Use of Frequency Control to Optimize Induction Axle Scan Hardening
1. International Union for Association of the Polish
Electricity Applications Electrical Engineers
MULTI-FREQUENCY AXLE
SCAN HARDENING
Prof. Valentin Nemkov
Eng. Robert Goldstein
Presented at the UIE Congress in Krakow, May 19-21, 2008
WWW.FLUXTROL.COM
2. Overview
• Background on Axle Hardening
• Current Coil Designs for Scan Hardening
• Axle Scan Hardening Design
– Traditional Hardening
– Optimized Single-frequency Hardening
– Optimized Multi-frequency Hardening
• Conclusions
3. Background on Axle Hardening
• Axles were one of the first parts
hardened by induction
• Original process was scan
hardening
• With improved power supplies
and increased production
demands, single shot method
emerged. Both technologies
used now, with scanning Photo Courtesy of
method as prevailing for long AjaxTOCCO Magnethermic
axles
4. Background on Axle Hardening ctd.
• Typical light truck axles and car hubs
5. Background on Axle Hardening ctd.
• Main difficulty is to achieve a combination of proper
fillet hardening with high scanning speed and proper
pattern in transient area
• Optimization of two-turn hardening coil at 1 and 3
kHz was presented at HES-07 Symposium in Padua
• Additional improvement due to frequency variation is
the new topic in current presentation
6. Current Scan Hardening Inductors
• Two major inductor types
used for scan hardening:
– Single turn coils
• Machined MIQ
• Formed tubing (1)
– Two turn profiled coils (2)
• Single turn coils provide better
pattern control in fillet area 1
2
• Two turn coils provide higher
scanning speed
7. Geometry for Simulation:
2-turn Standard Coil
• Deep case hardening –
Box for motion
10–14 mm
• Shaft diameter 48 mm
• Frequency used - 1 and 3
kHz
• Short term temperature
surface Tmax = 1100 C
• Available power 100 kW
per spindle
8. Standard 2-turn Coil: Temperature
Distribution at 1kHz
Color Shade Results Color Shade Results
Quantity : Temperature Deg. Celsius Quantity : Temperature Deg. Celsius
Time (s.) : 11 Pos (mm): 25 Phase (Deg): 0 Tim (s.) : 23 Pos (m ): 138.999 Phase (Deg): 0
e m
Scale / Color Scale / Color
20.10919 / 83.03406
20.00348 / 86.64462
83.03406 / 145.95891
86.64462 / 153.28577 145.95891 / 208.88379
153.28577 / 219.92691 208.88379 / 271.80862
219.92691 / 286.56805 271.80862 / 334.73349
286.56805 / 353.2092 334.73349 / 397.65836
397.65836 / 460.58319
353.2092 / 419.85034 460.58319 / 523.50806
419.85034 / 486.49149 523.50806 / 586.43292
486.49149 / 553.13269 586.43292 / 649.35779
553.13269 / 619.7738 649.35779 / 712.28265
619.7738 / 686.41492 712.28265 / 775.20752
775.20752 / 838.13239
686.41492 / 753.05603 838.13239 / 901.05725
753.05603 / 819.69727 901.05725 / 963.98212
819.69727 / 886.33838 963.98212 / 1.02691E3
886.33838 / 952.97949
952.97949 / 1.01962E3
1.01962E3 / 1.08626E3
9.5 mm/sec scan –
limited by power
Drawbacks of this coil:
• Marginal pattern control
After 10 sec dwell + “Jump” • Difficult setup
+ 1sec scan at 20mm/sec • Low efficiency and scan speed
9. Steps to Improve the Inductor
• Change the profile of both turns
• Apply Fluxtrol A magnetic flux controller to
the lower turn to improve heating of the
fillet area
• Bring the upper turn closer to the part to
increase preheating and facilitate faster
scanning
11. 1 kHz Summary
• At 1 kHz, available power was the factor that
limited the scan speed
• Dwell time required for the standard 2-turn coil
was longer in order to achieve sufficient depth in
the fillet without overheating the area above
• Special process start-up was required for the
standard 2-turn coil
• For the same inductor power, the new optimized
inductor could scan over 15% faster
13. New Approach
New machines with variable frequency may be built
and we can:
• Use higher frequency (3 kHz) for initial heating
(dwelling)
• Use lower frequency (1 kHz) for scanning
• Optimize power, speed and frequency switch time
for transition from dwelling to regular scanning
• Design the coil for these conditions:
– Reduce bottom face width
– Bring the upper turn closer to the axle and make it shorter to
increase preheating and scanning speed
User-guided computer simulation had been used for
optimal design (Flux 2D program)
14. New 2-turn Coil: Temperature
Distribution after Preheating at 3kHz
Color Shade Results
Quantity : Temperature Deg. Celsius
Tim (s.) : 4.5 Pos (m ): 0 Phase (Deg): 0
e m
Scale / Color
20.00003 / 83.80288
83.80288 / 147.60571
147.60571 / 211.40857
211.40857 / 275.21143
275.21143 / 339.01428
339.01428 / 402.81711
402.81711 / 466.6191
466.6191 / 530.42279
530.42279 / 594.22565
594.22565 / 658.0285
658.0285 / 721.83136
721.83136 / 785.63422
785.63422 / 849.43707
849.43707 / 913.23993
913.23993 / 977.04272
977.04272 / 1.04085E3
After 4.5 s dwell @ 3 kHz
15. New 2-turn Coil: Temperature
Distribution after Dwell
Color Shade Results
Quantity : Temperature Deg. Celsius
Tim (s.) : 8 Pos (m ): 0 Phase (Deg): 0
e m
Scale / Color
20.00021 / 84.95692
84.95692 / 149.91364
149.91364 / 214.87035
214.87035 / 279.82703
279.82703 / 344.78375
344.78375 / 409.74048
409.74048 / 474.6972
474.6972 / 539.65387
539.65387 / 604.6106
604.6106 / 669.56732
669.56732 / 734.52399
734.52399 / 799.48071
799.48071 / 864.43744
864.43744 / 929.39417
929.39417 / 994.35089
994.35089 / 1.05931E3
After dwell: 4.5 s @ 3 kHz + 3.5 s @ 1 kHz
16. New 2-turn Coil: Temperature
Distribution during Regular Scanning
Color Shade Results
Quantity : Temperature Deg. Celsius
Time (s.) : 20 Pos (mm): 134.499 Phase (Deg): 0
Scale / Color
20.04147 / 81.04058
81.04058 / 142.03967
142.03967 / 203.03879
203.03879 / 264.0379
264.0379 / 325.03702
325.03702 / 386.0361
386.0361 / 447.03522
447.03522 / 508.03433
508.03433 / 569.03345
569.03345 / 630.03259
630.03259 / 691.03168
691.03168 / 752.03076
752.03076 / 813.02991
813.02991 / 874.02899
874.02899 / 935.02814
935.02814 / 996.02722
Video
11.6 mm/sec scan @ 1 kHz – limited by power supply
17. Comparison of Results
Freq-cy, Coil Dwell Scan Scan Coil
kHz Time Start-up Speed, Current,
sec mm/sec kA
Stand. 10 Special 9.5* 13.5
1.0 Optim. 8 Normal 11* 13.5
Stand. 12 Normal 6.5** 7.0
3.0 Optim. 12 Normal 9** 8.0
3.0+1.0 Optim. 8 Easy 11.6* 13.0
* Power limited ** Temperature limited
18. Conclusions
• Significant improvements of axle scan hardening
are possible by optimization of coil design and
frequency selection and combination
• Optimal coil design at 1 kHz gives the following
improvements:
– More reliable results for the fillet area hardening
– Higher scan speed due to better efficiency
– Lower current demand and therefore reduced losses in
and size of the supplying circuitry
• Preheating at 3kHz makes possible to meet the
most challenging specs for the fillet area
treatment in combination with high scan speed
• Case-dependable optimal coil and process design
may be made by user-guided computer simulation