 |
 |
Influence of Coarse Fraction on Swelling Characteristics
Associate Professor, Dept of Civil Engineering, Sri Venkateswara University, Tirupati Post Graduate student, Dept of Civil Engineering, Sri Venkateswara University, Tirupati Research Scholar, Dept of Civil Engineering, Sri Venkateswara University, Tirupati
|
ABSTRACT
Expansive soils are known to cause severe damage to structures resting on it. Safe and economic designs of foundations on these soils require the knowledge of swelling characteristics viz., swelling pressure, swell potential, and swelling index. The direct determinations of swelling characteristics require considerable time and money involving the services of trained technicians and expert geotechnical engineers. Ever since the expansive soils are identified, the geotechnical engineering community is engaged in developing empirical regression models for estimation of swelling characteristics from easily determined Atterberg’s limits and placement conditions. This investigation aims at studying the influence of fraction coarser than 425µ on swelling characteristics which received little attention but plays vital role as Atterberg’s limits are determined on fraction passing 425µ sieve. The study reveals that the coarse fraction does influence all the swelling characteristics viz., swell potential, swell pressure and swelling index. Empirical regression models have been established to estimate the modified swelling characteristics having known the fraction coarser than 425µ.
INTRODUCTION
The swell shrink behaviour exhibited by expansive soils due to seasonal moisture variation is known to cause distress to structures resting on them. Swelling pressure, swell potential, and swelling index are identified as three important swelling characteristics which are required for safe and economic design for structures resting on them. In turn several investigators have developed correlations to predict swelling characteristics basing on placement conditions viz., dry density, initial moisture content and Atterberg’s limits [Satyanarayana and Ranganatham (1969), Komarnik and David (1969), Vijayvergiya and Ghazzaly (1973), Nayak and Christensen (1974), Mallikarjuna Rao (1988) Mowafy and Bauer (1985), Erzin and Erol (2004)]. Such correlations solve the purpose of providing quick estimates of swelling characteristics on one hand and also serve as cross check on field/laboratory tests conducted on limited number of soils on the other hand. However none of these correlations were found to be universally applicable (Mallikarjuna Rao, 1988). This may be attributed to the fact that the Atterberg’s limits are determined on soil fraction passing 425µ sieve where as actual soil may contain fraction coarser than 425 µ fraction also which could influence swelling characteristics. Hence an attempt has been made here to study the influence of fraction coarser than 425µ on swelling characteristics by conducting a series of tests on two different soils in the laboratory. The results of this series of tests are presented and discussed in the following sections.
MATERIALS AND METHODS
Soils Used
The soils used in this study have been obtained from Renigunta and Gajulamandyam tank near Renigunta near Tirupati in India. Disturbed but representative soils were collected from trail pits at a depth of about 2.0m from ground level. The soils obtained from the field are air dried, pulverised with a wooden mallet and sieved through 425µ sieve. The soil fraction passing through 425µ sieve is used in this investigation. The index properties of these soils along with classification are presented in Table1.These soils are designated as S1 and S2 respectively for convenience. Both the soils fall under the CH category i.e., clay of high compressibility as per I.S classification system (IS 1498-1970). The fine fraction has very high Liquid Limit and Plasticity Index. Based on Free Swell Index, Liquid Limit and Plasticity Index of the soil, they come under the category of High degree of Expansiveness (IS 1498-1970).
Table 1. Properties of the soils used
| ||
| Properties | Renigunta | Gajulamandyam |
| ||
| Specific Gravity | 2.68 | 2.72 |
| Gravel (%) | 0% | 0% |
| Sand (%)+425µ | 0% | 0% |
| Sand (%)-425µ | 36.3% | 33.7% |
| Silt + Clay (%)(-75µ) | 63.7% | 66.3% |
| Liquid Limit | 93% | 137% |
| Plastic Limit | 24% | 26% |
| Plasticity Index | 69% | 111% |
| Free Swell Index | 167% | 258% |
| I.S. classification | CH | CH |
| ||
Tests Conducted
Free Swell Oedometer tests are conducted on both S1 and S2 soil samples to determine the swelling characteristics viz., Swell Potential (Sp) Swelling Pressure (pS) and Swelling Index (Cs). In order to study the influence of coarse fraction (+425µ) on swelling characteristics a series of free swell Oedometer tests are conducted on S1 and S2 soils by mixing them with a sand coarser than 425µ in proportions of 20%, 40%, 60% and 80% by weight. Table 2 summarises the mixtures studied and Table 3 gives the grain size distribution of sand used. Since the sand used for mixing with soils is coarser than 425µ, the Atterberg’s Limits of mixtures studied are same as that of soils used i.e. S1 and S2. The free swell odometer tests are conducted on soil samples statically compacted at dry density of 17.5 kN/m3 and initial moisture content of 0% the free swell Oedometer tests are carried out essentially in accordance with B.I.S code of practice, I.S.2720 (part 41)-1997, wherein the sample is allowed to swell freely upon inundation.
Table 2. Clay-sand mixtures studied
| ||
| S.No. | Mixtures Studied | |
| ||
| S.No. | Soil 1(S1) | Soil 2 (S2) |
| 1 | 100%S1+0%Sand | 100%S2+0%Sand |
| 2 | 80%S1+20%Sand | 80%S2+20%Sand |
| 3 | 60%S1+40%Sand | 60%S2+40%Sand |
| 4 | 40%S1+60%Sand | 40%S2+60%Sand |
| 5 | 20%S1+80%Sand | 20%S2+80%Sand |
| ||
Table 3. Grain Size Distribution of Sand (Coarse Fraction)
| |||||
| Sieve size(mm) | Weight retained (gr) | % of weight retained | Cumulative weight retained (%) | % finer | |
| |||||
| 4.75 | 0 | 0 | 0 | 100 | |
| 2.36 | 29 | 0.058 | 5.8 | 94.2 | |
| 1.18 | 113 | 0.226 | 28.4 | 71.6 | |
| 0.6 | 163 | 0.326 | 61 | 39 | |
| 0.425 | 195 | 0.39 | 100 | 0 | |
| |||||
RESULTS AND DISCUSSIONS
Swelling Characteristics
The e-log p plots obtained from free swell odometer tests for both the soils as well as the mixtures studied are presented in Fig. 1 and 2 for S1 and S2 soils respectively. The e-logp plots indicate that the coarse fraction has significant influence on swelling characteristics viz., swell pressure, swell potential and swelling index.
Figure 1. e-log p curves for S1 soil with varying coarse fraction.
Figure 2. e-log p curves of S2 soil with varying coarse fraction.
Swelling characteristics of all the soils and mixtures obtained from these tests are summarized in Table 4. All the swelling characteristics are observed to decrease with percent coarse fraction added.
Table 4. Swelling characteristics of ‘s1’ and ‘s2’ soils with varying proportions of sand
Effect of Coarse Fraction on Swelling Pressure
Fig. 3 shows the variation of swelling pressure with coarse fraction for soils S1 and S2. Swelling pressure is observed to decrease from 540 kPa to 42 kPa as sand coarser than 0.425mm size added is increased from 0% to 80% for soil S1.
Figure 3. Swell pressure vs coarse fraction
For S2 soil, the swelling pressure is decreased from 739 kPa to 35 kPa due to addition of the sand. The reduction in swelling pressure due to coarse fraction may be attributed to the fact that there is an overall reduction in fraction of soil that swells. Further, swelling should take place against the friction offered by the surface of the coarse fraction. Percentage reduction in swell pressure due to coarse fraction is evaluated using the following formula and the same is presented Table 4.
 |
(1) |
where
Spr = percentage reduction in swelling pressure
Spm = swelling pressure of soil sand mixture
Spo = swelling pressure of soil (i.e soil + 0% sand)
Figure 4. Percentage reduction in swell pressure vs coarse fraction
Fig. 4 depicts the variation of reduction in swell pressure against percent coarse fraction added for both the soils. The relationship is observed to be unique irrespective of soil type. The nonlinear regression analysis yielded the following relationship with a correlation coefficient of 0.9696
 |
(2) |
where
Spr= percentage reduction in swelling pressure
CF = percent coarse fraction added
The above correlation may be conveniently used to evaluate possible reduction in swell pressure due to fraction coarser than 425µ.
Effect of Coarser Fraction on Swell Potential
Fig. 5 shows the variation of swell potential with percent coarse fraction for soil sand mixtures of S1 and S2 soils. From Fig. 5 and Table 4, swell potential is observed to decrease from 27.15% to 0.56% for soil S1 and from 43.45% to 0.7% for soil S2, as the coarse fraction increased from 0% to 80%. The reduction in swell potential is expected as coarse fraction is non swelling in nature and it may counteract the swelling process by gravity and surface friction.
Figure 5. Swell potential vs coarse fraction.
This observation is in tune with the observations made in the case of swell pressure. Here also percent reduction in swell potential (Spr) is calculated using the fallowing formula and the values so calculated are summarised in Table 4.
 |
(3) |
where
Csr = percentage reduction in swell potential
Csm = swell potential of soil sand mixture
Cso = swell potential of soil
Figure 6. Percentage reduction in swell potential vs coarse fraction
Fig. 6 depicts the variation of reduction in swell potential against percent coarse fraction added for both the soils. Here also the relationship is observed to be unique irrespective of soil type as in the case of swell pressure. The regression analysis yielded the following relationship with a correlation coefficient of 0.9515
 |
(4) |
where
Csr = percentage reduction in swell potential
CF = percent coarse fraction added
The above correlation may be conveniently used to evaluate possible reduction in swell potential due to fraction coarser than 425µ.
Effect of Coarser Fraction on Swelling Index
Fig. 7 shows the variation of swelling index with coarse fraction for soils S1 and S2. The swelling index is observed to decrease from 0.0399 to 0.0091 as coarse fraction is increased from 0% to 80% for soil S1. For S2 soil, the swelling index is decreased from 0.0771 to 0.0183. This observation is in tune with the observations made in the case of swell potential and swelling pressure.
Figure 7. Swell Index vs coarse fraction
Here also, the reduction in swelling index due to coarse fraction may be attributed to the overall reduction in fraction of soil that swells and possible reduction in swelling due to friction offered by the surface of the coarse fraction. Percent reduction in swelling index (Csr) is calculated using the following formula
 |
(5) |
where
Csr = percentage reduction in swelling index
CF = percent coarse fraction added
The percentage reduction swelling index evaluated using above formula are presented in Table 4. Fig. 8 depicts the variation of reduction in swelling index against percent coarse fraction added for both the soils. The relationship is observed to be unique irrespective of soil type.
Figure 8. Percentage redution in Swell Index (Cs) vs coarse fraction
A polynomial regression analysis yielded the following relationship with a correlation coefficient of 0.952.
 |
(6) |
where
Csr = percentage reduction in swelling index
Csm = swelling index of soil sand mixture
Cso = swelling index of soil
The above correlation may be conveniently used to evaluate possible reduction in swelling index due to fraction coarser than 425µ.
CONCLUSIONS
A series of tests are conducted on mixtures of sand and expansive soils to study the influence of the coarse fraction on swelling characteristics. The study reveals that the swelling characteristics are significantly influenced by fraction coarser than 425µ. Conventional correlations, available in literature for prediction of swelling characteristics using the Atterberg’s limits do not take into account this fact. Regression models were developed to predict percent reduction in any given swelling characteristic in terms of percent coarse fraction coarser then 425µ. These correlations may be used to get the reduced values of the swelling characteristics due to the presence of coarse fraction.
REFERENCES
 | |
| © 2006 ejge | |