Improving the Fatigue Resistance of Thermite Railroad Rail Weldments F. V. Lawrence Y-R. Chen J. P. Cyre 1
Outline! Fatigue problems with thermite welds! Improving the rail head! Improving the rail web and base 2
Metallic Fatigue ACELA A.M. Zarembski Bulletin 673, 1979, Volume 80 of AREA proceedings 3
Rolling contact fatigue Railroad car wheel moving over rail causes fatigue to occurinboththerail head and base. H Rail W B Thermite Weld 4
Fatigue crack initiation sites Internal Fatigue Crack Rail Head Rail Web! 40% of all service failures are due to thermite field welds.! 10% of all derailments are due to broken field welds. Web-to-base Fillet Fatigue Crack at Weld Toe in Fillet Rail Base Fatigue Crack at Weld Toe in Base 5
Fatigue crack in rail head Internal fatigue crack initiation in rail head 6
Fatigue crack in rail base Cold Lap Site of crack initiation Limit of fatigue crack growth 7
Thermite weld service failures 35.0% 30.0% 25.0% 20.0% 15.0% 10.0% 5.0% 0.0% Base Web-base fillet Web Location Head-web fillet Head? Record of 244 service failures on a Class 1 railroad involving thermite field welds. 8
Service failures or markouts?! Most field-weld service failures originate at web or base.! But defects detected and removed from the rail head before a service failure can occur ( markouts ) exceed service failures by 2:1! 9
Implications! Fatigue cracks in the web and base are less frequent but are the principal cause of service failures since they are difficult to detect. Crack initiation occurs at external stress concentration.! Fatigue cracks in head are more frequent but are generally removed. Crack initiation occurs at internal stress concentration. 10
Outline! Fatigue problems with thermite welds! Improving the rail head! Improving the rail web and base 11
Porosity types Shrinkage Gas (Spherical) Thermite welds studied contain about 1.5% shrinkage porosity. 12
Porosity initiates fatigue Odario 1992 Running surface. Formation of shells in tangent track at interdendritic shrinkage porosity. 13
Interdendritic shrinkage porosity 14
Possible solution?! Eliminate weld metal! (?)! Developed a modified thermite welding process called Squeeze Welding in which ends of joint forced together to expel most of the thermite weld metal. 15
Squeeze welding Rail Ends Moved Together While Weld Metal Still Molten Expelled Imputities Force Force Final Weld Thickness Rail Cross-section 16
Weld longitudinal-sections Standard Squeezed WM HAZ BM Fry 1992 17
Laboratory test results Withee 1998 1000 Maximum Stress, S max (MPa) 100 Withee - Standard Withee - Squeezed Withee - Vibrated Liu - Standard Liu - Squeezed Liu - Vibrated Liu - Squeezed/Vibrated 10 10 4 10 5 10 6 10 7 10 8 Fatigue Life, N f (cycles) Fatigue behavior of small specimens taken from head of 6.35 mm weld shows some improvement. 9.27 mm 19 mm R 208 mm 19 mm 87 mm 18
But distribution unchanged! 1.0 0.8 Cumulative Probability 0.6 0.4 0.2 Standard Weld Squeezed Weld Vibrated Weld size range of pores initiating failure imputed from SEM images 0.0 10 0 10 1 10 2 10 3 10 4 10 5 Pore Size, area ( µm 2 ) Pore size distribution unchanged! Withee 1998 19
Largest pore size controls! Initial Stress Intensity Factor, K o (MPa*m 1/2 ) 100.0 10.0 19 mm 6.35 mm R 208 mm 87 mm 1 3 19 mm 9.27 mm B2 B3 B1 Standard Weld (C) Squeezed Weld (B) Vibrated Weld (D) Regression Analysis 1.0 10 4 10 5 10 6 10 7 B4 C5 D5 C1 Fatigue Life, N f (cycles) Withee 1998 Single relation for all treatments depending only on pore size (and applied stress). 20
Implications! Reducing the size of the largest pores and/or the volume of weld metal should increase in the (average) fatigue life.! Largest pore per unit volume (porosity) and the volume of weld metal jointly determine the fatigue strength. 21
Theoretical study Σ(t) p (t) Fry 1995 22
Stress MPa) Stress history experienced 23
Fatigue occurs at critical depth Depth below running surface, Y (mm) No residual stress Considering residual stress Worst depth Worst locations on pore Fatigue damage parameter Fry 1995 24
Effects of pore shape? RAHELS Predictions Ratio of pore's longitudinal and transverse axes, λ X / λ Z 0 0.5 1 1.5 2 2.5 3.5 Favors vertical split-head Shelling Vertical Split Head Detail Fracture FBY 3 λ Y /λ Z 2 2.5 Favors Detail fracture PCT 15 20 25 30 40 10 8 4 6 PCV FBX 2 1.5 1 Fry 1995 60 50 0.5 72 2 0 Favors shelling FBZ PCH Sphere 25
Model predictions! Critical depth for fatigue crack initiation ( 15mm) determined by wheel-contact-induced residual stresses.! Model predicted that shelling, vertical split heads and detail fracture could all initiate at shrinkage pores depending upon the pore shape. 26
New measurement technique specimen film Central portion of weldment machined and ground flat to 12.7 mm thickness. Stepped penetrameter. Chen 2000 27
Typical radiograph L1!?! Difference in contrast due to micro-porosity (shrinkage porosity. Porosity not uniformly distributed! Chen 2000 28
Radiographs of field welds F1 F2 F3 29
Optical determination of porosity 0.52% interface 1.72% 30
Radiographic image density 0% 0.9% porosity 39000 37000 35000 33000 31000 29000 27000 25000 0 500 1000 1500 2000 2500 3000 3500 4000 Distance BM F1 WM Measured changes in grey scale in photoshop. Penetrameter with 0.11 mm steps indicate at least 1% sensitivity 31
Porosity in 10 thermite welds Chen 2000 1.4 1.2 1 Porosity (%) 0.8 0.6 0.4 0.2 0 B-3 B-4 A-6 A-7 A-8 A-1 L-1 B-1 A-2 A-3 Average porosity in 10 markouts varies considerably! 32
Developing detail fracture Detail fracture in head of rail appears to be developing in association with an area with a high concentration of shrinkage porosity? 33
Conclusions! Large variation in porosity from weld to weld. Porosity not uniformly distributed.! Porosity clusters at weld centerline frequently seen. Fatigue cracks in head often associated with associated with porosity clusters. 34
Why? A-1 A-2 A-3 A-4 A-8 A-9 A-10 B-1 L-short L-long 100mm Apparently there are large variations in thermal conditions during thermite welding. Observed variations in melt-back (weld profile) on radiographs. 35
Outline! Fatigue problems with thermite welds! Improving the rail head! Improving the rail web and base 36
Thermite weld service failures 35.0% 30.0% 25.0% 20.0% 15.0% 10.0% 5.0% 0.0% Base Web-base fillet Web Location Head-web fillet Head? Record of 244 service failures on a Class 1 railroad involving thermite welds. 37
Web-to-base fillet!?!! Why does this happen?????! Answer:! Residual stresses!! Weld toe geometry!! Flank angle.! Cold laps. 38
Residual stresses Compression Tension Neutral Axis Critical locations: Web-to base fillet Rail base Webster et al. 39
Weld toe flank angle 85 Flank Angle 40
Weld toe geometry Flank Angle (θ) Toe Radius (r) Roughness (R) Weld Metal Base Metal Improve by: Flank angle Fatigue Severity = 1+0.27 tanθ 0.25 t ( 1+0.1054S u R) 1 r 41
Current Orgo-thermit mold profiles Measured profiles of Orgo-thermit molds AA A-A BB A-A 45 B-B C-C 30 D-D CC E-E Mold Rail and weld DD 42
Modified Orgo-thermit mold profiles Modified Current Suggested modifications to Orgothermit molds AA BB CC DD 43
Nature of critical defects? Sand burn in Inclusion in Head Hot Pull-apart Grind Burn Columnar Grains in Head Lack of Fusion Porosity Hot Tear Slag Cold Lap 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0% Analysis of 244 service failures on a Class 1 railroad involving thermite welds. 44
Cold laps - Dimitrakis Coldlap Nocoldlap Cold laps greatly reduce the fatigue life of a weldment 45
Cold laps at thermite weld toe Cold Lap Base Metal 46
Weld toe cold laps θ D φ r Weld Metal Weld Toe Location Without Cold-Lap Defect Heat Affected Zone Curved Path Vertical Path Base Metal Loading Direction 47
Effect of cold laps Condition Percentage of Fatigue Life Flank angle (θ) = 30Þ 100% Flank angle (θ) = 45Þ 56% Flank angle (θ) = 60Þ 44% Cold lap depth (D) = 0 100% Cold lap depth (D) = 1mm 20% Cold lap depth (D) = 2mm 15% 48
Causes of cold laps! Gap between mold and rail in the critical web-to-base fillet area.! Inadequate melt back causing incomplete fusion at the weld toe? 49
Variations in melt-back A-1 A-2 A-3 A-4 A-8 Melt back varies considerably in the location of the web-to base fillet A-9 A-10 B-1 L-short L-long 100mm 50
Melt back dimensions 0 1 Height of Rail (in.) 2 3 4 Weld Sample #2 Weld Sample # 3 Weld Sample #1 5 6 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Length of Melt Back (in.) 51
Melt back at web-to-base fillet 50 Collar width defined by mold 40 Bad! Good! 30 20 10 0 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61-65 66-70 Melt-back width (mm) 52
UIUC Experimental Program Rail Weld Metal Rail Melt back Standard 1 thermite weld: Large flank angle and cold laps. UIUC modified 1 thermite weld molds are used with a 1.4 rail gap. Mold sealed at weld toe with refractory paste. And: Reduced flank angle! 53
UIUC Experimental Program Sealing paste from Railtech w/ Brazing Flux. Lutting paste from Railtech. 54
UIUC Experimental Program Leecote mold wash and Uni Ram Blu refractory paste. Uni Ram Blu refractory paste. 55
3 2 1 Weld Fabrication 4 5 6 56
Fatigue testing Standard 4-point bending test. 57
Modified Weld Specimen #28 Crack initiation points Limit of fatigue crack growth 58
Effect of modifications Cold lap formation beyond sealing paste 59
Standard weldments 10,000 1 3 Process A Process B Process C Process D TAMU 1,000 100 1,000 10,000 100,000 1,000,000 10,000,000 Cycles to Failure, N f 60
UIUC experimental welds 10,000 1,000 1 3 Process A Process B Process C Process D Modified UIUC TAMU 100 1,000 10,000 100,000 1,000,000 10,000,000 Cycles to Failure, N f 61
UIUC experimental welds 10,000 1,000 1 3 Process A Process B Process C Process D Modified UIUC TAMU 100 1,000 10,000 100,000 1,000,000 10,000,000 Cycles to Failure, N f 62
Summary - Head failures! Head failures caused by internal defects notably porosity and high concentration areas of porosity.! Thermal conditions during solidification may cause one weldment to be good and another to be bad? 63
Summary - Web-base failures! Web and base failures aggravated by severe external geometry and cold laps.! Thermal conditions during solidification play a role in web-base fatigue problems?! Fatigue life can be increased by modifications of external weld geometry. 64