Rail Stress Management Webinar #3

Transcription

1 Rail Stress Management Webinar #3 Improving Hot Weather Speed Restrictions and Curve Stability Management September 19, 2013 Sponsored by James P. Aten President, L.B. Foster - Salient Systems Dublin, OH USA Dr. Andrew Kish Kandrew Inc. Consulting Services Peabody, MA USA

2 Improving Hot Weather Speed Restrictions and Curve Stability Management Overview Overview Key Questions Addressed: When and where should speed restrictions be imposed? [Topic #1] How to improve curve stability through more effective neutral temperature (RNT) management? [Topic #2]

3 Improving Hot Weather Speed Restrictions and Curve Stability Management Talking Points Topic #1: Improving Hot Weather Speed Restrictions What is the neutral temperature (RNT) problem relevant to speed restrictions? How to best determine/estimate neutral temperature? What is a new speed restriction formula based on buckling safety criterion and risk acceptance? Topic #2: Improving Curve Stability Management What is the neutral temperature problem in curves and what are the key adjustment issues? How to best manage curve stability through more effective destressing? 99-03

4 Improving Hot Weather Speed Restrictions Why Speed Restrictions? ANSWER: to reduce the risk of track buckling caused derailments After track maintenance (weakened ballast) At elevated temperatures (high thermal force conditions) Approach: apply slow orders while restoring track lateral resistance to increase buckling temperature. Approach: apply slow orders to reduce train dynamics thereby reducing buckling risk. If dynamic track stabilization (DTS): no or limited slow orders. If no dynamic track stabilization: slow order traffic for 0.1 MGTs or more. When (at what temperatures) to impose speed restrictions? Where (localized or blanket)? localized blanket where weaker RNT conditions exist territory wide as preventive measure

5 Improving Hot Weather Speed Restrictions Speed restrictions at elevated temperatures (high thermal force conditions) Where (localized or blanket)? localized where weaker RNT conditions exist RR s CWR Policy blanket territory wide for heat protection RR s CWR Policy Applies slow orders* to locations where RNT has NOT been readjusted after rail breaks or defect removals Applies blanket speed restrictions at stipulated threshold levels (typically at RLT + 10 to 20 deg F) Rail Break/Cut Temperature ( F) Rail Temps ( F) at Which to Apply Slow Orders if not Readjusted * 25mph with NO daily inspections; or 40 mph WITH daily inspections May be conservative! Can be improved through a new speed restriction rationale based on buckling safety criteria and RNT knowledge

6 Speed Reduction Issues and Concepts How Do Speed Restrictions Reduce the Derailment Risk? Speed restrictions reduce the consequence of a derailment i.e. less damage severity. Theory suggests lower speeds (less train energy) reduces buckling potential (i.e. increases buckling temperature when operating in the buckling regime ) T Bmax (Zero Buckling Energy) Low speed Temperature increase above neutral T Bmin (Max. Buckling Energy) Buckle Safe temperature Safe temperature increase increase (T all ) Buckling regime High speed Track Buckling Strength Determines Slow-Order Temperatures δ o Deflection Slow speeds (i.e. reduced energy input into track) raises the track s buckling temperature within the buckling regime

7 Buckling Margin of Safety (BMS) Criterion BMS Risk Acceptance for Speed Restrictions T T Bmax T Bmin BMS = T all T SR BMS is the reserve buckling strength in terms of additional temperature increase to cause a buckling risk T all T SR RNT Deflection T all, RNT, and BMS determine speed restriction temperatures, T SR!

8 Slow Order Temperature Criterion Question: When (at what temperatures) to impose slow orders? Answer: At At a temperatures at at which the buckling probability potential of buckling is is acceptably is acceptably small. small. Determined by T all and RNT Determined by BMS risk acceptance T T Bmax BMS ACTION T Bmin BMS Larger than 20 F OK; no action T ALL all T SR Less than 20 F Advisory on slow orders RNT Buckle Amplitude T SR Equation T SR = T all BMS min + RNT Example: 130 F 80 F 20 F 70 F How to Know/Determine T all and RNT?

9 How to Determine Safe and Critical Temperatures? US DOT/Volpe Model CWR-SAFE evaluates buckling temperatures: CWR-SAFE T R T Bmax Version 2000 CWR-BUCKLE (deterministic) T Bmin BMS Buckling regime T all Property of the USDOTs Federal Railroad Administration. Developed by the USDOT s Volpe Center and Foster-Miller, Inc. T N Buckle Amplitude T SR BMS based slow orders CWR-SAFE Property of the USDOTs Federal Railroad Administration. Developed by the USDOT s Volpe Center and Foster-Miller, Inc. Version 2000 CWR-RISK (probabilistic) Risk Index based slow orders Probability of buckling, PB (%) Risk Index Critical Temp, T c Rail Temperature ( F)

10 How to Determine Safe Temperature Increase, T all? CWR-SAFE T R T Bmax BMS based slow orders Version 2000 CWR-BUCKLE (deterministic) T Bmin BMS Buckling regime T all T SR Property of the USDOTs Federal Railroad Administration. Developed by the USDOT s Volpe Center and Foster-Miller, Inc. T N Buckle Amplitude CWR-SAFE Parametric Studies Ballast lateral resistance Alignment defect Curvature + Buckling Tests + Industry Guidance 70 WEAK 90 AVERAGE STRONG ALL TRACK TYPES and CONDITIONS 110 Buckling Strength, T all ( F) Track Condition Based T all Rule of Thumb weak, average or strong 60 F 80 F 100 F Track buckling strength, T all, can be calculated directly by CWR-SAFE or approximated by the 60:80:100 rule

11 How to Determine Safe Temperature Increase, T all? Buckling Strength (T all ) Characterization: Weak, Average or Strong? CWR-SAFE Parametric Studies Ballast lateral resistance Alignment defect Curvature + Buckling Tests + Industry Guidance 70 WEAK 90 AVERAGE ALL TRACK TYPES and CONDITIONS Buckling Strength, T all ( F) Buckling Strength Rule of Thumb Descriptors STRONG 110 Track Condition Based T all Rule of Thumb weak, average or strong 60 F 80 F 100 F Weak (60 F) Average (80 F) Strong (100 F) Higher degree curves (4 deg+); line defects exceeding US Class 4; weak to average lateral resistance (at least partially consolidated, 2000#/tie +) Tangent to low deg curves; Class 3-5 line defects; typical standard lateral resistance 3000#/tie + Tangent, well consolidated (4000#/tie +); Class 5-6 line defects Note: for details on buckling strength characteristics and parametric studies, refer to Kish & Samavedam: Track Buckling Prevention: Theory, Safety Concepts, and Applications [DOT/FRA/ORD-13/16, March 2013]

12 How to Determine Neutral Temperatures (RNT)? RNT Determination for Slow Orders Slow order Equation T SR = T all BMS min + RNT Estimate Calculate Measure Industry Experience + Test Data on RNT Industry CWR Policy + RNT Recording Direct RNT Monitoring via RSM Technology RNT Reduction Factor: 30 F from RLT Example: rail installed (RLT) at 100 F, it s effective RNT is 70 F! Exception: in rail breaks, defect removals, and curves CWR Policies require recording/cataloguing RNTs when rail breaks or when cut for defect removal Such data could be used to estimate subdivision s RNT for slow orders Install RSMs either localized or territory wide Use actual RSM measured rail temp and RNT data to dictate slow orders

13 Illustrative Example 1: US Railroad Case Study When RNT Is Measured by RailStress Monitor (RSM ) IntelliTrack Data Management System Slow order Equation T SR = T all BMS min + RNT At what temperature to impose hot weather speed restrictions? RNT avg =103 F(40 C) High Quality Track: assume buckling strength = AVERAGE or T all = 80 F 20 F RNT avg =75 F(24 C) T SR = 80 F 20 F F = 163 F T SR = 80 F 20 F + 75 F = 135 F Note: this railroad s current hot weather speed restriction is at 90 F ambient (120 F rail) may be conservative

14 Illustrative Example 2: US Railroad Case Study 16 Months Data on a Major US Railroad Track RNT avg =100 F Neutral Temperature, F Tangent, concrete tie track 43.3 MGT for 2010 Class 4 ½ mile track segment (5 pairs of RSMs at 500 ft apart)

15 Illustrative Example 2: US Railroad Case Study 16 Months Data on a Major US Railroad Tmax=140 F COMPRESSION Track RNT avg =100 F COMPRESSION COMPRESSION Neutral Temperature, F T E N S I O N T E N S I O N Max Daily Temp Swing: 80 F Tmin=15 F RSM measured rail temperatures are beneficial to dictate speed restrictions i.e. they are exact, and NOT having to rely on ambient temp weather forecasts and conversions!

16 Illustrative Example 2: US Railroad Case Study 16 Months Data on a Major US Railroad Track RNT avg =100 F Neutral Temperature, F Tangent, concrete tie track 43.3 MGT for 2010 Class 4 ½ mile track segment (5 pairs of RSMs at 500 ft apart) T SR = T all BMS + RNT Current 105 F amb 180 F 100 F 20 F 100 F high quality concrete tie track [high buckling strength + high RNT] (100 F) (100 F) does NOT need blanket speed restrictions

17 Illustrative Example 3: US Railroad Case Study When to Apply Speed Restrictions on This Subdivision? Month # of Service Defects T BR (ºF) RNT PB (ºF) ΔT Causing Break (ºF) RNT Change from install TOTAL Avg Rail Break Temp Avg Subdivision RNT Avg Temp Difference Causing Break Avg RNT Decrease From RLT When (what temp) to impose blanket heat orders based on the above data? This subdivision is HIGH quality concrete tie track i.e. has at least AVERAGE buckling strength of 80ºF T SR = T all BMS + RNT = 132 F Current SR s at 90 F amb Note: above railroad data is also in line with RNT safety factor of 30 F reduction from installation RNT of 100 F!

18 Improving Current Speed Restrictions Are Current Hot-Weather Speed Restrictions Conservative? Typical SR temperatures: RLT F If RLT=100 F, then T SR =110 F, and for average track: RNT min = T SR -T all + BMS = = 50 F (a decrease of 50 F from RLT=100 F) Max RNT decrease* from RLT=100 F Strong [100 F] 70 F (highly conservative) Track Buckling Strength, T all Average [80 F] 50 F (moderately conservative) Weak [60 F] 30 F (minimally conservative) *Note: conservatism is even HIGHER since the additional 20 F margin of safety Hot weather speed restriction temperatures ARE conservative, and could be improved by effective application of the SR formula: T SR = T all BMS + RNT

19 Improving Speed Restrictions Hot Weather Speed Restriction Summary Methodology is in place to determine when/where to impose speed restrictions T SR Equation T SR = T all BMS + RNT Track Buckling Strength Rail Operator Defined Track Neutral Temperature weak, average or strong Risk Acceptance Based 60 F 80 F 100 F 20 F Estimated Calculated Measured Judicious application of methodology helps improve current speed restrictions!

20 Improving Curve Stability Management Topic #2: Improving Curve Stability Management How to improve curve stability through more effective neutral temperature (RNT) management? Discussion Topics Curved track stability issues Curve movement versus RNT behavior Curve RNT readjustment issues/difficulties How to best manage curve stability through more effective destressing?

21 Curve Stability Issues CWR Curve Stability Issues Curve s buckling strength is lower than tangent (a 6 deg curve buckles at 15 deg F lower than tangent, all other parameters being equal) 120 Influence of Curvature on Buckling Temperatures [CWR-SAFE Model] Temperature increase above neutral ( F) Temperature increase above neutral to buckle track Track Parameters wood tie track 136# rail 20 in. tie spacing Lat.Res=100 lb/in. Class 4 line defects Curvature (deg) Curve s tendency to string- line (move-in-out) with temperature is a major influence in its RNT condition, hence on in its geometry retention and stability Major impediment in managing curve stability is its UNKNOWN RNT condition [how to manage something we can t evaluate/measure?] Apply current best practice guidelines Manage via stress/rnt diagnostics

22 What Happens to RNT in Curves? Curve Mechanics/Rail Movement Issues Curved rails breathe (move in-and-out) under temperature changes. When the rail temperatures are below the RNT (in tension) curves pull in reducing RNT, and when rail temperatures are above the RNT (in compression) curves move out increasing RNT + RNT - RNT As tracks move in-out, ballast at tie ends get voided, thereby reducing lateral resistance. Also there can be line defects formed due to non-uniform curve movement.

23 What Happens to RNT in Curves? Curve Movement Impacts RNT! 18 month TTCI Curve RNT Test on a Major US Railroad 30 F Reduction curve realignment 45 F Reduction buckling prone conditions *from AAR/TTCI Report # ½ deg (320m radius) curve: RNT reduction of 45 F 2½ deg (728m radius) curve: RNT reduction of 30 F 5½ deg curve can be buckling prone in May/June hot temperatures Realignment restored 5½ deg curve to 75 F (but NOT to the desired RLT of 100 F)

24 What Happens to RNT in Curves? Curve Shift and Lining Track In-Out Influence on RNT Change 80 Curve Shift vs RNT Change Neutral temperature change at center ( F) RNT - RNT δ=curve movement δ = 6 in. δ = 5 in. δ = 4 in. δ = 3 in. δ = 2 in. δ = 1 in Degree of curvature Curve movement reduced RNT + reduced lateral resistance + increased line defects Buckling Prone Track Managing curve movement is critical to maintaining curve stability!

25 How to Manage Curve Stress/RNT Condition? The 3 Key Curve Stress/RNT Change Events Curve breathing due to temperature change Curve lining-in due to maintenance when rail in tension Rail breaks and defect removals Readjust reduced RNT by destressing Question: How to destress, when RNT is not known? Apply best practice guidelines: TTCI Software CWR-Adjust RR s CWR Policies Kish/AREMA 2013 paper Kish/Webinar #2 (May 23, 2013) Answer 1 (preferred): monitor curves with RailStress Monitors (RSM) Answer 2 (if no RSM monitoring): MUST have reliable curve movement data to enable any destressing

26 Curve Destressing Issues How to Destress Curves After Movement? Answer: either through curve realignment (easier) or rail removal (more difficult). Lining curves back is one method, but MUST have reliable curve movement data (usually via curve staking): Note: lining curves back to pre-movement geometry restores RNT ONLY to what it was before, hence NOT necessarily to target RLT! Destressing curves by cutting rail out is also an option, but: How much rail to cut out? [Depends on curve length and knowing/measuring curve pull-in requires effective curve staking]. When destressing, gap closing can be problematic [Rail pullers cause rail misalignments, especially in cold temperatures and in sharper curves.] Both rails have to be destressed

27 Curve Destressing Issues How Much Rail to Cut Out When Destressing Curves? Answer: depends on curve movement and length of curve When a curve pulls in, the effect is adding rail (i.e. to reducing RNT) Curve center maximum pull-in can be related to rail length change, therefore to the effective amount of rail added It is this effective rail added that must be removed for readjustment Key Requirement: Must know/measure curve movement: requires effective curve staking and data acquisition Determination of Rail Added versus Curve Movement: See Tables, Charts, Equation

28 Curve Destressing Issues Curve Movement Versus Length Change (i.e. Rail Added ) S=Rθ δ R-δ Curve ends don t move R R Change in S = S-s=Rθ (R-δ)θ = δθ or ΔS = δθ (where θ is in radians, and ΔS is per 100 feet of curve length) [1 degree=.0175 rad or 1 rad=57.3deg] θ EXAMPLE: a 1000 ft 5 curve pulling in 5 inches will change length by inches [i.e. 5 =0.0873rad; so 5"x x 10 = 4.375"] ΔS (in) = δ (in) x Deg x x curve length(ft)/100(ft) Equivalent rail added to remove

29 Curve Destressing Issues Guidelines on Curve Destressing How Much Rail to Cut Out? EFFECTIVE RAIL LENGTH ADDED TO CURVE PER 1000 FT. OF CURVE (INCHES) DEGREE CURVE DISTANCE CURVE HAS CHORDED INWARD (INCHES) 1/ / / / / / /8 1/8 1/8 1/4 1/4 1/4 3/8 3/8 3/8 1/2 1/ /8 1/8 1/4 3/8 1/2 1/2 5/8 3/4 3/4 7/ /8 1/4 3/8 1/2 5/8 3/4 7/ /8 1-1/4 1-1/2 1-5/ /8 3/8 1/2 3/4 7/ /4 1-3/8 1-1/2 1-3/4 1-7/8 2-1/ /4 3/8 5/8 7/8 1-1/8 1-1/4 1-1/2 1-3/4 1-7/8 2-1/8 2-3/8 2-5/ /4 1/2 3/ /4 1-5/8 1-7/8 2-1/8 2-3/8 2-5/8 2-7/8 3-1/ /4 5/8 7/8 1-1/4 1-1/2 1-7/8 2-1/8 2-1/2 2-3/4 3-1/8 3-3/8 3-5/ /8 3/ /8 1-3/4 2-1/8 2-1/2 2-3/4 3-1/8 3-1/2 3-7/8 4-1/ /8 3/4 1-1/8 1-5/ /8 2-3/4 3-1/8 3-1/2 3-7/8 4-3/8 4-3/ /2 7/8 1-1/4 1-3/4 2-1/8 2-5/ /2 3-7/8 4-3/8 4-3/4 5-1/4 Note: (1) equation and Tables apply to max. central inward movements (2) must have good curve staking data

30 Curve Destressing Issues Guidelines on Curve Destressing How Much Rail to Cut Out? DEGREE CURVE Guidelines EFFECTIVE on RAIL Curve LENGTH Destressing ADDED TO How CURVE Much PER Rail to 1000 Cut FT. Out? OF CURVE (INCHES) DISTANCE CURVE HAS CHORDED INWARD (INCHES) 1/ / / / / / / /2 1-7/8 2-3/8 2-7/8 3-3/8 3-7/8 4-1/4 4-3/4 5-1/4 5-3/ / /8 2-1/8 3-5/8 3-1/8 3-5/8 4-1/8 4-5/8 5-1/4 5-3/4 6-1/ /8 1-1/8 1-3/4 2-1/4 2-7/8 3-3/ / /8 6-1/4 6-3/ /8 1-1/4 1-7/8 2-1/ /8 4-1/4 4-7/8 5-1/2 6-1/8 6-3/4 7-3/ /8 1-1/ /8 3-1/4 3-7/8 4-5/8 5-1/4 5-7/8 6-1/2 7-1/4 7-7/ /4 1-3/8 2-1/8 2-3/4 3-1/2 4-1/4 4-7/8 5-5/8 6-1/ /8 8-3/ /4 1-1/2 2-1/ /4 4-1/2 5-1/4 5-7/8 6-5/8 7-3/8 8-1/8 8-7/ /4 1-1/2 2-3/8 3-1/8 3-7/8 4-3/4 5-1/2 6-1/ /8 8-5/8 9-3/ /8 1-5/8 2-1/2 3-3/8 4-1/ /4 6-5/8 7-3/8 8-1/4 9-1/ /8 1-3/4 2-5/8 3-1/2 4-3/8 5-1/4 6-1/ /4 8-3/4 9-5/8 10-1/2 NOTE: for in-between TABLE values interpolate; if can t, use the equation below: ΔS (in) = δ (in) x Deg x x curve length(ft)/100(ft) Effective rail added (inches) Max central curve movement (inches) Degree of curve (deg) Example: if a 1500ft - 6 deg curve pulls in by 8 in, what is the effective rail added? [12.6"]

31 How to Manage Curve Stress/RNT Condition? Managing Stress/RNT After Curve Movement SUMMARY Curve breathing due to temperature change Curve lining-in due to maintenance when rail in tension Readjust reduced RNT by destressing MUST have reliable curve movement data Method 1 Method 2 Destress by lining curve back to original position restores RNT to what it was before the pull-in Destress by cutting rail out amount given by Curve Movement Tables both rails need destressing restores RNT to what it was before the pull-in might need alignment correction Question: how to manage if there is NO curve movement data?

32 How to Manage Curve Stress/RNT Condition? Question: how to improve or manage WITHOUT curve movement data? ANSWER Deploy RSMs (RailStress Monitors) in curves to monitor RNT Manage curve stability by real time RNT data Destress as per RNT change dictates Monitor destress quality and post /RNT behavior via RSMs

33 How to Manage Curve Stress/RNT Condition? Example #1: how RSMs indicate curve movement through RNT change? Curve Shift RNT Data from a Major US Railroad Low Rail RSMs High Rail RSMs RNT ( F) F RNT decrease in a 2 deg curve pulling-in 2½ in RSMs can supplement/replace staking of curves for a more exact RNT condition determination

34 How to Manage Curve Stress/RNT Condition? Example #2: how rail stress/rnt measurements promote destress quality and post-destress RNT condition management? 18 month TTCI Curve RNT Test on a Major US Railroad curve realignment Indicates destress quality: restored only to 75 F (and not to the 100 F target) 45 F Reduction buckling prone conditions Monitors post-destress RNT condition Salient System s RSM has a Buckling Hazard Warning Capability *from AAR/TTCI Report # 1003 Refer to Webinar #1

35 Conclusions on Improving Hot Weather Speed Restrictions Webinar # 3 Summary Notes: Topic #1 Hot weather speed restrictions depend on track buckling strength, track RNT, and a risk based safety factor in accordance with the formula: T SR = T all BMS + RNT Track Buckling Strength Rail Operator Defined Risk Acceptance Track Neutral Temperature Track buckling strength can be estimated by the 60:80:100 rule for weak, average or strong tracks in terms safe temperature increase above neutral Track RNT can be estimated based on rail break/defect removal data or by direct RSM (RailStress Monitor) measurement Application of speed restriction formula helps to evaluate/improve current heat restriction policies, thus promoting increased network capacity without compromising safety

36 Conclusions on Improving Curve Stability Management Webinar # 3 Summary Notes: Topic #2 Curve movement has a large influence on lateral stability Highly impacts track RNT Also impacts lateral resistance and alignment condition The knowledge of RNT is critical for effective curve stability management Curve realignment and cutting rail out are the current best practice procedures for RNT management, but: Both require reliable curve movement data Restores RNT to only what it was before movement (i.e. NOT necessarily to target RNT) The only correct destressing of curves is via RNT measurements such as Salient System s RSM technology

37 Rail Stress Management Webinar # 3

38 Lecturer s CV ANDREW KISH, Ph.D. Railway Technology Consultant Kandrew Inc. Consulting Services (978) kandrewinc@aol.com Dr. Kish received his Ph.D. from New York University in Applied Mechanics in 1974 under Prof. Arnold D. Kerr. He has been at the US DOT s Volpe Center for over 30 years where as a senior technical expert on track structures and mechanics managed a multitude of research programs in the field of track mechanics including track stability, CWR maintenance, track buckling prevention and high speed rail safety. On the international front, from he served as the US Department of Transportation s representative to the European Rail Research Institute s (ERRI) committee on track stability where he was instrumental in developing new CWR safety codes for the Union of International Railways (UIC). His work has resulted in over 100 publications on CWR stability and safety and related topics and he has been a three-time recipient of the US Department of Transportation's Superior Achievement Award. Dr. Kish also held the position of Adjunct Assistant Professor of Aerospace and Mechanical Engineering at Boston University and has also taught graduate courses in applied mechanics at Northeastern University. Over the years, he has conducted many track buckling workshops and seminars for the industry and is a recognized international expert in the field of track stability, CWR maintenance and track buckling prevention. Dr. Kish retired from the Volpe Center in 2004 to form his consulting company Kandrew Inc. Consulting Services and remains active in providing railway technology consulting services to the research community and the railway industry.

39 Applicable Bibliography Recent relevant publications: Kish and Samavedam, Track Buckling Prevention: Theory, Safety Concepts and Applications, DOT/FRA/ORD- 13/16, Final Report, March 2013 Kish, Best Practice Guidelines for CWR Neutral Temperature Management, to be presented at 2013AREMA/Railway Interchange Conference, September 29-October 2, Indianapolis, USA Kish and Aten, A Smart-Systems Approach for Better Managing CWR Thermal Forces at Extreme Temperatures, presented at 2012 CORE, September 10-12, Brisbane, Australia Kish, On the Fundamentals of Track Lateral Resistance, presented at 2011AREMA/Railway Interchange Conference, September 18-21, Minneapolis, USA Kish and Clark, Heat Impact on CWR: Track Buckling Prevention Through Risk Based Remedial Actions, presented at the International Heavy Haul Association (IHHA) Technical Conference, June 19-22, 2011, Calgary, Canada Kish, McWilliams and Harrison, Track Buckling Hazard Detection and Rail Stress Management, presented at the World Congress for Railway Research (WCRR), May 22-25, 2011, Lille, France Kish and Clark, Track Buckling Derailment Prevention Through Risk-Based Train Speed Reductions, 2009 AREMA Annual Conference, Chicago, September 2009 Kish, Track Lateral Stability, in Guidelines to Best Practices for Heavy Haul Railway Operations: Track, International Heavy Haul Association publication, 2009 Kish, Destressing/Restressing for Improved CWR Neutral Temperature Management, in Guidelines to Best Practices for Heavy Haul Railway Operations: Track, International Heavy Haul Association publication, 2009