Assessment of Micro-pile Group Capacity in Soft Clay Soils Using Closed-Form Machine Learning Models

Gnananandarao, Tammineni; Chukka, Naga Dheeraj Kumar Reddy; Kumar, B.A.V. Ram; Muktinutalapati, Jayatheja; Maralapalle, Vedprakash; Kumar, Mummidivarapu Satish; Ajay, CH.; Uma Maheswara Rao, P.

doi:10.22115/scce.2025.501243.2023

Assessment of Micro-pile Group Capacity in Soft Clay Soils Using Closed-Form Machine Learning Models

Document Type : Regular Article

Authors

Tammineni Gnananandarao ¹

Naga Dheeraj Kumar Reddy Chukka ¹

B.A.V. Ram Kumar ²

Jayatheja Muktinutalapati ³

Vedprakash Maralapalle ⁴

Mummidivarapu Satish Kumar ¹

CH. Ajay ¹

P. Uma Maheswara Rao ¹

¹ Department of Civil Engineering, Aditya University, Surampalem, Andhra Pradesh, India, 533437

² Department of Civil Engineering, GMR Institute of technology, Rajam, Andhra Pradesh, India

³ RICS School of Built Environment, Amity University Maharashtra, Mumbai, 410206, India

⁴ Associate Professor, L. S. Raheja School of Architecture, Bandra, Mumbai, 400051, India

10.22115/scce.2025.501243.2023

Abstract

The computation of the in-situ load-bearing capacity of micropiles for different soil types using conventional methods is often difficult and time-consuming. Machine learning models, on the other hand, offer promising outcomes by predicting real-time solutions without requiring extensive traditional calculations. In this study, models such as Artificial Neural Networks (ANN) and the M5P model tree are developed using a dataset of 434 experimental results obtained from previous studies. The accuracy of the ANN model is evaluated using the correlation coefficient (𝑅²), which was observed to be 0.99 for both training and testing, with mean absolute percentage errors (MAPE) of 18.38 and 7.67, respectively. Similarly, the M5P model tree achieved an 𝑅² of 0.99, with corresponding MAPE values of 23.92 and 10.01, which were slightly higher than those of the ANN model. The equations developed by both models demonstrated a strong correlation factor with minimal errors. The primary influencing parameters identified were Se/b and 𝑛. Although both models performed well, ANN was found to be more effective than the M5P model, as it exhibited lower relative errors and higher accuracy in predictions.

Graphical Abstract

Assessment of Micro-pile Group Capacity in Soft Clay Soils Using Closed-Form Machine Learning Models

Highlights

· ANN and M5P used to predict micropile capacity in soft clay

· Dataset of 434 cases with six dimensionless inputs

· ANN showed higher prediction accuracy than M5P

· Models produced usable predictive equations

· Sensitivity analysis identified key input effects

Keywords

Micro-piles

Soft clay soil

Artificial neural networks

M5P model

Subjects

Artificial Neural Networks

[1] PJ S, T A, P G, Keeley J, B. T. Micropile Design and Construction. https://rosap.ntl.bts.gov/view/dot/50231. Fed Highw Adm Washingt 2025.

[2] Tsukada Y, Ichimura Y. Micropiles in Japan: Present status and future prospects. Int. Workshop on Micropiles. Seattle, Japan 1997.

[3] Bruce DA, DiMillio AF, Juran I. Micropiles: the state of practice Part 1: Characteristics, definitions and classifications. Proc Inst Civ Eng - Gr Improv 1997;1:25–35. https://doi.org/10.1680/gi.1997.010104.

[4] Borthakur N, Dey AK. Experimental Investigation on Load Carrying Capacity of Micropiles in Soft Clay. Arab J Sci Eng 2018;43:1969–81. https://doi.org/10.1007/s13369-017-2894-3.

[5] Han J, Ye S-L. A field study on the behavior of micropiles in clay under compression or tension. Can Geotech J 2006;43:19–29. https://doi.org/10.1139/t05-089.

[6] Kershaw KA, Luna R. Full-Scale Field Testing of Micropiles in Stiff Clay Subjected to Combined Axial and Lateral Loads. J Geotech Geoenvironmental Eng 2014;140:255–61. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000968.

[7] Abd Elaziz AY, El Naggar MH. Group behaviour of hollow-bar micropiles in cohesive soils. Can Geotech J 2014;51:1139–50. https://doi.org/10.1139/cgj-2013-0409.

[8] Fayyad U, Piatetsky-Shapiro G, Smyth P. The KDD process for extracting useful knowledge from volumes of data. Commun ACM 1996;39:27–34. https://doi.org/10.1145/240455.240464.

[9] Shahin MA, Jaksa MB, Maier HR. Artificial neural network applications in geotechnical engineering. Aust Geomech 2001;36:49–62.

[10] Barkhordari MS, Fattahi H, Armaghani DJ, Khan NM, Afrazi M, Asteris PG. Failure mode identification in reinforced concrete flat slabs using advanced ensemble neural networks. Multiscale Multidiscip Model Exp Des 2024;7:5759–73. https://doi.org/10.1007/s41939-024-00554-9.

[11] Lee S., Lee S., Kim Y. An approach to estimate unsaturated shear strength using artificial neural network and hyperbolic formulation. Comput Geotech 2003;30:489–503. https://doi.org/10.1016/S0266-352X(03)00058-2.

[12] Zhang C, Zhu Z, Liu F, Yang Y, Wan Y, Huo W, et al. Efficient machine learning method for evaluating compressive strength of cement stabilized soft soil. Constr Build Mater 2023;392:131887. https://doi.org/10.1016/j.conbuildmat.2023.131887.

[13] Gnananandarao T, Dutta RK, Khatri VN, Kumar MS. Soft computing based prediction of unconfined compressive strength of fly ash stabilised organic clay. J Soft Comput Civ Eng 2022;6:43–58.

[14] Onyelowe KC, Gnananandarao T, Ebid AM. Estimation of the erodibility of treated unsaturated lateritic soil using support vector machine-polynomial and -radial basis function and random forest regression techniques. Clean Mater 2022;3:100039. https://doi.org/10.1016/j.clema.2021.100039.

[15] Noorzaei J, Hakim SJS, Jaafar MS. An approach to predict ultimate bearing capacity of surface footings using artificial neural network. Indian Geotech J 2008;38:515–28.

[16] Dutta RK, Gnananandarao T, Khatri VN. Application of soft computing techniques in predicting the ultimate bearing capacity of strip footing subjected to eccentric inclined load and resting on sand. J Soft Comput Civ Eng 2019;3:30–42.

[17] Gnananandarao T, Kumar Dutta R, Khatri VN. Tsunamigenic Seismic Activity (Earthquakes) Prediction from III- Component Seismic Data, 2021, p. 351–60. https://doi.org/10.1007/978-981-15-9976-7_31.

[18] Kohestani VR, Hassanlourad M, BAZARGAN LMR. Prediction the ultimate bearing capacity of shallow foundations on the cohesionless soils using M5P model tree 2016.

[19] López Paredes CR, García C, Onyelowe KC, Zuniga Rodriguez MG, Gnananandarao T, Andrade Valle AI, et al. Evaluating the impact of industrial wastes on the compressive strength of concrete using closed-form machine learning algorithms. Front Built Environ 2024;10. https://doi.org/10.3389/fbuil.2024.1453451.

[20] Gnananandarao T, Naik S V, Onyelowe KC, Panwar V. Forecasting Bearing Capacity, Error Analyses and Parametric Analysis of Circular Footing Seating on the Limited Thick Sand-Layer with Eccentric-Inclined Load. Civ Eng Infrastructures J 2025;58:87–102.

[21] Borthakur N, Dey AK. Evaluation of group capacity of micropile in soft clayey soil from experimental analysis using SVM-based prediction model. Int J Geomech 2020;20:4020008. https://doi.org/0.1061/%28ASCE%29GM.1943-5622.0001.

[22] Armaghani DJ, Liu Z, Khabbaz H, Fattahi H, Li D, Afrazi M. Tree-Based Solution Frameworks for Predicting Tunnel Boring Machine Performance Using Rock Mass and Material Properties. Comput Model Eng Sci 2024;141:2421–51. https://doi.org/10.32604/cmes.2024.052210.

[23] Yang B, Jahed Armaghani D, Fattahi H, Afrazi M, Koopialipoor M, Asteris PG, et al. Optimized Random Forest Models for Rock Mass Classification in Tunnel Construction. Geosciences 2025;15:47. https://doi.org/10.3390/geosciences15020047.

[24] Huat CY, Moosavi SMH, Mohammed AS, Armaghani DJ, Ulrikh DV, Monjezi M, et al. Factors Influencing Pile Friction Bearing Capacity: Proposing a Novel Procedure Based on Gradient Boosted Tree Technique. Sustainability 2021;13:11862. https://doi.org/10.3390/su132111862.

[25] Jahed Armaghani D, Hayati M, Momeni E, Dowlatshahi MB, Asteris PG. Estimation of powder factor in mine blasting: feasibility of tree-based predictive models. Multiscale Multidiscip Model Exp Des 2025;8:138. https://doi.org/10.1007/s41939-024-00725-8.

[26] Guide MU. Neural network toolbox. The MathWorks 2002;40.

[27] Ince R. Prediction of fracture parameters of concrete by Artificial Neural Networks. Eng Fract Mech 2004;71:2143–59. https://doi.org/10.1016/j.engfracmech.2003.12.004.

[28] Dutta RK, Rani R, Gnananandarao T. Prediction of ultimate bearing capacity of skirted footing resting on sand using artificial neural networks. J Soft Comput Civ Eng 2018;2:34–46.

[29] Acharyya R, Dey A, Kumar B. Finite element and ANN-based prediction of bearing capacity of square footing resting on the crest of c - φ soil slope. Int J Geotech Eng 2020;14:176–87. https://doi.org/10.1080/19386362.2018.1435022.

[30] Acharyya R, Dey A. Assessment of bearing capacity for strip footing located near sloping surface considering ANN model. Neural Comput Appl 2019;31:8087–100. https://doi.org/10.1007/s00521-018-3661-4.

[31] Dutta RK, Dutta K, Jeevanandham S. Prediction of Deviator Stress of Sand Reinforced with Waste Plastic Strips Using Neural Network. Int J Geosynth Gr Eng 2015;1:11. https://doi.org/10.1007/s40891-015-0013-7.

[32] Onyelowe KC, Kamchoom V, Gnananandarao T, Arunachalam KP. Developing data driven framework to model earthquake induced liquefaction potential of granular terrain by machine learning classification models. Sci Rep 2025;15:21509. https://doi.org/10.1038/s41598-025-07494-5.

[33] Panwar V, Dutta RK. Bearing capacity of rectangular footing on layered sand under inclined loading. J Achiev Mater Manuf Eng 2021;108:49–62. https://doi.org/10.5604/01.3001.0015.5064.

[34] Das SK, Basudhar PK. Undrained lateral load capacity of piles in clay using artificial neural network. Comput Geotech 2006;33:454–9. https://doi.org/10.1016/j.compgeo.2006.08.006.

[35] Ferreira FPV, Shamass R, Limbachiya V, Tsavdaridis KD, Martins CH. Lateral–torsional buckling resistance prediction model for steel cellular beams generated by Artificial Neural Networks (ANN). Thin-Walled Struct 2022;170:108592. https://doi.org/10.1016/j.tws.2021.108592.

[36] Panwar V, Dutta RK. Application of machine learning technique in predicting the bearing capacity of rectangular footing on layered sand under inclined loading. J Soft Comput Civ Eng 2022;6:130–52.

[37] Quinlan JR. Learning with continuous classes. 5th Aust. Jt. Conf. Artif. Intell., vol. 92, World Scientific; 1992, p. 343–8.

[38] Goh ATC, Kulhawy FH, Chua CG. Bayesian Neural Network Analysis of Undrained Side Resistance of Drilled Shafts. J Geotech Geoenvironmental Eng 2005;131:84–93. https://doi.org/10.1061/(ASCE)1090-0241(2005)131:1(84).

[39] Garson DG. Interpreting neural network connection weights 1991.

[40] Olden JD, Jackson DA. Illuminating the “black box”: a randomization approach for understanding variable contributions in artificial neural networks. Ecol Modell 2002;154:135–50. https://doi.org/10.1016/S0304-3800(02)00064-9.

[41] Onyelowe KC, Gnananandarao T, Nwa-David C. Sensitivity analysis and prediction of erodibility of treated unsaturated soil modified with nanostructured fines of quarry dust using novel artificial neural network. Nanotechnol Environ Eng 2021;6:37. https://doi.org/10.1007/s41204-021-00131-2.