NuMI horn stripline failure, analysis, and recovery or The Case of the Cracked Stripline P. Hurh* Fermilab April *On behalf of Target Systems Department Engineering and Operations Team
The setting In June, 2015 the NuMI target facility was turned off due to an abnormal electrical pulse response in the NuMI Horn circuit First horn 1 (PH1-04) designed for NOvA era (700 kW primary proton beam power) Previous horns were designed for MINOS era (400 kW) PH1-04 accumulated ~27 million pulses out of a desired 50 million PH1-04 never operated above ~450 kW April P. Hurh | High Power Targetry Workshop 20162
The crime Crack in outer conductor on the underside (looking from bottom up) April P. Hurh | High Power Targetry Workshop Appears to have been initiated on inside radius or clamp bolt hole
Modus Operandi Stripline Flags were moved outward to lessen beam heating and enhance convective air cooling by target chase air flow Results in longer unclamped distance on lower stripline Fatigue failure from magnetic forces? April P. Hurh | High Power Targetry Workshop kW Design700 kW Design
The investigation Vibration measurement and analysis consultant brought in on the case April P. Hurh | High Power Targetry Workshop Many, many tests conducted including “live fire” with test power supply and “mallet blow” No striplines were harmed during interrogation (& no water-boarding)
Culprit identified? April P. Hurh | High Power Targetry Workshop lightly damped mode at 424 Hz (takes a long time to decay) highly damped mode at 788 Hz (decays quickly)
Crime reenactment (424 Hz) April P. Hurh | High Power Targetry Workshop 20167
The sting April P. Hurh | High Power Targetry Workshop Clamp large unsupported section of stripline to reduce or eliminate 424 Hz mode, 6” up from restraint centerline (point 7) Clamp large unsupported section of stripline to reduce or eliminate 424 Hz mode, 6” up from restraint centerline (point 7)
The culprit apprehended April P. Hurh | High Power Targetry Workshop Clamp restraint kills local 424 Hz normal mode Clamp restraint kills local 424 Hz normal mode
The confession (well, sort of) Failure mode appears to be very high cycle fatigue of aluminum Longer, unclamped length resulted in stress concentrations at the clamp location Lightly damped mode likely contributed 10 to 100 times more cycles than the number of actual pulses Vibration testing of the 400 kW design showed much less active participation with relatively heavy damping April P. Hurh | High Power Targetry Workshop kW design400 kW design
Prevention The failure occurred about 1 month before a scheduled 2 month shutdown –Not enough time to diagnose, re-design, and fabricate a more robust stripline package –New NOvA horn 1 (700 kW design) had same vulnerability –Spare MINOS horn 1 (400 kW design) was available Decision was made to qualify the 400 kW design for 700 kW –Modifications to increase cooling –Re-calculation of thermal performance utilizing realistic convective heat transfer calculations Long-term solution; new stripline package for 700 kW horn based upon 400 kW stripline package design April P. Hurh | High Power Targetry Workshop
400 kW design at 700 kW April P. Hurh | High Power Targetry Workshop Analyses assumed maximum HTC at stripline of h= 5 W/m 2 K Analyses assumed maximum HTC at stripline of h= 5 W/m 2 K 400 kW design with no modification 400 kW design with improved water cooling collar Max operating temperature for Al stripline less than ~130 C to avoid softening
Target Chase mock-up for HTC measurements April P. Hurh | High Power Targetry Workshop PH1-03 Spare Horn PH1-03 Spare Horn Current mockup has flowrate of ~ 10,000 cfm Beamline flowrate measurements were ~15,000 cfm Current mockup has flowrate of ~ 10,000 cfm Beamline flowrate measurements were ~15,000 cfm
HTC Measurements April P. Hurh | High Power Targetry Workshop Heated Aluminum coupons Thermally insulated mounts Measured heater current and voltage Measured coupon temperature Benchmarked to be within 25% accuracy Many cases, including with “air scoop” diverter to increase flow to inner flags Results indicated realistic HTC for inner flags of: 26 W/m 2 /K without diverter 36 W/m 2 /K with diverter Results indicated realistic HTC for inner flags of: 26 W/m 2 /K without diverter 36 W/m 2 /K with diverter
Super-scooper April P. Hurh | High Power Targetry Workshop With air diverter, predicted temperatures are under 100 C Retro-fitted PH1-03 was installed in September and has been running at up to 580 kW primary beam power since.
Additional improvements For the next 700 kW design spare horn 1, additional improvements are being made now –Larger radius bend to increase material at stress concentration –Aluminum alloy 6013 for flag regions higher fatigue and fracture toughness than 6101 –Extruding flags through radius Achieves better grain alignment transverse to crack propagation direction April P. Hurh | High Power Targetry Workshop Piece 6013-T6 Forged Flags 1-Piece 6013-T6 Forged Flags 6101-T61 Conductor 6101-T61 Conductor Larger radius fillet Analysis indicates fatigue safety factor minimum of 2.2 for 100 million cycles
Case closed Pending fabrication and testing of new stripline design, –The case of the cracked stripline has been solved! Special thanks and acknowledgements to: –Kris Anderson, Dave Larson, Keith Anderson, Cory Crowley, Mike Geelhoed, Marcellus Parker, Kavin Ammigan, Abhishek Deshpande, Frank Schneider, Mike Andrews, Jim Hylen, Bob Zwaska, Frankie Kelly and the rest of the TSD operations team April P. Hurh | High Power Targetry Workshop