All Issue

2025 Vol.24, Issue 2 Preview Page

Research Article

Quantitative Evaluation of AI-Based Field Pile Capacity Prediction: Analyzing the Effectiveness of Rebound Value

AI 기반 말뚝 현장 지지력 예측의 정량적 검토: 리바운드 값의 실효성 분석

Gunwoong Kim1*
김 건웅1*
1Member, Postdoctoral Researcher, Korea Institute of Civil Engineering. and Building Technology, KICT, 283 Goyang-daero, Ilsanseo-gu, Goyang-si, Gyeonggi-do 10223, Republic of Korea
*Corresponding Author
30 June 2025. pp. 15-24
PDF XML Citation
Abstract

Most domestic pile construction sites rely on set values (penetration per blow) to manage pile installation, but this approach fails to provide a reliable basis for evaluating structural safety. Managing pile capacity through predicted bearing capacity offers greater reliability. The Hiley formula, a representative method for capacity estimation, includes rebound values and often delivers high prediction accuracy. However, manual rebound measurements in the field often vary with site conditions and operator technique, reducing reliability and consistency. High-speed cameras or precision sensors enable more accurate rebound measurements, yet field deployment faces cost and practicality constraints. This study investigates whether AI models can accurately predict pile capacity without rebound data. By applying K-Nearest Neighbors (KNN) and Random Forest model, the analysis compares prediction accuracy between models with and without rebound input. Results show that models excluding rebound still achieve strong performance. Existing input variables appear to capture much of the rebound’s underlying information, minimizing its added value in practice. Although rebound holds theoretical significance, machine learning techniques can compensate for its absence. The findings support AI-based prediction of pile capacity without depending on unreliable or hard-to-measure field variables. This approach promotes practical deployment of smart construction technologies in pile foundation management.

Keywords
PHC pile
Rebound
Pile driving formula
KNN
Random forest
Bearing capacity

국내 말뚝 시공 현장에서는 대부분의 말뚝에 대해 관입량(set)을 기준으로 관리하고 있으나, 이는 공학적 안전성을 정량적으로 판단하는 데 한계가 있다. 이에 따라 더욱 신뢰성 있는 지지력 값을 활용한 관리 방식이 필요하다. 지지력 산정을 위한 대표적인 항타 공식인 Hiley 공식은 리바운드(rebound) 값을 포함함으로써 예측 정확도가 높지만, 현장에서 수기로 측정되는 리바운드 값은 현장 조건과 작업자에 따라 편차가 크며, 계측 신뢰성과 일관성이 떨어진다. 정확한 리바운드 계측을 위해 고속 카메라나 고정밀 센서가 요구되지만, 비용과 장비 측면에서 현실적인 적용이 어려운 실정이다. 본 연구는 이러한 실무적 제약을 고려하여, 리바운드 값을 제외한 입력 변수만으로도 말뚝 지지력을 신뢰성 있게 예측할 수 있는 가능성을 검토하였다. 이를 위해 K-최근접 이웃(K-Nearest Neighbors, KNN) 및 랜덤 포레스트(Random Forest) 모델을 활용하여 AI 기반 예측 모델을 구축하고, 리바운드 값의 포함 여부에 따른 예측 정확도를 정량적으로 비교·분석하였다. 분석 결과, 리바운드 값을 제외하더라도 높은 예측 성능을 확보할 수 있었으며, 이는 기존 입력 변수들이 리바운드 값의 정보를 상당 부분 내포하고 있음을 시사한다. 즉, 리바운드는 이론적으로 유의미한 변수이나, AI 기법을 통해 그 영향을 충분히 보완할 수 있다. 본 연구는 현장에서 확보가 어렵거나 신뢰성이 낮은 변수에 의존하지 않더라도, 정확도 높은 지지력 예측이 가능함을 입증하였다. 이는 향후 스마트 건설 기술을 기반으로 한 실용적 말뚝 시공 관리 시스템의 구축에 기여할 수 있을 것으로 기대된다.

References
1

Altman, N. S. (1992), "An introduction to kernel and nearest-neighbor nonparametric regression", The American Statistician, Vol.46, No.3, pp.175-185.

10.1080/00031305.1992.10475879
2

Breiman, L. (2001), "Random forests", Machine Learning, Vol.45, No.1, pp.5-32.

10.1023/A:1010933404324
3

Burba, F., Ferraty, F. and Vieu, P. (2009), "k-nearest neighbour method in functional nonparametric regression", Journal of Nonparametric Statistics, Vol.21, No.4, pp.453-469.

10.1080/10485250802668909
4

Cao, X. L., Dai, G. L., Gong, W. M., Zhu, M. X. and Tang, J. A. (2020), "Experimental study on the seismic behavior of new PHC piles", Arabian Journal of Geosciences, Vol.13, p.778.

10.1007/s12517-020-05819-4
5

Cho, C. W. and Lee, M. W. (2001), "The application of new pile driving formulas", Journal of the Korean Geotechnical Society, Vol.17, No.5, pp.157-164.

6

Goh, A. T. C. (1995), "Back-propagation neural networks for modeling complex systems", Artificial Intelligence in Engineering, Vol.9, No.3, pp.143-151.

10.1016/0954-1810(94)00011-S
7

Gregorutti, B., Michel, B. and Saint-Pierre, P. (2017), "Correlation and variable importance in random forests", Statistics and Computing, Vol.27, No.3, pp.659-678.

10.1007/s11222-016-9646-1
8

Gupta, R., Kewalramani, M. A. and Goel, A. (2006), "Prediction of concrete strength using neural-expert system," Journal of Materials in Civil Engineering, Vol.18, No.3, pp.462-466.

10.1061/(ASCE)0899-1561(2006)18:3(462)
9

Ha, I. S., Han, S. S., Han, S. G. and Kim, M. M. (2003), "Measurement of pile rebound and penetration using laser sensor", Proceedings of the Korean Society of Civil Engineering Conference, Daegu, Republic of Korea, October 24-25, pp.3257-3262.

10

Halder, R. K., Uddin, M. N., Uddin, M. A., Aryal, S. and Khraisat, A. (2024), "Enhancing K-nearest neighbor algorithm: A comprehensive review and performance analysis of modifications", Journal of Big Data, Vol.11, No.113.

10.1186/s40537-024-00973-y
11

Heo, S. J. and Jeong, R. (2021), "Cases of Artificial Intelligence Development in the Construction field According to the Artificial Intelligence Development Method", Proceedings of the Fall Conference of the Korean Institute of Building Construction, Vol.21, No.2, pp.217-218.

12

Huynh, V. H., Nguyen, T., Nguyen, D. P., Nguyen, T. S., Huynh, T. M. D. and Nguyen, T. C. (2022), "A novel direct SPT method to accurately estimate ultimate axial bearing capacity of bored PHC nodular piles with 81 case studies in Vietnam", Soils and Foundations, Vol.62, 101163.

10.1016/j.sandf.2022.101163
13

Jang, Y. E., Jung, J., Han, J. T. and Yu, Y. (2022), "Deep learning based estimation of depth to bearing layer from in-situ data", Journal of the Korean Geotechnical Society, Vol.38, No.3, pp.35-42.

14

Kim, B. J., Ho, J. K. and Hwang, Y. C. (2014), "Prediction of Shear Wave Velocity on Sand Using Standard Penetration Test Results : Application of Artificial Neural Network Model", Journal of the Korean Geotechnical Society, Vol.30, No.5, pp.47-54.

10.7843/kgs.2014.30.5.47
15

Kim, G., Seo, S., Kim, J., Choi, C. and Chung, M. (2024), "Enhancing quality management of PHC piles through improved driving formulas: A comprehensive review and analysis", Applied Sciences, Vol.14, No.15, p.6770.

10.3390/app14156770
16

Kim, J. Y., Jeong, I. G. and Byun, S. O. (2017), "Application of Smart Construction Technology for Innovation of Construction Works", Journal of the Korean Society of Civil Engineers, Vol.65, No.6, pp.20-23.

17

Kim, R. and Kim, D. (2019), "Analysis of Bearing Capacity of Rock Socketed Pre-Bored Super Strength Piles Based on Dynamic Load Test Results", Journal of the Korean Geosynthetics Society, Vol.18, No.3, pp.89-100.

18

Kou, H. L., Chu, J., Guo, W. and Zhang, M. Y. (2016), "Field study of residual forces developed in pre-stressed high-strength concrete (PHC) pipe piles", Canadian Geotechnical Journal, Vol.53, pp.696-707.

10.1139/cgj-2015-0177
19

Lee, J. K., Choi, J. S., Kim, T. H. and Geem, Z. W. (2022), "Calculation of Shear Strength of Rock Slope Using Deep Neural Network", Journal of the Korean Geosynthetics Society, Vol.21, No.2, pp.21-30.

20

Lee, S., Kim, Y., Kim, Y., Park, J. and Ji, B. (2025), "Advanced rock type classification using deep learning: A data-driven approach to tunnel stability monitoring and maintenance", Journal of the Korean Geosynthetics Society, Vol.24, No.1, pp.89-100.

21

Oh, S.W. and Lee, J. D. (2003), "Characteristics of bearing capacity and reliability-based evaluation of pile-driving formulas for H pile", Journal of the Korean Institute of Intelligent Systems, Vol.18, No.1, pp.81-88.

22

Shahin, M. A., Jaksa, M. B. and Maier, H. R. (2001), "Artificial neural network applications in geotechnical engineering," Australian Geomechanics Journal, Vol.36, No.1, pp.49-62.

23

Toloşi, L. and Lengauer, T. (2011), "Classification with correlated features: unreliability of feature ranking and solutions", Bioinformatics, Vol.27, No.14, pp.1986-1994.

10.1093/bioinformatics/btr30021576180
24

Yun, H. S., Seok, J. W., Hwang, D. J. and Kim, M. M. (2005), "Quality control of driven piles using DPRMs", Proceedings of the Korean Society of Civil Engineering Conference, Jeju, Republic of Korea, October 20-21, pp.3945-3948.

25

Zhu, T. (2020), "Analysis on the applicability of the random forest", Journal of Physics: Conference Series, Vol.1607, No.012123.

10.1088/1742-6596/1607/1/012123
Information
  • Publisher :Korean Geosythetics Society
  • Publisher(Ko) :한국지반신소재학회
  • Journal Title :Journal of the Korean Geosynthetics Society
  • Journal Title(Ko) :한국지반신소재학회 논문집
  • Volume : 24
  • No :2
  • Pages :15-24
  • Received Date : 2025-04-22
  • Revised Date : 2025-05-14
  • Accepted Date : 2025-05-20
  • DOI :https://doi.org/10.12814/jkgss.2025.24.2.015
Journal Informaiton Journal of the Korean Geosynthetics Society Journal of the Korean Geosynthetics Society
  • NRF
  • KOFST
  • KISTI Current Status
  • KISTI Cited-by
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close