Application of dimensional analysis and regression tools to estimate swell pressure of expansive soil in Tebessa (Algeria)

2018 
In the Tebessa area, the nature of geological formations in the subsurface is influenced considerably the superstructures built on them by the shrink-swell phenomenon. In many places this phenomenon has caused severe damage to dwellings, roads, and sewage networks. Cracks on all walls, staircases widening, disturbed windows and doors can be observed in many constructions throughout the area of study. The problem is worsening as urbanization is more and more extended to less adequate places. The phenomenon of swelling in soils is affected by many parameters such as the initial state of hydration, the percent of the fine fraction, and the stress history expressed in terms of the soil density and the preconsolidation state. In this research work, we expect to obtain a reliable model from the use of the dimensional analysis technique to predict the swelling pressure generated by the clayey soil in this area. The derivation of the dimensionless number was carried out by introducing a maximum number of parameters (independent or dependent), which are likely to affect the behavior of the clayey soil of the region. Statistical analysis tools were introduced as a simple technique in order to take into account the combined effects of the different soil properties on swelling behavior and deal with the models of swelling pressure. In the selection of independent variables, regression analyses were carried out, and a multicollinearity problem was considered. Hence, two independent variables were found to have a strong correlation with the swelling pressure. When extending to multiple regression analyses, regressors were added to obtain the final model. The present models show better results when compared with previous models cited in the literature. The model obtained using dimensional analysis fits better Tebessa’s soil conditions and will do so for all soil conditions, as it can easily be applied to any soil.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    29
    References
    9
    Citations
    NaN
    KQI
    []