Skip to main content
AkademIndex

Products

For developers

AkademBasesoonOpen API for the ecosystem
Latin
English
Article

Artificial Neural Network-Based Prediction of Soil Settlement

Archit PriyadarshiTula’s Institute,Dept of Civil Engineering,Dehradun,IndiaMuyassar AllaberganovaUrgench State University,Department of Data Transmission Networks and Systems,Urgench,UzbekistanAnant DeogaonkarNagpur Symbiosis International (Deemed) University,Symbiosis Institute of Business Management,PuneChandan VichorayRamdeobaba University,Dept. of Computer Science & Engg,Nagpur,IndiaM. P. S. BishtTula’s Institute,Dept of Civil Engineering,Dehradun,IndiaRishabh Mandal
2025
ABI

Abstract

The estimation of sediment settlement on river basins is a challenging and poorly understood subject. Numerous formulae based on various theoretical or practical techniques have been included in the geotechnical literature to produce an exact or nearly accurate prediction of such settling. However, these approaches consistently fall short when it comes to precise settlement prediction. Current techniques for predicting settlements mostly concentrate on estimating the greatest amount of ground settlements by employing static influencing elements, such as soil properties. However, these methods cannot be used directly to forecast daily ground settlements since various time-dependent influencing factors can have an impact on daily ground settlements, and such factors should be taken into account for an accurate daily ground settlement prediction. This paper presents a daily ground settlement forecast approach based on artificial neural networks to address this issue. The optimal artificial neural network takes into account both time-dependent and static effect elements, as well as historical settlement monitoring data. Statistical metrics were employed to assess the various artificial neural networks that were built using the Gradient Descent and Levenberg-Marquardt training methods. The Levenberg-Marquardt training algorithm has produced more accurate results than the Gradient Descent training algorithm, according to statistical research.

Topics

Identifiers

Citations and references

Cited by 023 references
Metrics — AkademScholar · Coming soon