Prediction of the Horizontal Displacement of Mechanically Stabilized Earth Wall using Soft Computing

Hamza, Saloua; Lafifi, Brahim; Nemissi, Mohamed; Moussaoui, Abdelkrim; Rouaiguia, Ammar

doi:10.22115/scce.2025.1959

Prediction of the Horizontal Displacement of Mechanically Stabilized Earth Wall using Soft Computing

Document Type : Regular Article

Authors

Saloua Hamza ¹

Brahim Lafifi ²

Mohamed Nemissi ³

Abdelkrim Moussaoui ⁴

Ammar Rouaiguia ²

¹ Ph.D. Candidate, Laboratory of Civil Engineering and Hydraulics, 8 may 1945 Guelma University, Guelma, Algeria

² Professor, Laboratory of Civil Engineering and Hydraulics, 8 may 1945 Guelma University, Guelma, Algeria

³ Professor, Laboratory LabSTIC, 8 may 1945 Guelma University, Guelma, Algeria

⁴ Professor, Department of Electrical and Automatic Engineering Laboratory (LGEG), 8 may 1945 Guelma University, Guelma, Algeria

10.22115/scce.2025.1959

Abstract

Mechanically Stabilized Earth (MSE) walls are a construction method that incorporates artificial reinforcement elements like geosynthetics within the soil. This geotechnical engineering technique is favored for reinforced soil structures due to its cost-effectiveness and ease of implementation. The primary goal of this study was to predict the horizontal displacement (U_x) of MSE walls by evaluating the performance of five AI-based machine learning models: multilayer perceptron (MLP), support vector machine (SVM), k-nearest neighbor (KNN), lazy k-star (LKS), and random forest (RF). A dataset of 712 numerical models was created using the finite element software Plaxis 2D and was randomly divided into a 70% training set and a 30% testing set. Each sample contained eight input variables: reinforcement length (L), the vertical spacing of reinforcement (S_v), wall height (H), length of the embankment (B), angle of inclination of the embankment (α), stiffness of reinforcement (EA), friction angle (φ) of the embankment, and unit weight (γ), along with one output response, the horizontal displacement (U_x). The accuracy of the models was assessed using ten statistical metrics. The findings indicated that the MLP model performed the best with a higher coefficient of determination of 0.9498 and a lower mean absolute error of 0.0136, while the SVM model performed the worst (R²=0.7067, MAE=0.0438). Sensitivity analysis for the MLP model evaluated the relationships between the input parameters and the output response, revealing that wall height (H) had the highest correlation with U_x. In contrast, the stiffness of reinforcement (EA) showed relatively lower correlations with U_x.

Graphical Abstract

Prediction of the Horizontal Displacement of Mechanically Stabilized Earth Wall using Soft Computing

Highlights

· The horizontal displacement of the MSE wall system was computed using PLAXIS 2D

· Five different machine learning algorithms were implemented

· Ten different statistical metrics were used to evaluate the models' accuracy

· The MLP model reached the best performance with R² = 0.9498, while SVM had R² = 0.7067

Keywords

Mechanically stabilized

Earth (MSE) wall

Horizontal displacement

Plaxis

Prediction

Machine learning

Subjects

Machine Learning

[1] No NHIC. Design and Construction of Mechanically Stabilized Earth Walls and Reinforced Soil Slopes – Volume I. vol. I. 2009.

[2] Rajagopal G, Thiyyakkandi S. Numerical Evaluation of the Performance of Back-to-Back MSE Walls with Hybrid Select-Marginal Fill Zones. Transp Geotech 2020:100445. https://doi.org/10.1016/j.trgeo.2020.100445.

[3] Portelinha FHM, Bueno BS, Zornberg JG. Performance of geotextile reinforced soil wall in unsaturated poorly draining backfill soil conditions. Proc 5th … 2012;5:455–9.

[4] Muteb HH, Falah MW. Mechanically Stabilized Earth MSE Walls Applications: Review. IOP Conf Ser Earth Environ Sci 2021;856. https://doi.org/10.1088/1755-1315/856/1/012031.

[5] Chaabani W, Remadna MS, Abu-Farsakh M. Numerical Modeling of the Ultimate Bearing Capacity of Strip Footings on Reinforced Sand Layer Overlying Clay with Voids. Infrastructures 2023;8:1–20. https://doi.org/10.3390/infrastructures8010003.

[6] Christopher BR, Stulgis RP. Low Permeable Backfill Soils in Geosynthetic Reinforced Soil Walls: State-of-the-Practice in North America. Proc North Am Geo-Synthetics Conf (NAGS 2005) 2005:14–6.

[7] Hossain MS, Kibria G, Khan MS, Hossain J, Taufiq T. Effects of Backfill Soil on Excessive Movement of MSE Wall. J Perform Constr Facil 2012;26:793–802. https://doi.org/10.1061/(ASCE)CF.1943-5509.0000281.

[8] Koerner RM, Koerner GR. The importance of drainage control for geosynthetic reinforced mechanically stabilized earth walls. J Geoengin 2011;6:3–13. https://doi.org/10.6310/jog.2011.6(1).1.

[9] Abu-farsakh M, Hanandeh S, Mohammad L, Chen Q. Geotextiles and Geomembranes Performance of geosynthetic reinforced / stabilized paved roads built over soft soil under cyclic plate loads. Geotext Geomembranes 2016. https://doi.org/10.1016/j.geotexmem.2016.06.009.

[10] Lafifi B, Rouaiguia A, Soltani EA. A Novel Method for Optimizing Parameters influencing the Bearing Capacity of Geosynthetic Reinforced Sand Using RSM, ANN, and Multi-objective Genetic Algorithm. Stud Geotech Mech 2023;45:174–96. https://doi.org/10.2478/sgem-2023-0006.

[11] Blivet JC, Gourc JP, Villard P, Giraud H, Khay M, Morbois A. Design method for geosynthetic as reinforcement for embankment subjected to localized subsidence. 7th Int Conf Geosynth 2002:341–4.

[12] Shukla SK. Geosynthetics and Their Applications. Thomas Telford Publ London, UK, 430 P 2002;3:31173. https://doi.org/10.1680/gata.31173.

[13] Chen Q, Hanandeh S, Abu-Farsakh M, Mohammad L. Performance evaluation of full-scale geosynthetic reinforced flexible pavement. Geosynth Int 2018;25:26–36. https://doi.org/10.1680/jgein.17.00031.

[14] Koseki J, Shibuya S. Mitigation of Disasters by Earthquakes, Tsunamis, and Rains by Means of Geosynthetic-Reinforced Soil Retaining Walls and Embankments. Transp Infrastruct Geotechnol 2014;1:231–61. https://doi.org/10.1007/s40515-014-0009-0.

[15] Skinner GD, Rowe RK. Design and behaviour of geosynthetic-reinforced soil walls constructed on yielding foundations. Geosynth Int 2003;10:200–14. https://doi.org/10.1680/gein.2003.10.6.200.

[16] Sayed S, Dodagoudar GR, Rajagopal K. Reliability analysis of reinforced soil walls under static and seismic forces 2008. https://doi.org/10.1680/gein.2008.15.4.246.

[17] Yu Y, Bathurst RJ. Probabilistic assessment of reinforced soil wall performance using response surface method. Geosynth Int 2017;24:524–42. https://doi.org/10.1680/jgein.17.00019.

[18] Nasser Sekfali BL. Mechanically Stabilized Earth Wall Reliability Analysis Using Response Mechanically Stabilized Earth Wall Reliability Analysis Using Response Surface Methodology , ANN , ANFIS and Multi-objective Genetic Algorithm 2024. https://doi.org/10.3311/PPci.22830.

[19] Umer M, Khan A, Kumar S. Transportation Geotechnics Numerical investigation of the structural response of a conduit buried within a soil slope. Transp Geotech 2021;30:100614. https://doi.org/10.1016/j.trgeo.2021.100614.

[20] Kim MK, Cho SH, Yun IJ, Won JH. Three-dimensional responses of buried corrugated pipes and ANN-based method for predicting pipe deflections. Int J Numer Anal Methods Geomech 2012;36:1–16. https://doi.org/10.1002/nag.986.

[21] Adombi AVDP, Chesnaux R, Boucher MA. Comparing numerical modelling, traditional machine learning and theory-guided machine learning in inverse modeling of groundwater dynamics: A first study case application. J Hydrol 2022;615. https://doi.org/10.1016/j.jhydrol.2022.128600.

[22] Kaloop MR, Bardhan A, Kardani N, Samui P, Hu JW, Ramzy A. Novel application of adaptive swarm intelligence techniques coupled with adaptive network-based fuzzy inference system in predicting photovoltaic power. Renew Sustain Energy Rev 2021;148:111315. https://doi.org/10.1016/j.rser.2021.111315.

[23] Abualigah L, Alkhrabsheh M. Amended hybrid multi-verse optimizer with genetic algorithm for solving task scheduling problem in cloud computing. J Supercomput 2022. https://doi.org/10.1007/s11227-021-03915-0.

[24] Abualigah L, Alkhrabsheh M. Amended hybrid multi ‑ verse optimizer with genetic algorithm for solving task scheduling problem in cloud computing. J Supercomput 2021. https://doi.org/10.1007/s11227-021-03915-0.

[25] Mirjalili S, Gandomi AH, Mirjalili SZ, Saremi S, Faris H, Mirjalili SM. Salp Swarm Algorithm: A bio-inspired optimizer for engineering design problems. Adv Eng Softw 2017;114:163–91. https://doi.org/10.1016/j.advengsoft.2017.07.002.

[26] Goudjil K, Arabet L. Assessment of deflection of pile implanted on slope by artificial neural network. Neural Comput Appl 2021;33:1091–101. https://doi.org/10.1007/s00521-020-04985-6.

[27] Belaabed F, Goudjil K, Arabet L, Ouamane A. Utilization of computational intelligence approaches to estimate the relative head of PK-Weir for submerged flow. Neural Comput Appl 2021;33:13001–13. https://doi.org/10.1007/s00521-021-05996-7.

[28] Lafifi B, Hamrouni A, Khoualdia T, Gheris A, Rouaiguia A. Prediction and optimization of the bearing capacity of strip footing resting on soft soil improved with stone columns using RSM, ANN, and multi-objective GA. Innov Infrastruct Solut 2024;9. https://doi.org/10.1007/s41062-024-01452-2.

[29] Goudjil K, Boukhatem G, Boulifa R, Bekkouche SR, Djihane D. Prediction and interpretation of limit pressure of clayey soils using ensemble machine learning methods and shapely additive explanations. Stud Eng Exact Sci 2024;5:e5567. https://doi.org/10.54021/seesv5n2-017.

[30] Hatami K, Bathurst RJ. Development and verification of a numerical model for the analysis of geosynthethic-reinforced soil segmental walls under working stress conditions. Can Geotech J 2005;42:1066–85. https://doi.org/10.1139/t05-040.

[31] Ali T, Haider W, Ali N, Aslam M. A Machine Learning Architecture Replacing Heavy Instrumented Laboratory Tests: In Application to the Pullout Capacity of Geosynthetic Reinforced Soils. Sensors 2022;22. https://doi.org/10.3390/s22228699.

[32] Abdelouhab A, Dias D, Freitag N. Geotextiles and Geomembranes Numerical analysis of the behaviour of mechanically stabilized earth walls reinforced with different types of strips 2011;29:116–29. https://doi.org/10.1016/j.geotexmem.2010.10.011.

[33] Flavigny E, Desrues J, Palayer B. Note technique : le sable d’Hostun « RF ». Rev Française Géotechnique 1990:67–70. https://doi.org/10.1051/geotech/1990053067.

[34] Ajibona AI, Taiwo BO, Afeni TB, Akinbinu VA. Development of Efficient Empirical Models for the Prediction of Oil Well Fracture Pressure Gradient. Daffodil Int Univ J Schience Technol 2022;17:1–9.

[35] Lawal AI, Kwon S, Young G. Prediction of the blast ‑ induced ground vibration in tunnel blasting using ANN , moth ‑ flame optimized ANN , and gene expression programming. Acta Geophys 2021;69:161–74. https://doi.org/10.1007/s11600-020-00532-y.

[36] Sadowski L, Wang D. Pull-off adhesion prediction of variable thick overlay to the substrate 2018. https://doi.org/10.1016/j.autcon.2017.10.001.

[37] Thai B, Tien D, Prakash I, Dholakia MB. Catena Hybrid integration of Multilayer Perceptron Neural Networks and machine learning ensembles for landslide susceptibility assessment at Himalayan area ( India ) using GIS. Catena 2017;149:52–63. https://doi.org/10.1016/j.catena.2016.09.007.

[38] Gao W, Alsarraf J, Moayedi H, Shahsavar A, Nguyen H. Comprehensive preference learning and feature validity for designing energy-efficient residential buildings using machine learning paradigms. Appl Soft Comput J 2019;84:105748. https://doi.org/10.1016/j.asoc.2019.105748.

[39] Moayedi H, Nguyen H, Safuan A. Novel metaheuristic classification approach in developing mathematical model ‑ based solutions predicting failure in shallow footing. Eng Comput 2019. https://doi.org/10.1007/s00366-019-00819-9.

[40] Burr T. Pattern Recognition and Machine Learning. J Am Stat Assoc 2008;103:886–7. https://doi.org/10.1198/jasa.2008.s236.

[41] Cortes C, Vapnik V. Support-vector networks. Mach Learn 1995;20:273–97. https://doi.org/10.1007/BF00994018.

[42] Hsu C, Chang C, Lin C. A Practical Guide to Support Vector Classification 2016;1:1–16.

[43] Barakat N, Bradley AP. Neurocomputing Rule extraction from support vector machines : A review. Neurocomputing 2010;74:178–90. https://doi.org/10.1016/j.neucom.2010.02.016.

[44] Hasilci B, Mumcu TV. Parameter Estimation of BLDC Motors by SVM for UAV Propulsion Systems İnsansız Hava Aracı Tahrik Sistemleri İçin SVM ile BLDC Motorların Parametre Kestirimi 1 . Introduction UAVs , which are an important part of the rapidly developing world technologies , 2022;15:406–19. https://doi.org/10.18185/erzifbed.930222.

[45] Altman NS. An introduction to kernel and nearest-neighbor nonparametric regression. Am Stat 1992;46:175–85. https://doi.org/10.1080/00031305.1992.10475879.

[46] Ho TK. Random decision forests. Proc. Int. Conf. Doc. Anal. Recognition, ICDAR, vol. 1, 1995, p. 278–82. https://doi.org/10.1109/ICDAR.1995.598994.

[47] Breiman LEO. Random Forests 2001:5–32.

[48] Cleary JG, Trigg LE. K*: An Instance-based Learner Using an Entropic Distance Measure. Mach. Learn. Proc. 1995, vol. 5, Elsevier; 1995, p. 108–14. https://doi.org/10.1016/B978-1-55860-377-6.50022-0.

[49] Webb G. Lazy Learning. In: Sammut C, Encyclopedia of Machine Learning. 2011.

[50] Taylor KE. in a single diagram 2001;106:7183–92.

[51] Saltelli A, Annoni P, Hombres BD. Sixth International Conference on Sensitivity Analysis of Model Output How to avoid a perfunctory sensitivity analysis. Soins Aides-Soignantes 2010;2:7592–4. https://doi.org/10.1016/j.sbspro.2010.05.133.

[52] Rajagopal G, Thiyyakkandi S. Influence of Retained Zone Width on the Behaviour of Back ‑ to ‑ Back Mechanically Stabilized Earth Walls upon Rainwater Infiltration. Int J Geosynth Gr Eng 2022;8:1–14. https://doi.org/10.1007/s40891-022-00416-9.

[53] Hamrouni A, Dias D, Sbartai B. Reliability analysis of a mechanically stabilized earth wall using the surface response methodology optimized by a genetic algorithm. Geomech Eng 2018;15:937–45. https://doi.org/10.12989/gae.2018.15.4.937.

[54] Rezaei M, Asadizadeh M, Majdi A, Hossaini MF. International Journal of Mining Science and Technology Prediction of representative deformation modulus of longwall panel roof rock strata using Mamdani fuzzy system. Int J Min Sci Technol 2015;25:23–30. https://doi.org/10.1016/j.ijmst.2014.11.007.

[55] Ghorbani B, Arulrajah A, Narsilio G, Horpibulsuk S, Bo MW. Development of genetic-based models for predicting the resilient modulus of cohesive pavement subgrade soils. Soils Found 2020;60:398–412. https://doi.org/10.1016/j.sandf.2020.02.010.