A laboratory study of railway ballast behaviour under traffic loading and tamping maintenance

Since it is difficult to conduct railway ballast testing in-situ, it is important to simulate the conditions experienced in the real track environment and study their influences on ballast in a controlled experimental manner. In this research, extensive laboratory tests were performed on three types of ballast, namely granites A and B and limestone. The grading of the tested ballast conforms to the grading specification in The Railway Specification RT/CE/S/006 Issue 3 (2000). The major laboratory tests in this research were used to simulate the traffic loading and tamping maintenance undertaken by the newly developed Railway Test Facility (RTF) and large-scale triaxial test facility. The Railway Test Facility is a railway research facility that is housed in a 2.1 m (width) x 4.1 m (length) x 1.9 m (depth) concrete pit and comprises subgrade material, ballast, and three sleepers. The sleepers are loaded with out of phase sinusoidal loading to simulate traffic loading. The ballast in the facility can also be tamped by a tamping bank which is a modified real Plasser tamping machine. Ballast breakage in the RTF was quantified by placing columns of painted ballast beneath a pair of the tamping tines, in the location where the other pair of tamping tines squeeze, and under the rail seating. The painted ballast was collected by hand and sieved after each test. It was found from the RTF tests that the amount of breakage generated from the tests was not comparable to the fouling in the real track environment. This is because the external input (such as wagon spillage and airborne dirt) which is the major source of fouling material was not included in the tests. Furthermore, plunging of the tamping tines caused more damage to the ballast than squeezing. The tested ballast was also subjected to Los Angeles Abrasion (LAA) and Micro-Deval Attrition (MDA) tests. It was found that the LAA and MDA values correlated well with the ballast damage from tamping and could indicate the durability of ballast. The large-scale triaxial test machine was specially manufactured for testing a cylindrical ballast sample with 300-mm diameter and 450-mm height and can perform both cyclic and monotonic tests with constant confining stress. Instead of using on-sample instrumentations to measure the radial movement of the sample, it measures sample volume change by measuring a head difference between the level of water that surrounds the sample and a fixed reference water level with a differential pressure transducer. The test results from cyclic tests were related to the simulated traffic loading test in the RTF by an elastic computer model. Even with some deficiencies, the model could relate the stress condition in the RTF to cyclic triaxial test with different confining stresses and q/p' stress ratios