Design & Development of PSC Sleeper and Philosophy of Design for High Speed

A. K. Singhal * Rajeev Verma **

Synopsis : The PSC sleepers were introduced on Indian Railways in the late 60s and since then have been accepted as essential component of modern track, to meet the requirement of the increasing traffic, heavier axle load and higher speed. Indian Railway is one of the largest user of concrete sleepers amongst the world Railways. More than 100 million prestress concrete sleepers equivalent to 60,000 track kilometers have been produced in Indian Railway so far. Present rate of production is about 10 million per annum. Design of Concrete sleeper differs from normal structural design in a way that the loading and support conditions cannot be assessed to a high degree of certainty. This paper summarises design requirements, design philosophy and factors influencing the design of PSC for high speed track.

1.0 Introduction The purpose of railway track is to transfer train loads to the formation. Railway track is a discrete system consisting of rail, sleeper and ballast laid over formation. Load transfer works on the principle of stress reduction, layer by layer. The maximum stress occurs between wheel and rail and reduces as the contact area increases. Sleeper is a very important component of track as load from the rail is transferred through rail pad to the ballasted bed and ultimately to the formation. Sleepers are members generally laid transverse to the rails, on which the rails are supported and fixed through fasteners. The most preferred material for sleeper has been timber since the beginning of railway constructions but due to indiscriminate deforestation and rapid dwindling of wood all over the world, a need for developing a strong alternative to wooden sleeper was realized by Railway Engineers, thus heralding the era of concrete sleepers.
* Executive Director/Track, RDSO, Lucknow ** Deputy Chief Engineer/Con/Design, Western Railway, Churchgate

27

Life of the sleeper normally depends upon the quality of material i.e. high strength concrete and prestressing wire used. Lateral and longitudinal stability of sleepers is provided by the weight, shape and length of the sleepers. Maintainability of track mostly depends upon the formation conditions and the formation pressure caused due to transmission of ballast pressure from the interface of sleeper bottom to formation through ballast. Recently, RDSO has carried out a simulation study with Charmers University of Sweden for factors influencing design of concrete sleeper and manufacturing. 2.0 Basic design requirement The design of concrete sleepers is determined by fundamental requirements arising from the role that the sleeper component plays in the entire track structure. The dimensional requirements for monoblock sleepers are as follows: The bottom plan area of sleeper must be such that the average ballast pressure should not exceed a certain value, normally 6 Kg/cm2. The end faces should preferably be of a size and shape to provide maximum resistance to lateral movement. As the prestressing wires (tendons) are straight, the cross section can be varied so that the geometry of a particular section may be used to provide prestress eccentricity to best match the bending moments, positive or negative. Consequently the shape of the sleeper body is tapered from both ends to the centre in height and sometimes width with gradual change of sectional profile to avoid stress concentrations.

The role of the sleeper is not only to distribute the vertical load from the train, but also to take care of the lateral movement induced by the train and in the track itself due to temperature variations. The main factor involved in lateral stability is friction between the sleeper base and the supporting ballast. There is also a significant contribution from compacted ballast shoulders at the sleeper ends and compacted ballast in the cribs.

28

A sleeper without any crack is desirable because cracks in the concrete caused by tensile bending moments lead to large increase in the stress range of the prestressing steel, which could cause a fatigue failure of the tendons. The most critical bending moment along the sleeper body often occurs at the rail seat, causing tensile stresses at the bottom of the sleeper. The design and the degree of prestressing are thus most focused on the conditions at the rail seat. The aim is to develop full prestressing at the position of the rail seat in order to prevent bending cracks in the serviceability state. Consequently, the requirements of the anchorage capacity of the tendons are tough and create a complex stress situation at the ends of the sleeper. The centre of the rail seat of sleepers is generally placed at a distance of 500 mm or less from the sleeper end, as shown in figure due to other considerations. It is thereby obvious

Fig. 1 : A Concrete Sleeper

that a short transmission length is essential for pre-tensioned sleepers. 3.0 Design philosophy Railway sleepers are probably unique in the way in which they act as structures. First, they are not subjected to self-load stresses during their working life as the static self-weight of the rail is only of the order of 0.1% of the total design load. They are subjected mostly to dynamic loads. Sleepers rest on the ballast but are not tied down to it, thus the impact load is able to make the sleepers oscillate. The points of application of load are defined, but on the support side, a wide range of support reactions are found. These depend on the nature of the ballast, state of compaction and formation below it as well as the form of ballasting and the quality of maintenance. The point of loading also defines the points of maximum moment and shear as coincidental. The vehicle (wagon/coach) applies its load to the track through axle and wheels. Axle load depends on the design of the vehicle and its

29

maintenance. Ascertaining the load coming on an individual sleeper and consequent ballast pressure involves the consideration of track stiffness, sleeper spacing, ballast packing, maintenance condition of the track and the bottom profile of sleeper. The load coming on the sleeper must also account for dynamic augment due to speed, rail wheel irregularity and track defects etc. The design sleeper moments are assessed from the sleeper loads and ballast pressure. Finally, a sleeper is designed with given material strengths. Many factors are difficult to be determined, quantified and maintained in the long term in practice, therefore, much design development relies on a probabilistic approach taking into account what has gone before. 4.0 Design There are a number of steps in the design i.e. assessment of load on the sleeper, ballast pressure distribution, selection of dimensions of sleeper followed by bending moment calculations at rail seat & centre of sleeper and finally checking the stresses in sleeper and calculating factor of safety and load factor. 4.1 Loading It has been an internationally accepted fact that the combination of a number of elastic media between wheel and formation coupled with the dynamics of track and rolling stock irregularities introduce many indeterminates that preclude determination of exact rail seat design load. The first step towards an economical design is therefore to find out the rail seat design load correctly to the possible extent. Rail seat load depends upon number of factors such as wheel load shared by sleepers adjacent to the particular wheel, dynamic loading due to wheel, properties/condition of rubber pad and maintenance practices of the track. 4.1.1 The wheel load coming on a rail is shared by many sleepers. The maximum load on a particular sleeper is derived by load distribution factor multiplied by wheel load. Load distribution factor of 0.50-0.55 for PSC sleeper (sleeper spacing 60 cm) design has been adopted by Indian Railways. 4.1.2 Dynamic augment is a factor which is multiplied to static load to calculate the design load. Dynamics of the rail wheel interaction is very complex phenomenon. Due to moving load,

30

vibrations of very low frequencies of the order of 1 Hz to high frequencies of the order of 2000 Hz are generated. These varying frequencies vibrations are to be absorbed by different component of the bogie and track. Concrete sleeper are susceptible to impact load especially in the frequency ranging from 25-300 Hz. Considering all the factors, dynamic augment for speed and rail wheel irregularities has been adopted as 2.5 by Indian Railways. 4.1.3 The effect of softer pads of 5 mm/6.5 mm thickness vis--vis harder EVA pads on loading has also been studied. This study shows that properties/quality of rubber pad do have effect on dynamic loading condition of sleeper. 4.1.4 Ballast pressure plays an important role in the design of concrete sleeper. Presently, a length of 1040mm from sleeper end is being considered as a zone of ballast reaction on IR. This is in near ideal track condition. With passage of time, track deteriorates and central portion of the sleeper also imparts partial reaction, which is considered as 40% of that imparted by end zones.

Fig. 2 : Different Support Conditions for Sleepers

4.2 Bending Moments 4.2.1 The bending moments (BMs) at rail seat and centre are functions of the bottom profile (plan) of the sleeper. For a given length of sleeper, the BMs at the two sections are influenced by bottom width at various sections. 4.2.2 BM at rail seat is influenced by the overhang of the sleeper and hence, by its length and track gauge. Relief in BM at rail

31

seat is offered by the uniformly distributed rail seat load up to the longitudinal neutral axis of sleeper. 4.2.3 BM at centre section is controlled by the distance between rail seats (gauge), coefficient of centre binding and the width at centre bottom and increases with all the three parameters. 4.2.4 The Indian Railway sleeper suffers in comparison to its European/ American/Australian counterpart in the matter of increased centre-moments/centre top tension because of the 17% wider track gauge. Even with a coefficient of centre binding of 0.5, the magnitude of hogging BM at centre remains comparatively smaller for the standard gauge sleepers than for BG. Centre of gravity of the set of HTS wire will depend upon two moments i.e. at rail seat & at centre and, therefore, it is very important to decide CG of set of HTS wire accordingly. 4.2.5 Longer length of sleeper reduces formation pressure but increases BM at rail seat with more demand of steel though centre section BM is reduced. Reduction in length has a reverse effect on the BMs at the two critical sections. In addition the margin of transmission length reduces in case of fully bonded pre-tensioned designs. However, the length of the sleeper is to be decided based on the aim to establish full prestressing at the position of the rail seat in order to prevent bending cracks in the serviceability state. 4.3 Section of Sleeper 4.3.1 Bottom width Larger width at the rail seat reduces ballast pressure. The area in the tamping zone should be large enough to restrict ballast pressure at the sleeper ballast interface to 5-6 kg/cm 2, the width being constrained by considerations of sleeper spacing and mechanised tamping. Very large widths causes problem of tamping and may result in broken edges, uneven ballast pressure and torsional stresses. The width at centre section should be as small as possible but the transition from rail seat to centre should be gradual to avoid stress concentration. 4.3.2 Top width Nature of fittings like dowels, inserts, slide chairs, bearing plates/rubber pads etc. dictate the choice to top width at rail 32

seat. Top width at centre section, being the smallest, should be adequate to accommodate HTS wire in top tier and should have adequate cover for durability. 4.3.3 Depth Increased depth means greater relief in bending stresses at rail seat, higher section modulus at rail seat bottom and hence larger factor of safety against the imposed BM and larger load factor against failure. On the other hand smaller depth result in reduced section modulus and very high bending tensile stress that cannot be overcome by prestressing forces, already weakened due to a much smaller number of HTS wires having to be placed in the reduced sectional area. The depth at centre section has to be decided considering that the section has adequate modulus to restrict bending tensile stresses at top which can be effectively brought within the permissible compressive and tensile stresses in concrete by the superimposed prestressing forces. The transition from rail seat section to centre section should be gradual to avoid stress concentration and cracking under impact of moving loads. 4.4 Permissible stresses 4.4.1 Concrete Final stresses in concrete (i.e. algebraic sum of bending stress and stress due to prestress) at critical sections i.e. at rail seat and centre must remain well within its permissible limits. 0.4 Fc and 0.04 Fc (where Fc is the specified concrete strength) are the present permissible limits in bending compression and tension respectively. 4.4.2 HTS Wires 0.2% proof stress for HTS wires has a generally stipulated value of not less than 85% of UTS. The initial prestressing force varies, according to the available information in different designs of the world from 70% to 75% of UTS. But it is not the initial prestress that really matters. It is the residual prestress after losses that influence not only the crack resisting capacity of the section but also its durability under pulsating loads.

33

Very high permanent prestressing force is likely to cause longitudinal splitting of concrete also. 4.4.3 Loss of Prestress of HTS Wire Losses in prestress occur due to relaxation of steel, creep, shrinkage and elastic shortening of concrete. Presently normal relaxation steel is being used. Low relaxation steel will help in reducing losses. In general, the better the compressive strength of concrete, the lower the losses. A high level of quality control on materials, concrete manufacturing and curing is essential to attain consistent high strength leading to lower losses. Practices in assumption of losses in prestress depend on production techniques. Assumption of losses is one thing but to ensure that no more losses take place than assumed is a measure of technology and quality control in the plant.

4.5 Factor of safety (FOS) & load factor (LF) 4.5.1 The designed MR at rail seat bottom is equal to BM x FOS. The FOS indicates that crack may occur under working load of the order of rail seat design load x FOS. A high factor of safety ensures greater immunity from cracks except under very high loads with a very low probability of incidence. A crack free section imparts longevity because it precludes ingress of moisture or chemicals that cause stress corrosion in HTS. With a smaller factor of safety the margin between the permissible tensile stress in concrete and its modulus of rupture reduces and cracks may appear at lower and more frequent loads. 4.5.2 Load factor is the ratio of failure moment (MF) capacity and imposed BM. MF is a function of the quantity of HTS in tensile zone, concrete strength and effective depth of section. A load factor of about 3 is considered adequate. Load factor indicates the capacity of the section to resist failure under rarely occurring very high impact loads as in case of derailment. 5.0 5.1 Recent Simulation Study Recently, RDSO has done a project in consultation with Chalmers University of Sweden on 'Optimisation of Prestressed Concrete Sleeper' under UIC Asia Regional 34

Cooperation Programme. In the project, various design parameters of the sleeper have been studied by computer simulation and recommendations have been made for improvement in design philosophy. As per study, load distribution factor and dynamic augmentation factor depend on track irregularities, wheel flat, stiffness of rubber pad, ballast pressure distribution etc. and Dynamic augmentation factor is largely affected by speed, wheel flat and rubber pad stiffness. Thus, detailed analysis and design of sleeper through simulation studies becomes very important for heavy haul and high speed traffic. 5.2 DIFF Model has been developed by CHARMEC using FEM analysis for a length of 70 sleepers' bays. Rail has been modeled with 16 beams element in each sleeper bay. Only half sleeper / track model is developed owing to symmetry of track structure. Four wheel loads (2 from each adjoining bogies) have been considered in the model. Uniform ballast bed stiffness distribution has been taken with value of ballast bed stiffness equal to 225 MN/m/m2 and ballast bed viscous damping factor of 1.0. In the modeling study, the DIFF model has been analysed to assess influence on Dynamic Augment factor due to rail wheel irregularity, rail pad stiffness, rail type, ballast bed stiffness and wheel flat etc. Based on studies, following important conclusions have been drawn -

5.3

5.4

5.4.1 The value of load distribution factor is different for different track conditions such as rail corrugation, rail joint, wheel flats, rail pad stiffness etc. Load distribution factor (i.e. ratio between rail seat load to wheel load) is 0.40-0.43 under static condition which is presently being taken as 0.55 in RDSO design. 5.4.2 The value of dynamic magnification is different for different track conditions such as rail corrugation, rail joint, wheel flats, rail pad stiffness etc. Maximum dynamic magnification factor for rail seat bending moment is 1.73 for BOXNHL and 2.24 for LHB coach for the case of wheel flat.

35

5.4.3 Influence of Dynamic rail pad stiffness and vehicle speed on sleeper Bending Moments a) There is moderate increase in rail seat & centre bending moment and rail seat load due to increase in rail pad stiffness. b) Rail pads with higher stiffness lead to a significant increase in the maximum contact forces, rail seat loads and sleeper bending moments.

Fig. 3 : Influence of dynamic rail pad stiffness and vehicle speed on (left) maximum sleeper bending moment at rail seat and (right) minimum sleeper bending moment at centre. Reference track model and reference freight vehicle model BOXNHL, see Section 3, and 60mm wheel flat (depth 0. 40 mm)

c)

d)

As per studies, Maximum rail seat load for axle load of 22.9 T with wheel flat & with Soft rail pad (stiffness 100 kN/mm) = 0.6 x 229/2 = 68.7 kN Medium rail pad (stiffness 200 kN/mm) = 0.8 x 229/2 = 91.6 kN Hard rail pad (stiffness 400 kN/mm) = 1.1 x 229/2 = 126 kN Whereas, as per RDSO existing practice, rail seat load = 0.55 x 2.5 x 229/2 = 157 kN Thus, it can be concluded that load distribution factor and dynamic augment factor are not constant but vary with the track conditions viz. rail corrugation, wheel flat, rail joint, ballast stiffness distribution, rail pad stiffness and speed etc.

36

5.4.4 Increased ballast bed stiffness leads to increased rail seat load and rail seat bending moment but decreased centre bending moment. 5.4.5 Different types of ballast stiffness distribution (based on distribution factor ) have been considered depending upon packing condition, as shown below: By considering the above distributions in analysis, it can be said that all possible conditions occurring in actual field conditions have been covered in design. 5.4.6 Rail seat bending moment is maximum for the distribution corresponding to ballast shoulder at ends and centre bending moment is maximum for distribution ballast shoulder at centre.

Fig. 4 : Different load distributions

5.4.7 Influence of ballast bed modulus distribution and vehicle speed a) Effect of geometrical properties of sleeper on bending moments has been studied. Uniform cross section leads to reduced rail seat bending moment. Sleeper with narrow crosssection at centre leads to reduced sleeper centre bending moment. Sleeper response for dynamic loading is different from static loading. During dynamic loading, maximum sleeper

b)

37

Fig. 5 : Influence of ballast bed modulus distribution and vehicle speed on (left) maximum sleeper bending moment at rail seat and (right) minimum sleeper bending moment at centre. Reference track model and reference freight vehicle model BOXNHL, see Section 3, and 60mm wheel flat (depth 0. 40 mm)

c)

displacement and maximum sleeper curvature / bending moment are not synchronized. As per European standards, the characteristic bending moment is representing the required capacity at the end of the service life of the sleeper, i.e. 50 years. The test bending moment is calculated by adding the bending moment capacity reduction given by the expected loss of prestress and reduction of tensile capacity of the concrete from the time of testing to the end of the service life of the sleeper. The test bending moment is normally 1.25 - 1.4 times the characteristic bending moment depending on the geometry of the sleeper and the prestress level. The fundamental idea is that the sleeper should have a capacity that is higher than the characteristic bending moment requirements during its entire service life. The bending capacity of the sleeper is believed to vary during the service life due to prestress losses and changes in the tensile capacity of the concrete.

38

5.5 Study for Manufacturing Process As per study, following improvements in sleeper production process can be carried out for an effective sleeper production system that produces high quality sleepers suitable for high speed. a) Prestressing There should be Parallel control of both the elongation and the applied hydraulic pressure. There should be long beds in order to avoid large influence of wire length and to provide means for an effective quality control. Simultaneous prestressing of all wires in the bed is desirable. b) Casting The majority of the compaction of the concrete should be accomplished by use of internal poker vibrators in order to ensure proper compaction of the concrete along the entire length of each sleeper in the bed. Compacting the concrete only by vibration of the entire mould should be avoided in order to ensure that the geometry of the moulds remains maintained and that the compaction of the concrete does not vary due to vibration nodes of the mould. c) Curing The Cement fineness (blaine) of at least 440 m/kg should be used. The moulds should be covered by tarpaulins directly after casting in order to utilise the heat created by the hydration process. The maximum allowed temperature in the concrete during the curing should be limited to 60 C. Anchorage of prestress Both single wire strands and three-wire strands can be used. An indented strand surface should be used in order to ensure the bond capacity of the strands. Cutting of sleepers The cutting of the sleepers should be performed by mean of a diamond disc cutter that simultaneously goes through both the concrete and the wires. No corrosion protection should be applied at the sleeper ends in order to enable future inspection of possible wire slippage at the ends of the sleeper.

d)

e)

39

f)

Cleaning and oiling A controlled cleaning and oiling of the moulds must be done in order to obtain good surfaces of the sleeper and avoiding build up of concrete in parts of the mould that will lead to the need of a higher demoulding force.

6.0 High speed - Issues involved High Speed is a relative term. There is no defined speed at which a train can be called a high speed train. We can say anything at or over 100mph (160 kmph) is fast and should be thought of as high speed. Viewing the advantage that high speed train offers over road and air travel in intercity journeys, particularly between metros, Indian Railways have decided to conduct pre-feasibility studies for running high speed trains at 300-350 kms per hour in all the four regions of the country. This will mean heavy investment in construction of new dedicated tracks for high speed trains, development and acquisition of high speed rolling stock and development & installation of Signalling and OHE. High speed trains hope to capitalise on rapid growth of Indian economy, rising industrialisation, urbanisation and unprecedented growth in intercity travel. Load distribution factor, dynamic augmentation factor depend on track irregularities, wheel flat, stiffness of rubber pad, ballast pressure distribution etc. and dynamic augmentation factors is largely affected by speed, wheel flat and rubber pad stiffness. Thus, detailed analysis and design of sleeper through simulation studies becomes very important for heavy haul and high speed traffic. The dynamic coefficient partially consists of a dynamic factor due to wheel defects. Typically, this dynamic factor is taken as 150% for speed below 200 kmph and 175% for speed above 200 kmph. AREMA code specifies an impact factor of 200% and a separate speed factor between 0.7 and 1.2 providing a dynamic factor from 140% to 240% respectively. AS1085.14 specifies 150% for a vehicle speed of 80 kmph and 200% for a vehicle speed of 115 kmph. As per AREMA, Rail seat characteristic bending moment Mdr = B x V x T B - unfactorised bending moment V - Speed factor T - Tonnage factor

40

Tonnage per annum - Million Gross Tons (MGT)

Fig. 6 : Function of the tonnage and the speed factors

As per UIC, speed factor for design of sleeper on soft pad & hard pad are 1.39 & 1.50 for 25T axle load (upto speed 120 kmph), 1.59 & 1.75 for high speed standard ( 22.5 T axle load & speed 200 kmph) & TGV (18 T axle load & speed 300 kmph). 7.0 Conclusion z The design of concrete sleeper differs from normal structural design as loading and support condition cannot be assessed to a high degree of certainty. PSC Sleeper is designed for NO crack condition as crack reduces the fatigue life of sleeper apart from decrease in load carrying capacity.
z

On the basis of past experience, Indian Railways have specified design criteria considering the certain degree of uncertainty. Due to variations in track and rolling stocks, correct assessment of design loading is not possible and analysis of the dynamic behavior of sleeper is still a matter of research. Imposed bending moment largely depends upon base area & length of sleeper and resisting moment is a property of section at different locations. Therefore, profile of the sleeper plays a major role in the strength of sleeper against cracking and failure. Revision of design criteria is an ongoing process and Indian Railways is continuously engaged in rationalising the various design parameters through research as well as experience gained.

41

Recent study shows, load distribution factor, dynamic augmentation factor depend on track irregularities, wheel flat, stiffness of rubber pad, ballast pressure distribution etc. The simulation study and determination of these factors is must for design of sleeper for high speed. Rail wheel irregularity to the minimum possible extent, joint less track, good quality of welds and track maintenance of high standard are essential prerequisite for sleeper for high speed track. Improvement in manufacturing process of concrete sleeper is imperative to achieve high quality in sleeper production suitable for high speed as elaborated in para 5.5.

*****

42