out starting from the closure point defined in the technical file, allowing the determination of the catchment area associated with each dam. This processing was performed using QGIS and Whitebox Tools.
4.3 Extraction of geomorphological variables and design parameters
The following variables were obtained for each micro-basin:
These variables represent the physical characteristics and respond to known non-linear relationships between the morphometry of the basin and the geometry of the reservoir, which justifies the relevance of using ANNs.
The statistical parameters used as model inputs are presented in Table 2.
Table 2. Statistical variables used as inputs to the model
|
No. |
East (m) |
North (m) |
Share |
Area (m²) |
Volume (m³) |
|
1 |
550719.67 |
8526072.20 |
4089.22 |
242915.2 |
17868.36 |
|
2 |
677388.27 |
8444857.11 |
4331.31 |
1185886.7 |
9274.14 |
|
3 |
677168.20 |
8445184.11 |
4335.58 |
1001653.9 |
10226.64 |
|
… |
… |
… |
… |
… |
… |
|
241 |
703075.76 |
8324764.51 |
4797.73 |
123500.0 |
13359.06 |
|
242 |
703452.03 |
8325525.19 |
4778.56 |
121900.0 |
19653.72 |
4.4 Design of the ANN
A multilayer perceptron (MLP) ANN with a feedforward architecture was used. The selection of this architecture is justified by its ability to model nonlinear functions between multiple independent and dependent variables.
The final architecture was defined through experimental iterations:
The structure and workflow of the proposed ANN model are shown in Figure 1.
Figure 1. Program flowchart with ARN
4.5 Model training and validation
The database was divided into 70% for training and 30% for validation, using stratified random sampling. MinMaxScaler normalization was applied to standardize the scales of the input variables.
The training was conducted over 500 epochs with dynamic adjustment of the learning rate. To prevent overfitting, early stopping and cross-validation were applied. The main training results are shown in Table 3.
Table 3. Training results
|
No. |
East 1 |
North 1 |
Level 1 |
This 2 |
North 2 |
Installment 2 |
L |
H |
|
1 |
550735.23 |
8526059.64 |
4090.70 |
550708.00 |
8526081.63 |
4090.70 |
35.00 |
1.55 |
|
2 |
677389.85 |
8444832.16 |
4333.00 |
677386.11 |
8444891.04 |
4333.00 |
59.00 |
1.75 |
|
3 |
677153.03 |
8445158.23 |
4336.90 |
677183.87 |
8445210.86 |
4336.90 |
61.00 |
1.30 |
|
… |
… |
… |
… |
… |
… |
… |
… |
… |
|
241 |
703073.44 |
8324799.43 |
4799.20 |
703078.30 |
8324725.99 |
4799.20 |
73.60 |
1.50 |
|
242 |
703469.59 |
8325483.75 |
4779.40 |
703441.02 |
8325551.15 |
4779.40 |
73.20 |
0.85 |
4.6 Performance evaluation
The model obtained R² values > 0.90 for the length and height predictions, with mean errors of MAE = 3.44 m in length and MAE = 0.09 m in height, demonstrating consistent predictive capacity with respect to the values documented in technical files. A detailed comparison between predicted and observed values is presented in Table 4.
Table 4. Comparison between predicted and observed values
|
No. |
This 1 Pred |
This 1 Real |
North 1 Pred |
North 1 Royal |
Elev 1 Pred |
Elev 1 Real |
|
1 |
547621.25 |
547592.32 |
8531093 |
8531027.41 |
4492.16 |
4479.9 |
|
2 |
547543.25 |
547583.85 |
8527010 |
8527001.98 |
4494.58 |
4507.6 |
|
3 |
552416.19 |
552458.83 |
8537658 |
8537681.16 |
4406.32 |
4438 |
|
10 |
585927.5 |
585877.09 |
8583604 |
8583537.68 |
4253.19 |
4261.4 |
|
20 |
642431.38 |
642483.85 |
8379206 |
8379115.74 |
4569.09 |
4544 |
The predicted and actual dam dimensions and their associated errors are summarized in Table 5.
Table 5. Predictions of the dam dimensions
|
No. |
Predicted Length |
Actual Length |
Predicted Height |
Actual Height |
Error Length (%) |
Alternative Error (%) |
|
1 |
49.42 |
44.5 |
1.74 |
1.7 |
11.06 |
2.35 |
|
2 |
97.4 |
94.8 |
1.26 |
1.25 |
2.74 |
0.8 |
|
3 |
49.42 |
37.6 |
2.93 |
2.75 |
31.44 |
6.55 |
|
10 |
35.92 |
32.1 |
2.5 |
2.5 |
11.9 |
0.04 |
|
20 |
58.79 |
57.3 |
1.51 |
1.4 |
2.6 |
7.86 |
The results obtained allow us to evaluate the capacity of the ANN model to predict the design parameters of cyclopean concrete dams in high Andean micro-basins, comparing the estimated values with those specified in the technical documents. The evaluation focused on three main parameters: (i) dam length, (ii) maximum height, and (iii) location of the closure axis.
5.1 Performance of the model with ANNs
The model achieved coefficient of determination values greater than R² > 0.90 for the length and height variables, indicating a high correlation between the predictions and the actual values. The average mean area under the curve (MAE) was 3.44 m for length and 0.09 m for height, demonstrating acceptable accuracy in terms of preliminary design. The detailed error metrics for each variable are listed in Table 6.
Regarding the location of the axis, the discrepancies remained within a range compatible with operational tolerances for small and medium-scale projects, considering that the exact definition is subsequently adjusted in the field through a complementary topographic survey.
Table 6. Error metrics of the model with ANN
|
Variable Statistics |
MAE |
MSE |
RMSE |
R2 |
|
East coordinate 1 |
52.4438 |
3269.1321 |
57.1763 |
1.0000 |
|
North coordinate 1 |
52.7511 |
3676.1906 |
60.6316 |
1.0000 |
|
Elevation 1 |
28.7834 |
1163.3739 |
34.1083 |
0.9760 |
|
East coordinate 2 |
45.2600 |
2740.0157 |
52.3452 |
1.0000 |
|
North coordinate 2 |
52.4019 |
3695.0130 |
60.7866 |
1.0000 |
|
Elevation 2 |
32.9644 |
1525.4094 |
39.0565 |
0.9686 |
|
Length |
3.4365 |
21.9413 |
4.6842 |
0.9243 |
|
Height |
0.0854 |
0.0094 |
0.0970 |
0.9664 |
5.2 Comparative analysis with empirical methods
Traditional empirical methods rely on professional experience and subjective criteria associated with the geomorphological interpretation of the terrain. Comparing the model results with dimensions estimated using these procedures revealed a 34.8% reduction in the variability of projected values, resulting in greater consistency between similar projects and, consequently, more efficient planning.
Figure 2 shows the linear regression between predicted and actual East 1 coordinates. As illustrated in Figure 3, the correlation between predicted and actual North 1 coordinates was also strong. Figure 4 presents the regression between predicted and actual elevation values for the closure axis. The relationship between predicted and observed dike lengths is displayed in Figure 5. The regression between predicted and observed dam heights is illustrated in Figure 6.
Figure 2. Linear regression of the East 1 coordinates: Predicted values versus actual values
Figure 3. Linear regression of North 1 coordinates: Predicted values versus actual values
Figure 4. Linear regression of Elevation 1: Predicted values versus actual values
Figure 5. Linear regression of dike length: Predicted values versus actual values
Figure 6. Linear regression of the dam height: Predicted values versus actual values
5.3 Influence of topographic variables on prediction
Synaptic weight and model sensitivity analyses showed that the variables with the greatest influence on prediction were:
This coincides with hydrological and structural studies that establish that the shape and energy of the relief determine the optimal geometry of the reservoir [16, 22].
5.4 Engineering interpretation of the results
The observed behavior confirms that the relationship between basin morphometry and dam dimensions is non-linear, which explains why:
Furthermore, the model performs better in micro-basins with moderate slopes and medium volumes, while in very narrow and deep basins, the error tends to increase slightly due to greater geometric variability, which is expected and does not compromise applicability in preliminary design.
5.5 Implications for water harvesting and storage projects
The proposed model allows:
This contributes to improving the sustainability and efficiency of public investments in hydraulic infrastructure in vulnerable areas.
The use of ANNs proved to be an effective tool for the preliminary design of cyclopean concrete dams in high-altitude Andean micro-basins, as it allows for the modeling of nonlinear relationships between topographic and hydrological variables and the geometric parameters of the structure. The results showed R² values greater than 0.90, indicating high accuracy in predicting the length and height of the dams.
The integration of data from DEMs and technical reports allowed for the construction of a standardized database for model training, reducing reliance on empirical and subjective criteria during the project formulation stage. The quality of the DEM directly influenced the model's stability, confirming the importance of its selection in environments with high geomorphological variability.
The comparison between the proposed model and conventional empirical methods showed a 34.8% reduction in design variability, suggesting that the RNA-based approach contributes to greater dimensional consistency and can optimize time and resources in the planning of water harvesting and storage projects.
The variables identified as having the greatest influence on the predictions were the micro-basin area, the estimated storage volume, and the average slope of the closure section, which aligns with hydrological and geomorphological principles reported in the literature. This validates the appropriateness of the model's input variable selection.
The proposed model is a useful tool for supporting decision-making during the pre-design phase, particularly in contexts where technical and implementation resources are limited. Its use allows for the standardization of criteria in the formulation of hydraulic projects in high Andean regions, strengthening the efficiency and sustainability of these interventions.
The authors gratefully acknowledge the institutional support provided to this research by the National University of San Cristóbal de Huamanga. They also acknowledge the collaboration of the Executive Unit of the Sierra Azul Fund (UEFSA) for facilitating access to the technical files of the planting and water harvesting projects, which formed the core database of this study.
[1] Biazar, S.M., Golmohammadi, G., Nedhunuri, R.R., Shaghaghi, S., Mohammadi, K. (2025). Artificial intelligence in hydrology: Advancements in soil, water resource management, and sustainable development. Sustainability, 17(5): 27. https://doi.org/10.3390/su17052250
[2] Goyburo, A., Rau, P., Lavado-Casimiro, W., Buytaert, W., Cuadros-Adriazola, J., Horna, D. (2023). Assessment of present and future water security under anthropogenic and climate changes using WEAP model in the Vilcanota-Urubamba catchment, Cusco, Perú. Water, 15(7): 1439. https://doi.org/10.3390/w15071439
[3] Sharma, A., Roy, M., Jha, V., Kumar, B., Singh, V.P. (2022). Velocity distribution in seepage-affected alluvial channels using Renyi entropy. Journal of Hydrologic Engineering, 27(6): 04022008. https://doi.org/10.1061/(ASCE)HE.1943-5584.0002180
[4] Muñoz-Carpena, R., Carmona-Cabrero, A., Yu, Z., Fox, G., Batelaan, O. (2023). Convergence of mechanistic modeling and artificial intelligence in hydrologic science and engineering. PLoS Water, 2(8): e0000059. https://doi.org/10.1371/journal.pwat.0000212
[5] Luo, D., Wang, K., Wang, D., Sharma, A., Li, W., Choi, I.H. (2025). Artificial intelligence in the design, optimization, and performance prediction of concrete materials: A comprehensive review. npj Materials Sustainability, 3(1): 14. https://doi.org/10.1038/s44296-025-00058-8
[6] Li, Z., Yoon, J., Zhang, R., Rajabipour, F., Srubar III, W.V., Dabo, I., Radlińska, A. (2022). Machine learning in concrete science: Applications, challenges, and best practices. npj computational materials, 8(1): 127. https://doi.org/10.1038/s41524-022-00810-x
[7] Wu, W., Su, H., Feng, Y., Zhang, S., Zheng, S., Cao, W., Liu, H. (2024). A novel artificial intelligence prediction process of concrete dam deformation based on a stacking model fusion method. Water, 16(13): 1868. https://doi.org/10.3390/w16131868
[8] Filo, G. (2023). Artificial intelligence methods in hydraulic system design. Energies, 16(8): 3320. https://doi.org/10.3390/en16083320
[9] Krishnan, S.R., Nallakaruppan, M.K., Chengoden, R., Koppu, S., Iyapparaja, M., Sadhasivam, J., Sethuraman, S. (2022). Smart water resource management using Artificial Intelligence—A review. Sustainability, 14(20): 13384. https://doi.org/10.3390/su142013384
[10] Plevris, V., Papazafeiropoulos, G. (2024). AI in structural health monitoring for infrastructure maintenance and safety. Infrastructures, 9(12): 225. https://doi.org/10.3390/infrastructures9120225
[11] Sieg, P., Posadzińska, I., Jóźwiak, M. (2023). Academic entrepreneurship as a source of innovation for sustainable development. Technological Forecasting and Social Change, 194: 122695. https://doi.org/10.1016/j.techfore.2023.122695
[12] Quiroz, J.A., Garcia-Chevesich, P.A., Martínez, G., Martínez, K., Tejada-Purizaca, T., Murray, K.E., McCray, J.E. (2023). Water resources evaluation and sustainability considering climate change and future anthropic demands in the Arequipa region of southern Peru. Sustainability, 15(23): 16270. https://doi.org/10.3390/su152316270
[13] Eghbalian, M., Pouragha, M., Wan, R. (2023). A physics-informed deep neural network for surrogate modeling in classical Elasto-plasticity. Computers and Geotechnics, 159: 105472. https://doi.org/10.1016/j.compgeo.2023.105472
[14] Lagaros, N.D. (2023). Artificial neural networks applied in civil engineering. Applied Sciences, 13(2): 1131. https://doi.org/10.3390/app13021131
[15] Malekloo, A., Ozer, E., AlHamaydeh, M., Girolami, M. (2022). Machine learning and structural health monitoring overview with emerging technology and high-dimensional data source highlights. Structural Health Monitoring, 21(4): 1906-1955. https://doi.org/10.1177/14759217211036880
[16] Dávila-Hernández, S., González-Trinidad, J., Júnez-Ferreira, H.E., Bautista-Capetillo, C.F., Morales de Ávila, H., Cázares Escareño, J., Ortiz-Letechipia, J., Robles Rovelo, C.O., López-Baltazar, E.A. (2022). Effects of the digital elevation model and hydrological processing algorithms on the geomorphological parameterization. Water, 14(15): 2363. https://doi.org/10.3390/w14152363
[17] Mukherjee, S., Joshi, P.K., Mukherjee, S., Ghosh, A., Garg, R.D., Mukhopadhyay, A. (2013). Evaluation of vertical accuracy of open source Digital Elevation Model (DEM). International Journal of Applied Earth Observation and Geoinformation, 21: 205-217. https://doi.org/10.1016/j.jag.2012.09.004
[18] Gunathilake, M.B., Karunanayake, C., Gunathilake, A.S., Marasingha, N., Samarasinghe, J.T., Bandara, I.M., Rathnayake, U. (2021). Hydrological models and artificial neural networks (ANNs) to simulate streamflow in a tropical catchment of Sri Lanka. Applied Computational Intelligence and Soft Computing, 2021(1): 6683389. https://doi.org/10.1155/2021/6683389
[19] Galletti, A., Avesani, D., Bellin, A., Majone, B. (2021). Detailed simulation of storage hydropower systems in large Alpine watersheds. Journal of Hydrology, 603(Part D): 127125. https://doi.org/10.1016/j.jhydrol.2021.127125
[20] Lin, K.R., Chen, H.Y., Xu, C.Y., Yan, P., Lan, T., Liu, Z.Y., Dong, C.Y. (2020). Assessment of flash flood risk based on improved analytic hierarchy process method and integrated maximum likelihood clustering algorithm. Journal of Hydrology, 584: 124696. https://doi.org/10.1016/j.jhydrol.2020.124696
[21] Jasim, A.M., Abdulaal, A.H., Albaker, B.M., Alwan, M.S. (2024). High-gain cubic boost converter analysis with hybrid ANN-incremental conductance MPPT for solar PV systems. Mathematical Modelling of Engineering Problems, 11(12): 3379-3390. https://doi.org/10.18280/mmep.111216
[22] Bekri, E.S., Economou, P., Yannopoulos, P.C., Demetracopoulos, A.C. (2021). Reassessing existing reservoir supply capacity and management resilience under climate change and sediment deposition. Water, 13(13): 1819. https://doi.org/10.3390/w13131819
[23] Mata, J. (2011). Interpretation of concrete dam behaviour with artificial neural network and multiple linear regression models. Engineering Structures, 33(3): 903-910. https://doi.org/10.1016/j.engstruct.2010.12.011
[24] Wang, X.X., Jin, Z.Q., Liu, J.P., Chen, F.X., Feng, P., Tang, J.H. (2021). Research on internal monitoring of reinforced concrete under accelerated corrosion, using XCT and DIC technology. Construction and Building Materials, 266(Part A): 121018. https://doi.org/10.1016/j.conbuildmat.2020.121018
[25] Prayitno, A., Indrawati, V., Yusuf, Y.G. (2024). Synchronization of heterogeneous vehicle platoon using distributed PI controller designed based on cooperative observer. Mathematical Modelling of Engineering Problems, 11(6): 1558-1566. https://doi.org/10.18280/mmep.110616
