Report Case Study

Published on January 2017 | Categories: Documents | Downloads: 32 | Comments: 0 | Views: 214
of 17
Download PDF   Embed   Report

Comments

Content

1.0

INTRODUCTION

The standard joint is often comprised of two joint bars sandwiching the rail on both sides of the web and connected by a series of bolts. Typically, joints are used to connect strings before they are welded, hold the rail plugs in place when defects are removed from the track before the plugs are welded onto the main rail, relieve thermal tension stresses in the rail and prevent rail pull-apart or track buckles, connect different sized rails (compromise joints) and provide isolation for signal blocks.

In addition, the joint is expected to provide safe passage of trains by providing semicontinuous rail girder action, minimizing overall vertical deflection of track, reducing movement of rail ends and being simple, user-friendly, effective and of sound design. At present, there are different types of joint: the six-hole joint, the five-hole compromise joint, the four-hole joint, the mechanical-insulated, and glued-insulated joint. Each of these joints has its use and place in today’s railway track.

Regardless of its form and style, the current railway joint constitutes a weakness in the system providing only a fraction of the bending strength of the continuous rail section at the joint. As a result, there have been joint failures that caused the spillage of dangerous commodities polluting the environment, taken human lives, led to evacuation of communities, and made news headlines. The industry has not adequately addressed this weakness, which, in recent times, has been identified by the Federal Railroad Administration as one of its primary focuses on safety. According to TTCI, the FRA reported that “railwayjoint-related accidents cost an average of $867,000 per annum in North America alone.”

In July 2008, Akhtar et al.2 reported that there are more than 286,000 standard joints in service in North American main line track. The Akhtar report also provided the failure statistics drawn from 3,786 failed joints. They showed that more than 80 percent of the failed joints were of the standard variety followed by compromise joints and lastly by insulated joints. The bar failures were more predominant in the winter months showing the contributions of tensile thermal stresses to their eventual demise.

The low failure figure for insulated joints is because they are removed from track as soon as the insulation fails rather than due to bar fracture. As a result, very few of the insulated joint bars experience fatigue failure.

Examination of the statistics reported by Akhtar et al. states that the failures originated from different locations along a typical joint bar. There were origins at the top of the bar, the toe, the inside and outside bolt holes, etc. This points to the fact that there does not appear to be any rhyme or reason for the failures. Without the benefit of performing any rigorous analysis, it is obvious from such indiscriminate failure origins that there are a lot of high stress areas along the bars and quite possibly the current joint bar is simply an overstressed component whose design may be inadequate for its current use.

2.0

RAILWAY JOINT’S COMPONENT

Figure 1.0 Railway joint

a. Nuts - Hollow cylinder of metal whose lining is threaded to screw onto a corresponding bolt. b. Fishplate – Long steel plate that is fitted into the two sides of the rail webs to join them end to end. c. Fishplate bolt – Bolt and screw assembly that fastens a fishplate to rails.

d. Dating nail – Nail attached to a tie that bears the two last numbers of the year in which the tie was laid. e. Tie plate – Metal plate placed between the rail head and the tie that helps to distribute the weight of the train on the tie. f. g. h. Running surface – Top of the rail head on which the train wheels roll. Spike – Large nail with a hooked head that fastens the base of the rail to the tie. Expansion space – Space left between two joining rails to absorb expansion due to heat.

3.0

JOINT FAILURES

Railway joints are often made with holes that are slightly larger than the diameter of the bolt for ease of installation. As a result, cumulated tolerances can lead to gaps as much as 1/4 inch. Add to this the bending and permanent deformation of the bolts under thermal tension stresses and the gap can be as wide as one inch. As the wheels pass over such gaps, they generate dynamic loads that cause the ballast to pump and suction mud up to the surface. The contamination fowls the ballast, leading to more pumping and joint displacement. The bars eventually develop cracks at the top due to high-compression stresses as the wheels pass and fretting from contact with the free end of the rail.

Reverse bending stresses as the wheels leave the joint propagate the crack in fatigue mode. Eventually, failure of the joint will occur. In these cases, the crack originated at the top, bottom and boltholes of the bars.

It has also been reported by TTCI that the inner bar is the often most likely to develop a crack and break first. This might be related to high lateral loads from rock and roll of the train due to joint stagger. In this paper, we will attempt to rationalize why there are numerous and sometimes indiscriminate failure origins. Given these random failures, the questions that arise are:

Why does the joint bar fail by fatigue crack initiation? Why are there more failures initiated at the top than at the bottom of the bars? Why are there more inside bolt hole crack initiation than at the outside of the bolt holes? Why are there more failures initiated in the inside bars than in the outside bars?

Hence, there is an urgent need to understand and rationalize joint behaviour and the anatomy of joint failure before we can define what changes and improvements are needed. In the short term, the primary objectives must be to reduce the stresses in the joint bars and provide some form of redundancy against bar failure in order to protect the nearly 300,000 bars currently in our main lines.

In the long term, the objectives should include developing a better joint bar design with improved material properties that matches the bending rigidity of the rail and provides redundancy against bar failure.

Answering these questions would bring a better understanding to joint bar behaviour and lay the foundation for future improvements in bar designs. It must be noted that the bar failures that TTCI analyzed did not include the new Hi-relief bar. To the author’s knowledge, there has not been a reported failure of a Hi-relief bar in the literature. Therefore, all references to the Hi-relief bar system in the analyses to follow are hypothetical.

Figure 1. West rail portion showing the fractured ends of the joint bars, a fracture between the first two bolt holes of the south joint bar, and rail batter.

Figure 2. East rail portion showing the fractured end base of the north joint bar.

4.0

DEFECT AND LOCATION

Broken rail There are many effects that influence rail defects and rail failure. These effects include bending and shear stresses, wheel/rail contact stresses, thermal stresses, residual stresses and dynamic effects. Defects due to contact stresses or rolling contact fatigue (RCF):
  

tongue lapping head checking (gauge corner cracking) squats - which start as small surface breaking cracks

Other forms of surface and internal defects:
     

corrosion inclusions seams shelling transverse fissures wheel burn

One effect that can cause crack propagation is the presence of water and other liquids. When a fluid fills a small crack and a train passes over, the water becomes trapped in the void and can expand the crack tip. Also, the trapped fluid could freeze and expand or initiate the corrosion process.

Parts of a rail where defects can be found:
     

head web foot switchblades welds bolt holes

A majority of the flaws found in rails are located in the head, however, flaws are also found in the web and foot. This means that the entire rail needs to be inspected.

5.0

MATERIAL PROPERTIES

First, let us examine the material properties of the rail, joint bars and bolts. It has three components and their minimum strength and carbon content requirement. Made with plain carbon steel, the strength of the elements is determined by the carbon content. The higher the percentage of carbon, the stronger and more harden-able the material is. This also makes the steel more brittle. The rail needs the high carbon content to withstand the high contact stresses generated by the wheel contact and minimize wear.

There is a minor compromise in the ductility of the rail steel as reflected in the minimum percent elongations. Otherwise, the yield strengths of the non-treated steels are similar. So how does the strength and material property affect the resistance of the bars to fatigue crack initiation?

The answer lies in the fatigue curve for the steel, which defines the number of cycles at any given stress level to initiate fatigue failure. Also associated with the fatigue curve is the concept of endurance limit defined in this paper as the stress below which the material will sustain more than 10 million cycles without failure.. It will be assumed that the joint bar and bolt steel will mimic the fatigue curve of standard carbon rail steel.

Having established the relationship between stress magnitude and number of cycles to fatigue crack initiation, the next logical step is to define the stress distribution and magnitude at the locations where these failures initiate in various joint components under typical North American wheel loading over a typical railway joint.

6.0

STRESS ANALYSIS

There are several analytical tools available for defining the stresses in a system such as a rail joint. Such tools must be able to probe into materials and reveal the non-obvious goings on in the system. One such tool is the Finite Element method. This technique for structural and stress analysis has been used effectively in design and analysis of structural systems for decades. In the hands of an expert, the FE method has been validated and proven by comparison to experimental work to be an accurate and effective way of peeking into structures and the way they distribute stress. The FE method will be used in this report to rationalize the seemingly unpredictable origin of cracks in joint bars.

Schematics of the two joint bar designs approved for use in North American heavyhaul continuously welded rail systems. The high relief joint bar designed by Portec Rail Products, Inc., allows for more lateral and vertical railhead wear in curves. The length of the contact between the standard bar designs and the rail head fillet is 0.659 inches compared to the contact of 0.826 inches provide by the Hi-relief bar.

The average section modulus of the SB is seven percent higher than that of the HB. The Finite Element Models were developed for each joint using 136-pound rails. A vertical load of 44 kips was used in each analysis. The entire load is placed at the edge of one of the rails in the joint. Winter loading conditions were simulated by adding a longitudinal force of 200 kips to simulate a temperature differential from neutral of 80 F. No lateral loads were used in the analyses.

Preliminary structural analysis of the global system was performed using Beam on Elastic Foundation theory 4 to determine the length of rail that would be used in the detailed FE analysis. It was determined that a 480-inch length of rail, including the joint, would be adequate for FE modelling of both joints. This length of model also allowed for locating loads applied away from the joint to cause reverse bending at the joint. A 480-inch model of continuous rail on similar supports was also analyzed to provide comparative displacement values. No joints analyzed and presented in this paper were in any way supported directly at their midspan. Boundary conditions were imposed on the models such that the different components could not penetrate each other, but could separate. The entire model was

supported on clusters of linear elastic springs placed at 24-inch tie spacing to simulate the track stiffness. The stiffness of the rail support springs was varied to simulate winter and summer track conditions. A track modulus of 1,500 lb/in/in was used for summer and 4,500 lb/in/in was used for the winter conditions.

Because of the bearing-only contact between the rails, bolts and joint bars the analysis preceded in an iterative manner until the stiffness matrix converged within specified tolerances. The analysis utilized linear elastic properties of the materials. Friction coefficient between materials was limited to 0.42.

7.0

STRESS ANALYSIS RESULTS

Joint Displacements. Displacement at the joint was computed under summer loading conditions. As can be seen, both joint vertical displacements were higher than that of the continuous rail. The Hi-relief joint had a slightly lower displacement than the standard joint due to its better support of the rail head.

Joint Stresses. As the wheel passes over the joint, the ends of the rail are driven downwards and the midspan of the two bars attempt to bow out laterally to accommodate the rail. This creates high contact stresses at the top of the midspan of the joint bars. Add friction to this and combined vertical and lateral bending stresses and you have a very complex stress field. Therefore, prior work that has attempted to show stress reduction or benefits by considering only one stress tensor at the rail/bar interface is incorrect.

In order to account for the complex stress state in a joint system, the von Mises stress will be used throughout the remainder of this paper to demonstrate the individual component reactions to the applied loads. The von Mises criterion has been shown to predict failure by yielding when the octahedral shearing stress (τoct) at any point reaches a particular value, the yield strength of the material. The octahedral shearing stress theory enables us to apply the distortion energy theory while dealing with stress rather than energy. The von Mises stress value is determined by the relationship of τoct to σYP obtained from a simple tension test, and given as

τoct = 0.47σYP

The octahedral shearing (von Mises) stress is expressed in terms of individual stress tensors in all orthogonal axis as: τoct=1/3[(σx - σy)2 + (σy - σz)2 + (σz -σx)2 + 6(τ2xy + τ2xz + τ2yz)2]

where σx and σy and σz are the normal stresses in the x, y and z directions, respectively, and τxy and τyz and τxz are the shear stresses in the x and y faces and along the y and z directions.

In order to examine the stresses in the rail, bolts and inside of the joint bars, elements of the system have to be made transparent depending on what is targeted. There is a concentration of stresses on the joint bar at the centerline of the joint. The magnitudes of this stress concentration are exposed in the cross-section view. The maximum stress in the SB system is 138 ksi versus 87 ksi in the HB system. This shows that the improved contact under the railhead fillet provided by the HB system reduces the stresses on the bar.

Of greater interest is the fact that similar to what happens at the wheel/rail contact, the railhead-fillet/top-of-bar contact stress is higher than the yield strength of the bar material. Therefore, as with rail, it would be reasonable to assume that the bars would suffer contact fatigue.

Rails are ground to remove contact fatigue damage, but the bars are not and we know what happens when we do not grind the rail. Hence, it is the contact-fatigue-damaged material at the top of the bar that, when subjected to reversing bending cycles, leads to the crack initiation that eventually causes the bar to fail. This is why the bars fail more from the top than from the bottom.

The damage is worse in the summer months when the track is softer and when joint pumping is most likely with larger downward deflection of the joint. In the winter, thermal stresses combine with the reverse loading to cause upward bending of the joint to open and propagate the cracks in transverse fatigue, hence the predominant winter month’s failure.

The propensity for this phenomenon is higher in the SB system than in the HB system. The stresses at the bottom of the SB system reach 55 ksi while that of the HB system reach 60 ksi. The high stress concentrations also show that the joint bar response does not conform to simple banding theory that in the past has been erroneously used to evaluate joint bar adequacy to carry wheel loading.

8.0

BOLT STRESSES

Examination of the bolt stresses and deformations show that there is significant curvature sustained by each bolt under thermal loads. The bolt stresses and deformations are compared to deformed bolts from an actual joint failure. The bolt stresses are virtually identical, showing that the thermal load and not the bending of the joint created the stresses. From the deformed shape of the bolts, it can be concluded that both joint designs subject the bolts to a combined bending and shear loading. This does not have to be so.

To further examine the interaction between the bolts, bars and rails, a horizontal plane slice along the mid-plane of the bolts is made through the joint. There are high contact stresses between the bolts and the rails and the bolts with the inner edge of the bar hole.

A close-up of one of the bolts is clear from this image that the high stresses in the inside edge of the bar hole will cause crack initiation and propagation to failure. This explain why there are more inside bolt hole cracks than outside ones. Additionally, fatigue cracks can develop from either edge of any of the bolt holes in the rail.

9.0

FATIGUE RESISTANCE OF THE BARS

Having computed stresses that exceeded the yield strength of the bar material in critical areas, it can be expected that fatigue cracks can form at any of these locations without the aid of impurities or surface discontinuities. At the toe of the bar, stress levels are of the order of 55 to 60 ksi in the SB and HB, respectively.

If the assumption is made that the fatigue limit for rail steel is 55 percent of the yield strength, it means that fatigue crack initiation can be expected at any location where the stress exceeds 38.5 ksi for the SB system and 48.4 ksi for the HB system. These locations can be identified by setting the limit on the stress legend to these values and replotting the stress contours on each bar.

10.0

DISCUSSION

It has been shown in this report that the observed failure of joint bars from multiple locations is explained. These failures arise from the simple fact that the joint design is simply inadequate for today’s heavy-haul railway track.

The stress analysis presented herein has revealed basic structural deficiencies in current joint systems that are not joint-only bar material property issues, but also joint bar geometrical shape issues. Compression stresses on the top of the bars at the middle of the joint, coupled with high contact stresses from the free end of the rail, lead to yielding and early fatigue crack initiation.

Bolt bending and severe compression contact with inside edge of the bar holes can initiate cracks under thermal loading. Either way, thermal loading plays an important part in the formation and propagation of the fatigue cracks. Overall, the high number of hot spots in both joint systems dictates that the industry revisit the joint system design with today’s loading in mind.

Meanwhile, there are some 300,000 joint bars in use in North America. Automated bar inspections are being carried out, but how effective are the inspections and at what cost? These in-service bars need to be protected to safeguard the trains that pass over them. Is there a simple fix to providing this safe guard? How this can be achieved will be presented in the next segment of this series.

11.0

REFERENCES

1. Transportation Technology Center, Inc., LD/TS-08-032001, August 2008. 2. Akhtar, M. N., Davis, D. D., and Riehl III, W. S. “Performance Evaluation of Mechanical Rail Joints,”Railway Track & Structures, July 2008. 3. Parker, Y. J., and Fletcher, F.B., “Fatigue Behaviour and Fracture Toughness of Standard Carbon and High Strength Rail Steels,” Second International Heavy Haul Conference, Colorado Springs, Colo., September 1982. 4 Hetenyi, M., “Beams on Elastic Foundation,” The University of Michigan Press, 1961.

12.0

EVIDENCE

Accident Summary: The Waterfall Accident •7:14 on Jan 31 2003, City Rail passenger train service C311 overturned at high speed and collided with stanchions and a rock cutting approximately 2 km south of Waterfall NSW •The train was carrying 47 passengers and 2 crew (Normally up to 800 with university students) •The driver and six passengers were killed with many more injured.

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close