LARGE-SCALE , CYCLIC TRIAXIAL TESTING OF RAIL BALLAST

Today’s railway freight infrastructure needs to sustain heavier loads and perform satisfactorily between maintenance cycles to be economically and technically sustainable. The frequency of railway maintenance is integral to the quality of railway ballast, which changes continually due to particle breakage and generation of fine particles (i.e., ‘fouling’). In this study, the plastic deformation of ballast under cyclic loading for different fouling materials, such as mineral fouling from crushed ballast and coal fouling from surface spillage, was studied using a large-scale triaxial apparatus. Increasing fouling content (percentage of particles < 4.75 mm) resulted in increased plastic deformation of ballast. A content of mineral fouling greater than 5% significantly changed the deformational characteristics of the ballast in high moisture condition (water content > 10%). The water content of fouling materials (increasing from 5 to 25%) had a significant effect on accelerating the plastic deformation and changing the shape of deformational behavior of ballast from a small rate of plastic deformation (i.e., ‘plastic creep’) to a high rate of plastic deformation (i.e., ‘incremental collapse’). Life expectancy of ballast in various fouling conditions was derived. Fouling can reduce the life expectancy of ballast to less than 2 months under extreme conditions of coal dust fouling (10%) and wet environment conditions (25% water content). INTRODUCTION After 60 years of research, the United States (US) has developed the world’s most advanced highway and aviation systems. These systems now face escalating congestion and rising environmental costs. The current US transportation system consumes 70% of the nation’s oil demand and contributes 28% of the nation greenhouse gas emissions (1). In parallel, there has been an increase in volume of rail traffic including additional freight and passenger volume and heavier freight loads carried over the railroad network (2,3). This increased rail capacity has required substantial reconstruction and maintenance of existing rail corridors into a comprehensive heavy freight and, in places, shared high-speed-intercity passenger rail network. The increased volume of rail traffic and freight tonnage is stressing rail substructure, including ballast, subballast, and subgrade (Fig. 1), to levels not experienced to date. Ballast has a significant role in dissipating and effectively distributing the load from the track surface to the underlying bearing subsurface. Unlike granular layers in pavement structures, railway ballast begins to break and deteriorate under heavy freight loads and the passing of high speed trains, deviating from original specifications and transforming into “fouled ballast.” Ballast fouling starts by internal degradation due to fracture and abrasion between the ballast particles (i.e., ‘mineral fouling’), infiltration from underlying layers (i.e., ‘subgrade fouling’), and surface spillage (e.g., coal fouling’) (3,4,5,6,7,8,9). The main sources of ballast fouling in the US and England are presented in Table 1. The amount of ballast deterioration increases with traffic load repetition and load intensity from the freight cars. Fouling impacts track performance mainly by changing the substructure properties, including (1) loss of effective drainage, (2) formation of “mud-holes”, (3) deterioration of track resiliency, (4) lack of resilience to lateral and longitudinal forces, (5) poor durability after maintenance, and (6) increased rate of deterioration (4,7,9,10). Failure may reoccur in a track shortly after maintenance if initial problems, such as fouling and limited drainage, are not identified and resolved properly. Jefferies and Johnson (11) showed that fouling accumulation is located in the area loaded by traffic, where tamping tines operate, and at the bottom of the effective ballast layer. Cribs and shoulders show less contamination by fines. Approximately 50% of sites investigated by Jefferies and Johnson (11) showed that crib ballast is acceptable for continued service without cleaning. Lee (3) showed the major fouling is created beneath and at the edge of ties due to stress concentrations. The fouling accumulates in voids from the bottom of ballast layer upwards due to particle breakage and fines penetration from the subgrade. Fouling location from different sources, such as particle breakage, subgrade infiltration, and surface spillage, is summarized in Fig. 1. A major concern facing the freight rail transportation industry in the US is increasing maintenance costs due to heavier freight loads and substandard track substructure (3). Surfacing and maintenance expenses of the ballast layer (i.e., the large-size aggregate material of railroad substructure) over the past few years has substantially increased; e.g., Burlington Northern Santa Fe (BNSF) Railway has spent approximately $200 million US annually, about 17% of their capital budget (3). The objective of this paper is to correlate fouling levels to the long-term deformational characteristics of ballast. Two sources of fouling material are examined, that from internal ballast particle crushing due to loading (mineral fouling) and that from surface spillage of coal dust. The critical combinations of fouling and moisture that may cause significant deterioration of railway track performance are determined by evaluation of plastic deformations using a large-scale cyclic triaxial testing (LSCT). The stress combinations to be used in the LSCT test were determined by matching the plastic deformations with those obtained from a full-size prototype track model experiment (FSTM). The life expectancy of ballast in various fouling conditions is determined by using the LSCT laboratory data. MATERIALS Granitic ballast and subballast, as natural aggregates, were provided from a quarry in Wyoming by BNSF. The particle size distribution of the ballast and subballast (in accordance with ASTM D6913), in comparison to ballast specification #24 by AREMA (12), are shown in Fig. 2. The particle size distribution of the as-received BNSF ballast is slightly coarser than that of the AREMA specification. The ballast has a maximum particle size of 60 mm and a minimum particle size of 25 mm, while subballast has a maximum particle size of 25 mm. Most of the particles of ballast and subballast have irregular shapes with particle aspect ratio (ratio of the largest and the smallest dimensions of a particle) between 1.5 and 3.5. Based upon visual inspection of thin sections and accompanying X-ray diffraction (XRD), the ballast is 35% granitic and 45% rhyolitic. Highly fouled ballast was obtained from a stockpile in a track yard of the Wisconsin and Southern Railroad Company in Madison, WI. The fouled ballast was sieved and separated into different particle sizes. Particles with grain size < 4.75 mm were designated as fouling materials per the definition of Selig and Waters (4). The fouling material from the fouled ballast (called ‘mineral fouling’) were 70% dolomitic (attributed to particle crushing) and 30% quartzite (attributed to subgrade intrusion), based on XRD analysis. The mineral fouling had low plasticity with liquid limit (ASTM D4318) of 20. The coal source was from the University of Wisconsin-Madison Charter Street Power Plant (a Wyoming-based coal) and was ground to produce the particle size distribution similar to the one observed in the field (8). Coal fouling was also a low plasticity material with liquid limit of 36. METHODS Full-Size Prototype Track Model (FSTM) Experiment A full-size prototype track model experiment (FSTM) was developed to evaluate the scale dependency of long-term deformational behavior of rail ballast under cyclic load. In the FSTM testing, scale effects and the contribution of underlying layers on the deformational performance of the ballast were studied. The result of FSTM testing was used to verify the plastic deformation of the ballast and subballast predicted on the basis of the large-scale cyclic triaxial (LSCT) testing. The prototype FSTM (box size of 24 x 70 in [0.6 x 1.8 m]) experiment is shown in Fig. 3. A typical tie distance in rail track is approximately 24 in (0.6 m) and only half of the track between two ties was simulated due to symmetry. The FSTM consisted of different layers of track substructure including ballast, subballast, and subgrade. The ballast was compacted at to a maximum dry unit weight (d = 97 lb/ft, 15.3 kN/m) by controlling the volume and weight of the material. A ‘shoulder’ of 1 ft (0.3-m) width was also created. The subballast was compacted by a plate compactor to 100% maximum dry unit weight (standard Proctor, ASTM D 698) at the optimum water content (Table 2), which was confirmed by nuclear density gauge testing. A soft subgrade was simulated by a block of styrofoam with a modulus of elasticity (E) of 14500 psi (100 MPa). Vertical stress (transferred load from a locomotive to the tie) was applied through a servo-hydraulic system (MTS, loading capacity of 20 kips [100 kN]). The deformation during loading repetitions was measured through instrumentation in each layer with linear variable differential transducers (LVDTs). A wooden tie with cross section of 7 x 6 x 21 in (0.18 x 0.15 x 0.53 m) was used, which is in the range of typical tie size for railway track (4,13). The length of the wooden tie was chosen to mimic the length of the rail bearing area (RBA) of a tie, which is between 20 and 24 in (0.5 and 0.6 m) (13). Large-Scale Cyclic Triaxial (LSCT) Equipment A prototype LSCT testing apparatus was developed to test a specimen with 12-in (305-mm) diameter and 24-in (610-mm) length. The LSCT was designed, manufactured, and constructed to perform cyclic testing at different confining pressures, frequencies, pulse shapes, and drainage conditions. Air pressure was used to apply confining pressure between 5 and 29 psi (35 and 200 kPa). Both the top and bottom plate caps we