Prediction of Swelling Parameters of Two Clayey Soils from Algeria Using Artificial Neural Networks

Abstract

The phenomenon of swelling is one of the more complicated geotechnical problems that the engineer have to deal with. However, its quantification is essential for the design of structures and various methods can be applied to the identification of this phenomenon. Some, such as mineralogical identification and direct measurements of swelling, are more or less long and require very specific equipment. However, there are other methods that offer the advantage of being relatively fast and lesser expensive: they are based on soil mechanics parameters. Using these parameters, several authors have introduced soil swelling prediction models, mostly in the form of classifications and empirical formulas. This work concerns in the first part the identification and classification of the swelling potential of two clays located in north-western Algeria. Followed by a statistical analysis carried out to test the reliability of the observations for the estimation of the pressure and the swelling amplitude using a multiple linear regression. A second part is devoted to the development of a prediction method by artificial neural networks allowing the estimation of swelling parameters (pressure and amplitude) by minimizing the difference between the experimental measurements and the numerical results. Modeling by artificial neural networks is of great interest in the field of prediction. The application of two networks makes it possible to obtain good forecasts of the swelling parameters.

Novel Insights in Soil Mechanics: Integrating Experimental Investigation with Machine Learning for Unconfined Compression Parameter Prediction of Expansive Soil

“…Effective applications of various machine learning (ML) models have been found in geotechnical analysis. These models cover a broad range, including support vector machines (SVMs) [39], multivariate adaptive regression splines (MARS) [40,41], relevant vector machines (RVMs) [42], decision tree regression (DTR) [43], gradient boosting regression (GBR) [43], K nearest neighbor regression (KNR) [43], particle swarm optimization (PSO) [44], random forest regression (RFR) [43], extreme gradient boosting (XGBoost) [38,[45][46][47], extreme learning machines (ELMs) [40], symbolic regression (SR) [48,49], and artificial neural networks (ANNs) [50][51][52][53][54]. Table 1 summarizes several studies that employed machine learning in geotechnical engineering.…”

“…Throughout both training and testing phases, the ANN-GMDH and ANN models demonstrated remarkable consistency in their performance, with ANN-GMDH, ANN and SVR delivering the most accurate predictions. However, because of the widespread use of ANN in geotechnical engineering [50][51][52]54,66,[68][69][70][71][72] and consistently superior performance in achieving better metrics. ANNs build complex input-output models that can learn intricate relationships within multidimensional data.…”

Neural networks based linear (PCA) and nonlinear (ISOMAP) feature extraction for soil swelling pressure prediction (North East Algeria)

“…Artificial Neural Networks (ANNs) mimic human brains, learning relationships and interdependencies, and extracting information from data. They have potential for predicting soil properties [ 1 , 2 , 11 , [21] , [22] , [23] , [24] , [25] , [26] , [27] ].…”

Identificação de solos expansivos e colapsíveis no Nordeste brasileiro a partir de Redes Neurais Artificiais geradas em Pernambuco

“…As Redes Neurais Artificiais (RNA), além de outras aplicações, são excelentes mecanismos computacionais que, com base no aprendizado neural biológico, e a partir da análise de um banco de dados, podem realizar previsões, classificações, reconhecimento de padrões e aproximações de funções (Pessoa et al, 2021). Trabalhos como os de Basma, Barakat & Omar (2003), Moosavi Yazdanpanah & Doostmohammadi (2006 e Doris, Rizzo & Dewoolkar (2008), Ashayeri & Yasrebi (2009), Ikizler et al (2010) e Merouane & Mamoune (2018) desenvolveram RNA para capturar tendências no movimento vertical da superfície do solo e prever a porcentagem de expansão livre e a tensão de expansão de solo, comparando a relação entre os valores de tensões medidos e os valores obtidos pelas RNA, todas essas pesquisas alcançaram percentuais acima de 90% de correlação na previsão.…”