Prediction of Consolidation Parameters of Soils Based on index Properties Using Regression Analysis & Geneti Programming
Record type:
paper
Creator:
Nig
Editor:
Ele
Publisher:
Nat
Host:
The determination of consolidation parameters like compression and recompression
indices in the laboratory is costly and time-consuming. Due to this, existing empirical
equations have been used by practicing engineers to determine these parameters.
However, these empirical equations are site specific and it is erroneous to use them for
soils of other localities for which they are not formulated. The prime objective of this study,
therefore, is to predict the compression and recompression indices of Addis Ababaβs red
clay soil based on the basic index properties. To develop the proposed models a laboratory
tests like consolidation and basic index tests have been conducted as per ASTM standard
on 16 red clay soil samples. Based on the classification test results an activity and
plasticity charts were established and from the chart it was found that the soils under the
current study were inactive soil and the majorities of them are located above the A-line
(inorganic clays with high plasticity).
In order to achieve the intended objectives, a conventional regression and a GP models
have been performed by using a software package named SPSS-25 and Eureqa
computational tool respectively. In developing an empirical equations the independent
variables considered are LL, PL, PI, π€π, πΎπ and ππ and the variables to be predicted are πΆ
π and πΆπ. From the conventional a strong correlation was established between πΆπ and the
three predictor variables LL, π€π and ππ: πΆπ = 0.0027(πΏπΏ - 0.481π€π + 32.6ππ - 13.7) with
a high value of (π 2 = 0.857) and πΆπ with same predictor variables: Cr =
0.000625(LL - 0.496wn + 21.6eo - 9.6) with a high value of (π 2 = 0.75). On the other
hand the best solution of the GP model of compression index is found as:Cc = 0.0123 + 0.0268e o + 0.00215LL + 0.00304 1.52e3-50.1wn+ -0.00356 50.1wn-1.47e3+ 0.00254
3.66wn-111-22.2eowhich, has a high value of π 2 = 0.99 and that of GP model of recompression index is:Cr = 0.0044LL + 13.2
LL + 0.00588LLe
o 2 + 1.74e-6LLw
n2 - 0.217 - 0.466e
o - 1.95e-6wneoLL 2 which has a high value of
π 2 = 0.94. These two newly developed equations were then compared with the existing
ones and it was found that both conventional and GP model equations are better in
estimating both πΆπ & πΆπ of Addis Ababaβs red clay soil. Moreover, the comparison was
made between the conventional & GP model equation and it was found that the GP model
of πΆπ and πΆπ outperforms the conventional regression model.
indices in the laboratory is costly and time-consuming. Due to this, existing empirical
equations have been used by practicing engineers to determine these parameters.
However, these empirical equations are site specific and it is erroneous to use them for
soils of other localities for which they are not formulated. The prime objective of this study,
therefore, is to predict the compression and recompression indices of Addis Ababaβs red
clay soil based on the basic index properties. To develop the proposed models a laboratory
tests like consolidation and basic index tests have been conducted as per ASTM standard
on 16 red clay soil samples. Based on the classification test results an activity and
plasticity charts were established and from the chart it was found that the soils under the
current study were inactive soil and the majorities of them are located above the A-line
(inorganic clays with high plasticity).
In order to achieve the intended objectives, a conventional regression and a GP models
have been performed by using a software package named SPSS-25 and Eureqa
computational tool respectively. In developing an empirical equations the independent
variables considered are LL, PL, PI, π€π, πΎπ and ππ and the variables to be predicted are πΆ
π and πΆπ. From the conventional a strong correlation was established between πΆπ and the
three predictor variables LL, π€π and ππ: πΆπ = 0.0027(πΏπΏ - 0.481π€π + 32.6ππ - 13.7) with
a high value of (π 2 = 0.857) and πΆπ with same predictor variables: Cr =
0.000625(LL - 0.496wn + 21.6eo - 9.6) with a high value of (π 2 = 0.75). On the other
hand the best solution of the GP model of compression index is found as:Cc = 0.0123 + 0.0268e o + 0.00215LL + 0.00304 1.52e3-50.1wn+ -0.00356 50.1wn-1.47e3+ 0.00254
3.66wn-111-22.2eowhich, has a high value of π 2 = 0.99 and that of GP model of recompression index is:Cr = 0.0044LL + 13.2
LL + 0.00588LLe
o 2 + 1.74e-6LLw
n2 - 0.217 - 0.466e
o - 1.95e-6wneoLL 2 which has a high value of
π 2 = 0.94. These two newly developed equations were then compared with the existing
ones and it was found that both conventional and GP model equations are better in
estimating both πΆπ & πΆπ of Addis Ababaβs red clay soil. Moreover, the comparison was
made between the conventional & GP model equation and it was found that the GP model
of πΆπ and πΆπ outperforms the conventional regression model.
Similar
Prediction of Consolidation Parameters Using Artificial Neural Network Model & Regression Analysis: A Case study of Lege Tafo AreaIndex swelling prediction of clayey soilsPrediction of the soil properties of Malagasy rice soils based on the soil color and magnetic susceptibilityAI-driven prediction of shear strength parameters in clay soils using experimental data<b>Investigating the competency of some soft computing techniques for prediction of selected lateritic soilsβ strength from index properties</b>Development of an Optimized Support Vector Regression Model Using Hyper-Parameters Optimization for Electrical Load Prediction