Poisson 9s ratio of the different layers for the finite element model were taken from the experimental investigation and the evalu- ated properties obtained by backcalculation computer program, BACKGA. The elastic moduli of subbase were determined using K- ¸ model and Park and Fernando model. Using the same FE model, material properties (k1, k 2 & k 3 ) of the subbase were analysed to evalu- ate the stress dependency of granular subbase layer using Park and Fernando model for characteri- zation of granular material.<br><br> The subbase moduli obtained by using Park and Fernando equation were found to be close to the backcalculated values using BACKGA, and these were fitted in the model to determine the equivalent elastic modulus of the plastic cell filled concrete block layer through iterative process. The equivalent elastic moduli for different load repetitions were found to be slightly higher for the block pavement of 75 mm thick and for the block pavement of 100 mm thick; the elastic moduli as determined by FE Model are found to be as close to the backcalculated moduli. A design catalogue is presented for selection of pavement crust for new construction.<br><br> 1. INTRODUCTION Plastic cell filled concrete block pavement originated in South Africa [Visser, 2003; 1999; 1994; Visser and Hall, 1999]. This type of pavement consists of a formwork of polyethylene square cells 150 mm x 150 mm in size with thickness of 50 mm or greater (See Figure 1).<br><br> The plastic cells are stretched over the carriageway under tension, filled with concrete, and compacted. During compac- 9th. International Conference on Concrete Block Paving.<br><br> Buenos Aires, Argentina, 2009/10/18-21 Argentinean Concrete Block Association (AABH) - Argentinean Portland Cement Institute (ICPA) Small Element Paving Technologists (SEPT) 2 tion, the cell walls get deformed (See Figure 2) bringing about interlocking among the concrete blocks. Alternatively, the cells are filled with single size aggregates [Visser and Hall, 1999] and rolled with static or vibratory roller. A mortar of cement and sand is vibrated into the voids by a plate compac- tor and the road can be opened to light traffic within 24h of the construction.<br><br> The cell-filled con- crete blocks behave as a flexible layer which will conform to the deformed shape of the underlying subbase if there is any secondary compaction or settlement of the underlying layers. This type of construction is labour intensive, cost effective and appears to be highly suitable for rural areas of India where mechanized construction and strict quality control may not be possible. Figure 1.<br><br> A formwork of Plastic cell after ten- sioning (150mm x 150mm) (Ryntathiang, 2005). Figure 2. Deformed plastic cell walls after compaction bringing about Interlocking (Ryntathiang, 2005).<br><br> The paper deals with a simple 2-D Finite Element study on determining the Equivalent Elastic Modulus of Plastic Cell Filled Concrete Block Pavements that were laid on the bed of accelerated pavement test facility. A brief detail of the Test Set Up, Construction of test pavements that were laid on the bed of accelerated pavement and testing procedure is also presented. Test data that were reported in the paper by Ryntathiang et al.<br><br> [2005] were used for the study so that the equivalent elastic modulus of plastic filled concrete block pavement at different standard axle load repetitions can be determined and compared with the backcalculated equivalent elastic modulus. 2. EXPERIMENTAL INVESTIGATION: 2.1 Test set up The test setup consisted of a dual wheel set, which can be loaded up to 60 kN.<br><br> The wheel can be moved to and fro by using 20 kW motor over a linear track of length 15 m. The numbers of cycles of loading were recorded in a counter. In their investigation, the loading bin was filled with steel rail sections, so that the load coming on the dual tires was about 40 kN.<br><br> A tire pressure of 586 kPa was maintained and load on the wheel was checked by a portable weigh-bridge available in the laboratory. 2.2 Construction of test pavement The test setup consists of two test pavements with thicknesses 75 mm and 100 mm of plastic cell filled concrete blocks, which were constructed at the test area of APTF, IIT Kharagpur. The test pavement was kept at 2.1 m x 2.0 m.<br><br> The sectional details of the both test pavements are shown in Figure 3. The subgrade for both test sections consisted of compacted sand and the foundation for the sand subgrade was formed by the existing undisturbed lateritic soil having a CBR value between 3% and 4%. Over the granular subbase, the formwork of plastic cells was stretched in all four di- rections so that each cell was close to a square of 150 mm sides, depths being 75 mm and 100 mm.<br><br> 9th. International Conference on Concrete Block Paving. Buenos Aires, Argentina, 2009/10/18-21 Argentinean Concrete Block Association (AABH) - Argentinean Portland Cement Institute (ICPA) Small Element Paving Technologists (SEPT) 3 Figure 3.<br><br> Plastic cell filled concrete block pavement of 75 mm and 100mm. 2.3 testing procedure The entire operation of APTF is controlled by a microprocessor. After the electrical switch is put on, the dual wheel starts moving to and fro without any lateral wander.<br><br> In the investigation, the number of repetitions of the wheel load per hour was found to be about 240. Surface levels of the pavement were recorded before the start of the test and after every 1 000 repetitions of the wheel load. The deflection profile of the test was determined by using falling weight deflectometer (FWD).<br><br> The deflections were measured at 0, 300, 600, and 900 mm from the center of the loading plate of the FWD. A backcalculated computer program developed by IIT Kharagpur laboratory (BACKGA) is used to evaluate the modulus values of different layers and these values are com- pared with the values determined by FE Modelling. 3.<br><br> 2-D FINITE ELEMENT STUDY 3.1 Model analysis Axial symmetric 2-D finite element modelling was carried out using commercial FE software, AN- SYS (6.0) to compare equivalent elastic modulus of the model with those of the experimental re- sults. The finite element model dimensions were taken similar to that of the experimental investiga- tion [Ryntathiang et al., 2005] and are shown in Figure 4. The thickness of the plastic cell filled concrete block, subbase and subgrade layers considered in the FE modelling are 75/100 mm, 102/96 mm and 990/850 mm respectively.<br><br> For the FE model, the width of the pavement section adopted is 1050 mm. Figure 5 shows a typical FE mesh idealization using 4-nodded quadrilateral axial- symmetric elements. In this model analysis, the top plastic cell filled concrete block layer has been modelled as continuous layer for simplification.<br><br> The load applied to the pavement structure is of 40kN with a radius of contact 150 mm. The pave- ment was treated as a three-layer system of which, the first or top layer consists of plastic cell filled concrete blocks, second layer consists of subbase layer and third consists of subgrade wherein each layer is having an elastic property. The Po isson 9s ratios were taken as 0.25 for the top layer, 0.35 for the second and third layer respectively.<br><br> The elastic moduli of subgrade were taken from values obtained by Ryntathiang et al. [2005]. Equivalent elastic moduli of subbase for different repetitions of load were determined by using K- ¸ model [Naidu and Pandey, 1994] and Park-Fernando model [1998].<br><br> The elastic moduli as determined by Park and Fernando model for different repetitions of 75mm Plastic Cell Concrete Paving Units Edge Restraint Plastic 990mm Sand Sub rade Guide Rail Soil Soil Brick Wall Brick Wall Soil PI=8.44% 102mm WMM Subbase Plastic Sheet 100mm Plastic Cell Concrete Paving Units Edge Restraint Plastic 850mm Sand Sub rade Guide Rail Soil Soil Brick Wall Brick Wall Soil PI=8.44% 96mm WMM Subbase Plastic Sheet 9th. International Conference on Concrete Block Paving. Buenos Aires, Argentina, 2009/10/18-21 Argentinean Concrete Block Association (AABH) - Argentinean Portland Cement Institute (ICPA) Small Element Paving Technologists (SEPT) 4 standard load were found to be close to the backcalculated moduli and these moduli were further used to determine the equivalent elastic modulus of the concrete block layer.<br><br> Fig.-4. Basic geometry of the pavement section Because of symmetric conditions, one half of the pavement was analyzed. The nodes at the bottom of the model cannot move in z-direction whereas the nodes at the edge of the model can only move in the z-direction.<br><br> These boundary conditions are simulated so that they represent the experimental test setup. The validity of the present model has been checked by comparing with the results of F- PAVE [Das, 1998] a multi-layer elastic program developed by civil engineering department, IIT Kharagpur. Fig.<br><br> 5: FE meshes of the pavement section Modeled as Continuous Block layer Subbase Sub grade Load (P/2) for 2-D 9th. International Conference on Concrete Block Paving. Buenos Aires, Argentina, 2009/10/18-21 Argentinean Concrete Block Association (AABH) - Argentinean Portland Cement Institute (ICPA) Small Element Paving Technologists (SEPT) 5 3.2 characterization of equivalent elastic modulus of granular subbase A two dimensional finite element model as shown in Figure 5 was used for the characterization of equivalent elastic modulus of granular layer.<br><br> The two models used for the present study to evaluate the equivalent elastic modulus of the granular layer are Park and Ferdinand Model and K- ¸ model. Park and Fernando model was used to evaluate the material parameters of the subbase layer. For the study, stresses were computed at the mid-depth of the subbase layer vertically below the centre of the loaded plate.<br><br> 3.2.1 Park and Fernando equation Park and Fernando (1998) developed a non-dimensional Equation for modelling the modulus of granular materials. The equation is given as: ¡Error! No se pueden crear objetos modificando códigos de campo.<br><br> (1) Where MR = Resilient modulus of granular material Atm = Atmospheric pressure (0.1MPa) ¸ = Sum of principal stresses (Ã1+2Ã3) Äoct = Octahedral shear stress = { a2(Ã1 3Ã3)}/3 k1, k2 & k3 = Material parameters. The above equation was used to evaluate the values of the material parameters from the known val- ues of elastic modulus, ¸ and Ä oct . The values of ¸ and Ä oct at mid depths of the subbase were used in Equation 1.<br><br> These material parameters were further used to determine the subbase modulus. 3.2.2 K- ¸ model The K- ¸ model [Naidu and Pandey, 1994] is given as, ¡Error! No se pueden crear objetos modificando códigos de campo.<br><br> (2) Where: MR = Resilient modulus of granular material ¸ = Sum of principal stresses (Ã1+ 2Ã3), kPa k1 and k2 = Material parameters = 3.47 and 0.7375 for WMM The values of k 1 & k 2 were evaluated by Naidu and Pandey [1994] for granular material having gra- dation close to that of the WMM used in the present study. 4. RESULT AND DISCUSSION The FE model analysis was conducted for various load repetition to obtain the material parameters for characterization of elastic modulus of granular subbase layer using Park- Fernando model and to find the modulus of the subbase by using K- ¸ models.<br><br> 9th. International Conference on Concrete Block Paving. Buenos Aires, Argentina, 2009/10/18-21 Argentinean Concrete Block Association (AABH) - Argentinean Portland Cement Institute (ICPA) Small Element Paving Technologists (SEPT) 6 Table 1.<br><br> Evaluated material parameters from Park and Fernando equation with different Load Repetitions. LOAD REPE- TITIONS GRANULAR SUBBASE ELASTIC MODULUS OF 75mm PLASTIC CELL(MPa)75 (MPa) GRANULAR SUBBASE ELAS- TIC MODULUS OF 100mm PLASTIC CELL(MPa) K 1 K 2 K 3 K 1 K 2 K 3 2 000 8.8153 4.1851 1.4552 189.942 0.6999 3.9477 3 000 23.3651 3.7466 0.5205 191.0426 0.6985 3.9444 4 000 153.6274 2.7585 -1.0604 189.942 0.6999 3.9477 5 000 8.8153 4.1851 1.4552 191.0426 0.6985 3.9444 6 000 128.2295 2.9655 -1.1404 191.0426 0.6985 3.9444 Table 1 shows the evaluated values of material parameters k 1 , k 2 & k 3 obtained for different load repetitions. These material parameters were then used in Equation 1 for evaluating the granular subbase modulus through iterative procedure.<br><br> Table 2. Comparison of Equivalent elastic modulus of granular subbase layer with different load repetitions as evaluated by Park-Fernando equation and K- ¸ model with the Backcalcu- lation computer program, BACKGA experimental modulus. LOAD REPE- TITIONS GRANULAR SUBBASE ELASTIC MODULUS OF 75mm PLASTIC GRANULAR SUBBASE ELASTIC MODULUS OF 100 mm PLASTIC PARK AND FER- K 3?<br><br> MODEL BACKCALCU- LATION PARK AND FER- K 3? MODEL BACKCALCU- LATION 2 000 175 214 181 75 162 82 3 000 195 212 202 70 160 80 4 000 205 209 216 75 162 82 5 000 175 214 184 70 160 80 6 000 170 214 180 70 160 72 Table 2 shows the comparison of the modulus of granular subbase obtained by FE analysis with the Backcalculation computer program, BACKGA of experimental investigation [Ryntathiang et al., 2005]. From the above table, it can be seen that granular subbase moduli at different repetitions of load evaluated by Park and Fernando equation are closer to the backcalculated moduli as compared to moduli computed by K- ¸ model.<br><br> It is also observed that the granular subbase moduli obtained by k- theta model for 100 mm block thickness is about two times the moduli obtained through backcalcu- lation. The moduli obtained by Park and Ferdinand equation for the 75 mm and 100 mm thick blocks were further used to determine the equivalent elastic moduli of plastic cell filled concrete block layer at different load repetitions. 5.<br><br> EQUIVALENT ELASTIC MODULUS OF BLOCK LAYER To determine the equivalent modulus of the plastic cells filled concrete block layer at different load repetitions, the subbase moduli obtained from Park and Ferdinand Equation given in Table 2 (being close to the backcalculated modulus) were used in the FE model. An assumed modulus (say 50 MPa) was assigned to the block layer for the first iteration, and the iteration was stopped till the old modulus and the new modulus obtained converged to the first decimal place and the deflection of 9th. International Conference on Concrete Block Paving.<br><br> Buenos Aires, Argentina, 2009/10/18-21 Argentinean Concrete Block Association (AABH) - Argentinean Portland Cement Institute (ICPA) Small Element Paving Technologists (SEPT) 7 the pavement was calculated. A plot between equivalent elastic modulus of the block layer versus deflection is drawn (Figures 6 to 15) to find the equivalent modulus of the block layer [Panda, 2000] corresponding to the deflection obtained during Falling Weight Deflectometer test at 40 kN load over a 300 mm diameter circular load. The deflection at D/20 mm (Stress being maximum at this distance from the centre) distance was chosen to find the equivalent elastic moduli of the con- crete filled block pavement for different load repetitions.<br><br> Table 4: Equivalent elastic modulus of plastic cell filled concrete block pavement layer LOAD REPE- TITIONS COMPUTED ELASTIC MODULUS OF 75 mm PLASTIC CELL(MPa) 75 (MPa) COMPUTED ELASTIC MODULUS OF 100 mm PLASTIC CELL(MPa) FE Analysis Backcalculation FE Analysis Backcalculation 2 000 3 162 2 978 1 900 1 766 3 000 2 903 2 988 1 845 1 598 4 000 2 944 2 990 2 132 2 378 5 000 3 021 2 966 2 102 2 337 6 000 3 203 2 995 2 166 2 542 Table 4 above shows comparison between backcalculated modulus and the modulus obtained by FE model for the different load repetitions. It can be seen that the equivalent elastic modulus of the cell filled concrete block pavement at different load repetitions is more or less of the same values as cal- culated by backcalculation method. 6.<br><br> DEVELOPMENT OF DESIGN CATALOGUE Based on the available literature and the data obtained in the present study, an equivalent elastic modulus of 2 500 MPa is appropriate in consideration for the design. Design chart was first devel- oped and then design catalogue for low volume roads having 5x10 4 to 2x10 6 repetitions of standard axles are extracted from the design chart and proposed as a design for plastic cell filled concrete block pavement. As the performance of plastic cell filled concrete block pavement is found similar to that of Interlocking Concrete Block Pavement [Visser and Hall, 1999] where failure of the pave- ment is only through rutting, therefore, vertical subgrade strain ( µ z ) is the criterion for the design.<br><br> For the development of design chart, several combinations of pavement sections were analysed us- ing the computer program FPAVE developed by IIT Kharagpur [Das, 1998] to determine the verti- cal subgrade strain ( µ z )). Values of µ z were computed at the top of subgrade, vertically below the center of standard dual wheel loads (40 kN) for tyre pressure of 0.56 MPa. The Poisson 9s ratios of the top layer, subbase layer and subgrade were taken as 0.3, 0.35 and 0.35.<br><br> The subgrade modulus was calculated as given by Lister and Powell 9s Equations [Lister and Powell, 1987] and the subbase modulus was obtained from the SHELL 9s Equation [Claessen et al., 1977; Dormon and Metcalf, 1965]. The proposed design catalogue for selection of pavement crusts are given in Table 5 and Table 6. 9th.<br><br> International Conference on Concrete Block Paving. Buenos Aires, Argentina, 2009/10/18-21 Argentinean Concrete Block Association (AABH) - Argentinean Portland Cement Institute (ICPA) Small Element Paving Technologists (SEPT) 8 Figures 6 to 10. Modulus-Deflection Curve for 75 mm thick block with granular subbase layer of 102 mm.<br><br> Figure 6. Modulus-Deflection curve for 2 000 times load repetitions. 0 1000 2000 3000 4000 0.4 0.5 0.6 0.7 0.8 Block Modulus(MPa) Deflection(mm) 0 1000 2000 3000 0.4 0.5 0.6 0.7 0.8 Block Modulus(MPa) Deflection(mm) Figure 7.<br><br> Modulus-Deflection curve for 3 000 times load repetitions. 0 1000 2000 3000 0.4 0.5 0.6 0.7 0.8 Block Modulus(MPa) Deflection(mm) Figure 8. Modulus-Deflection curve for 4 000 times load re etitions.<br><br> 0 1000 2000 3000 4000 0.4 0.5 0.6 0.7 0.8 Block Modulus(MPa) Deflection(mm) Figure 9. Modulus-Deflection curve for 5 000 times load re etitions. Figure 10.<br><br> Modulus-Deflection curve for 6 000 times load re etitions. 0 1000 2000 3000 4000 0.4 0.5 0.6 0.7 0.8 Block Modulus(MPa) Deflection(mm) 9th. International Conference on Concrete Block Paving.<br><br> Buenos Aires, Argentina, 2009/10/18-21 Argentinean Concrete Block Association (AABH) - Argentinean Portland Cement Institute (ICPA) Small Element Paving Technologists (SEPT) 9 Figures 11 to 15. Modulus-Deflection Curve for 100 mm thick block with granular subbase layer of 96 mm. 0 1000 2000 3000 0.2 0.4 0.6 0.8 1 Block Modulus(MPa) Deflection(mm) 0 1000 2000 3000 0.2 0.4 0.6 0.8 1 Block Modulus(MPa) Deflection(mm) Figure 14.<br><br> Modulus-Deflection r Figure 13. Modulus-Deflection r Figure 12. Modulus-Deflection curve.<br><br> 0 1000 2000 3000 0.4 0.5 0.6 0.7 0.8 Block Modulus(MPa) Deflection(mm) Figure 11. Modulus-Deflection curve. 0 1000 2000 3000 0.4 0.5 0.6 0.7 0.8 Block Modulus(MPa) Deflection(mm) Figure 15.<br><br> Modulus-Deflection curve. 0 1000 2000 3000 0.2 0.4 0.6 0.8 1 Block Modulus(MPa) Deflection(mm) 9th. International Conference on Concrete Block Paving.<br><br> Buenos Aires, Argentina, 2009/10/18-21 Argentinean Concrete Block Association (AABH) - Argentinean Portland Cement Institute (ICPA) Small Element Paving Technologists (SEPT) 10 Table 5. Pavement codes of CISCBP for use with reference to Table 6. SUBGRADE CBR (%) VILLAGE ROAD STANDARD AXLE LOAD REPETITIONS Less than 5x10 4 5x10 4 to 1x10 5 1x10 5 to 2x10 5 2x10 5 to 4x10 5 4x10 5 to 9x10 5 9x10 5 to 2x10 6 2 A A A B C D 3 A A A A E F 4 A A A A A E 5 A A A A A A > 5 G G G A A A Table 6.<br><br> Pavement structures of CISCBP to be used with reference to Table 5. PAVEMENT STRUC- TURE PAVEMENT COMPO- SITIONS PAVEMENT STRUCTURE PAVEMENT COMPO- SITIONS A 100 mm concrete blocks Subbase layer=100 mm E 100 mm concrete blocks Subbase layer=140 mm B 100 mm concrete blocks Subbase layer=165 mm F 100 mm concrete blocks Subbase layer=200 mm C 100 mm concrete blocks Subbase layer=245 mm 75 mm concrete blocks Subbase layer=100 mm D 100 mm concrete blocks Subbase layer=320 mm 7. CONCLUSION The model analysis using axial symmetric 2-D finite element model was found to be satisfac- tory and reliable as the result obtained were close to the backcalculated analysis.<br><br> The same 2-D finite element was used to evaluate the material parameters of granular layer for characterization of its equivalent elastic modulus by using K- ¸ model and Park -Fernando model. The Park-Fernando model was found to be close to the backcalculation computer pro- gram, BACKGA. Subbase Concrete block layer 9th.<br><br> International Conference on Concrete Block Paving. Buenos Aires, Argentina, 2009/10/18-21 Argentinean Concrete Block Association (AABH) - Argentinean Portland Cement Institute (ICPA) Small Element Paving Technologists (SEPT) 11 The equivalent elastic modulus of plastic cell filled concrete block layer for the different repeti- tions of load for the two pavement thicknesses as determined by FE Model was found to have a difference of 0-15% as compared to backcalculated modulus. A simple design catalogue is proposed for the design of new roads.<br><br> 8. REFERENCES ANSYS (Version 6.0). Ansys, Canonsburg, PA CLAESSEN, AIM; EDWARDS, JM, SOMMER, P, AND UGE, P, 1977.<br><br> Asphalt Pavement De- sign-The SHELL Method. Proc. 4 th Int.<br><br> Conf. on Struct. Des.<br><br> of Asp. Pavements, Vol. 1, pp 39-74.<br><br> DAS, A, 1998. Analytical Design Of Bituminous Pavements Based On Field Performance. PhD Thesis, Department of Civil Engineering, Indian Institute of Technology Kharagpur, India.<br><br> DORMON, GM AND METCALF, CT, 1965, Design Curves For Flexible Pavements Based On Layered System Theory. Highway Research Board, Highway Research Record, No. 71, pp.<br><br> 69-84. LISTER, NW, AND POWELL, WD, 1987, Design Practice For Bituminous Pavements In The United Kingdom. 6th Int.<br><br> Conf. On Structural Design Of Asphalt Pavement s , Vol. 1, pp.<br><br> 220-231. NAIDU, PK, AND PANDEY, BB, 1994, Dynamic Modulus Of Non Bituminous Materials In Flexible Pavement Design. Highway Research Board Bulletin, No.<br><br> 50, Indian Road Congress, New Delhi 189-195. PANDA, BC, 2000. Structural Behaviour Of Concrete Block Pavements.<br><br> Department of Civil En- gineering, Indian Institute of Technology Kharagpur, India. PARK SEONG-WAN AND FERNANDO EG, 1998, Sensitivity Analysis Of Predicted Stress- Dependency and Plastic Behavior for Load Zoning. 5th Int.<br><br> Conf. On The Bearing Capacity of Road and Airfields, Trondheim, Norway, pp. 727-736.<br><br> RYNTATHIANG, TL, 2005. An Investigation On Precast And Cast In-situ Concrete Block Pave- ments for Low Volume Roads. PhD Thesis, Department of Civil Engineering, Indian Institute of Technology Kharagpur, India.<br><br> RYNTATHIANG, TL; MAZUNDAR, M AND PENDEY, BB, 2005, Structural Behavior Of Cast In-Situ Concrete Block Pavement, Jnl of Transp. Engg. September, pp.<br><br> 662-668. VISSER, AT, 1994, A Cast In-Situ Block Pavement For Labour-Enhanced Construction. Concrete Beton, Concrete Society of South Africa, No.<br><br> 71, February, pp. 1-8. VISSER, AT, 1999, The Response Of Flexible Portland Cement Concrete Pavements Under Ultra Heavy Loading.<br><br> Concrete Beton, Concrete Society of South Africa, pp. 11-18. VISSER, AT, 2003, Innovative And Cost Effective Solutions For Roads In Rural Areas And Diffi- cult Terrain.<br><br> Low Volume Road Conference, Reno, Nevada, pp. 1-14. VISSER, AT, AND HALL S, 1999, Flexible Portland Cement Concrete Pavement For Low- Volume Roads.<br><br> Transportation Research Record 1652, Washington DC, pp. 121-127. VISSER, AT, AND HALL S, 1999, Flexible Portland Cement Concrete Pavement For Low- Volume Roads.<br><br> Transportation Research Record 1652, Washington DC, pp. 121-127. <br><br>

less