TRACK SUB-STRUCTURE ANALYSIS FOR A HIGH SPEED TRAIN IN MALAYSIA

Transcription

1 TRACK SUB-STRUCTURE ANALYSIS FOR A HIGH SPEED TRAIN IN MALAYSIA Affendi Abdullah 1, A. Arulrajah 2 and Siew Ann Tan Harry 3 1 Affendi & Partners Consult, 50 Jalan Jasa 1, Taman Jasa, Batu Caves, Malaysia. draffendi52@yahoo.co.uk 2 Swinburne University of Technology, Faculty of Engr and Ind Sciences (H38), Hawthorn, VIC 3122, Australia. AArulrajah@groupwise.swin.edu.au 3 National University of Singapore, Dept. of Civil Engineering, 1 Eng Drive 2, E1A 07-03, Singapore cvetansa@nus.edu.sg A high-speed railway project in Malaysia for trains of speeds of up to 160 km/hr has been constructed between Rawang and Bidor (110 km long) in Peninsular Malaysia. The project requirement for track substructure analysis for the above stretch of works was studied by Ranhill Consulting. The geotechnical factors which are of importance for the subgrade performance are the strength and stiffness of the subgrade soil. Laboratory tests carried out, include grain size distribution, California Bearing Ratio and cyclic triaxial load test to determine the Threshold stress and the Resilient Modulus. Dynamic amplification factor and static deflection are influenced by the stiffness of the subgrade. Finite element analyses were carried out using a Young s modulus of 25 Mpa for subgrade to indicate a corresponding static deflection that is acceptable. Develop relationship between CBR and cyclic triaxial resilient modulus, undrained shear strength and resilient modulus as well as comparison of strength and resilient modulus measure in cyclic triaxial test. Based on this requirement, it was safely assessed that low embankment construction using Removal and Replacement Method have met the project s required criteria. Keywords: Treshhold stress, Resilient modulus, Cyclic triaxial test, Track sub-structure analysis. 1. INTRODUCTION As part of the double-tracking railway works from Rawang to Bidor in Malaysia, Ranhill has studied the requirement for track substructure analysis for the above stretch of works. The geotechnical factors which are of importance for the subgrade performance are the strength and stiffness of the subgrade soil. Technical requirements for the track subgrade in this high speed railway project requires to achieve soaked CBR of the subgrade value of 5% and the field dry density should be at least 95% of the maximum dry density of the modified Proctor test. Proceedings of the International Conference on Ground Improvement and Ground Control Edited by Buddhima Indraratna, Cholachat Rujikiatkamjorn and Jayan S. Vinod Copyright 2012 by Research Publishing Services. All rights reserved. ISBN: :: doi: /

2 1668 Proceedings of the International Conference on Ground Improvement and Ground Control It can be safely assumed that the soaked CBR of 5% will equate to an undrained shear strength of 150 kpa. Due to the effect of suction for compacted soil above the water table, the actual strength of the subgrade material could exceed the 150 kpa value. The principal of load transfer equates to an expected stress level of 50 kpa on the subgrade (Modern Railway Track) based on an axle load of 200 kn. As such, the ratio of applied stress/strength of the subgrade is about 1/3. Li and Selig (1998) reported that an acceptable threshold level ratio of 1/2 can be applied and as such a ratio of 1/3 is on the conservative side. Dynamic amplification factor and static deflection are influenced by the stiffness of the subgrade. Finite element analyses carried out by the Nanyang Technological University using a Young s modulus of 25 Mpa for subgrade has indicated a corresponding static deflection of 6 mm which is acceptable. However, at 3 meters depth, the corresponding Eu value reduces to about 14 MPa. This would correspond to a required Cu of 70 kn/m 2 based on the equation Eu = 200*CuandtoaSPT N value of blows. As such, so long the sub-soil profile at the Removal and Replacement (R/R) ground treatment areas achieve this value at the base of the R/R works, the subgrade would have achieved the desired requirements. 2. SOIL SAMPLES AND TESTING 2.1. Soil Samples Soil Samples were collected by the contractors and consultants at eleven locations along the Electrified Double Track Train Route between Rawang and Ipoh, Malaysia. Method of sampling was by block sampling. One block soil sample at each chainage was obtained in-situ that is, after compaction in the field from the actual compacted track bed (at subgrade level). This was achieved by excavating a pit and then extracting a block of soil. The block samples were about mm in size, and were waxed after excavation and were sent to National University of Singapore Laboratory Outline of Tests Samples used in testing were cut sample from block samples. The tests conducted included grain size distribution curves, CBR and Cyclic triaxial tests. Standard grain size distribution (particle size distribution) tests were conducted on disturbed samples from block samples. Wet sieving method was used, as there were significant amount of fines (silt and clay) in the material. The samples for CBR tests, (152 mm diameter and 127 mm height), were prepared by carefully cutting the block sample using a saw. CBR soil samples were soaked for 4 days, and a 15 kpa surcharge was applied during soaking which represents weight of ballast etc. The triaxial samples were prepared by pushing 50 mm diameter thin-walled tubes into the block sample. It was assumed that water content in the block sample did not change very much since the in-situ compaction, due to natural drying, sampling and subsequent time lapse between sampling and testing. A total of 11 particle size distribution, 15 CBR tests and 11 cyclic triaxial tests were carried out. Four additional cyclic triaxial tests were also carried out to determine threshold value at two locations. The four additional (15 instead of 11) CBR tests were carried out as a result of uncertainty in the water content.

3 Track Sub-Structure Analysis for a High Speed Train in Malaysia GRAIN SIZE DISTRIBUTION A total of eleven particle size distribution curves were plotted. Summary on particle size distributions shows that except for soil samples from Ch , Ch , and Ch , all other soil samples have very high fine content. The soil samples from Ch , Ch , Ch and Ch have particularly very high fine content generally more than 60%. Presence of significant amount of fines may pose problems in the long-run, as material gets weaker due to accumulation of pore pressures due to cyclic loading from train loading. 4. CBR TESTS CBR samples were cut from block samples were soaked for 4 days under a surcharge of 15 kpa representing weight of ballast and superstructure on the sub-grade. The CBR test was carried out according to the standard procedure. A plunger was penetrated into soil sample at a constant rate of 1 mm/min. Proving ring and dial gauges were used for measuring load and displacements respectively. CBR1 and CBR2 values at 2.5 mm and 5.0 mm penetration were calculated, and the higher of the two values was selected as the CBR value. At chainages Ch and , it was possible to cut an undisturbed soil sample for CBR test. It was also possible to prepare some CBR samples from Ch , Ch , Ch , Ch , and Ch Therefore, at these chainages it was possible to carry out CBR tests on the field undisturbed samples. At locations Ch , Ch , Ch , and Ch it was not possible to cut an undisturbed CBR soil samples from the block sample. Therefore, CBR soil samples were prepared at 10% and 15% water contents in the laboratory, as optimum water content is not known. It should be noted that, since optimum water content is not known, the values were used with caution. 5. CYCLIC TRIAXIAL TESTS Cyclic triaxial tests were carried out on 50 mm diameter and 100 mm long diameter soil samples. The objectives of the cyclic tests are to obtain variation of resilient modulus with number of cycles. The equipment is capable of applying cyclic deviatoric load up to 2 Hz. All the cyclic triaxial experiments reported in this report were conducted at 1 Hz. The number of cycles carried out is 10,000. Soil samples were placed in triaxial system and 20 kpa of cell pressure was applied. The 20 kpa was based on pressure due to ballast and sub-ballast (0.6 m*24 kn/m3 = 14.4 kpa) plus load due to sleeper, rail etc. The magnitude of deviatoric loading applied during cyclic loading was 72 kpa. Three dimensional FE analyses showed that static stress on sub-grade was 38 kpa. Dynamic Amplification Factor (DAF) is 1.94 according to American railway Engineering Association (AREA). Therefore, maximum deviatoric loading is 38*1.94 = 72 kpa. After application of cell pressure of 20 kpa, 10,000 cycles at 1 Hz were applied. Additional tests were carried out to determine threshold stress values Results of Soil at Ch Out of the eleven chainages studied, results obtained on samples from Ch are shown in Figures 1 a, b & c to relate the cyclic stress strain behaviour, resilient modulus and

4 1670 Proceedings of the International Conference on Ground Improvement and Ground Control a) b) c) Figure 1. a) Cyclic stress strain behaviour, b) Resilient Modulus vs number of cycles, Unconsolidated Undrained Test. Table 1. Resilient Modulus and Strength values measured on soil samples. Location Resilient Modulus, E r, MPa Shear Strength, C u,kpa E r /C u Ch Ch Ch Ch Ch Ch Ch Ch Ch Ch Ch stress strain behaviour. The accumulated plastic strain after 10,000 cycles is about or 0.8%. Resilient modulus versus number of cycles varied from 46 MPa to 55 MPa. Residual deviatoric stress is about 250 kpa and therefore undrained shear strength is about 125 kpa. Summary for other chainages are presented in Table 1.

5 Track Sub-Structure Analysis for a High Speed Train in Malaysia 1671 Figure 2. Comparison of Strength and Resilient Modulus measured in Cyclic Triaxial Test. 6. THE RELATION BETWEEN STRENGTH AND RESILIENT MODULUS The undrained shear strength and resilient modulus measured in various triaxial tests is compared in Figure 2. The data seems to be scattered, therefore it is difficult to determine one relation which is applicable to all soil conditions. Two lines has been drawn based on the data available and are as follows: and E r = 0.67C u (for granular material) (1) E r = 0.45C u (for clayey material) (2) 6.1. Present Experimental Work Relation between triaxial resilient modulus and CBR is shown in Figure 3. CBR values used are an average value for a given soil. Since resilient modulus (Er) and CBR values were not measured on the same soil sample, some differences could be expected. Some Figure 3. Relation between CBR and Cyclic Triaxial Resilient Modulus.

6 1672 Proceedings of the International Conference on Ground Improvement and Ground Control scatter is expected as water content has significant influence on strength and stiffness. A relation in which resilient modulus is 10 times CBR (Er is MPa, CBR in %) was deduced from the present data, as shown in Figure 3. The variation in data is result of difference in soil types. 7. RELATION BETWEEN STRENGTH AND CBR An attempt is made to obtain relation between undrained strength and CBR. The CBR test and undrained shear strength were carried out on different samples at different times. Undrained strengths were carried on 95% maximum dry density whereas CBR tests were carried out 100% maximum dry density. A relation was attempted based on average values. No unique C u /CBR ratio was obtained, but ranged between RESEARCH WORK BY OTHERS For roads applications, the modulus test on subgrade soils is a well-researched topic. A number of researchers attempted to relate CBR and Resilient modulus. For fine-grained subgrade soils, several correlations with CBR have been established. The correlation obtained by Huekelom and Klomp (1962) was: E r = 1500 (CBR) in psi or approximately 10(CBR) in MPa. The present experimental work (Figure 3) shows some scatter, but E r = 10 CBR seems a reasonable fit similar to the relationship of E r = 10 CBR suggested by other researchers. It was therefore recommended that this correlation be used for this project. 9. SUMMARY AND CONCLUSIONS Grain size distribution show that fine content (silt and clay) is more than 25% at all the locations. Based on the laboratory work conducted so far, resilient modulus (Er) could be related to CBR using Er = 10 CBR, where Er is in MPa and CBR is in %. This result is consistent with the values reported in the literature. The measured resilient modulus at all the locations is more than 38 MPa. The values varied from 38 to 120 MPa. The minimum value of 38 MPa is higher than the design requirement of 30 MPa. Tests conducted in the double track project between Ipoh and Bidor have conclusively proved the following: The Threshold stress is well above the expected static and dynamic stresses on the subgrade in this project. The resilient modulus of the material of the subgrade is much higher than the modulus adopted in the calculation to verify the adequacy of the thickness of ballast and sub-ballast for the high speed track. REFERENCES 1. Van Til, C.J., Mc Cullough, B.F., Vallerga, B.A. and Hicks, R.G. (1972). Evaluation of AASHO Interim Guides for Design of Pavement Structures, NCHRP Report Earnest T. Selig, John M. Waters, (1984). Track Geotechnology and Substructure Management, Thomas Telford publications, London. 3. Li, D. and Selig, E.T., 1998, Method for Railroad Track Foundation Design I: Development, Journal of Geotechnical and Geoenvironmental Engineering, Vol. 68, No. 7 8, pp W. Heukelom, A.J.G. Klomp (1962). Dynamic Testing as means of Controlling Pavements During and After Cconstruction, Proceedings of the 1 st Intnl. Conf. on the Structural Design of Asphalt Pavements., Ann Arbor, Aug 20 24, 1962.