Behavior Of Ballast Research Articles

Effect evaluation of drainage condition and water content on cyclic plastic deformation of aged ballast and its estimation models

With the ballast aging, the changes in the size and shape of the ballast particle reduce the drainage capacity, as well as cause a greater permanent deformation of the railroad ballast. Therefore, it is meaningful to investigate the effect of aging on the mechanical behavior of unsaturated ballast, and to estimate the cyclic plastic deformation by considering the aging effects. Here, “aging effect” means the increase in fine fraction content and the particle shape becomes rounded and smooth as compared with fresh ballast. In this study, the influence of aging on the cyclic plastic deformation of unsaturated ballast was evaluated through a series of cyclic loading triaxial compression tests. Test results indicate that the cyclic plastic deformation of ballast is seriously affected by water content, fine fraction content and drainage condition, and the increasing trend becomes more remarkable at the water-rich aged ballast under the fully undrained condition since the effective confining pressure decreases due to the generation of excess pore water pressure. Furthermore, the applicability of two types of estimation models (i.e., a semi-empirical model named University of Illinois at Urbana-Champaign (UIUC) model, and an elasto-plastic model named subloading surface extension (SSE) model) to the prediction for cyclic plastic deformation of unsaturated ballast is also verified in this study by comparing with results of cyclic loading triaxial compression tests. As the result, it is revealed that the UIUC model is suitable for predicting the cyclic plastic deformation of fresh ballast (or slightly aged ballast) with different water contents under the fully drained condition, and the SSE model shows good potential to estimate the cyclic plastic deformation of fresh and aged ballasts by considering the effects of water content and drainage condition. The findings of this study indicate that drainage conditions have a significant effect on predicting the cyclic plastic deformation of aged ballast, and appropriate test conditions for triaxial compression tests (i.e., CD and CU tests) should be selected according to hydraulic properties (i.e., permeability and water retentivity) of the aged ballast.

In-situ ballast condition assessment by tamping machine integrated measurement system

The condition of the ballast bed is one of the most important parameters to be determined in order to assure a safe and economical operation of railway systems. Better knowledge of ballast condition provides an advantage in defining the optimum time for ballast bed cleaning or renewal. State of the art ballast condition quantification methods all require in-situ ballast sampling followed by laboratory tests, reducing track availability and making additional track closure necessary, thus making the determination of ballast condition a time-consuming and challenging task. The tamping process is the core maintenance activity in ballasted track and it is crucial for the economical service life of the track and essential in restoring the track geometry for safe train operations. During the tamping process, the tamping tines interact with the ballast matrix, transferring the displacement caused by the dynamic excitation to the ballast, compacting it under the sleeper. This interaction is observed and measured in-situ within the framework of a research project presented in this paper. Lateral forces and compaction energy, together with the loading and unloading response of the ballast matrix during compaction are used to determine the ballast condition. Serving as a confirmation of the ballast condition definitions made based on the in-situ measurements, a semi-analytical model of the tamping unit-ballast matrix interaction has been developed. The mechanical model is used to simulate different ballast conditions in order to optimize the ballast life cycle and improve the understanding of ballast behaviour under cyclic loading.

Geomechanical assessment of an inert steel slag aggregate as an alternative ballast material for heavy haul rail tracks

The geomechanical behaviour of railway ballast is a key aspect in the performance of ballasted tracks, mainly under increased axle loads and/or train speeds. Recently, and driven by the circular economy paradigm, some researchers have evaluated the use of alternative ballast materials as replacement for traditional natural crushed rocks. This work presents a comparison between the geomechanical behaviour of a granite aggregate and the one of an electrical arc furnace steel slag aggregate, with the trade designation of ‘Inert Steel Aggregate for Construction’ (ISAC). Both materials were designed for heavy haul railway track applications and showed similar particle size distribution (PSD) curves established by the standard gradation AREMA N. 24. Laboratory tests were performed on scaled down ballast specimens to evaluate both the macro-structural behaviour under cyclic loading triaxial tests (long-term permanent deformation and resilient modulus under various stress paths) and the micro-structural behaviour (particle breakage in cyclic loading and single-particle crushing strength). With a view to analise the structural response of the railway track and to compare the influence of the two aggregates, numerical simulations by FEM are presented, using static loading and non-linear elastic material models for both the ballast and the sub-ballast layers. The models were calibrated on basis of laboratory test results, to obtain the elastic vertical displacement on the rail top level and the stress distribution on the track foundation layers. The results suggest that the use of steel slag aggregate as ballast material is promising. When compared with the granite aggregate, the ISAC demonstrated to have a higher crushing strength of its particles; a greater tendency to stabilize the permanent deformation (PD); a lower particle breakage after PD tests; and a higher resilient modulus. The results of the numerical modeling showed that the rail presented less maximum elastic vertical displacements (approximately 3% on average) and that the foundation soils underwent less maximum vertical stresses (approximately 4% for a 32.5 t/axle loading and 9% for a 40 t/axle loading), in the model with ISAC in the ballast layer.

Numerical exploration of the behavior of coal-fouled ballast subjected to direct shear test

Ballast undergoes gradual and irreversible fouling process during its service life. To explore the influence of fouling materials on the mechanical properties of ballast aggregates, a series of large-scale direct shear tests were carried out on coal-fouled ballast. Laboratory results show that increasing fouling fines decreases both the shear strength and the volume dilation of ballast aggregate. The corresponding DEM numerical models were established, with the help of which some microscopic characteristics, including the coordination number, the contact force, the force anisotropy and the rotation of ballast particles, were investigated. Using the calibrated DEM model, the size gradation effect of fouling fines on the mechanical behavior of ballast aggregates were explored. The results show that fouled ballast aggregates exhibit lower shear stresses, volume dilations, peaking shearing resistance angles and dilation angles than clean ballast, since they generate smaller coordination numbers, average contact forces, force anisotropies and particle rotations as observed by numerical simulations.

Influence of aging on hydro-mechanical behavior of unsaturated ballast

To better understand adverse impacts of aging on track maintenance, the hydro-mechanical behavior of unsaturated fresh and aged ballasts was investigated by means of saturated permeability tests, water retention tests, and monotonic loading triaxial compression tests. Firstly, the aged ballast was prepared through the Los Angeles abrasion (LAA) test to reproduce the gradation and particle shape for in-situ aged ballast extracted from actual railway tracks in Japan. Moreover, the results of saturated permeability tests and water retention tests prove that aging induces a remarkable decrease in the saturated coefficient of permeability and a significant increase in water retentivity compared with fresh ballast, due to the presence of fine fraction content. For the shear strength, the effective shear strength parameters of fresh ballast under fully drained conditions (CD test) are close to those under fully undrained conditions (CU test), regardless of water content. However, for the aged ballast, due to the low permeability, it is difficult to simulate fully drained conditions with laboratory tests, even at a very low axial strain rate (0.01%/min). In this case, the effective shear strength parameters of aged ballast are more suitable to be determined by the CU test. Besides, the results of triaxial compression tests prove that aging reduces the effective internal friction angle of ballast, while it increases the total cohesion compared with fresh ballast, especially under the high saturation condition. Lastly, the shear strength of both materials has a decreasing tendency with increasing water content, and the declining trend of shear strength is more noticeable for aged ballast. Therefore, the effects of aging and moisture content on the hydro-mechanical behavior of the railroad ballast should be synergistically considered in the railway track design.

Experimental Investigation of the Cyclic Behavior of Steel-Slag Ballast Mixed with Tire-Derived Aggregate

AbstractIn this study, the cyclic behavior of steel slag ballast materials mixed with tire-derived aggregate (TDA) is experimentally investigated. To this end, a series of ballast box tests are car...

Time-domain Markov chain Monte Carlo–based Bayesian damage detection of ballasted tracks using nonlinear ballast stiffness model

This article reports the development of a methodology for detecting ballast damage under a sleeper based on measured sleeper vibration following the Bayesian statistical system identification framework. To ensure the methodology is applicable under large amplitude vibration of the sleeper (e.g. under trainload), the nonlinear stress–strain behavior of railway ballast is considered. This, on one hand, significantly reduces the problem of modeling error, but, on the other hand, increases the number of uncertain model parameters. The uncertainty associated with the identified model parameters of the rail–sleeper–ballast system may be very high. To overcome this difficulty, the Markov chain Monte Carlo–based Bayesian model updating is adopted in the proposed methodology for the approximation of the posterior probability density function of uncertain model parameters. Owing to the nonlinear behavior of the system, the model updating is performed in the time domain instead of the modal domain. The applicability of the proposed damage detection methodology was first verified numerically using simulated impact hammer test data in two damaged cases perturbed with Gaussian white noise. Second, impact hammer tests of in situ sleepers in the full-scale in-door ballasted track test panel were carried out to collect data for the experimental verification of the proposed methodology. Artificial ballast damage was simulated under the target concrete sleeper by replacing normal-sized ballast particles (∼60 mm) by small-sized ballast particles (∼15 mm). The proposed methodology successfully identified the location and severity of ballast damage under the sleeper. From the calculated posterior marginal probability density functions of model parameters, one can quantify the uncertainties associated with the damage detection results. The proposed methodology is an essential step in the development of a long-term railway track health monitoring system utilizing train-induced vibration.

A new degradation model for life cycle assessment of railway ballast materials

This study presents a correlation between two specific aspects of railway ballast experiments for better understanding and assessment of materials degradation. Firstly, standard tests for quality control of ballast aggregates and related degradation indices/methods have been investigated and accordingly, an Improved Ballast Quality Index (IBQI) has been developed. Secondly, a series of ballast box test has been performed to identify the aggregates long-term behavior such as settlement, fouling condition, and particles breakage. Box tests have been done at three stages of 100,000, 300,000, and 500,000 load cycles. At each stage, by performing durability tests and calculating the developed index, the quality of ballast materials has been determined in parallel. Analysis of the results shows that IBQI can predict the long-term behavior of ballast with high levels of compatibility (less than 10% average difference between the results in a conservative manner). Hence, it can be used as a simple and practical method for estimation of ballast degradation and its life cycle prediction. Finally, a life cycle degradation model has been proposed to demonstrate the capability and applicability of IBQI.

Investigation of the reversible response of track ballast considering complementary approaches

This paper aims at showing that the reversible behavior of track ballast can be reasonably modeled by a homogeneous elastic layer in the computation of the reversible dynamic response of railway structures subjected to moving train loads. To accomplish this, 2D simulations by means of the discrete element method (DEM) and triaxial tests were carried out. Importantly, these two approaches were performed placing a soft support under the granular specimen. A small number of loading cycles has been considered in both approaches since the long-term settlement of the ballast layer is not the focus of this study, unlike most of the existing research on this topic. In the DEM part, the transfer of the mechanical loads applied to the sample through force chains in the granular medium has been illustrated and discussed based on the numerical responses obtained. In addition, the analysis of the overall response of the grain samples using a homogenized constitutive law shows that the reversible behavior of a ballast layer can be modeled by a continuum approach. However, investigating the effect of the subbase stiffness on the mechanical response of the ballast layer indicates that the homogenized behavior is not intrinsic but depends on the sublayer. These results were confirmed by triaxial tests performed in the laboratory on samples of crushed stones resting on an elastomer mat with different values of the stiffness. Finally, the results from the DEM simulations and the triaxial tests attest that the reversible response of railways can be computed using an elastic layer for the ballast bed provided that an appropriate choice for its apparent modulus is made. The approach that consists of deducing a reversible homogenized behavior of track ballast based on DEM simulations and/or triaxial tests taking into account the support layer looks promising. In future research, more realistic DEM simulations could be envisaged to develop more accurate constitutive laws that could integrate nonlinear and/or anisotropic aspects.

Effect of geogrid on deformation response and resilient modulus of railroad ballast under cyclic loading

Laboratory investigations were carried out using process simulation test (PST) apparatus to assess the role of geogrid on the deformation response and resilient modulus of railroad ballast at different loading frequencies (f) varying from 10 to 40 Hz. The materials used for testing involved fresh granite ballast and subballast having mean particle sizes (D50) of 42 mm and 3.5 mm, and geogrids with different aperture shapes and sizes. The results from the laboratory investigations reveal that the degradation and deformation behavior of ballast is influenced by the loading frequency (f). The use of geogrids is shown to reduce the extent of deformations in ballast (both lateral and vertical) and its degradation (Bg: ballast breakage), but the effect of geogrid diminishes with the increase in frequency (f). The assessment of the lateral strain profiles along the ballast depth indicates that the geogrid influence zone (GIZ) is dependent on both the type of geogrid used and the loading frequency (f). The GIZ is found to vary from 38 mm (0.90 D50) to 237 mm (5.64 D50), with it being the lowest at f = 40 Hz and for geogrid having relatively larger apertures. Further, the geogrids are found to significantly enhance the resilient modulus (Mr) of ballast. In addition, the Mr is established to increase with f and Bg. The inclusion of geogrids also reduced the extent of vertical stress (σv) including the dynamic amplification factor (DAF).

Modelling railway ballasted track settlement in vehicle-track interaction analysis

The geometry of a ballasted railway gradually deteriorates with trafficking, mainly as a result of the plastic settlement of the track-bed (ballast and sub-base). The rate and amount of settlement depend on a number of factors, and for various reasons are difficult to predict or estimate analytically. As a result, various empirical equations for estimating the rate of development of plastic settlement of railway track with train passage have been proposed. A review of these equations shows that they (i) do not reproduce the form of settlement vs number of load cycles relationships usually seen in the field; (ii) do not reflect current knowledge of the behaviour of soil subgrades in cyclic loading; and (iii) are often critically dependent on the curve fitting parameters used, which in turn depend on the circumstances in which the calibration data were obtained. To address these shortcomings, this paper develops a semi-analytical approach, based on the known behaviour of granular materials under cyclic loading, for the calculation of plastic settlements of the trackbed with train passage. The semi-analytical model is then combined with a suitable vehicle-track interaction analysis to calculate rates of development of permanent settlement for different initial trackbed stiffnesses, vehicle types and speeds. The model is shown to be able to reproduce recursive effects, in which a deterioration in track geometry causes an increased variation in dynamic load, which feeds back into a further deterioration in track geometry. The new model represents a significant improvement on current empirical equations, in that it is able to reproduce observed aspects of railway track settlement on the basis of the known behaviour of soils and ballast in cyclic loading.

Analysing the effect of principal stress rotation on railway track settlement by discrete element method

Principal stress rotation induced by moving loads from trains significantly influences railway track settlement accumulation. The stationary cyclic loading commonly adopted to study railway ballast behaviour under repeated train loading cannot fully represent the effects of principal stress rotation, which needs to be properly considered in both laboratory tests and numerical simulations for a better understanding of ballast deformation behaviour. This paper focuses on studying railway ballast deformation behaviour with an emphasis on particle scale interactions under two different loading scenarios – namely, stationary cyclic and moving wheel loading. A ballasted track model consisting of five sleepers was established based on the discrete-element method (DEM) with realistic polyhedron-shaped elements. The numerical model was validated first based on the testing results from a full-scale high-speed railway testing facility at Zhejiang University. Numerical results clearly indicated that moving wheel loading induced larger principal stress rotation than stationed cyclic loading did. Larger principal stress rotation mobilised higher particle rotation and displacement, which further increased particle rearrangements through individual particle rolling and sliding, and potentially could cause accelerated ballast degradation. It is recommended to consider principal stress rotation in ballast settlement predictions to prevent possible underestimation by stationary cyclic loading and its limitations.

Investigation of deformation and degradation response of geogrid-reinforced ballast based on model track tests

A series of large-scale cyclic tests was conducted using process simulation test ( PST) apparatus to capture the influence of loading frequency ( f) on the deformation and degradation behavior of ballast with and without geogrids. Fresh granite ballast and subballast having mean particle diameter ( D50) of 42 mm and 3.5 mm, respectively and five geogrids of different aperture shapes and sizes were used in this study. The tests were conducted at f ranging from 10 to 40 Hz and up to 250,000 load cycles. The test results from the laboratory investigations confirmed that the deformation and degradation behavior of ballast is highly influenced by loading frequency ( f). Moreover, the results showed that geogrid reinforcement stabilizes ballast by reducing the extent of lateral spreading ( ld) and vertical settlement ( Sv) and by minimizing particle breakage ( BBI). The inclusion of geogrids also reduces the extent of vertical settlement in subballast layer ( Svsb). In addition, the extent of volumetric ( εv) and shear strains ( εs) including void ratio ( ef) and density ( γbf) were found to be highly influenced by f. It is further seen that lateral spread and vertical settlement reduction index ( LSRI and VSRI) of all the geogrids decreases with the increase in f. Moreover, it is revealed that with the decrease in LSRI, the extent of Sv and BBI increases. A correlation is developed between the deformations ( ld and Sv) and interface efficiency factor ( α). The effectiveness of geogrid in keeping the ballast in place is evaluated in terms of a new parameter ‘geogrid efficiency factor ( Gef)’. The value of Gef for geogrids used is found to vary from 0.02 to 0.27. It is further shown that Gef under cyclic loading conditions is a function of the interface efficiency factor ( α) evaluated under direct shear conditions.

Effectiveness of geogrid reinforcement in improvement of mechanical behavior of sand-contaminated ballast

Vertical stiffness and shear strength of ballasts are significantly degraded when contaminated with sands. There is a lack of solutions/studies related to strengthening ballast against sand contamination. Addressing this limitation, a comprehensive laboratory investigation was made on effectiveness of geogrid reinforcement for improvement of mechanical properties of sand-contaminated ballast. To this end, large-scale direct shear tests as well as plate load tests were conducted on geogrid-reinforced ballast samples prepared with different levels of sand contamination. The obtained results indicate that geogrid reinforcement considerably improves shear strength and vertical stiffness of contaminated ballast. A bandwidth was obtained for contamination levels in which ballast reinforcement is effective. Through examining geogrid with different aperture sizes and locations in the ballast layer, the best performance conditions of geogrid reinforcement were derived. The results were used to propose an effective method of ballast reinforcement and an efficient ballast maintenance approach in sandy areas.

Numerical investigation of the behavior of stone ballast mixed by steel slag in ballasted railway track

Recently, implementing steel slag ballast has been proposed as an appropriate material to substitute stone ballast. In this regard, one of the technical concerns is the behavior of steel slag ballast in both time and frequency domains that needs to be assessed, properly. Furthermore, the combination of stone ballast and steel slag is unavoidable in steel slag ballasted tracks during track maintenance concerning the limitation of steel slag resources. Therefore, this paper suggests an optimal stone ballast-steel slag (SB-SS) combination regarding the dynamic behavior of five SB-SS combinations as 0%SS, 25%SS, 50%SS, 75%SS and 100%SS by weight of ballast using a finite element method (FEM) model of a 50-meter test track. Moreover, using elasticity modulus and Moher-coulomb parameters obtained via a series of plate load and shear strength tests for each SB-SS combination turns FEM model to be more close to the real test track results.Experimental results show that adding steel slag particles to stone ballast increases elasticity modulus and friction angle of ballast layer resulting in the improvement of mechanical behavior of railway track. Consequently, the maximum deflections and root mean square (RMS) of accelerations decrease by increasing steel slag content. Analyzing free vibration of ballast layer combinations reveals that damping ratios of 100%SS ballast layer is the maximum value as 0.25 followed by 75%SS, 50%SS, 25%SS and 0%SS combinations. Moreover, the dominant frequencies of each ballast layer combinations determine that 0%SS, 25%SS and 50%SS coincides within the track excitation frequency range made by wheel sets, while 75%SS and 100%SS are out of which. Finally, according to all results, 75%SS ballast layer is proposed as the optimal SB-SS combination.

Behaviour of ballast under principal stress rotation: Multi-laminate approach for moving loads

Abstract Railway tracks are subjected to millions of loading cycles over time and at high speeds, these moving trains induce dynamic amplification of vertical stresses and rotation of principal stress axes in the track layers. It is important to predict and analyse the behaviour of ballast under these loads with complex stress paths involving principal stress rotation. In this paper, a constitutive model based on a multi-laminate framework is used to predict the deformation and degradation of ballast under complex stress paths. The yield and plastic potential surfaces are developed based on a non-linear critical state and bounding surface plasticity concepts. The proposed model is validated with independent test data to capture the influence of confining stress, loading frequency, Cyclic Stress Ratio ( CSR ) and Shear Stress Ratio ( η τ ) on the permanent strain response. Furthermore, the response of ballast under traffic loading stress paths with different CSR and η τ is analysed. These model predictions show that higher CSR and η τ values lead to exacerbated particle breakage of ballast, large and unstable axial strains and dilatant volumetric strains. Furthermore, a stability surface is proposed based on model predictions, to estimate the allowable CSR and η τ for a stable response.

Degradation-induced evolution of particle roundness and its effect on the shear behaviour of railway ballast

Abstract Ballast degradation has a strong effect on railway track performance (including accumulative deformation, ballast resistance, and drainage). The main objective of this research is to determine the evolution of roundness during ballast degradation and investigate the mechanical properties of real ballast shape at different abrasion stages. To this end, the Los Angeles (LA) abrasion test is conducted on three groups of fresh ballast materials with different sizes (i.e., 25–30 mm, 40–50 mm, and 50–60 mm). Then, a homemade image-acquisition platform is used to extract ballast outlines from multiple projection directions. Next, the corner sharpness of the degraded ballast is quantified by particle roundness. The LA abrasion test shows that particle roundness gradually increases with increasing abrasion. Subsequently, biaxial shear tests are performed on ballasts with different roundness characteristics based on the discrete element method (DEM). The shear behaviours of degraded ballasts with different degrees of roundness are investigated through macroscopic (e.g., the shear strength and dilatancy response) to microscopic (e.g., the mean coordination number, the ratio of sliding contacts, and the mean particle contact moment) analyses. The results show that the shear strength slightly increases in the peak state, while a substantial reduction in shear strength is observed in the residual state with increasing roundness. Moreover, the microscopic analysis shows that greater roundness results in a smaller mean coordination number, as well as decreases in the ratio of sliding contacts and the mean particle contact moment.

DEM simulation of the pullout behavior of geogrid-stabilized ballast with the optimization of the coordination between aperture size and particle diameter

Based on the well-established constitutive models and particular configurations used in the discrete element method (DEM) simulation of the geogrid-stabilized ballast system, this paper presents the pull-out behavior of a triangular geogrid embedded in ballast under special consideration of the particle size. A representative particle size that equals to the diameter of the inscribed circle of the triangular aperture was selected to visualize the axial force distribution and the interface responses regarding the contact force. Meanwhile, fabric anisotropy on the normal component of contact force was introduced to explain the development of macroscopic strength. Results show that the first transverse rib makes the most contribution to the pull-out force. The V-shaped strong contact force chains were mainly grown in the geogrid-ballast substantial interaction region R1, and this interaction mode is not affected by particle size. Moreover, the maximum number of particles trapped in an aperture, Nmax, is used as the quantized indicator for numerical pull-out tests on 6 controlled conditions, and results unveil that the optimal reinforcement can be achieved when Nmax = 2, by evaluating the pull-out resistance, the interlocking effect, the energy dissipation, and the volumetric response.

Performance evaluation of polyurethane-stabilized railroad ballast under direct shear conditions

The behavior of Elastan-stabilized and unstabilized ballast was explored using the large-scale direct shear apparatus under different normal stresses (σn) and shearing rates (Sr). Fresh granite ballast and Elastan polyurethane having a density of 1100 kg/m3 was used for the study. Further, a triangular aperture geogrid was also used to compare its benefits with that of Elastan treatment of ballast. Elastan-stabilized ballast samples were tested after a curing period of 3 days. Test results established that the shear strength of ballast was profoundly affected by both the applied normal stress and shearing rate. The friction (φ) and dilation (ψ) angles of unstabilized ballast were found to decrease from 65° to 59° and from 21° to 10°, respectively with an increase of σn and Sr. The results further confirmed that the Elastan treatment significantly enhanced the shear strength of ballast. The friction (φ) and dilation (ψ) angles of Elastan-stabilized ballast increases from 65° to 75° and 21° to 22°, respectively. The stabilization efficiency factor (Sef), defined as the fraction of the shear strength of stabilized ballast to that of unstabilized ballast, varies from 1.6 to 1.75 for Elastan-stabilized ballast. The breakage of ballast (Bg) increased from 7.62 to 11.72% with an increase of σn and Sr. Further, the geogrid reinforcement reduced the Bg from 11.72 to 4.24% when compared to unstabilized ballast. On the other hand, visual inspection of the Elastan-stabilized ballast indicated no degradation of ballast particles.

Evaluation of particle shape on direct shear mechanical behavior of ballast assembly using discrete element method (DEM)

Abstract Ballast layer is a part of railway foundation that its behavior is affected by the properties of constituent particles. The shape of the particles, which affects ballast behavior, is an important characteristic that is represented by angularity and sphericity indexes. In this study, the macro- and micro-mechanical shear behaviors of railway ballast are investigated by using Discrete Element Method (DEM). For this purpose, a three-dimensional (3D) program based on DEM has been modified and verified with the experimental results to simulate different real particle shapes for the direct shear test. Eight assemblies with different particle shape indexes were prepared. The results showed that the shear strength of ballast layer increased significantly with an increase in the angularity index; however, the strength initially increased with increasing sphericity and then dropped in as the sphericity decreased in the assembly of particles. Also, micro-mechanical responses showed that particle shape affected the formation of shear band.