1. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 1 FORCE-BASED ASSESSMENT OF WELD GEOMETRY Coenraad Esveld Delft University of Technology Esveld Consulting Services Delft University of Technology Esveld Consulting Services

2. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 2 DAMAGE DUE TO POOR WELD GEOMETRY

3. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 3 EXISTING WELD GEOMETRY STANDARDS For example Versine: 0 < p < 0.3 mm For example Versine: 0 < p < 0.3 mm p < 0.3 mm Grind off top

4. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 4 The dynamic contact force as a function of the first time derivative: VELOCITY APPROACH

5. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 5 l QI ≤ 1: Accepted l QI > 1: Rejected l QI ≤ 1: Accepted l QI > 1: Rejected QUALITY INDICES (QI)

6. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 6 FORCE-BASED STANDARDS VelocityF Dyn Inclination 40 km/h 5 kN3.2 mrad 80 km/h15 kN2.4 mrad 140 km/h 35 kN 1.8 mrad 200 km/h65 kN0.9 mrad 300 km/h140 kN0.7 mrad 100 km/h50 kN1.4 mrad QI=1 Conventional HSL HH Implemented in RAILPROF Total force in principle 225 kN

7. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 7 NEW VERSUS OLD NORM Velocity Versine [mm] Inclination [mrad] 40 km/h0.963.2 80 km/h0.722.4 140 km/h 0.54 1.8 200 km/h0.270.9 300 km/h0.210.7 Old Norm0.301.0 For 80 km/h the new norm is 2.4 times more favorable than the old norm, provided short waves have been ground off.

8. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 8 LATERAL GEOMETRY STANDARDS VelocityVersine 40 km/h1.0 mm 80 km/h0.7 mm 140 km/h0.5 mm 200 km/h0.5 mm 300 km/h0.5 mm QI=1 Implemented in RAILPROF

9. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 9 ASSESSMENT OLD AND NEW ON PRORAIL RP002432 RP002945 RP002949 RP003125 Old norm: Rejected, New: OK Old norm: OK, New: Rejected Old norm: Rejected, New: Rejected

10. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 10 SELECTION ON PRORAIL 1.8 mrad (140 km/h) Limit at 80 km/h 100 welds per group

11. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 11 OLD VERSUS NEW STANDARDS

12. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 12 Low correlation force and versine Low correlation force and versine High correlation force and QI High correlation force and QI CALCULATED DYNAMIC FORCES

13. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 13 Dynamic force linear with QI Dynamic force linear with train speed Dynamic force linear with train speed CALCULATED DYNAMIC FORCES

14. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 14 AXLE BOX ACCELERATIONS Dynamic amplification less than 2.2

15. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 15 Procedure: l Sample weld geometry with digital straightedge l Filter measured signal l Determine 1 st derivative (inclination) l Normalize with intervention value for line speed l Calculate QI. l QI < 1: OK, otherwise: grinding. Procedure: l Sample weld geometry with digital straightedge l Filter measured signal l Determine 1 st derivative (inclination) l Normalize with intervention value for line speed l Calculate QI. l QI < 1: OK, otherwise: grinding. PRACTICAL IMPLEMENTATION

16. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 16 PRACTICAL IMPLEMENTATION

17. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 17 PDA SCREEN V = 140 km/h QI = 1.06 V = 140 km/h QI = 1.06 QI uniquely shows where to grind

18. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 18 DESKTOP SOFTWARE  All data and graphs can be shown on a PC;  Results in pdf-format can directly be emailed to customer.  All data and graphs can be shown on a PC;  Results in pdf-format can directly be emailed to customer.

19. ECS Weld Geometry Standards www.esveld.com www.rail.tudelft.nl 19 CONCLUSIONS 1.Theory based on first derivative works fine in practice; 2.Steel straightedge is absolutely inadequate; 3.Instead electronic straightedges with QI (RAILPROF); 4.High correlation of force and QI, low correlation with versine; 5.With RAILPROF QI measurement:  You see what you do;  Higher quality;  Less rejections provided short waves are ground properly (also negative welds allowed);  Extension of life cycle. 1.Theory based on first derivative works fine in practice; 2.Steel straightedge is absolutely inadequate; 3.Instead electronic straightedges with QI (RAILPROF); 4.High correlation of force and QI, low correlation with versine; 5.With RAILPROF QI measurement:  You see what you do;  Higher quality;  Less rejections provided short waves are ground properly (also negative welds allowed);  Extension of life cycle.