This action might not be possible to undo. Are you sure you want to continue?
YAP, S. P.1, SALMAN, F. A.2 & SHIRAZI, S. M.3 Department of Civil Engineering, Faculty of Engineering, University of Malaya, 50603, Kuala Lumpur, Malaysia.
1 2 3
email@example.com, contact: +6012 635 4222
firstname.lastname@example.org, contact: +603 7967 7649
email@example.com, contact: +603 7967 7650
Abstract: Determination of distribution and magnitude of active earth pressure is crucial in retaining wall designs. A number of analytical theories on active earth pressure were presented. Yet, there are limited studies on comparison between the theories. In this study, comparison between the theories with finite element analysis is done using the PLAXIS software. The comparative results showed that in terms of distribution and magnitude of active earth pressure, Rankine’s theory possesses highest match to the PLAXIS analysis. Parametric studies are also done to study the responses of active earth pressure distribution to varying parameters. Increasing soil friction angle and wall friction causes decrease in active earth pressure. In contrast, active earth pressure increases with increasing soil unit weight and height of wall. Concluding from this study, Rankine’s theory has the highest compatibility to finite element analysis among all theories, and utilization of this theory leads to proficient retaining wall design. Keywords: active earth pressure; retaining wall; PLAXIS; comparative study; Rankine’s theory; Coulomb’s theory
INTRODUCTION The major concern in retaining wall analysis and design is the determination of
magnitude and distribution of lateral earth pressure (active and passive) on the retaining wall. In retaining wall problems, active earth pressure contributes to the failure of the wall. Determination of active earth pressure distribution is a very important step in analysis of a retaining wall and subsequently in its design. Conventional theories from Rankine and Coulomb are widely used in retaining wall design to calculate magnitude and distribution of active earth pressure[1,2]. However, the linear distribution of the active earth pressure assumed
Dubrova is among the early researchers who incorporate arching effect into distribution of active earth pressure. Roscoe pointed out that Dubrova’s method to determine earth pressure distribution consider translational mode of wall displacement is open for criticism. Wang’s earth pressure is culvilinearly distributed. Therefore. Ko.. should be between the coefficient of active earth pressure. Ka. Wang  proposed an analytical method to determine a theoretical result for the earth pressure on a retaining wall on the basis of Coulomb’s theory. Paik & Salgado pointed out that Wang’s formulation gives a total active force equal to that calculated by Coulomb’s theory. The disadvantage of Dubrova’s method is that it provides only solution for straight surfaces but cannot be used for slope surfaces. and thus further investigation is needed. Paik & Salgado assumed that translational movement of wall and linear failure plane which has an angle of 45o + ϕ/2 to the horizontal.by Rankine and Coulomb has been pointed out by Terzaghi and Roscoe to be inaccurate. They checked the accuracy of the new coefficient of active earth pressure developed with the value of new active lateral stress ratio proposed by Paik & Salgado. K. No further discussion is provided by Wang. Kawn matches the values of Rankine’s active earth 2 . in reality. She considered that the wall rotates about the midheight. However. Coulomb’s solution for total active force is not exact. Distribution of active earth pressure is non-linear and maximum active earth pressure does not occur at toe of the wall. the coefficient of lateral earth pressure. and coefficient of earth pressure at-rest. According to Wang . Arching is a condition where stress redistribution of a part of soil mass with higher stress to a soil mass with lower stress. further researches were focused on obtaining the nonlinear distribution of active earth pressure by incorporating arching effect[5-8].
Different theories are available to be used in retaining wall design. 3 . Comparison between finite element analysis and analytical theories shows the compatibility of the theories to the actual behavior of active earth pressure. Hence. An overdesign of a retaining wall will lead to wastage of construction materials whereas an underdesign will lead to a higher risk of failure. instead of circular arch assumed by Paik & Salgado. there are limited studies on comparing the theories for their accuracy or fallacy in determining the active earth pressure.pressure coefficient for δ = 0 (smooth wall). Besides. Goel & Patra’s theory is an improvement of Paik & Salgado’s method. they compare their theory with other theories. analytical theories are studied in order to compare these theories. numerical modeling using PLAXIS software is conducted to check the accuracy of the theories. They compared their results to Paik & Salgado and showed that their theory is more accurate. Other than that. They concluded that planar failure surface with parabolic arch shape predicts closest to the experimental results. However the famous numerical method to be used is finite element analysis. there is a necessity to determine the most accurate theory to be used in retaining wall design. It is showed that Paik & Salgado’s method matches the experimental results with better compatibility than other theories. Other than the analytical theories discussed. In this study. The methods are Sokolovski’s method and Smear Shear Band Method by Hazarika & Matsuzawa . However. The well-established PLAXIS as finite element analysis software has its advantages to represent the actual behavior of active earth pressure distribution. different numerical methods are available to determine the distribution of lateral earth pressure. An accurate distribution of active earth pressure enables an efficient design of retaining wall which reduces the chance of over or underdesign of the walls.
finite element analysis creates partial differential equations to be solved numerically. L 10 Fixities Vertical fixity at toe of wall Prescribed displacement 0 Unit kNm2/m kN/m Unit m m - 4 . the accuracy of finite element analysis approximately matches the actual condition of retaining wall problems. EI 2.II METHODOLOGY: NUMERICAL MODELING USING PLAXIS Finite element analysis is one of the most accurate numerical methods to find an approximate solution for engineering problems. E Poisson’s ratio. EA 3 x 107 (c) GEOMETRICAL INPUTS FOR RETAINING WALL Geometrical Inputs Control Values Height of wall. υ Rinter Control Values 16. The model by Yang & Liu is referred with some modifications. ϕ Cohesion. In this study. With the aid of finite element software which can perform high number of iterations. Table 1 and Figure 1 show the parameters and geometries used in PLAXIS modelling.c Young’s modulus. numerical modelling using PLAXIS 8.2 (denoted as PLAXIS in later discussions) is conducted in order to compare the analytical theories to the finite element analysis and to carry out parametric studies. H 5 Width of backfill. In short.5 x 106 Normal stiffness.3 0.667 Unit kN/m3 Degree kN/m2 kN/m2 - (b) MATERIAL PROPERTY FOR RETAINING WALL Plate Properties Control Values Bending stiffness.4 36 0 30000 0. Table 1: Parameters used for numerical modelling of the retaining wall using PLAXIS (a) SOIL PROPERTIES Soil Properties Soil Unit weight (Dry) Friction angle.
5 . While Terzaghi assumed that the failure surface is approximately parabolic in nature and zero stress occurs at the base of the wall. underestimate the height of the center of pressure. They stated that the assumption by Rankine and Coulomb assume a planar failure surface that regardless of wall friction puts maximum pressure at the base of the wall.Figure 1: PLAXIS model for comparison between different theories on active earth pressure and FEA III RESULTS AND DISCUSSIONS (a) Distribution of active earth pressure Distribution of active earth pressure is non-linear and maximum pressure does not occur at the bottom of the wall (Figure 2). explained by partial support of the soil arching. Therefore the results verified the statement made by Goel & Patra.
From Figure 3.2.Figure 2: Distribution of active earth pressure of PLAXIS model (b) Comparison of analytical theories with PLAXIS modeling Figure 3 shows a graph combining all analytical theories on active earth pressure[1.5-8] and PLAXIS analyses including the present study and Yang & Liu’s analysis. After 80% of wall height. the distribution of active earth pressure matches to Rankine’s theory and remains linear. Goel & Patra’s theory gives higher active earth pressure than the present study at the upper midheight of the wall. For this portion. theories from Coulomb[1. The distribution of active earth pressure is linear at the upper midheight of the wall (from top of the wall to 50% of wall height). 2.2]. the distribution of active earth pressure is nonlinear. This matches the results from all analytical theories except Goel & Patra. most of the distribution of active earth pressure matches to Rankine’s theory. followed by other theories. 6 . Dubrova. the distribution of active earth pressure becomes parabolic until the bottom of the wall (100% of wall height). Therefore it can be concluded that up to 80% of wall height. the following discussions can be made: 1. For the lower midheight of the wall. From 50% to 80 % of wall height.
The height of application of maximum active earth pressure of PLAXIS analysis is within the range of 90 to 100% of wall height. but with lower magnitude of maximum active earth pressure. and only Rankine’s theory[1. maximum active earth pressure from Paik & Salgado is slightly higher than PLAXIS result.2] and Dubrova. Even though Rankine’s theory is one of the most conventional theories and its linear distribution has been pointed out by Terzaghi as a fallacy.2] shows the most compatible results to the PLAXIS analysis than these theories. Result from Goel & Patra shows the highest deviation from PLAXIS result. the compatibility of magnitude of Rankine’s of active earth pressure to FEA has proven its effectiveness to be used widely in retaining wall problems. at which its height of application of maximum active earth pressure is located in the range of 80 to 90% of wall height. Coulomb[1. Therefore.Wang. After that. the increasing order of degree of compatibility to FEA is Coulomb[1. Paik & Salgado and Goel & Patra give lower active earth pressure than present study. 4. result from Wang’s theory falls under this range as well. Therefore it can be concluded that Wang’s method provides the most compatible result to FEA for height of application of maximum active earth pressure. From Figure 3. Finally for theories from Rankine.2]. 7 . Goel & Patra. the maximum active earth pressure is located at bottom of the wall. it can be concluded that Rankine’s theory gives the most accurate results compared to other theories for both magnitude and distribution of active earth pressure. For other theories except Rankine’s theory. and Paik & Salgado. Wang. 3. from PLAXIS analysis. However. Dubrova.
H = 5m Figure 3: Different theories and PLAXIS analysis on active earth pressure (NOTE: Results from Coulomb’s theory overlapped with Dubrova’s theory) (c) Parametric Studies 1. the soil friction angle used in previous section is 36o. 8 . ϕ The first parametric study is carried out on varying soil friction angle. The values of ϕ used in this part are 0. The case of ϕ = 0 represents liquid behavior of the soil. 30o. Figure 4 shows the distribution of active earth pressure for different ϕ.Height of wall. ϕ. 36o. 20o. 10o. similar to Yang & Liu. and 40o. by using the same model used in previous section. From Table 1. Therefore model of ϕ = 36o is used as the control model in this parametric study. Soil Friction Angle.
and the active earth pressure is directly proportional to the vertical stress acting on the soil. This leads to a linear distribution of active earth pressure as shown in Figure 4. For ϕ = 0. there is no shear strength in the soil (zero cohesion and zero friction angle). the soil does not provide resistance to shear. the active pressure acting on every depth of the retaining wall decreases and the height of application of the maximum active earth pressure (towards top of the wall) increases. when the friction angle increases from ϕ = 0. For this case. Hence less active earth pressure can develop. the distribution of active earth pressure changes from linear to non-linear. when soil friction angle is more than 30o. the difference in distribution of active earth pressure is less when soil friction 9 . Moreover. the distribution of active earth pressure is linear with the highest value located at the bottom of wall. Therefore in liquid condition.Height of wall. As the friction angle increases. H = 5m Figure 4: Change of active earth pressure distribution with ϕ From Figure 4. The lowest active earth pressure and highest height of application of the maximum active earth pressure (towards top of the wall) was observed in case ϕ = 40o. This is due to the increasing internal shear strength within the soil with increasing soil friction angle.
4 kN/m3. 18 kN/m3.angle changes. 15 kN/m3. Height of wall. 2. PLAXIS model in previous section with unit weight of 16. This situation matches the previous discussion at which soil with a higher unit weight will exert a higher active earth pressure on the retaining wall. Refer to 10 . 16. γ The second parametric study is carried out on varying soil unit weight.4 kN/m3 is used as control model. H = 5m Figure 5: Graph of active earth pressure with γ From Figure 5. Similar to first parametric study. the active earth pressure acting on each depth throughout the retaining wall increases with increasing soil unit weight. the unit weights used are 12 kN/m3. Next. Distribution of active earth pressure with varying unit weight is shown in Figure 5. γ. Soil Unit Weight. These changes in response to varying in values of ϕ matches the parametric study carried out by Paik & Salgado . the height of application of maximum active earth pressure (towards top of the wall) does not change significantly in response to the change in soil unit weight. In this parametric study. compared to soil friction angles less than 30o.
Figure 6 shows the results from PLAXIS analysis. Therefore this can be explained that when soil unit weight increases. the height of wall does not affect the shape of the distribution of active earth pressure. Height of Wall. It means that linear portion of 11 . the values of height of wall used are 4. Moreover. H. the height of application of active earth pressure increases with a very small amount of depth towards the top of the wall only. 3. In this parametric study. when soil unit weight increases.Figure 5. the height of wall used is 5 meters. the distribution remains linear and has similar magnitude along the whole height of retaining wall except the last half meter from bottom of the wall. In control model from previous section. For the shape of active earth pressure distribution. Figure 6: Graph of active earth pressure with H From Figure 6. vertical stress acting on a soil mass increases. soil with higher unit weight requires a lesser wall displacement for development of active earth pressure. but have effect on its magnitude. H The varied parameter in the third parametric study is height of wall. 5 and 6 meters. When soil unit weight increases. the active earth pressure acting on retaining wall increases. and eventually causes lateral active stress to increase.
12 . the wall friction used is 26o. In the last half meter of the wall height. or in other words. which gives a Rinter value of 0. From Figure 7. δ The fourth parametric study is carried out on wall friction. The results matches to parametric study from Paik & Salgado shown in Figure 7. and 36o. In this parametric study.5 meters for 5 meters wall and at top 5. Figure 7: Change of active earth pressure distribution with normalised depth to wall height ratio 4. δ. the distribution of active earth pressure is in parabolic shape.5 meters for 4 meters wall. 10o. Wall Friction.5 meters for 6 meters wall. Results are showed in Figure 8.667.distribution of active earth pressure occurs at top 3. at top 4. the friction angle between soil and retaining wall. All values for wall friction used in this parametric study are 0. distribution of active earth pressure increases at every normalized depth to wall height ratio ( when height of wall increases. 26o. as given by Yang & Liu. 20o. and magnitude increases for increasing wall height.
both maximum active earth pressure and height of application of maximum active earth pressure towards top of wall increase. as wall friction increases. H = 5m Figure 8: Graph of active earth pressure with δ. when cohesion exists within the soil. Refer to Figure 8. a zone with zero active earth pressure occurs at the top of the wall. 13 . In Model II. The result matches with tensile crack zone given by Rankine’s and Coulomb’s theory. At upper zone of the wall. the distribution of active earth pressure is linear. cohesion of 1 kN/m2 is used. From the results (Figure 9). (d) Effect of cohesion Another PLAXIS model (denoted as Model II) is done with reference to Yang & Liu in order to investigate the effect of cohesion on distribution of active earth pressure.Height of wall. which is consistent to Rankine’s theory. for smooth wall (δ = 0). the magnitude of active earth pressure is almost equal with minimum deviation as wall friction increases. the distribution of active earth pressure changes from linear (for δ = 0) to non-linear. As wall friction increases. While for lower zone of the wall (which includes 0.5 meter from bottom of the wall).
the highest shear forces are located at the quarterheight from top of the wall. These results are important for design of retaining wall. the maximum bending moment is located at the middle height of the retaining wall. While Figure 10(b) and 10(c) show the bending moment and shear force of the retaining wall respectively. the middle height zone has to be designed to support high bending moment.Figure 9: Comparison of active earth pressure on (a) Model II and (b) PLAXIS model (e) Effective Normal Stress. highest reinforcement must be located near the middle height for reinforced concrete retaining walls. Therefore. Shear Force and Total Displacement of Retaining Wall Figure 10 shows the different forces acting on the retaining wall and total displacement of retaining wall in Model II. denoted as effective normal stresses acting on retaining wall in PLAXIS output. While from Figure 10 (c). From Figure 10 (b). In other words. and decreased bending moment near top and bottom of the wall. Bending Moment. Figure 10(a) shows distribution of active earth pressure acting on retaining wall. and near bottom of the wall. 14 . high shear reinforcement has to be applied at these zones for reinforced concrete retaining walls. Hence for retaining wall designs.
In addition.Then refer to Figure 10 (d). (c) shear force. Figure 10: (a) Effective normal stress. and (d) total displacement of retaining wall 15 . (b) bending moment. The direction of displacement of the wall is the same as the rotation about the bottom of wall. the magnitude of the total displacement is only 2.31mm. which is very small. total displacement of retaining wall is shown.
129-170. Geotechnique. McGraw-Hill. R. 643-653. 16 . which is Rankine’s theory as proven in this study. 50(1). • Maximum bending moment occurs at midheight of the retaining wall while maximum shear force is observed at quarter-heights at top and bottom of the wall. the active earth pressure increases. Singapore. On the other hand. 7th ed. Distribution of earth pressure on a retaining wall. Moscow. & Salgado. 20(2). H. K. (2003). A. Intersection of soil and structures. Geotechnique. B.IV • CONCLUSIONS From the comparison of active earth pressure calculated from analytical theories and finite element analysis.  Paik. (1996). (1963). 53(7).  Wang. Estimation of active earth pressure against rigid retaining walls considering arching effects. Principles of Foundation Engineering. International Thomson Publishing Asia. 71-88. Y. The influence of strains in soil mechanics. 83-88. Foundation Analysis and Design. Rechnoy Transport. (1936) A fundamental fallacy in earth pressure computations.  Dubrova. V REFERENCES  Das. E. Z. In summary. when the soil unit weight and height of wall increase. Journal of Boston Society of Civil Engineers. (1970). Izd. Geotechnique. J.  Bowles. • For cohesive soil. results from Rankine’s theory show the highest compatibility to the PLAXIS analysis. K. (2011).  Roscoe. the active earth pressure decreases. can lead to efficient retaining wall designs. 5th ed. tension zone with zero active earth pressure exists at top of retaining wall. 23. utilization of the most accurate theory. H. M.  Terzaghi. (2000). K. • While in parametric studies. for the magnitude and distribution of active earth pressure. when soil friction angle and wall friction increase. G.
V. K. (1996). Effect of arching on active earth pressure for rigid retaining walls considering translation mode. Finite Element Analysis of Earth Pressures for Narrow Retaining Walls. V. K. Goel. 19(3). H. N.  Sokolovski. S. & Patra.  Terzaghi. H. 17 . 8(2). London. 123133. & Liu. Journal of GeoEngineering. Butterworths Scientific Publications. 2nd ed. 193-219. Computer and Geotechnics. H. John Wiley and Sons. R. C. & Matsuzawa. (2008). 2(2).  Yang. Statics of Soil Media.  Hazarika. N. (1960). 43-52. International Journal of Geomechanics. (2007). Wall displacement modes dependant active earth pressure analyses using shear band method with two bands. (1954). New York. Theoretical Soil Mechanics.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.