ICHS 4 2011 September San Francisco Study of

ICHS 4 – 2011 September, San Francisco Study of potential leakage on several stressed fittings for hydrogen pressures up to 700 bar D Houssin-Agbomson 1, D Jamois 2, Ch Proust 2, J Daubech 2, S Ruban 1, S Jallais 1 sidonie. ruban@airliquide. com 1 Air Liquide, Centre de Recherche Claude-Delorme, 78350 Les Loges-en-Josas, France 2 INERIS, 60550 Verneuil-en-Halatte, France

Content I. III. IV. V. Context Tested configurations Experimental device Results Conclusions and perspectives H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 2

I. Context H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 3

Study description Context Several H 2 applications developed in "Hydrogène Horizon Energie" (H 2 E) project Define "predictable" and "accidental" leaks and be able to assess consequences in order to: • Design natural ventilation to avoid any dangerous accumulation of “predictable” leaks • Prove that an "important" leak (potentially inducing an explosion) has a very low probability Objectives Quantify "small" leaks which are probable Contribute to assess the probability of "important" leaks Means H 2 systems analysis by fault tree • Scope: fittings • Identification of leak causes and means of mitigation Tests for leaks • INERIS test bench: H 2, up to 700 bar, on double ring tube fittings and threaded ring fittings H potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 2 4

II. Tested configurations Studied fittings Stress configurations H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 5

Studied fittings Based on a preliminary study of potentially leaking elements, four kinds of commercial fittings usually employed on H 2 energy-based systems were chosen and tested Threaded ring Double ring Nut Body 6 -mm double ring tube fitting Nut Medium-pressure double ring tube fitting Threaded ring tube fitting External tube size Maximum working pressure recommended by the constructor Swagelok® double ring tube fitting (316 SS) 6 mm 420 bar Rotarex® double ring tube fitting (316 SS) 6 mm 420 bar 1/2’’ 1035 bar 9/16’’ 1500 bar Fittings Swagelok® medium-pressure double ring tube fitting (316 SS) Maximator® threaded ring tube fitting (316 SS) H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 6

Stress configurations (1/2) Existing literature Few results in open literature Conditions already studied(1) • • Moderate pressures up to 35 bar Mainly 1/4'' pipe fittings Main result • leak flow rate of 800 cm 3. s-1 at 20 bar for tightening by hand of the fitting Experimental conditions of the present study H 2, up to 700 bar Test matrix considered as maximal stress in real-life conditions of a hydrogen-based application Fatigue conditions tested beyond constructor tests and recommendations for users Stress configurations Two kinds of stresses • stresses applied at atmospheric pressure ( « static tests » ): assembly/dismantling cycles / tightening / "thermal effects" • stresses applied when the fitting were under pressure (dynamic tests) : rotation / flexion / traction (1) Gentilhomme, O. , Tkatschenko, I. , Joncquet, G. , Anselmet, F. , Outcomes from the French National Project DRIVE, Proceedings of the Seventh International Conference and Trade Fair on Hydrogen and Fuel Cell Technologies, 22 -23 October 2008, Hamburg, Germany H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 7

Stress configurations (2/2) Details of the tested configurations Stresses Static stresses Tested stress level considered as maximum in situation Around 20 cycles, for 6 mm double ring tube fitting 100 cycles < 20 cycles, for high pressure fittings (double and threaded ring tube fittings) 50 cycles 1 1/4 turn, for 6 mm double ring tube fitting 1 turn 1 1/4 turn, for 1/2’’ double ring tube fitting 3/4 turn and 1 turn 80 N. m with silicone grease on conical sealing surfaces, for 9/16’’ threaded ring tube fitting 60 N. m with silicone grease 1 1/4 turn, for 6 mm double ring tube fitting 1 1/2 turn +60 to -20°C Crimping at 100°C and leakage test at 30°C 0 turn 1/8 turn compared to initial position Flexion 0° 10° from fitting initial axis Traction 0 N 320 N in fitting axis direction Assemblydismantling cycles Undertightening Overtightening Thermal effects Dynamic stresses Normal conditions of use Counterclockwise rotation H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 8

III. Experimental device Main test bench "Dynamic stress" module H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 9

Main test bench A simple but accurate AND safe test bench… Characterization of H 2 high pressure leaks Gas for tests: H 2 Pressure range: up to 700 bar Sensitivity: flow rate higher than 10 -2 cm 3. s-1 Main equipment A H 2 compressor High pressure reservoirs Pneumatic valves Accurate pressure transducers Thermocouples Climatic enclosure Scheme of the main test bench Climatic enclosure with device for tests H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 10

Module for "dynamic stress" Characteristics Module set-up in order to apply stress on pressurized fittings thanks to a pneumatic jack Pressure range: up to 700 bar Tested "dynamic" stresses • • • Flexion Rotation Traction Only applied on 6 -mm tube fitting "Dynamic" module Main equipment: a pneumatic jack (up to 320 N) Inserted inside the climatic enclosure of the main test bench H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 11

IV. Results H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 12

Results of the study Only two test configurations showed small leakages leak flow rates experimentally determined appear close to the threshold sensitivity value of 10 -2 cm 3. s-1 Fitting Swagelok mediumpressure double ring tube fitting Maximator threaded ring tube fitting External tube size Tested stress Pressure conditions Measured leakage rate 1/2’’ After 50 assembly-dismantling cycles 700 bar 0. 05 cm 3. s-1 9/16’’ Under-tightening with silicone grease: 60 N. m instead of 80 N. m 700 bar 4. 3 cm 3. s-1 These values are very low compared to the flow rates usually considered for risk analyses of potentially hazardous systems (i. e. around 900 cm 3. s-1 for the "0. 1 mm-200 bar" hypothesis), It would be detected by normal control procedures particularly for these types of stresses Leak visualisation in water The other stress scenarii did not exhibit detectable leakage given experimental installation sensitivity H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 13

V. Conclusions and perspectives H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 14

Conclusions A specific experimental installation was designed and set-up by INERIS in order to study potential leakages on fittings commonly used in H 2 existing systems This test bench enables – – to accurately quantify, when existing, leakages with a flow rate above 10 -2 cm 3. s-1 to carry out the tests with hydrogen in safe conditions to reach test pressures up to 700 bar to apply stresses on fittings under H 2 pressure (i. e. “dynamic stresses”) Very few scenarios gave rise to quantifiable leaks (given test bench detectable thresholds) Only two stress configurations showed measurable but low leak flow rates compared to values usually employed for risk assessment – 0. 05 cm 3. s-1 for the Swagelok® 1/2’’ medium-pressure double ring tube fitting after 50 cycles of assembly-dismantling – 4. 3 cm 3. s-1 for a 25% under-tightening of the Maximator® 9/16’’ threaded ring tube fitting A good reliability of the fittings used in H 2 -energy based systems is shown For hydrogen systems considered for the H 2 E project, the present leakage rates are much too low to produce an accumulation of a flammable atmosphere; a leakage rate of a few tens of cm 3. s-1 would be required A different conclusion may be given for other types of confinement H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 15

Perspectives This study is a first step and several other tests can be considered – Determining the critical stress levels which give rise to significant leaks – Testing other stresses (e. g. vibration effects) – Combining different types of stress (e. g. under-tightening and traction) – Testing other potentially leaking elements of H 2 -based systems H 2 potential leakage on stressed fittings – ICHS 4, 2011, San Francisco 16

ICHS 4 – 2011 September, San Francisco Study of potential leakage on several stressed fittings for hydrogen pressures up to 700 bar Thanks for your attention sidonie. ruban@airliquide. com