CODEN (USA): IJCRGG, ISSN: 0974-4290, ISSN(Online):2455-9555 Vol.10 No.10, pp 10-26, 2017
Abstract : Expansive Sub-grade soil improvement is one of the primary and major processes in the construction of any highway. Roads on black cotton soil often fail due to swelling and shrinking of such soil which makes stabilization mandatory. As flyash is available at very lower cost it can be used for stabilization of expansive soils for various uses. This present research aims to utilize the fly ash in road application. In this research index, engineering, chemical properties of virgin soil has been studied. In addition, chemical analysis is done for soil and fly ash mixture. Flyash is added to the soil with 10%, 20%, 30%, 40%, and 50% by weight of soil. The soil falls under CI category. It has 50% free swell index. The CBR of soil is 13.6% and it reduces to 2.66% when soaked. Shear strength of soil is 42.06 kpa at its optimum moisture content of 15% with maximum dry density of 1.658 g/cc. This study indicates that plasticity index, free swell index, pH, and cation exchange capacity, are decreasing with the addition of fly ash and total soluble solids, calcium carbonate content are increasing with the addition of fly ash. To ascertain the soil composition, XRD analysis has been done. Keywords : Subgrade, Flyash, Plasticity index, Free swell index, pH, Cation exchange capacity, XRD analysis, Total soluble solids, Calcium carbonate content, Geotextile, CBR, MRA.
In general Clays exhibit undesirable engineering properties. These to have low shear strengths and to lose shear strength further upon wetting or other physical changes. They can be plastic and compressible and they expand when wetted and shrink when dried. Some types expand and shrink greatly upon wetting and drying – a very undesirable characteristics. Clayey soils can creep over time under a load, especially when the shear stress is depending its shear strength, making them prone to sliding. They develop large lateral pressures. These tend to have low resilient modulus values. For these reasons, clays are generally poor materials for foundations. The annual cost of damage done to non-military engineering structures constructed on expansive soils is estimated at $220 million in the United Kingdom and many billions of dollars are worldwide. Flyash was successfully used for stabilizing expansive clays. The strength characteristics of flyash stabilized clays are measured by means of unconfined compressive strength (CBR) or California Bearing Ratio (CBR) values. Based upon the soil type, the effective flyash content for improving the engineering properties of the soil varies
1,2,3,4,5,6,7
between 15 to 30%.
Geosynthetics is a class of geomaterials that are used to improve soil conditions for a number of applications. They consist of manufactured polymeric materials used in contact with soil materials or pavements to act as a separator and reinforcing material like steel bars in concrete. Geosynthetics has been increasingly used in geotechnical and environmental engineering for the last 5 decades. Over the years, these had helped designers and contractors to solve several types of engineering problems where the use of conventional construction materials would be restricted or considerably more expensive. There are a significant number of
8,9,10,11,12,13,14,15
geosynthetic types and geosynthetic applications in geotechnical and environmental engineering.
In this study, characteristics of soil stabilized with flyash are studied. In addition reinforcing effects of geosynthetics are also studied. AMultiple Regression model that could predict CBR value based on Atterberg’s
16,17,18,19,20
limit, OMC, MDD, Flyash content, and number of layers of geotextile has also been developed.
Methodology mainly consists of three parts. Part 1 includes identification of problems in construction of roads in black cotton soil, review of literatures, collection of soil and flyash. Part 2 and part 3 are laboratory works which are mainly focused on determination of index, engineering, chemical properties of soil and chemical properties of soil and flyash mixture. Methodology is graphically represented by figure 1.
Figure1. Methodology
Soil for this research work is collected from Cheranmaanagar, Coimbatore, Tamilnadu state, India. The locations are 11.05420, 77.01830. The soil sample was present at depth of 4 feet from ground level. The soil sample used for analysis is clay. The laboratory investigations confirm that the soil falls under the category Clay with Intermediate Compressibility.
Fly ash is obtained from Mettur Thermal Power Plant, Tamilnadu state which belongs to class F.
Mineral composition of fly ash was obtained from mettur thermal power plant and is as follows.
Table 1.Composition of flyash Table 2. Properties of geotextile
S.no | Constituents | % |
---|---|---|
1 | MgO | 0.57 |
2 | Al2O3 | 24.12 |
3 | SiO2 | 52.55 |
4 | K2O | 0.965 |
5 | P2O5 | 0.72 |
6 | CaO | 2.65 |
7 | Loss on Ignition | 18.18 |
Property | Values |
Tensile Strength ( MD) kN/m | 28.5 |
Tensile Strength ( CD) kN/m | 26.5 |
Elongation ( MD) % | 30 |
Elongation ( CD) % | 27 |
Trapezoid Strength (MD) N | 320 |
Trapezoid Strength (CD) N | 320 |
Puncture Strength (N) | 370 |
Apparent Opening Size (mm) | 0.075 |
Water Permeability (Flow Rate) | 9.5 l/m2/s |
Mass Per unit Area (gsm) | 140 |
Geotextile for this research is manufactured by Techfab India with the following specifications.
3. Laboratory Investigations on Index and Engineering Properties
This elaborates the various index and engineering properties of soil namely natural moisture content, specific gravity, liquid limit, plastic limit, shrinkage limit, grain size distribution, optimum moisture content, maximum dry density, unconfined compressive strength, CBR test and free swell test etc.Table.3 gives the index and Engineering properties of the soil.
Table 3. Properties of soil 3.1Standard Proctor’s Compaction Test (light compaction)
S.no | Properties | Result | Remarks |
---|---|---|---|
1 | Natural Moisture Content | 8.69% | - |
2 | Specific Gravity | 2.71 | - |
3 | Sieve Analysis % of Gravel % of sand % of Silt % of Clay | 2.1% 30.5% 22.1% 45.3% | - |
4 | Differential Free Swell Index | 50% | Degree of expansion is high |
5 | Liquid Limit (WL) | 47% | |
Plastic Limit (WP) | 17% | ||
Shrinkage Limit (WS) | 12% | Degree of expansion is marginal | |
Flow Index ( If) | 22 | ||
Plasticity Index ( IP) | 30% | Swelling potential is high | |
Toughness Index ( It) | 1.36 | Since (It) > 1.0 Soil nor friable at Plastic state. | |
Liquidity Index( IL ) | -27 % | Since ( IL) < 0 Very Stiff | |
Consistency Index ( IC) | 127.7 % | Since ( IC) >100 Very Stiff | |
Soil Classification | CI | Clay of intermediate Compressibility | |
Activity (A) | 0.53 | A < 0.75 Soil is Inactive | |
6 | Optimum Moisture Content | 15% | - |
Maximum Dry Density | 1.658 g/cc | - | |
7 | Unconfined Compressive Strength (qu) | 84.12 kN/m2 | - |
Cohesion (Cu) | 42.06 kN/m2 | - | |
8 | CBR unsoaked CBR soaked | 12.88% 2.68 % | - |
The Optimum Moisture Content (OMC) and Maximum Dry Density (MDD) is determined by conducting standard proctor’s test as per IS: 2720 (Part 7) – 1980. This test has been conducted for soil with 0%, 10%, 20%, 30%, 40%, and 50% of flyash along with single and double layer of geotextile. Geotextile has cut into plan dimension of proctor mould and placed at the end of each soil layer. Table 4 gives the comparison of OMC and MDD.
Table 4.Comparisons of OMC and MDD
S.no | % of Flyash added | Zero layer of reinforcement | Single layer of reinforcement | Double layer of reinforcement | |||
---|---|---|---|---|---|---|---|
OMC in % | MDD in g/cc | OMC in % | MDD in g/cc | OMC in % | MDD in g/cc | ||
1 | 0 | 15 | 1.658 | 14.41 | 1.649 | 14.36 | 1.621 |
2 | 10 | 14.93 | 1.678 | 14.11 | 1.669 | 14.08 | 1.657 |
3 | 20 | 13.45 | 1.642 | 13.09 | 1.6304 | 13 | 1.618 |
4 | 30 | 12.2 | 1.628 | 12 | 1.617 | 12.17 | 1.612 |
5 | 40 | 10.76 | 1.611 | 10.4 | 1.603 | 10.28 | 1.584 |
6 | 50 | 9.7 | 1.593 | 9.26 | 1.580 | 8.967 | 1.564 |
Table 5. Comparison of UCC Strength
S.no | Flyash content | UCC in KN/sq.m | Cohesion in KN/sq.m |
---|---|---|---|
1 | 0 | 84.12 | 42.06 |
2 | 10 | 97.27 | 48.635 |
3 | 20 | 126.85 | 63.425 |
4 | 30 | 106.82 | 53.41 |
5 | 40 | 84.02 | 42.01 |
6 | 50 | 78.81 | 39.405 |
3.2Unconfined Compressive Strength
The unconfined compressive strength and cohesive strength is obtained by conducting Unconfined Compressive Strength test. The test is conducted as per IS:2720(Part 10)-1991 for soil with 0%, 10%, 20%, 30%, 40%, and 50% of fly ash. Table 5 gives the comparison of UCC strength and cohesion.
3.3Determination of CBR
For any pavement design, CBR is the prime factor which determines the thickness of each pavement layer. CBR (unsoaked& soaked) test is done as per IS2720 part 16. For soil with 0%, 10%, 20%, 30%, 40%, and 50% of flyash mixtures soaked and unsoaked tests have been done. For soil with 0%, 10%, 20%, 30%, 40%, and 50% of flyash mixtures along with the inclusion of single and double layer of geotextiles, soaked CBR tests have been done. Spacing in-between the geotextiles is kept arbitrarily 42mm (±5mm). Table 6 and 7 gives the results of unsoaked and soaked CBR.
Table 6.Results of UnsoakedCBR test Table7.Results of Soaked CBR test
S.no | Flyash content in % | CBR in % |
---|---|---|
1 | 0 | 12.88 |
2 | 10 | 16.1 |
3 | 20 | 17.17 |
4 | 30 | 16.46 |
5 | 40 | 15.38 |
6 | 50 | 14.31 |
S.no | % of Flyash added | Soaked CBR without reinforcement | Soaked CBR with single layer of reinforcement | Soaked CBR with double layer of reinforcement |
---|---|---|---|---|
1 | 0 | 2.68 | 3.57 | 6.44 |
2 | 10 | 3.22 | 3.75 | 6.8 |
3 | 20 | 4.29 | 6.08 | 10.73 |
4 | 30 | 3.57 | 4.83 | 7.87 |
5 | 40 | 2.86 | 3.93 | 5.72 |
6 | 50 | 2.68 | 3.22 | 5.37 |
4. Laboratory investigation on chemical properties
Chemical analyses are very much important to ascertain the mechanism behind the stabilization and also to know influence of various chemical parameters such as total soluble solids, pH, Cation Exchange Capacity, calcium carbonate content, and soluble sulphates. The mechanism can be found from XRD analysis.
These analyses are carried out for soil with 0%, 10%, 20%, 30%, 40%, and 50% addition of flyash.
Total soluble solids indicate the amount of presence of soluble salts and other soluble materials present in soil. This test is done in accordance with IS 2720 part 21 (Gravimetric analysis) and also indirectly determined using TSS analyzer. Results from both the tests are tabulated. It is observed that there is no significant change in soluble solids concentration from both the tests.
The CaCO3 content can be found from volumetric analysis of soil-flyash mixture blended with 0.1 N HCL against 1N NaOH as per IS 2720 part 23 (1976).
Table .9 shows the amount of calcium carbonate present.
Table .8 total soluble solids
S.no | % of Flyash added | Soluble solids (ppm) (analyzer) | Soluble solids (ppm) (IS method) |
---|---|---|---|
1 | 0 | 102.2 | 101 |
2 | 10 | 110.9 | 110 |
3 | 20 | 125.5 | 125 |
4 | 30 | 133.7 | 133 |
5 | 40 | 140.1 | 141 |
6 | 50 | 145.9 | 146 |
Table .9 Amount of calcium carbonate present 4.3 Determination of pH
S.no | % of Fly ash added | Calcium carbonate (% by weight) |
---|---|---|
1 | 0 | 20 |
2 | 10 | 20.7 |
3 | 20 | 21.3 |
4 | 30 | 22 |
5 | 40 | 22.8 |
6 | 50 | 23.5 |
The pH of the samples were determined using the method of Eades and Grim specified by IS 2720 part 26, which involves mixing the solids with pure water (1:5 solid: water), periodically shaking samples, and then testing with a pH meter after 1 hour.
Table .10 pH of soil with Fyash
S.no | % of fly ash added | pH |
---|---|---|
1 | 0 | 8.89 |
2 | 10 | 8.72 |
3 | 20 | 8.52 |
4 | 30 | 8.43 |
5 | 40 | 8.38 |
6 | 50 | 8.33 |
Table .11Cation Exchange Capacity of soil with flyash
S.no | % of Fly ash added | CEC (meq/100g) | |
---|---|---|---|
1 | 0 | 89.683 | |
2 | 10 | 80.011 | |
3 | 20 | 76.884 | |
4 | 30 | 74.573 | |
5 | 40 | 73.169 | |
6 | 50 | 72.931 |
CEC represents the exchangeable cations present in soil. There are two methods available to determine CEC namely Chapman method (IS 2720 part 24: 1974) and Soil Society of America (compulsive exchange) method. Tests are conducted based on above mentioned methods and there are no much difference is observed between the results from both methods.
Three methods are specified by IS 2720 part 27 to determine total soluble sulphates namely precipitation method, volumetric method, and calorimetric method. The last two methods are being subsidiary methods; precipitation method is used in this analysis.
It is observed that sulphate present in the soil is 0.012% by mass. This shows only a trace of sulphate is present and there is no sulphate present in flyash. Hence this test not conducted for soil-flyash mixture as the influence of sulphate on the stabilization process of this particular soil is nil.
XRD analysis was done to know the mineralogical composition of soil. This test was carried out in Avinasilingam University, Coimbatore, tamilnadu state, India. The mineralogical composition of soil is as shown in table .12 and the XRD output is shown in figure 2.
Figure .2 Output of XRD Analysis
Details of minerals present and their D spacing is given in table 12.
S.no | Mineral | D Spacing (10 -10m) |
---|---|---|
1 | Quartz | 3.34 |
2 | Mica | 3.20 |
3 | Felds | 3.02 |
4 | Kaolinite | 4.2 |
5 | Illite | 2.23 |
6 | Chlorite | 1.38 |
The variation of optimum moisture content and maximum dry density with the addition of flyash and inclusion of geotextile layers can be observed from figure 3 to 5.
Figure 3.Variation of OMC&MDD with addition of Flyash
The geotextile is chemically inert and has low specific gravity than soil and flyash mixtures. Thus inclusion of geotextile does not cause change in OMC but MDD is decreasing when compared with soil-flyash mixture.
Figure 4.Variation of OMC &MDD with addition of flyash and single layer of geotextile
Figure 5. Variations of OMC &MDD with addition of flyash and dual layers of geotextile
The variation of unconfined compressive strength with fly ash content is given in Figure 6 for standard proctor density.
140 120
UCC IN kN/sq.m
100 80 60 40 20 0
FLYASH CONTENT IN %
Figure 6. Variations in UCC strength with the addition of flyash
The reactions behind the variations in CBR values with the addition of flyash are same as that of the variations in UCC strength. The increases in CBR values with the inclusion of geotextile are mainly due to the reinforcing effects of geotextile. Lateral restraint and tensioned membrane effects of geotextile contribute to the increase in CBR. Figure 7 and 8 shows that the variations in CBR.Table13 give the percentage increase in CBR with the inclusion of geotextiles.
Figure 7 Variations in CBR with the addition of flyash and single layer of geotextile
12 10
CBRin %
8 6 4 2 0
Figure 8. Variations in CBR with the addition of flyash and double layer of geotextile Table 13. % increase in CBR with inclusion of geotextile with respect to the flyash stabilized subgrade
S.no | % of flyash added | % increase in soaked CBR with single layer of reinforcement | % increase in soaked CBR with double layer of reinforcement |
---|---|---|---|
1 | 0 | 33.20 | 140 |
2 | 10 | 16.45 | 111.8 |
3 | 20 | 41.72 | 150.1 |
4 | 30 | 35.29 | 120.44 |
5 | 40 | 37.41 | 100 |
6 | 50 | 20.14 | 100.37 |
Variation of Atterberg’s limits with addition of flyash can be observed from figures 9 to 11. Liquid limit, plastic limit, plasticity index are decreasing with the addition of flyash.
50
LIQUID LIMIT (%)
40 30 20
liquid limit %
10 0 0 20 40 60
Figure .9 Variation of liquid limit with the addition of fly ash
20
PLASTIC LIMIT (%)
15 10 plastic limit %
5 0 0 20 40 60
Figure .10 Variation of Plastic limit with the addition of fly ash
40
PLASTICITY INDEX %
30 20 10
plasticity index % 0 0 20 40 60
Figure .11 Variation of Plasticity Index with the addition of fly ash
Table 14 summarizes the percentage reduction in Atterberg’s limit and graphically represented by figures 12, 13 and 14.
S.no | %of Flyash added | % Reduction in Liquid limit | % Reduction limit | in Plastic | % Reduction in Plasticity index |
---|---|---|---|---|---|
1 | 0 | 0 | 0 | 0 | |
2 | 10 | 14.89 | 23.52 | 10 | |
3 | 20 | 34 | 35.3 | 33.33 | |
4 | 30 | 44.68 | 45.8 | 44 | |
5 | 40 | 51.06 | 52.9 | 50 | |
6 | 50 | 55.31 | 58.23 | 53.66 |
% REDUCTION IN LIQUID LIMIT
60 50 40 30
% REDUCTION IN 20 LIQUID LIMIT
10 0
0 102030405060
Figure .12 Percentage Reduction in Liquid Limit
70
% REDUCTION IN PLASTIC
LIMIT
60 50 40 30
% REDUCTION IN 20 PLASTIC LIMIT
10 0 0 102030405060
Figure .13 Percentage Reduction in Plastic Limit
60
% REDUCTION IN PLASTICITY
INDEX
50 40 30
% REDUCTION IN 20 PLASTICITY INDEX
10 0
Figure .14 Percentage Reduction in Plasticity Index
Total soluble solids increases with addition of flyash. The increased soluble solid content with addition of flyash indicates that amount of flyash available for cementing actions. This gives a positive result, which shows in figure 15.
140
solubility in ppm
120
100 80 60
Soluble solids (ppm)
40 20 0
0 102030405060
% of flyash added
Figure .15 variations of total soluble solids with the addition of fly ash
Calcium carbonate acts as a binding material and it increases with the increase in fly ash content in soil. This content may vary with respect to time since cementaneous process is a long time chemical reaction.
The variation of CaCO3 with the addition of flyash is sown in figure 16.
24
calcium carbonatepercentage by mass
23.5 23
22.5 22
21.5 21
20.5 20
19.5 0 102030405060
Figure .16 Variation of caco3 with the addition of flyash
The pH of soil is an indirect measure of Cation Exchange Capacity of soil. pH is directly proportional to CEC in alkaline state. CEC and pH are indirectly proportional to strength of soil.
Figure .17 Variation of ph with the addition of flyash
Cation exchange capacity indicates amount of exchangeable ions adsorbed on clay surface. CEC fixes the double layer thickness of clay. Plasticity index is directly proportional to the double layer thickness. A decrease in CEC is observed with the addition of flyash to soil.
Figure .18 Effect of flyash on cation exchange capacity of soil
Differential free swell has a trend of decreasing due to decrease in plasticity index. It indirectly indicates that swell pressure may also be reduced.
The variation of DFS with the addition of flyash is shown in figure.19.
50
40
DFSIN %
30 20 10
0 0 102030405060
% OF FLYASH ADDED BY WEIGHT
Figure .19 variation of dfs with the addition of flyash
6. Multiple Regression analysis modelling
The ultimate motto of regression analysis is to develop an equation which could have the capacity to find CBR value based upon some input parameters.
The ambition of ANN modelling is to develop an ANN model to predict CBR value and also to study the effect of number of neurons in hidden layer with different algorithms.
Multiple regressionsare regression with two or more independent variables on the right-hand side of the equation. Multiple regressioncan be adopted if more than one cause is associated with the effect we wish to understand. For the development MRA model, MS-Excel 2007 software has been used.
5.2.1Inputs and output
Two models 1) Artificial Neural Network (ANN) model 2) Multiple Regression Analysis(MRA) model have been developed to predict the soaked CBR values of reinforced flyashstabilizedsoil .Both the models have been developed by taking, Atterberg’s limits, % of flyash added, OMC (%) and MDD (kN/m3), number of geotextile layers as input variables and soaked CBR (%) as output variable.
5.2.2Summary of output
Using MS-Excel 2007, regression analysis has been done and the following relationship is obtained with a co-efficient of correlation(R2) 0.8878.
CBR= 0.09895X1 -0.2171X2 +0.0451X3+2.737X4-55.785X5+1.979X6+63.483
Here,
X1= % of flyash added, X2= Liquid limit, X3=Plastic limit, X4= OMC, X5= MDD, X6= No. of geotextile layers.
5.3.2Effect of training algorithm
As stated earlier, at the end of training a set MSE and R values are obtained. This process is repeated until the maximum R value and minimum MSE has reached for the particular algorithm. The values are tabulated in table 15.
Table15.Effect of training algorithm
S.no | Algorithm | R value | MSE |
---|---|---|---|
1 | Quasi-Newton back propagation | 0.88712 | 1.083x10 -4 |
2 | Bayesian regulation back propagation | 0.85190 | 4.983x10 -5 |
3 | Conjugate gradient back propagation with Powell-Beale restarts | 0.94122 | 3.776x10 -7 |
4 | Conjugate gradient back propagation with Fletcher-Reeves updates | 0.81167 | 7.339x10 -6 |
5 | Conjugate gradient back propagation with Polak-Ribiére updates | 0.85819 | 2.964x10 -9 |
6 | Gradient descent back propagation | 0.94862 | 9.985x10 -9 |
7 | Levenberg-Marquardt back propagation | 0.98695 | 8.0242x10 -11 |
8 | One-step secant back propagation | 0.92335 | 1.388x1010 |
9 | Scaled conjugate gradient back propagation | 0.96904 | 1.946x10 -6 |
From table 6.1 we can infer that Levenberg-Marquardt back propagation shows maximum R value of 0.98695 and minimum MSE value of 8.0242e-11 and hence Levenberg-Marquardt back propagation can be used although it consumes higher memory usage compared to any other algorithm. When there is a constraint to memory usage Scaled conjugate gradient back propagation can be employed.
5.3.3Effect on number of neurons in hidden layer
Keeping the percentage of data allotted for training and testing as constant, numbers of neurons are varied and R value is noted. Randomness is observed between R value and numbers of neurons.The values can be read from table 16 and figure 20.
Table16.Relationship between number of neurons and r values
Number of | Percentage of data allotted and R value | ||
---|---|---|---|
neurons | 10%-10% | 15%-15% | 20%-20% |
2 | 0.9334 | 0.9736 | 0.8340 |
4 | 0.7812 | 0.8712 | 0.6725 |
6 | 0.8573 | 0.6698 | 0.9745 |
8 | 0.9223 | 0.7845 | 0.9421 |
10 | 0.8125 | 0.9869 | 0.7967 |
15 | 0.8823 | 0.4356 | 0.5698 |
20 | 0.9461 | 0.5739 | 0.8883 |
25 | 0.9388 | 0.7866 | 0.8934 |
30 | 0.7174 | 0.4358 | 0.8352 |
Figure 20.Effect on number of neurons in hidden layer
Based on the laboratory, experimental investigations on stabilization and computational modelling, the following conclusions can be drawn.
*****