With geocell reinforcement the fill surface is found to undergo settlement （Fig.6）indicating that the geocell mattress has sunk down, under footing penetration,and hence is pulled away from the wall of the test tank that the tank wall does not affect its performance.From Fig. 8 it can be observed that the zone where deformation of soil （indicated through deformation in the colored Line） dies down to practically negligible value lies much above the tank base. Through these observations it is now established that there is no influence of the test tank boundary on the performance of the footing.Hence the test tank-footing geometry used in the present study is adequate enough to overcome the boundary

effect.

Scale Effect   The important parameters in the geocell-reinforced foundation system can be assumed to be   and  , where   is the strength/stiffness of the reinforcement,G is the shear modulus of soil, and  is the unit weight of soil.Other symbols have been defined previously. The function   that governs the system can be written as

By using the scaling law proposed by Langhaar （1951） and dimensional

analysis of Buckingham （1914） it is found that the geometric parameters have a linear variation while the strength and stiffness parameters vary in second order. The strength/ stiffness of the reinforcement in the prototype reinforced soil foundation bed should be of   times the strength/stiffness of the reinforcement used in the model test, where N is the model scale. Details of the present scale analysis can be found in Sireesh et al.（2009）.

    In the present model tests the strength/stiffness of the geocell

joint is 4.75 kN/m, which is much lower compared to that of the geocell wall material （i.e., geogrid）.Hence, for the results from the present study to be applicable in practice,the prototype geocell should have minimum strength/stiffness of 4.75  kN/m.

Conclusions

    In this paper the influence of relative density of foundation soil on the performance of geocell-reinforced foundation beds has been studied through a series of laboratory model load tests. The deformation pattern on fill surface and in sand subgrade indicates that the relative size of test tank and footing, used in the present study, is adequate enough to overcome the boundary effect.

   It is observed that the beneficial effect of geocell reinforcement,

in terms of increase in stiffness,bearing capacity,and load dispersion angle of the foundation bed, is present over a wide range of relative density （ID=30–70%）;however,it is higher for dense condition of foundation soil.The value of subgrade modulus at 3% settlement   has increased from about 10   with ID=30% to about 40   with ID=70%, indicating that the stiffness of the geocell-reinforced foundation bed has increased by fourfold with increase in relative density of soil from 30 to 70%.At relatively lower settlement of footing （i.e., s /B  10%）,the value of   does not change much with change in relative density of soil,while,at higher settlement range （i.e., s /B  10%）, the rate of increase of bearing capacity factor   with increase in relative density （ID） is relatively rapid with dense soil （i.e., ID  50%） compared to that with loose soil. With geocell reinforcement offering three-dimensional confinement the dilation induced benefit is substantially high for dense soil fill. Therefore, for effective utilization of geocell reinforcement,the foundation soil should be compacted to higher density. In field,to achieve dense soil fill within geocells,it is suggested that light rolling compaction with some amount of overfilling (i.e., about 150 mm; Bush et al.1990) should be adopted. With repeated passage of rolling and filling,a dense and compact geocell structure can be achieved.