Determination Appropriate Model for Estimating Soil Water Characteristic Curve in Various Moisture Conditions in The One Compacted Clay Soil

ABSTRACT
clay soils in landfills are used as compacted and to reduce contamination. As regards, compacted clay liners are often unsaturated, estimating unsaturated hydraulic properties including soil water characteristic curve is essential. In this study, clay soil water characteristic curve (SWCC) in Various Moisture conditions (dry, Optimum, wet) and different compactions (reduced, standard, modified) were obtained using a pressure plate. Then, the five models of Tani, exponential, Russo, Fredlund and Xing five parameters, Van genuchten were used to estimation Swcc and then select the best model parameters were studied in order to investigate the behavior of clay. The result showed that Van genuchten model in Undisturbed sample and reduced compaction-optimum; Russo model in modified compaction-wet and Fredlund and Xing in other sample due to larger R2 ,smaller SSR and RMSE provided the best performance and Tani provided the weakest performance . Value of α in the models of Fredlund and Xing and Van genuchten in modified and standard compactions increased with increasing Moisture, while in the reduced compaction was different. Value of α in modified compaction with increasing moisture content to be a constant value.
1. Introduction
Compacted clay soils are commonly used as barrier materials in waste containment facilities. The selection of these soils as barrier materials is based on their saturated behavior (miller et al, 2002).
The objective of this study was to estimate SWCC by fitting models of Tani[20], Expotential[13], Russo[21], Fredlund And Xing Five Parameters (Fx5([18], Van Genuchten (VG) [22] and determination parameters of them using matlab software and select the best model and also investigate soil behavior as liner.
2. Materials and Methods
2.1. Soils
Soils used in the study were obtained from agricultural research center of Hamadan, Iran. Soil samples were obtained from depths of 0 – 30. Particle size distribution (PSD) was determined using hydrometer method, based on Stokes’ Law (ASTM D422) and Then Texture AutoLookup (TAL) software for windows (version 4.2) was used to determine the soil texture class based on United States Department of agriculture textural classification system and soil texture triangle (USDA). Nine disturbed samples (three compaction type under three moisture conditions) and one undisturbed samples were collected from each soil texture. Samples were prepared using three compaction efforts included reduced, standard, and modified compaction efforts. For each compaction effort, samples were prepared at three water content conditions included 2% dry of optimum, optimum and 2% wet of optimum water content.
2.2. Compaction
For the standard compaction, the hammer was dropped on the soil in the mold 25 times on each of three soil layers (ASTM D1557). The reduced compaction was similar to standard compaction with one exception; 15 blows/layer were used instead of 25 blows/layer (Daniel and Benson, 1990). The reduced effort was used to simulate poor quality compaction procedures in the field. The Modified compaction, the heavier hammer was also dropped 25 times on each of five soil layers (ASTM D698). Then compaction tests were performed over a range of soil moisture contents. The results were then plotted as dry density versus moisture contents. The maximum dry unit weight occurs at a water content that is called the optimum water content. The maximum dry unit weights and the corresponding optimum water contents were estimated. Compaction characteristics of the soils are presented in Table 2.
Table 1. Compaction Characteristics of Soil Samples
Clay Soil
reduced Standard modified Compaction effort
13.08 13.91 14.57 Maximum unit weight(KN/m3)
22.48 16.39 12.46 Optimum water content%
After calculation the optimum water content, Once again, soil samples were compacted. Then, soil samples were prepared at three compaction effort under three water content conditions included 2% dry of optimum, optimum and 2% wet of optimum water content. For each soil, nine samples were prepared.
2.3. Evaluation Procedures
To evaluate the models, two parameters; root-mean-square error (RMSE) and the coefficient of determination (R2) were used.
Which were calculated as follows:
(6)
(7)
Where:
θm and θp=the measured and predicted volumetric soil water content, respectively, in cm3 cm-3, n=the number of paired observations, and =the mean of measured values in cm3 cm-3.
2.4. Comparison between SWCC Models
An optimization routine was used to fit the parametric models to the measured data by altering the fitted parameters iteratively until the squared differences between the predicted and measured θ (ψ)data were minimized. The sum of the squared residuals (SSR) is defined as
(8)
3. Results and Discussion
3.1. Evaluation between SWCC Models
The R2 and RMSE calculated for the clay soil are close to 1 and 0, respectively, as shown in Table 3. Mean values of the RMSE for the five models: VG, FX5, Russo, Expotential and Tani are, respectively, 0.01872, 0.0189, 0.07755, 0.2323 and 0.04106 cm3cm-3. Therefore, in all models, amounts of both R2 and RMSE were able to estimate the moisture curve. All models provide the best performance in clay soil but the FX5 and Tani models provide the better and weakest performance compared to the other methods, respectively .
Table 3. SSR Parameters and R2 and RMSE of 3.1. Evaluation of Models for Clay Soil
Parameter/Model Tani Expotential Russo VG FX5
Witness SSR 0.02077 7.710-3 0.0139 7.710-4 7.510-4
R2 0.815 0.932 0.69 0.993 0.993
RMSE 0.0832 0.0507 0.084 0.028 0.194
modified compaction dry SSR 2.910-4 2.610-4 2.410-4 9.410-4 4.710-4
R2 0.984 0.986 0.987 0.949 0.975
RMSE 0.0099 0.0092 0.011 0.031 0.015
optimum SSR 8.110-4 2.510-4 4.310-4 1.6510-4 5.110-4
R2 0.97 0.991 0.985 0.994 0.982
RMSE 0.0165 0.0092 0.015 0.013 0.0131
wet SSR 1.910-3 4.810-4 110-3 6.9910-5 5.110-4
R2 0.95 0.988 0.974 0.998 0.987
RMSE 0.022 0.013 0.022 0.0084 0.0159
standard compaction dry SSR 3.2510-3 1.110-4 1.910-3 9.7510-5 1.110-3
R2 0.927 0.973 0.955 0.998 0.977
RMSE 0.029 0.0199 0.032 0.0099 0.022
optimum SSR 4.910-3 1.710-3 310-3 8.8810-5 1.110-3
R2 0.91 0.969 0.945 0.998 0.982
RMSE 0.035 0.0239 0.0389 0.0094 0.022
wet SSR 7.210-3 2.3410-4 4.410-4 9.610-5 9.510-4
R2 0.89 0.965 0.935 0.9986 0.986
RMSE 0.042 0.0023 0.047 0.0098 0.0218
reduced compaction dry SSR 9.410-3 3.5910-3 6.110-3 5.1610-4 1.510-3
R2 0.87 0.951 0.918 0.9931 0.979
RMSE 0.056 0.036 0.055 0.0227 0.0278
optimum SSR 0.011 3.510-3 6.610-3 2.510-3 1.810-3
R2 0.88 0.959 0.923 0.9699 0.979
RMSE 0.051 0.034 0.0575 0.051 0.03
wet SSR 0.0133 3.710-3 8.0410-3 3.3710-5 1.610-3
R2 0.867 0.963 0.919 0.9997 0.984
RMSE 0.66 0.035 0.063 0.0058 0.028
Conclusions:
In this study, using the field data and laboratory analysis of SWCC models were evaluated. The results of this study can be summarized as follows:
1. Among the fitting models to experimental models by Matlab, FX5 and VG models Provided the best performance compared to other models and Thani model presented the weakest performance.
2. There is poor correlation between the measured and predicted values of and in models FX5 and VG.
3. The amount of α by increasing compaction and also moisture have increasing trend and in high compaction with increasing moisture is a constant value.
survey and identify appropriate models to estimate SWCC is crucial in saving time and costs compacted clay soil, soil water storage, predicting the hydraulic conductivity and soil moisture, and generally survey the behavior of the clay liner in landfill.