REMEDIATION OF UNTESTABLE RAIL LOCATIONS DUE TO SURFACE CONDITIONS

Daniel Hampton Manager Contract Services II 500 Water Street Jacksonville, FL 32202 CSX Transportation 904-366-5876 Daniel_Hampton@CSX.com Charles Rudeen Chief Engineer, Rail Quality Loram Maintenance of Way 3900 Arrowhead Dr. Hamel, MN 55427 763-478-5931 Charles.P.Rudeen@loram.com NUMBER OF WORDS: 4939

Abstract

SSC (Shelly, Spalling, or Corrugation) rail defects are locations where the rail surface has deteriorated enough to interfere with ultrasonic testing for internal rail defects. Internal defects that occur in those locations go undetected and when left in the track can cause a broken rail. Frequency of ultrasonic testing may be set for as frequent as every 31 days. The railroad wants to detect any defects developing inside the rail to prevent rail breaks which can cause a derailment. SSCs are a lagging indicator of where on the network the metal removal demand has not been met to maintain the rail surface condition.

Addressing these locations as part of a preventive grinding program is difficult with shrinking maintenance windows. CSX has been able to reduce the total combined length of SSC rail defects on the CSX network by 59% in 2016, then 44% in 2017, and 5% in 2018. Every foot of rail unable to be ultrasonically tested can be considered an area of risk. Through strategic capital planning, data sharing, and new technologies and grinding techniques developed by CSX and Loram, the risk is being reduced. Topics covered include the reasons metal removal demand is often not met, and the corresponding program and technology changes made to address each underlying cause, which resulted in the reduction of these rail surface conditions that interfere with ultrasonic testing.

Introduction

Nearly 10 years ago ultrasonic testing classifications were changed to better quantify areas that are non-testable (NT). These non-testable areas were then classified to separate out non-testable areas due to rail surface conditions (codes SSC and SDZ) and other locations where a surface condition is not seen however, there is still a Loss of Expected Response (LER).

One condition for a valid ultrasonic test is the signal must penetrate the rail to the base and reflect back to the receiver. Locations designated as SSCs (Shelly, Spalling, or Corrugation) have a rail surface condition which inhibits transmitting or returning the signal. This loss of signal limits ability to conduct a successful test for internal defects. SDZs are SSCs that occur within the dead zone of a switch. The dead zone of switch refers to locations where different rails share the same plate meaning broken rails would not be detected by the signal system.

The FRA mandates rail testing which has the goal of identifying internal defects in the early stages of growth, so they can be remediated before potentially causing a broken rail. CSX establishes frequencies of rail inspection using a risk-based models that include factors such as tonnage, traffic type, rail quality, and rail age. Excluding continuous rail testing routes, the highest frequency is a 31-day testing cycle.

A regular preventive rail grinding program has become the industry best practice to extend rail life. At the same time rolling contact fatigue (RCF), which is a common cause of locations designated as SSCs, is also managed through an effective preventive grinding program.

Rail Grinding

Regular preventive rail grinding has become the industry best practice to extend rail life. Preventive grinding is accomplished by matching the capabilities of the rail grinder performing the work to the growth rate of RCF on the territory with a goal of grinding as much of the territory with a single grind pass. This approach has many benefits notably extending the life of the rail by removing less material each grind cycle and avoiding corrective grinding with multiple passes and high metal removal amounts. The preventive grinding strategy also reduces the development and occurrence of SSC and other surface defects which inhibit ultrasonic testing of the rail. By grinding areas before RCF reaches the point of requiring multiple passes this also means the rail surface condition is relatively clean and ultrasonically testable.

Preventative grind cycles are normally defined in terms of MGT and can vary based on rail grinder capability and the specifics of the particular railroad territory such as curvature or grade. Numerous variables can make certain locations behave differently than the rest of the system. These variables range from rail characteristics such as manufacturer, year rolled or chemistry to track geometry characteristics, curvature, track gauge, super elevation, rail cant or subsurface issues. If RCF develops deeper than the grinder can remove in a single pass in many locations or if the grinder does not adequately address the RCF that is present then changes need to be made to the grind cycle interval or depth of cut settings respectively.

If RCF is more severe in longer sections the production grinder is capable of making multiple passes to achieve the necessary metal removal. However, in specialty track locations or relatively short stretches of track removing severe RCF with the mainline grinder becomes less efficient or impossible. This is problematic because the goal is to keep the mainline grinder moving and every time it is required to back up it means less miles are covered each day, shrinking the amount of territory the grinder can cover in a preventive mode or making it impossible to get back in time for the next grind cycle which would require slower speeds and more multiple passes as RCF would have grown deeper.

CSX’s Rail Grinding Program

The challenge with optimizing a metal removal program is balancing efficiency and speed of getting around the network to arrive on cycle against obtaining a customized and detailed grind that more closely matches the demand at all locations. Combined rail grinding operations, utilizes both a mainline production grinder and a specialty grinder in the same work block, has been standard practice on the CSX since July 2015. By utilizing the same work block the specialty grinder receives more track time and therefore can cover a larger territory. Combining the mainline and specialty rail grinders in the same work block also alleviates concerns about using the production grinder to address small SSC locations as work can be shared between the machines allowing each grinder to concentrate on locations where their size and capabilities are the most effective at addressing. The specialty grinder is equipped with a set of grind patterns which mirror the production grinder allowing work to be shared between the two grinders. If the shorter SSC locations are in an open rail section the production grinder first applies a grind pattern to address the entire curve or tangent and the specialty grinder adds additional work only were it is needed. By utilizing the production grinder to grind the entire curve or tangent section to the desired transverse profile template the specialty grinder can focus exclusively on removing the RCF with an aggressive high metal removal full coverage pattern maintaining the desired shape.

Grinding SSC Locations

In order to reduce the associated risk of having non-testable locations, CSX partnered with rail grinding contractor Loram to address locations which are already non-testable and then develop a more robust system of preventing rail surface conditions from becoming severe.

Initial step was to determine whether they could be removed through rail grinding and the level of effort needed. CSX partnered with Loram and Sperry in 2015 to conduct joint field visits and analyze SSC locations along a route and develop corrective grinding techniques for removal. The study used a Loram 24-stone specialty grinder to grind the rail until visually clean and then retest the segment using a Sperry ultrasonic hand testing unit (walking stick). Severe rolling contact fatigue (RCF) in the center of rail was the primary cause for a location to be non-testable. RCF on the field or gauge corner typically didn’t interfere with the ultrasonic reflection to the base of the rail. Most locations could not be immediately retested after grinding until after several trains had passed to relieve the grinding marks. The specialty rail grinder typically required 3-4 full coverage grind patterns to achieve sufficient metal removal and an additional 1-2 grind patterns to achieve desired rail profile, for a total of sixteen to twenty-two grind passes. The resulting rail was visually inspected to be 70-100% clean of surface conditions, and some isolated short locations with deep pitting/spalling remained that would be too intensive to grind. Deep pits on the gage corner typically didn’t interfere with obtaining a valid test. Severity of the surface damage can vary along the length of the SSC. Any remaining deep pitting in the center of the rail which retested as a shorter length SSC would be replaced with plug rail. This testing indicated that a majority of the SSCs could be eliminated through targeted corrective grinding.

Meeting the Metal Removal Demand

After an SSC is removed through grinding or rail replacement, it can reappear at the same location over a period of time based on the cause of the SSC and how much of the metal removal demand for that location is met. The reasons for not meeting the metal removal demand at a specific location is attributable to 4 different scenarios:

1) Assets or track segments are not being ground (no metal removal maintenance being performed to counteract the rail surface damage)

2) Repeatedly late cycle or inadequate depth of cut (problems with executing the known and desired grind program)

3) Grinding on cycle for a route/subdivision, but the preventive grind frequency needed is higher than other track grinding segments on the route. (Example: sharp curves)

4) Grinding on cycle for a route/subdivision, but the depth of cut or preventive grind frequency needed is higher in a specific location than the majority of track within the same track grinding segment. (Example: portion of the entering spiral on a curve with directional traffic) Due to the size of the equipment, the rail is treated uniformly within a track grinding segment for efficiency. Example: Production Rail Grinder is 650 feet long and the average track grinding segment length on CSX 1,315 feet for curves and 2,444 feet for tangent. The Specialty Rail Grinder is 150 feet long and the typical specialty grind area treated uniformly is 300 to 400 feet when including the approaches. The average SSC length is less than 50 feet which means either the demand is higher in a shorter length area than the rest of the segment, or not grinding deep enough when there or are not grinding that area at all. If the whole segment is ground for the worst condition seen in a short area, then the rest of the segment is receiving more metal removal than needed. Currently the depth of cut is based on the damage to the majority of the rail segment.

Reasons for the higher metal removal demand in scenarios 3 and 4 can be due to:

• Lower quality rail or fatigued rail (near end of useable life)

• Higher damaging forces acting at that location.

Matching the metal removal demand is an optimization problem based on the amount of variability in the metal removal demand across a railroad network and the equipment capabilities and costs. The solution is a more customized program and reducing the variability in demand. This means developing a more customized metal removal plan using improved equipment capabilities and operating methods that provide full coverage and are more efficient, and lowering demand at locations with higher demand by improving rail materials, rail profiles, lubrication, and track geometry.

Common Locations for SSCs

An analysis of all SSC and SDZ defects from 2014 to 2018 was performed to determine the primary cause for their occurrence based on location, train handling, and rail steel. Figure 1 indicates the count and Figure 2 indicates total length of track for each identified cause, and both are displayed as a percentage of the total data set from 2014 to 2018. The length is a measure of the amount of risk, and the count is where these problems are occurring.

Figures 1 & 2 – SSC & SDZ Defects by Location

• Specialty Assets: o A large portion of the SSCs were found around specialty assets such as switches,

crossings, or defect detectors. These assets are skipped by the production grinders due to tight clearances and have to be ground by a specialty grinder that has smaller stones and moveable grinding motors. These assets cost 3-10 times as much as out of face rail to replace and have similar damage to mild curves but are typically not ground on a preventive cycle.

• Curves: o Due to the higher stresses which result in higher rail wear and RCF, curves need more

metal removal (closer grind cycles and/or increased depth of cut) to maintain profile and surface conditions. Curves suffer more surface damage than tangents for several reasons: wheel slip as the outer wheel travels further around the curve than the inner wheel, yaw between the wheelset and rail due to truck suspensions that prevent wheelsets from fully aligning in curves – even with perfect steering forces – resulting in creepage and lateral forces; train speeds lower than the designed super elevation putting additional forces on one rail (low rail typically); other track geometry variation in gauge or rail cant contribute to high contact stresses resulting in higher rail wear and RCF.

• Rail Quality & Fatigue: o Lower quality steel locations or fatigued rail due to accumulated stresses over the life of

the rail, receive the same surface damaging forces as other rails in a track segment but the forces more frequently exceed their lower material strength.

• Bridges: o Non-concrete ballast deck bridges such as timber open decks are mostly skipped by the

production grinders unless a specific plan, job briefing, and additional manpower are in place for fire watch.

• Tangents up to 200’ past the tangent to spiral (TS) point: o Tangents up to 200’ past a curve can receive surface damage similar to the full body of

the curve, as the truck struggles to realign under the car body. This area is often ground with the same level of metal removal as the rest of the tangent segment with normal damage. The metal removal demand more closely matches the curve but the curve’s targeted rail profiles (templates) for the high rail and low rail are both different than the tangent profile templates. Curves over 2 degrees use a two-point contact high rail profile as the wheel flange is in consistent contact, whereas the tangent profile is a single point of contact.

• Sidings: o Sidings can be skipped during the grind cycle due to parked cars or trains blocking the

track when the rail grinding team is there to perform the work or have lower quality older rail. When multiple grind cycles are missed, the RCF damage can accumulate to interfere with ultrasonic rail testing. Lastly the same starting and stopping damage to the rail surface occurs in the sidings.

• Steep Grades and Absolute Signals: o Steep grade & signal locations incur more RCF damage due to the forces of positive and

negative longitudinal creepage by repeated train acceleration and braking in the same locations.

Changes to the Grinding Program

CSX and Loram together developed and applied the same Gradual Preventive Strategy used by production grinders to specialty grinding. Three-pass grind patterns were developed for switches and crossings to address the profile and surface of each asset while allowing specialty grinder to maintain pace with the production grinder. Similar to the production grinder strategy, it enables addressing a high percentage of the network with limited resources, staying ahead of deteriorating surface conditions, and working towards the desired rail profile over multiple grind cycles. The operational efficiency savings enabled CSX to bring on an additional specialty grinder in mid-2015 to provide a specialty grinder to work with each production grinder. In the past year, the pre-grind rail inspection vehicle (RIV) now creates a specific grind plan for each specialty asset, best matching the demand at each asset. By using the same CSX templates for the specialty grinder and production grinder, it creates a uniform running band across specialty assets that matches the out of face rail. Most importantly, utilizing a CSX specific rail profile template that matches the most common CSX wheel shapes will reduce contact stresses in turnouts and road crossings, less often exceeding the material strength of the rail, resulting in less surface damage. Using the custom profile template on the high rail of a road crossing in a curve can be expected to

achieve the same 50-70% reduction in contact stress versus a standard new rail 8” radius (Magel, Sroba, Sawley, & Kalousek, 2004) that is typically used as the target profile when manually inspecting a specialty asset. This reduction in contact stress will reduce both rail wear and RCF on the specialty asset, thus significantly lowering occurrences of SSCs or SDZs.

SSC locations are only able to be addressed if the rail grinders have timely and accurate information on their locations. CSX has automated the data feed of latest SSC locations daily to the LORAM rail inspection vehicles and rail grinders. The inspection vehicles and grinders can access the SSC information visually on either a web or mobile based application to determine which locations they will encounter during the upcoming work shift. The rail inspection vehicles are able to adjust the grind plan to appropriately address these locations either by assigning additional grind effort to the production grinder if the SSC is long enough or create a sub section of a curve or tangent to include in a grind plan for the specialty grinder.

CSX provides GIS location data for tracks and assets such as control points, switches, crossings, bridges, as well as rail surface defect information. Loram has used the data to build management and operations tools to enhance visibility, planning, and job briefings. The information is displayed securely online through a web-based operations map, mobile apps, and through offline capable navigation software developed by Loram. Each application is a GIS view of track chart information combined with additional data sets such as rail defects and is used for planning and targeting grind locations. Before the grind shift grinders get updated information on locations of defects if an internet connection exists and the data is stored locally for using during the work shift even if internet connectivity is lost. Using the Loram developed navigation software, the grinders can target specific SSC locations and get live updates on the location of the grinder compared to the location of the SSC. This information greatly increases productivity and safety of the grinders eliminating the need for a crew member to exit the machine and locate the start and end of the SSC location on track as some locations may only be a few feet in length and grinding often takes place during non-daylight hours. While in the machine cab the operators receive a live feed counting down the distance to the start of the defect and then the distance to the end of the defect allowing the grinder to address the exact location where extra effort is required.

A properly implemented and maintained preventive grind program is the most effective approach to reducing the instances of SSC locations; however, due to the large number of variables influencing the growth of RCF some locations will still become ultrasonically untestable. The implementation of combined rail grinding operations on the CSX, the increase in specialty grinding work days, the establishment of bridge grinding procedures and dedicated manpower, along with the timely sharing of data and development of navigation tools has resulted in a steep decline in the total combined length of SSC rail defects on the CSX network.

Improvements in SSC and SDZ Defects on CSX Transportation

Since 2015 CSX and Loram have actively worked to not just efficiently remove SSC or SDZ defects once identified, but analyzed each primary cause area, and then made changes to the grinding program to match the metal removal demand or take preventive steps to lower the metal removal demand by reducing or eliminating the underlying drivers.

Figure 3 – Total Lengths of SSCs and SDZs by Location

Figure 3 shows the change in defect lengths over the past five years. The y axis is the length and the same scale for each graph. The changes made to the program or maintenance practice resulting in the reductions are described for each area.

Near Specialty Asset: defined as on or within 150 feet of a turnout, highway crossing at-grade, defect detector, or other specialty asset. The 150 feet out is typically where the production grinder begins to raise the grinding stones to sequence up and over the obstacle and then sequence down 150 feet past the obstacle. Over the five-year period, this category represents over a third of these surface defects found. CSX addressed by both increasing the number of units ground per year to grind a larger percentage of specialty assets on cycle, and more recently by using the RIV pre-inspection to reduce the contact stresses by applying a more appropriate transverse profile template.

Figure 4 – Total Lengths of SSCs and SDZs by Location

Year 2014 2015 2016 2017 2018 Specialty Units

Ground 2,303 13,972 23,419 21,067 18,899

Table 1 – CSX Specialty Grinding Units Ground by Year

05,00010,00015,00020,00025,000

2014 2015 2016 2017 2018 -

20,000 40,000 60,000 80,000

100,000

SSC SDZ Length vs. Specialty Units Ground

NEAR SPECIALTY ASSET (SSC LENGTH IN FT)

SPECIALTY UNITS GROUND (UNITS)

Rail Age /Fatigue: Since 2015, there has been an increased focus rail on replacement for repeat SSC defects on this older fatigued rail. Short sections receive plug rail by local forces and longer sections are included in the CSX capital rail replacement program which uses large production teams. SSC rail defects are included in the algorithm that generates the replacement evaluation scoring, and replacement rail segments manually added for locations where SSCs are reappearing over multiple grind cycles.

Bridges: In order to address surface and profile rail conditions on more bridges, a new bridge grinding policy was implemented in 2017 that set standard actions based on the material types of the substructures and superstructures. With the new policy, the only bridges not ground are timber substructure bridges due to the long lead time to place the track back in service if the pilings or abutments were to burn down. Timber ballast deck bridges receive a presoak and fire watch, and timber open deck bridges receive a pre-soak, post soak and fire watch.

Bridge Characteristics Miles % Timber Substructure 35 15% Timber Open Deck 97 42% Timber Ballast Deck 18 8% Concrete Deck 78 34% Total Active Undergrade Bridges 229 100%

Table 2 – CSX Bridge Characteristics

Prior to implementation, only 5%-10% of bridges were ground because as the grinder was working down the track they didn’t know the risk level or what type of bridge was coming up, so they would take the safest course of action and not grind. To make the implementation of the new policy effective, it was resourced with additional manpower and equipment for the fire watches and data shared to provide visibility of risk level and discuss all upcoming bridges during the job briefing. Additional fire risk factors are addressed in the job briefing such as bridge length, walkway access, tie condition, precipitation, water remaining, track window available, adjacent tracks, and time of day. The bridges were classified by categories 1-4 in the bridge database and fed to Loram along with our GIS data for mapping. The result is an electronic track chart view that shows the fire protection requirement for all the bridges for the grinding team and an operations map view for managers. With the changes implemented, we are now grinding 85% of our bridges, as the remaining 15% have timber substructures.

Figure 5 – Operations Map view of Bridge Protection Levels and SSC locations relative to grinding equipment locations

Curves > 3 degrees and Curves <= 3 degrees: Curves were broken up in the data at less than or equal to 3 degrees and greater than 3 degrees because we are able to grind curves less than 3 degrees on cycle, as they more closely match the subdivision/route preventive cycle. Not achieving the preventive grind frequency results in accumulated surface damage that is more likely to interfere with ultrasonic rail testing in these curves. The pre-grind inspection often determines a deeper depth of cut to remove the surface damage accrued, often requiring corrective multi-pass grinding. Tangents and curves less than 3 degrees make up 91% of the CSX network and thus strongly influence the targeted grind frequency for scheduling the rail bound grinding equipment. A potential solution to grind sharp curves on a preventive cycle is using non-rail bound grinding equipment, but a cost benefit analysis is needed.

In fall of 2018, CSX increased the depth of cut for moderate and severe RCF to more aggressively correct surface damage conditions. Other initiatives underway involve updating the targeted rail profiles to address recent changes in traffic to reduce the contact stresses, installing higher quality rail more resistant to damage, and correcting track geometry issues such as wide gauge or rail can’t that result in increased contact stresses or super elevation not matched with actual car speeds resulting in increased loading on one rail.

Tangents within 200’ of Spiral: The metal removal performed using the production grinder does not currently match the damage from the curve that extends out onto the tangent, as the tangent segment is treated for the average condition found in the full tangent segment. As of 2018, in order to grind these tangent segments past the spiral, the specialty grinder is used to treat the smaller location when observed by the RIV operator and then entered into the specialty grinding grind plan. Comparing the lengths of all locations, 19% of SSCs in the 2014-2018 CSX data set are in the spiral or 200 feet past the spiral. The damage is expected to be higher as the spiral takes the brunt of the lateral forces to begin steering the trucks through the curve or returning to tangent. Within all curves and out to 200’ past the spiral, only 39% are in the full body only.

Figure 6 – Total Lengths of SSCs and SDZs by Location

The project to update the CSX rail grinding profile templates also includes the creation of new “transition templates” to be applied within the beginning of the spiral and out onto the 200’ feet of tangent. This will enable a smoother transition between the two distinct rail profiles between the high rail and tangent and the low rail and tangent, as well as allow for separate grind patterns and passes to be called for this transition segment that has damage similar to the curve. This transition segment is planned to be ground by the production grinder for longer spirals and by the specialty grinder for shorter spiral lengths, due to production grinder equipment requirements for a minimum length needed to change grinding patterns.

Reducing Metal Removal Demand

For all locations where damage is higher, the levers to reduce Rolling Contact Fatigue to more closely match the grind cycle for the rest of the subdivision/route are:

• Install higher quality rail that has higher shear strength (resistance to contact stress damage) • Ensure effective and consistent lubrication for gauge face and top of rail (lower the traction

coefficient) • Correct rail profile to avoid non-conformal and closely conformal wheel rail profile matches

(resulting contact stresses are 4 times more sensitive to the profile than the car’s weight) • In Curves:

o Adjust track super elevation and/or change timetable speed to more closely match actual train speeds when possible

o Correct rail can’t with tie replacement or adzing o Correct gauge variation greater than ¾” from standard

Conclusions

Since 2015 CSX and Loram have worked together to make changes to the rail grinding program with the goal of reducing SSC and SDZ defects. Results of the changes can be seen in the overall length of track marked as a SSC or SDZ defect going back to 2014 in Figure 7.

Figure 7 – SSC & SDZ Defect Total Lengths by Location 2014-2018

Going forward CSX and Loram plan to continue expanding the benefits of the combined grinding operations to provide efficient work load leveling and increase the level of detail to address shorter variations in the demand. Traditionally each curve or tangent section was ground as a single unit based on the grinding needs of a majority of the section. As advancements in automated rolling contact fatigue evaluation and improved location tracking develop individual curve and tangent sections can be broken up based on grinding need either for transverse profile shape or rolling contact fatigue levels. Work is

underway to create a standard Rail Surface Quality Index (RSQI) that categorizes RCF in stages based on the visual assessment and known depths of cut. Each stage (score) is progressive and data collected will enable determining how much traffic (by tonnage) it takes to advance to the next stage for each curve. This rail surface scoring index data can be used to adjust grinding cycles and build a large data set enabling a move to predictive grinding. (Hampton, Magel, & Harris, 2017). By collecting the surface condition data at a more detailed level, better aggregation can be performed to immediately develop a customized grind plan. By working both the production and specialty grinders in the same block the utilization of each can be maximized by planning the production grinder to handle to majority of the demand and utilizing the specialty grinder to provide additional grind effort to smaller sections of track.

References

1. Hampton, D. M. (2017). Moving to Predictive Grinding. International Heavy Haul. Cape Town, South Africa.

2. Magel E, S. S. (2004). Control of Rolling Contact Fatigue of Rails.

Reducing Untestable Rail Due to Surface Conditions

Daniel W Hampton – Manager Contract Services II, CSX TransportationCharles P Rudeen – Chief Engineer Quality, Loram Maintenance of Way

Reducing Untestable Rail Due to Surface Conditions

- Definitions- Why important- Remediation- Causes- Changes to CSX Rail Grinding

Program- Steps you can take

- NT: Non testable area - SSC (Shelly, Spalling, or

Corrugation) Rail Defect: rail surface condition rough enough to interfere with ultrasonic testing for internal rail defects.

- SDZ is just a SSC in the dead zone of a switch.

Ultrasonic Testing Codes used when not able to obtain a valid test

Mount Carbon, WV DerailmentFebruary 16, 2015

SSC (Shelly, Spalling, or Corrugation) rail defects

- Area of risk: No visibility of internal defects that can cause a broken rail. Ultrasonic testing may be set for as frequent as every 31 days to prevent rail breaks which can cause a derailment.

- Measured in count (where occurring) and length (amount of risk)

- Lagging indicator of where on the network we have not met the metal removal demand to maintain the rail surface condition.

Objectives:1) Minimize the total risk

(reduce length)- Corrective actions:

replace rail or grind out

2) Eliminate occurrences (count)

- Understand- Prevent

2014 2015 2016 2017 2018

CSX SSC & SDZ Total Lengths by Cause per year

Why perform metal removal

Control rail shape (profile)BEFORE

AFTER

Control surface conditions

Metal Removal – Rail Profile

• Wheel Rail Interaction -> contact stresses -> rail wear and surface damage -> rail life

• Contact patch size -> fuel savings

Metal Removal – Surface Conditions

• Reduce or eliminate surface initiated rail defects -> Reduce broken rails (rail service failures) -> Reduce derailments

• Allow ultrasonic testing and maximize speed of testing

• Reduce vertical loading (dips, corrugation)

Transverse Detail Fracture (TDD) - A detail fracture is a progressive fracture starting from a longitudinal separation close to the running surface, or from shelling starting at the gage corner and spreading transversely through the head. This would be known as a Detailed Fracture from Shelling. Detailed fractures can also occur from head checking.

• Example of Surface Initiated Defect = Transverse Detail Fracture (TDD)

• 27% of Total Rail Defects• 29% of Rail Service Failures

Surface Initiated Defect: TDD

Rail Grinding EquipmentProduction Grinders – RG400 series 120 Stones – 650’ Long

FCC Water Car 1 Grind Car 1 GC2 GC3

GC4 Water Car 2 Water Car 3 Power Car Caboose

Forward Control Car (FCC) Grind Car w/Rear Cab

Specialty Grinders – RGS series 24 Stones– 135’ Long

• Out of face rail grinding at up to 20 mph

• Grinds specialty assets (switches, crossings, etc.) at 3-6 mph

Can Rail Grinding remove SSCs/SDZs?2015 field verification testing on high SSC route- 24 stone specialty grinder, ultrasound walking stick, test between grinding passes - 9-25 passes with 24 stone grinder- Typically need at least 2 trains to eliminate grinding mark interference- Approximately 80% of SSCs can be ground visually clean along the length - Use standard of pitting deeper than 1/4” on less than 40’ section, then replace

with plug rail

Can Rail Grinding economically remove SSCs/SDZs?

YES NO

How long will it take to reappear?

• It depends…• Rate of Damage vs. Grind Cycle Frequency and

depth of cut per cycle• Why weren’t we meeting the metal removal demand?

• Location of increased damage • Length of track receiving increased damage

• What are the mechanisms causing damage higher at this location than other locations?

Why weren’t we meeting the metal removal demand?1. Assets or track segments are not being ground2. Operational Issue: Repeatedly late cycle or

inadequate depth of cut3. Scalability: Grinding the subdivision on cycle, but…

• Track segment frequency higher than route/subdivision frequency

• Location within the track segment has a higher demand than rest of segment

Common SSC locations• Specialty Assets – 35%• Curves – 26%• Rail Quality & Fatigue – 17%• Bridges – 14% • Tangents within 200’ of

spiral – 3%• Sidings – 3%• Steep Grades and Absolute

Signals – 1%

- 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000

0

5,000

10,000

15,000

20,000

25,000

SSC SDZ Total Lengths vs. Specialty Units Ground

NEAR SPECIALTY ASSETSPECIALTY UNITS GROUND

Program Changes to match metal removal demand based on Location

Specialty Assets • Increase in specialty grinding• Pre-inspection of specialty

assets • Use of CSX rail profile

templates creates uniform running band, lowers contact stresses and resulting RCF damage

Program Changes to match metal removal demand based on Location

SSC Removal• Automated SSC/SDZ data from

CSX to Loram equipment within 2 hours of receipt from rail test company

• Developed GIS map view targeting system for grinder cab and phone app with display of defect information and distance from defect

• Eliminated lost time searching with man on ground for SSCs in the dark

Program Changes to match metal removal demand based on Location

Bridges• Previously only 5%-10% of

bridges ground • Unknown risk & info• No standard to follow

• Began grinding 80% - 85% of bridges based on structure types:• Do Not Grind when Timber

Substructure • Use different levels of Pre-

soak, post-soak & fire watch

Bridge Characteristics Miles %Timber Substructure 35 15%Timber Open Deck 97 42%Timber Ballast Deck 18 8%Concrete Deck 78 34%Total Active Undergrade Bridges 229 100%

Program Changes to match metal removal demand based on Location

Bridges• Developed GIS map view

for grinder cab and phones with display of bridge data and action required

• Assigned B&B manpower with a fire suppression truck to perform duties following grinding team

Program Changes to address Operational Issues• Increased Depth of Cut for moderate & severe RCF• CSX Technology Operations Research group created

a computer-based Schedule / Routing Optimizer • Minimizes travel and determines optimal routes for

grinding on cycle across the network based on priorities and tonnage frequencies

• Inspect tangent every cycle and algorithm determines what is skipped based on criteria.

Optimization Problem

• High variability in the metal removal demand across the network.

• Methods to address:• more accurately capture the demand (level of detail)• refine operational capabilities to achieve customized

metal removal that more closely matches the demand (work load leveling, right sized equipment)

• reduce the demand variability

Upcoming Changes to address ScalabilitySSCs/SDZs are lagging indicatorsNeed foot by foot evaluation of the rail instead of just recording condition for whole track segment, and better tracking of where on the rail head.Rail Surface Quality Index (RSQI)• Shows progressive growth of RCF -> Predictive

measure (MGT between stages)• Establishes depths of cut / metal removal

needed to remediate• Based on years of visual assessment and

remediation• Technologies that evaluate internal damage to

correlate existing output to RSQI so depth of cut can be determined

Program Changes to address Scalability• Demand in portion of track segment < full segment

• Requires a more detailed grind for short segments receiving higher damage. • Example: first hundred feet of absolute signal where engines are starting and stopping.

• Combined Rail Grinding Operations with pre-inspection for both specialty and production grinders provide full coverage of rail, increased efficiency, safety, and 65% less track time to accomplish the same work.

• Rail Inspection Vehicle creates a grind target for shorter length rail surface conditions for 150’ Specialty Grinder to address after 650’ Production Grinder.

Upcoming Changes to address Scalability – Complementary Grind Plans

• Cycle 1

• Cycle 2 when Cycle 1 without Specialty Grinder

• Cycle 2 when Cycle 1 with Specialty Grinder

ProductionGrinder

Profile orSurface Deviation

SpecialtyGrinder

LEGEND

• RESULTS: More closely match demand = rail life extension by minimizing grinding wear and surface initiated defects

Met

al

Rem

oval

De

man

dDistance

Levers to reduce metal removal demand1. Rail Profile - reduce

contact stresses (P0) 2. Rail Quality – increased

shear strength (K) to resist damage

3. Lubrication – (gage face and top of rail)when combined with surface roughness determines the traction coefficient (T/N).

Rail

Qua

lity

Lubrication

Damaging Contact

Non-Damaging Contact

P0

P0

Levers to reduce metal removal demand• Rail Profile shape can increase the contact

stress 4x greater than an empty vs. loaded car

Rail

Qua

lity

Lubrication

2

1

Fitzgerald MP 654Measured Hardness: 336 HBMeasured TOR COF: 0.404

Conformal ContactEmpty Tank Car Po= 93,000 psi

1

Conformal Contact Loaded Tank Car Po= 156,350 psi

2

Non - Conformal Contact Loaded Tank Car Po= 440,040 psi

3

3

Non-Damaging Contact

Diagrams & analysis peformed by Canadian National Research Council (CNRC) Alexander Woelfle for CSX

Non-Damaging Contact

Damaging Contact

Levers to reduce metal removal demand

• Next highest impact levers for Curves• Adjust track superelevation and/or change timetable

speed to more closely match actual train speeds when possible

• Correct rail cant with tie replacement or adzing • Correct gauge variation greater than ¾” from standard

Summary• SSCs and SDZs are lagging indicators of where we are not meeting the

metal removal demand and often have higher damage occurring. • Majority are less than 50’ long, while a 650’ long production grinder

grinds a track segment uniformly (average 1,315’ for curves and 2,444’ for tangent) requiring overgrinding to treat or a more detailed grind.

• They can be ground out but need to move from remedial to preventive.• Proactive actions:

• utilize more granular predictive surface condition scoring • develop program and equipment to efficiently match the variation in metal

removal demand• take actions to reduce the forces creating higher demand