Course of explosions in Propene/O 2 /N 2 and Methane/O 2 /N 2 mixtures in vessels of 20 l to 2500 l and load on the walls

Transcription

1 Course of explosions in Propene/O /N and Methane/O /N mixtures in vessels of l to 5 l and load on the walls Dr. Hans-Peter Schildberg BASF-AG GCT/S - L511 D-6756 Ludwigshafen hans-peter.schildberg@basf.com Tel: H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 1

2 Outline of talk Introduction Explosion characteristics Course of explosion in l sphere Pressure load P det on the wall of the vessel - quantification of P det for four scenarios occuring in a pipe (no precompression) - quantification of P det for four scenarios occuring in a pipe (with precompression) - quantification of P det for Propene/O /N in vessel at P initial = 5 bar abs, T initial = C Comparison: detonative regimes in l vessels 1, 5, 5 l vessels Predetonation distances of Propene/O in tubes Did ignition source directly trigger a spherical detonation? H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton

3 Information available in literature on explosive gas mixtures of type combustible/o /N Explosion characteristics (explosion limits, p ex, (dp/dt) ex ) combustible / air /N explosion limits p ex, (dp/dt) ex p initial 1 bar a p initial > 1 bar a + + explosion limits o -- combustible / air /O p ex, (dp/dt) ex Course of explosion (deflagrative, detonative, heat explosion) combustible / air /N combustible / air /O pipe (L/D > 1) vessel (L/D 1) pipe (L/D > 1) vessel (L/D 1) o - 1 p initial 1 bar a p initial > 1 bar a combustible [vol.-%] explosive range of combustible/air/o N [vol.-%] explosive range of combustible/air/n O [vol.-%] 5 4 combustible/air mixtures 3 1 => Much work needed to better understand combustible/air/o mixtures H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 3

4 Motivation for this work Trend in partial oxidation reactions: - replacing air or diluted air as oxidant by oxygen enriched air or by pure oxygen Advantages: - increase of yield in space and time - lower operating pressures Critical Issue when dealing with combustible/air/o -mixtures: Process Safety - vessel-like geometry - explosion characteristics - course of explosion (deflagration, detonation) Focus of this work - bubble column geometry (explosive gaseous bubbles dispersed in liquid) - mechanisms of flame propagation - course of explosion (deflagration, detonation) H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 4

5 Gaseous mixtures investigated Propene / O / N, very detailed exp. investigations Methane / O / N Hydrogen / O Cyclohexane / NO only for one value of P initial and T initial Cyclohexane / N O Cyclohexane / NO H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 5

6 Introduction Explosion characteristics Course of explosion in l sphere Pressure load P det on the wall of the vessel - quantification of P det for four scenarios occuring in a pipe (no precompression) - quantification of P det for four scenarios occuring in a pipe (with precompression) - quantification of P det for Propene/O /N in vessel at P initial = 5 bar abs, T initial = C Comparison: detonative regimes in l vessels 1, 5, 5 l vessels Predetonation distances of Propene/O in tubes Did ignition source directly trigger a spherical detonation? H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 6

7 Experimental setup Test vessel: l sphere l sphere, P max =6 bar, T max =5 C Ignition source: exploding wire according to EN1839, Bomb method, section (arc discharge, ca. A over ca. 3 ms) Ignition energy: ca. J Mixture preparation: partial pressure method, ideal gas characteristics assumed H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 7

8 References for complete documentation of explosion characteristics H.P. Schildberg in deliverables no. 8 and no. 13 of EU-Research Project SAFEKINEX, Contract No. EVG1-CT--7; download possible from Explosion characteristics of Propene/O/N at 1, 5, 1 and 3 bar abs, 5 C and C, Hans-Peter Schildberg in Process Safety and Industrial Explosion Protection, editor: ESMG European Safety Management Group e. V., Hamm, Germany (5), ISBN (proceedings of the International ESMG Symposium 5 in Nürnberg, Germany, 3 pages ) Determination of the Explosion Behaviour of Methane and Propene in Air or Oxygen at Standard and Elevated Conditions, A.A. Pekalski, H.-P. Schildberg, P.S.D. Smallegange, S.M. Lemkowitz, J.F. Zevenbergen, M. Braithwaite, H.J. Pasman, 11th International Symposium Loss Prevention and Safety Promotion in the Process Industries (4), Praha Congress Centre 31st May-3rd June 4; Thematic Section B, page (ISBN ) Determination of the Explosion Behaviour of Methane and Propene in Air or Oxygen at Standard and Elevated Conditions, A.A. Pekalski, H.-P. Schildberg, P.S.D. Smallegange, S.M. Lemkowitz, J.F. Zevenbergen, M. Braithwaite, H.J. Pasman, Process Safety and Environmental Protection, Volume 83, issue B5, (5) H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 8

9 Introduction Explosion characteristics Course of explosion in l sphere Pressure load P det on the wall of the vessel - quantification of P det for four scenarios occuring in a pipe (no precompression) - quantification of P det for four scenarios occuring in a pipe (with precompression) - quantification of P det for Propene/O /N in vessel at P initial = 5 bar abs, T initial = C Comparison: detonative regimes in l vessels 1, 5, 5 l vessels Predetonation distances of Propene/O in tubes Did ignition source directly trigger a spherical detonation? H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 9

10 Questions to be raised when considering explosions of combustible/air/o mixtures in vessel-like geometry What compositions will undergo a transition from Deflagration to Detonation (DDT) after ignition with a thermal ignition source? What is the largest conceivable pressure load on the vessel wall? How does the volume of the vessel affect the detonative range? How about the obscure effects reported mostly only in non-public communication? (Vessels believed to be explosion pressure proof did sometimes rupture, but when repeating the experiment at same initial pressure with even the worst case mixture (i.e. what one believed to be worst case), a mechanically identical vessel did withstand the load). Will combustible/air mixtures remain deflagrative for higher values of T initial and P initial? H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 1

11 Literature data on deflagrative and potentially detonative explosion regime of n-butane/o /N at 1 bar abs, C N [Vol.-%] n-butane [Vol.-%] 9 range of mixtures, where transition from deflagrative to detonative explosion is possible stoichiometric ratio n-butane : O = 1 : range of deflagrative explosion 6 O [Vol.-%] 5 binary mixture n-butane/air 4 Characteristic values of the explosive range [1]: Lower explosion limit in O : 1.4 vol.-% Upper explosion limit in O : 51 vol.-% Lower explosion limit in air: 1.4 vol.-% Upper explosion limit in air: 9.4 vol.-% Limiting oxygen concentration: 9.4 vol.-% Characteristic values of the potentially detonative range [,3]: Lower detonation limit in O :.5 vol.-% Upper detonation limit in O : 38 vol.-% Lower detonation limit in air: 1.98 vol.-% Upper detonation limit in air: 6.18 vol.-% [1] E. Brandes and W. Möller, Sicherheitstechnische Kenngrößen - Band 1: Brennbare Flüssigkeiten und Gase, Wirtschaftsverlag NW, Bremerhaven (3), ISBN: ; UEL in O is an estimate, basis is measured value of 56 Vol-% at C, ZET-report no N [] M. A. Nettleton, Fire Prev. Sci. and Tech. No 3, p.9 (198) [3] M. A. Nettleton in Gaseous Detonations, p. 76, Chapman and Hall Ltd, New York (1987) 3 1 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 11

12 What does potentially detonative regime mean? If a mixture lies outside the potentially detonative regime, an explosion will under all circumstances (e.g. type of ignition source, geometry) only be deflagrative If the mixture lies inside the potentially detonative regime and a detonative ignition source triggers an explosion, the flame front will propagate as detonation right from the beginning If the mixture lies inside the potentially detonative regime and a thermal * ignition source triggers an explosion, the flame front will initially propagate as deflagration and the transition to detonation (DDT) will not necessarily occur! There are further conditions that have to be fullfilled to enable the DDT (e. g. geometry of the containment must support flame acceleration). However, in practical applications it is mostly hard to assess to which degree these conditions are satisfied. There are absolutely no simulation tools (e.g. reactive CFD codes) available which could predict the DDT in vessels at least to some rudimentary extent. *: in process plants almost all ignition sources are thermal. The initial stage of the explosion they can trigger is always deflagrative H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 1

13 Literature data on deflagrative and potentially detonative explosion regime of Propen/O /N at 1 bar abs, C N [Vol.-%] Propene [Vol.-%] 9 8 range of mixtures, where transition from deflagrative to detonative explosion is possible stoichiometric ratio Propene : O = 1 : 4.5 Characteristic values of the explosive range [1,]: Lower explosion limit in O : vol.-% Upper explosion limit in O : 58 vol.-% Lower explosion limit in air: vol.-% Upper explosion limit in air: 11 vol.-% Limiting oxygen concentration: 11.5 vol.-% Characteristic values of the potentially detonative range [3]: Lower detonation limit in O :.5 vol.-% Upper detonation limit in O : 5 vol.-% Lower detonation limit in air: 3.55 vol.-% Upper detonation limit in air: 1.4 vol.-% [1] E. Brandes and W. Möller, Sicherheitstechnische Kenngrößen - Band 1: Brennbare Flüssigkeiten und Gase, Wirtschaftsverlag NW, Bremerhaven (3), ISBN: ; [] H.P. Schildberg in Process Safety and Industrial Explosion Protection, editor: ESMG European Safety Management Group e. V., Hamm, Germany (5), ISBN (proceedings of the International ESMG Symposium 5 in Nürnberg, Germany, 3 pages ) [3] M. A. Nettleton in Gaseous Detonations, p. 76, Chapman and Hall Ltd, New York (1987) H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 13 7 range of deflagrative explosion 6 O [Vol.-%] 5 binary mixture Propene/air 4 3 1

14 Literature data on deflagrative and potentially detonative explosion regime of H /O /N at 1 bar abs, C N [Vol.-%] range of deflagrative explosion H [Vol.-%] 9 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton range of mixtures, where transition from deflagrative to detonative explosion is possible stoichiometric ratio H :O = :1 O [Vol.-%] binary mixture H /air Characteristic values of the explosive range [1,]: Lower explosion limit in O : 4 vol.-% Upper explosion limit in O : 95. vol.-% Lower explosion limit in air: 4 vol.-% Upper explosion limit in air: 77 vol.-% Limiting oxygen concentration: 4.3 vol.-% Characteristic values of the potentially detonative range [3]: Lower detonation limit in O : 15 vol.-% Upper detonation limit in O : 9 vol.-% Lower detonation limit in air: 18.3 vol.-% Upper detonation limit in air: 58.9 vol.-% [1] E. Brandes and W. Möller, Sicherheitstechnische Kenngrößen - Band 1: Brennbare Flüssigkeiten und Gase, Wirtschaftsverlag NW, Bremerhaven (3), ISBN: [] V. Schröder in Process Safety and Industrial Explosion Protection, editor: ESMG European Safety Management Group e. V., Hamm, Germany (5), ISBN (proceedings of the International ESMG Symposium 5 in Nürnberg, Germany) [3] M. A. Nettleton in Gaseous Detonations, p. 76, Chapman and Hall Ltd, New York (1987) 1

15 Experimental setup to determine the potentially detonative range of explosive gas mixtures Booster-mixture, Acetylene/O, stoichiometric with respect to formation of CO and H O. It transits to detonation after a few millimeters and acts upon the mixture to be tested with a stronger detonation front than the one that occurs in the mixture when it undergoes a detontive explosion mixture to be tested for potential detonability (i.e. ability to propagate a detonation which was once initiated either by an external source or by an initially deflagrative explosion in the mixture to be tested, which succeeded in running up to a detonation) ca. 15*pipe diameters location of thermal ignition source for booster mixture thin membrane or nothing (after injecting the mixture to be tested the booster mixture is injected close to the ignition source and propagates almost as plug flow into the pipe) Signals of pressure and/or flame sensors give evidence whether a stable detonation could be triggered in the mixture Note: Pipe must be long enough for to allow that the initial state of the overdriven detonation can pass over to a stable final state, which can be either a stable detonation or a deflagration in the mixture to be tested. H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 15

16 Different explosion regimes (deflagration, heat explosion, detonation) found for Propene/O /N in a l sphere at yellow range: possibly heat explosion Propene C 3 H 6 [Vol.-%] soot is formed here (43 vol.-% up to 69 vol.-%) 6 5 O [Vol.-%] range of detonative explosion N [Vol.-%] range of deflagrative explosion, 4 3 stoichiometric C3H O -> 3 CO + 3 H stoichiometric C3H6 + 3 O -> 3 CO + 3 HO stoichiometric C3H O -> 3 CO + 3 HO 1 propene/air-mixtures H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 16

17 Different explosion regimes (deflagration, heat explosion, detonation) of Propene/O /N in a l sphere at P initial = 1, 5 and 1 bar abs, 5 C 1 bar abs bar abs 9 8 yellow range: yellow range: 3 possibly heat 3 possibly heat 7 explosion explosion range of 6 deflagrative explosion, 6 1 bar abs, 5 C range of detonative explosion O [Vol.-%] stoichiometric C3H O -> 3 CO + 3 H stoichiometric C3H6 + 3 O -> 3 CO + 3 HO stoichiometric C3H O -> 3 CO + 3 HO N [Vol.-%] propene/air-mixtures Propene C 3 H 6 [Vol.-%] Propene C 3 H 6 [Vol.-%] soot is formed here (43 vol.-% up to 69 vol.-%) 6 5 O [Vol.-%] range of detonative explosion N [Vol.-%] range of deflagrative explosion, 4 3 stoichiometric C3H O -> 3 CO + 3 H stoichiometric C3H6 + 3 O -> 3 CO + 3 HO stoichiometric C3H O -> 3 CO + 3 HO 1 propene/air-mixtures 1 bar abs bar abs 9 yellow range: possibly heat 8 explosion range of deflagrative explosion, 3 1 bar abs, 5 C (Heat explosion range and detonative range were not determined. It might be that the detonative range even extends down to the air line) range of detonative explosion (boundary at propene conc. > 1 vol.-% only estimated) O [Vol.-%] stoichiometric ratio Propene : O = 1 : 4.5 N [Vol.-%] propene/air-mixtures Propene C 3 H 6 [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 17

18 Different explosion regimes (deflagration, heat explosion, detonation) of Propene/O /N in a l sphere at 1, 5 and 1 bar abs, C 1 bar abs bar abs 9 yellow range: 8 yellow range: possibly heat 3 possibly heat explosion 7 range of detonative explosion explosion O [Vol.-%] stoichiometric C3H O -> 3 CO + 3 H stoichiometric C3H6 + 3 O -> 3 CO + 3 HO stoichiometric C3H O -> 3 CO + 3 HO N [Vol.-%] range of deflagrative explosion, 5 bar abs, C Propene C 3 H 6 [Vol.-%] 1 propene/air-mixtures Propene C 3 H 6 [Vol.-%] soot is formed here (41 vol.-% up to 75 vol.-%) O [Vol.-%] 6 range of detonative 5 N [Vol.-%] 4 3 stoichiometric C3H O -> 3 CO + 3 H stoichiometric C3H6 + 3 O -> 3 CO + 3 HO explosion stoichiometric C3H O -> 3 CO + 3 HO 9 range of deflagrative explosion, 5 bar abs, C 1 propene/air-mixtures 1 bar abs bar abs 9 yellow range: 8 possibly heat 3 explosion 7 (Heat explosion range and 4 6 detonative range were not 5 5 determined. It might be 6 that the detonative range 4 7 range of even extends down to the deflagrative explosion, 3 1 bar abs, C 8 air line) Propene C 3 H 6 [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 18 O [Vol.-%] range of detonative N [Vol.-%] stoichiometric C3H O -> 3 CO + 3 H stoichiometric C3H6 + 3 O -> 3 CO + 3 HO explosion stoichiometric C3H O -> 3 CO + 3 HO propene/air-mixtures

19 Different explosion regimes (deflagration, heat explosion, detonation) of Methane/O /N in a l sphere at 1 and and C 1 bar abs, 5 C yellow range: yellow range: 8 mixture reacts sometimes possibly heat 3 detonative and sometimes possibly heat 3 7 as heat explosion 7 range of detonative explosion explosion explosion Methane CH 4 [Vol.-%] Methane CH 4 [Vol.-%] O [Vol.-%] stoichiometric CH O -> CO + HO stoichiometric CH4 + O -> CO + HO N [Vol.-%] soot formation starts at 55 vol.-% CH 4 O [Vol.-%] range of deflagrative explosion, 1 bar abs, 5 C stoichiometric CH4 +.5O -> CO + H methane/air-mixtures yellow range: possibly heat explosion H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton Methane CH 4 [Vol.-%] O [Vol.-%] range of detonative explosion 5 stoichiometric CH O -> CO + HO stoichiometric CH4 + O -> CO + HO N [Vol.-%] N [Vol.-%] soot formation starts at 55 vol.-% CH 4 stoichiometric CH O -> CO + HO stoichiometric CH4 + O -> CO + HO methane/air-mixtures soot formation starts at 55 vol.-% CH 4 range of deflagrative explosion, 5 bar abs, C range of deflagrative explosion, stoichiometric CH4 +.5O -> CO + H stoichiometric CH4 +.5O -> CO + H 5 bar abs, C methane/air-mixtures

20 Pressure-time traces of Propene/O at 5 bar, 5 C (1) Propene =.5 vol.-% O = vol.-%, 5 File: PROP76-.5%.DAT Propene = 3 vol.-% O = 97 vol.-%, 1 3 Propene = 4 vol.-% O = 96 vol.-%, File: PROP8A;Propene=3, O=97 Vol.%; 5bar abs; 5C.DAT File: PROP14A-4%.DAT N [Vol.-%] stoichiometric C 3 H O -> 3 CO + 3 H O Propene C 3 H 6 [Vol.-%] range of detonative explosion stoichiometric C 3 H O -> 3 CO + 3 H O 7 6 range of deflagrative explosion, stoichiometric C 3 H O -> 3 CO + 3 H propene/air-mixtures top center bottom O [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton

21 Pressure-time traces of Propene/O at 5 bar, 5 C () File: PROP7A;Propene=5, O=95 Vol.%; 5bar abs; 5C.DAT Propene = 5 vol.-% O = 95 vol.-%, File: PROP6A;Propene=6, O=94 Vol.%; 5bar abs; 5C.DAT Propene = 6 vol.-%, O = 94 vol.-%, Propene = 7 vol.-%, O = 93 vol.-%, File: PROP9A;Propene=7, O=93 Vol.%; 5bar abs; 5C.DAT N [Vol.-%] stoichiometric C 3 H O -> 3 CO + 3 H O Propene C 3 H 6 [Vol.-%] range of detonative explosion stoichiometric C 3 H O -> 3 CO + 3 H O 7 6 range of deflagrative explosion, stoichiometric C 3 H O -> 3 CO + 3 H propene/air-mixtures top center bottom O [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 1

22 Pressure-time traces of Propene/O at 5 bar, 5 C (3) File: PROP84A-8%.DAT Propene = 8 vol.-%, O = 9 vol.-%, N [Vol.-%] File: PROP143A-13%.DAT Propene = 13 vol.-%, O = 87 vol.-%, Propene = vol.-%, O = 81.8 vol.-%, 5 1 File:PROP13A-18.%.DAT stoichiometric C 3 H O -> 3 CO + 3 H O Propene C 3 H 6 [Vol.-%] range of detonative explosion stoichiometric C 3 H O -> 3 CO + 3 H O 7 6 range of deflagrative explosion, stoichiometric C 3 H O -> 3 CO + 3 H propene/air-mixtures top center bottom O [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton

23 Pressure-time traces of Propene/O at 5 bar, 5 C (4) Propene = 3 vol.-%, O = 77 vol.-%, Propene = 8 vol.-%, O = 7 vol.-%, File: PROP131A-3%.DAT File: Prope13a.dat, 8 vol.-% Propene = 33 vol.-% O = 67 vol.-%, File:PROP137A-33%.DAT N [Vol.-%] stoichiometric C 3 H O -> 3 CO + 3 H O Propene C 3 H 6 [Vol.-%] range of detonative explosion stoichiometric C 3 H O -> 3 CO + 3 H O 7 6 range of deflagrative explosion, stoichiometric C 3 H O -> 3 CO + 3 H propene/air-mixtures top center bottom O [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 3

24 Pressure-time traces of Propene/O at 5 bar, 5 C (5) Propene = 35.5 vol.-% O = 64.5 vol.-%, File:PROP1A;Propene=35.5, O=64.5 Vol.%; 5bar abs; 5C.DAT Propene = 38 vol.-% O = 6 vol.-%, File: PROP136A-38%.DAT Propene = 39 vol.-%, O = 61 vol.-%, N [Vol.-%] File: Prope11A;Propene=39, O=61 vol.-%, 5 bar abs, 5C stoichiometric C 3 H O -> 3 CO + 3 H O Propene C 3 H 6 [Vol.-%] range of detonative explosion stoichiometric C 3 H O -> 3 CO + 3 H O 7 6 range of deflagrative explosion, stoichiometric C 3 H O -> 3 CO + 3 H propene/air-mixtures O [Vol.-%] 5 top center bottom H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 4

25 Pressure-time traces of Propene/O at 5 bar, 5 C (6) File: PROP1A;Propene=4, O=6 Vol.%; 5bar abs; 5C.DAT Propene = 4 vol.-%, O = 6 vol.-%, File: PROP13A;Propene=41, O=59 Vol.%; 5bar abs; 5C.DAT Propene = 41 vol.-%, O = 6 vol.-%, Propene = 4 vol.-%, O = 58 vol.-%, File: PROP14A;Propene=4, O=58 Vol.%; 5bar abs; 5C.DAT N [Vol.-%] stoichiometric C 3 H O -> 3 CO + 3 H O Propene C 3 H 6 [Vol.-%] range of detonative explosion stoichiometric C 3 H O -> 3 CO + 3 H O 7 6 range of deflagrative explosion, stoichiometric C 3 H O -> 3 CO + 3 H propene/air-mixtures O [Vol.-%] 5 top center bottom H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 5

26 Pressure-time traces of Propene/O at 5 bar, 5 C (7) Propene = 43 vol.-%, O = 57 vol.-%, Propene C 3 H 6 [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 6 O [Vol.-%] range of detonative explosion N [Vol.-%] range of deflagrative explosion, top stoichiometric C3H O -> 3 CO + 3 H stoichiometric C 3 H6 + 3 O -> 3 CO + 3 H O stoichiometric C 3 H O -> 3 CO + 3 H O File: PROP139A-43%.DAT propene/air-mixtures

27 Pressure-time traces of stoich. Propene/O /N at 5 bar, 5 C (1) Propene = 5.5 vol.-% O = 4.75 vol.-%, N = File: PROP15A; Propen=5.5, O=5, N=69.5 Vol.%;5bar abs;5c.dat Propene = 6 vol.-% O = 7 vol.-%, N = 67 (first test) File: PROP8A-7%O(erster Test).DAT Propene = 6 vol.-% O = 7 vol.-%, N = 67 (second test) File: PROP9A-7%O(zweiter Test).DAT bottom center top N [Vol.-%] Propene C 3 H 6 [Vol.- stoichiometric C 3 H O -> 3 CO + 3 H O 8 range of detonativ explosion ran deflagrati 5 bar a stoichiometric C 3 H O -> 3 CO + 3 H O stoichiometric C 3 H O -> 3 CO + propene/air H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 7

28 Pressure-time traces of stoich. Propene/O /N at 5 bar, 5 C () Propene = 6.9 vol.-% O = 31 vol.-%, N = 6.1 File:PROP16A;Propen=6.9, O=31, N=6.1 Vol.%;5bar abs;5c.dat File: PROP7A-35%O.DAT Propene = 7.7 vol.-% O = 35 vol.-%, N = 57.3 (first test) Propene = 7.7 vol.-% O = 35 vol.-%, N = 57.3 (second test) File: PROP17A;Propen=7.8, O=35, N=57. Vol.%;5bar abs;5c.dat bottom center top N [Vol.-%] Propene C 3 H 6 [Vol.-% stoichiometric C 3 H O -> 3 CO + 3 H O range of detonative explosion range deflagrative 5 bar abs stoichiometric C 3 H O -> 3 CO + 3 H O stoichiometric C 3 H O -> 3 CO + 3 propene/air-m H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 8

29 Pressure-time traces of stoich. Propene/O /N at 5 bar, 5 C (3) Propene = 8.8 vol.-% O = 4 vol.-%, N = 51. File: PROP18A;Propen=8.9, O=4, N=51.1 Vol.%;5bar abs;5c.dat Propene = 13.3 vol.-% O = 6 vol.-%, N = 6.7 File: PROP16A-6%O.DAT Propene = vol.-%, O = 81.8 vol.-%, 5 1 File:PROP13A-18.%.DAT H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 9 1 bottom center 9 top N [Vol.-%] Propene C 3 H 6 [Vol.-% stoichiometric C 3 H O -> 3 CO + 3 H O range of detonative explosion range deflagrative 5 bar abs stoichiometric C 3 H O -> 3 CO + 3 H O stoichiometric C 3 H O -> 3 CO + 3 propene/air-m

30 Different explosion regimes (deflagration, heat explosion, detonation) of Methane/O /N in a l sphere at 1 and and C 1 bar abs, 5 C yellow range: yellow range: 8 mixture reacts sometimes possibly heat 3 detonative and sometimes possibly heat 3 7 as heat explosion 7 range of detonative explosion explosion explosion Methane CH 4 [Vol.-%] Methane CH 4 [Vol.-%] O [Vol.-%] stoichiometric CH O -> CO + HO stoichiometric CH4 + O -> CO + HO N [Vol.-%] soot formation starts at 55 vol.-% CH 4 O [Vol.-%] range of deflagrative explosion, 1 bar abs, 5 C stoichiometric CH4 +.5O -> CO + H methane/air-mixtures yellow range: possibly heat explosion H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton Methane CH 4 [Vol.-%] O [Vol.-%] range of detonative explosion 5 stoichiometric CH O -> CO + HO stoichiometric CH4 + O -> CO + HO N [Vol.-%] N [Vol.-%] soot formation starts at 55 vol.-% CH 4 stoichiometric CH O -> CO + HO stoichiometric CH4 + O -> CO + HO methane/air-mixtures soot formation starts at 55 vol.-% CH 4 range of deflagrative explosion, 5 bar abs, C range of deflagrative explosion, stoichiometric CH4 +.5O -> CO + H stoichiometric CH4 +.5O -> CO + H 5 bar abs, C methane/air-mixtures

31 O [Vol.-%] Pressure-time traces of Methane/O at 5 bar, 5 C (1) Methane = 7 vol.-%, O = 93 vol.-%, yellow range: possibly heat explosion stoichiometric CH + O -> CO + H O 4 N [Vol.-%] Methane = 1 vol.-%, O = 9 vol.-%, Methane CH [Vol.-%] 4 Methane = 1 vol.-%, O = 88 vol.-%, H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 31 top center bottom

32 Pressure-time traces of Methane/O at 5 bar, 5 C () Methane CH [Vol.-%] 4 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 3 top center bottom O [Vol.-%] yellow range: possibly heat explosion Methane = 14 vol.-%, O = 86 vol.-%, stoichiometric CH + O -> CO + H O 4 N [Vol.-%] 15 1 Methane = 15 vol.-%, O = 85 vol.-%, nd test: Methane = 15 vol.-%, O = 85 vol.-%,

33 Pressure-time traces of Methane/O at 5 bar, 5 C (3) Methane CH [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 33 top center bottom O [Vol.-%] yellow range: possibly heat explosion Methane = 17.5 vol.-%, O = 8.5 vol.-%, stoichiometric CH + O -> CO + H O 4 N [Vol.-%] Methane = vol.-%, O = 8 vol.-%, nd test: Methane = vol.-%, O = 8 vol.-%,

34 Pressure-time traces of Methane/O at 5 bar, 5 C (4) Methane CH [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 34 top center bottom O [Vol.-%] yellow range: possibly heat explosion Methane = 5 vol.-%, O = 75 vol.-%, stoichiometric CH + O -> CO + H O 4 N [Vol.-%] nd test: Methane = 5 vol.-%, O = 75 vol.-%, rd test: Methane = 5 vol.-%, O = 75 vol.-%,

35 Pressure-time traces of Methane/O at 5 bar, 5 C (5) Methane = 4 vol.-%, O = 6 vol.-%, Methane = 35 vol.-%, O = 65 vol.-%, yellow range: possibly heat explosion N [Vol.-%] 4 stoichiometric CH + O -> CO + H O O [Vol.-%] top center bottom Methane = 45 vol.-%, O = 55 vol.-%, Methane CH [Vol.-%] 4 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 35

36 Pressure-time traces of Methane/O at 5 bar, 5 C (6) Methane CH [Vol.-%] 4 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 36 top bottom O [Vol.-%] yellow range: possibly heat explosion stoichiometric CH + O -> CO + H O 4 N [Vol.-%] Methane = 5 vol.-%, O = 5 vol.-%, Methane = 55 vol.-%, O = 45 vol.-%,

37 Pressure-time traces of stoich. Methane/O /N at 5 bar, 5 C (1) 4 N [Vol.-%] 3 bottom top yellow range: possibly heat explosion stoichiometric CH + O -> CO + H O Methane CH [Vol.-%] 4 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 37 Methane = 1 vol.-%, O = vol.-%, N = 7 vol.-%, Methane = 15 vol.-%, O = 3 vol.-%, N = 55 vol.-%,

38 Pressure-time traces of stoich. Methane/O /N at 5 bar, 5 C () yellow range: possibly heat explosion N [Vol.-%] bottom top stoichiometric CH + O -> CO + H O Methane = vol.-%, O = 4 vol.-%, N = 4 vol.-%, Methane = 5 vol.-%, O = 5 vol.-%, N = 5 vol.-%, Methane CH [Vol.-%] 4 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 38

39 O [Vol.-%] Pressure-time traces of Methane/O at 5 bar, C (1) 3 Methane = 1 vol.-%, O = 9 vol.-%,, 5 bar abs, C yellow range: possibly heat explosion stoichiometric CH + O -> CO + H O 4 N [Vol.-%] Methane CH [Vol.-%] 4 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 39 Methane = 1.5 vol.-%, O = 87.5 vol.-%,, 5 bar abs, C top center bottom Methane = 14 vol.-%, O = 86 vol.-%,, 5 bar abs, C

40 O [Vol.-%] Pressure-time traces of Methane/O at 5 bar, C () yellow range: possibly heat explosion stoichiometric CH + O -> CO + H O 4 N [Vol.-%] Methane CH [Vol.-%] 4 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 4 top center bottom Methane = 15 vol.-%, O = 85 vol.-%,, 5 bar abs, C Methane = 17.5 vol.-%, O = 8.5 vol.-%,, 5 bar abs, C Methane = vol.-%, O = 8 vol.-%,, 5 bar abs, C

41 O [Vol.-%] Pressure-time traces of Methane/O at 5 bar, C (3) yellow range: possibly heat explosion stoichiometric CH + O -> CO + H O 4 N [Vol.-%] Methane CH [Vol.-%] 4 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 41 top center bottom Methane = 5 vol.-%, O = 75 vol.-%,, 5 bar abs, C Methane = 4 vol.-%, O = 6 vol.-%,, 5 bar abs, C Methane = 45 vol.-%, O = 55 vol.-%,, 5 bar abs, C

42 O [Vol.-%] Pressure-time traces of Methane/O at 5 bar, C (4) yellow range: possibly heat explosion stoichiometric CH + O -> CO + H O 4 N [Vol.-%] Methane CH [Vol.-%] 4 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 4 top center bottom Methane = 5 vol.-%, O = 5 vol.-%,, 5 bar abs, C Methane = 55 vol.-%, O = 45 vol.-%,, 5 bar abs, C 3 1 Methane = 6 vol.-%, O = 38 vol.-%,, 5 bar abs, C 4 6 8

43 O [ Pressure-time traces of stoich. Methane/O /N at 5 bar, C (1) H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 43 Methane = 15 vol.-%, O = 3 vol.-%,, N = 55 vol.-%, 5 bar abs, C yellow range: possibly heat explosion N [Vol.-%] bottom stoichiometric CH + O -> CO + H O Methane = vol.-%, O = 4 vol.-%,, N = 4 vol.-%, 5 bar abs, C center top Methane CH 4 [Vol.-%] 1 1 nd test: Methane = vol.-%, O = 4 vol.-%,, N = 4 vol.-%, 5 bar abs, C

44 Pressure-time traces of stoich. Methane/O /N at 5 bar, C () Methane = 5 vol.-%, O = 5 vol.-%,, Methane = 3 vol.-%, O = 6 vol.-%,, N = 5 vol.-%, 5 bar abs, C N = 1 vol.-%, 5 bar abs, C yellow range: possibly heat explosion center bottom top N [Vol.-%] 4 stoichiometric CH + O -> CO + H O O Note:This is not on the stoichiometric line!! Methane = 15 vol.-%, O = 55 vol.-%,, N = 3 vol.-%, 5 bar abs, C Methane CH [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 44

45 Summary of the differences in the pressure time diagrams between explosions in vessels that were triggered by a thermal ignition and ended up as deflagration, heat explosion or detonation presence of pressure pulse width of pressure pulse at half height [µs] height of pressure pulse divided by pressure in mixture just before its occurrence precompression ratio at the moment when pressure pulse occurs "oscillations" in reaction gases after occurrence of pressure pulse acoustic sound heard outside the vessel course of explosion deflagration heat explosion detonation no yes yes (not applicable) (not applicable) (not applicable) no oscillations anyway 1 - < 5 typically 4 to 7 (equals deflagration pressure ratio of the precompressed unreacted mixture) 5 1 (largest value equals p ex /p initial of the corresponding gas mixture) no oscillations typically 15 to 5 (due to too small sampling rate the true maximum was presumably mostly missed) 1 (largest value equals p ex /p initial of the corresponding gas mixture) massive oscillations (due to shock front bouncing backwards and forwards in reaction gases) nothing click prolonged whistle H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 45

46 Heat explosion tentative explanation of the course of events Before heat explosion: Reaction rate sufficiently slow to allow for pressure equilibrium over entire sphere Pressure as function of radial position in l sphere (p initial = 1 bar abs) just before heat explosion occurs in the precompressed, unreacted mixture (R = 17 cm) pressure/p initial hot reaction gases flame front precompressed unreacted mixture radial position r/r 1 At heat explosion: Immediate temperature rise and hence also pressure rise in the shell so far unreacted, but heated by adiabatic compression. The induction time must be of the order of.1 ms or even less, otherwise the flame front of the initial deflagration would have run through the thin shell of unreacted mixture before the heat explosion could develop. After heat explosion: reaction gases flow to the central part of the sphere and pressure equilibrium is attained H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 46

47 Temperatures obtainable by adiabatic compression p final /p initial T final [ C] resulting for different values of γ in case that T initial = 35 C = 38 K T final [ C] resulting for differrent values of γ in case that T initial = C = 473 K γ = 1,4 γ = 1,3 γ = 1, γ = 1,1 γ = 1,4 γ = 1,3 γ = 1, γ = 1,1 1,5 7,8 65, 56,5 46,6 58,1 46,4 33,1 17,8 1,5 88,4 7,7 55, 33,6 8, 57,9 3, ,6 13,9 96,9 67,3 374,4 336,5 95, 49, ,7 151,1 115,1 76,4 49,9 378,3 3,9 63,5 5 14,8 173,5 19,8 83,5 476,1 41,7 345,5 74,5 6 4,9 19,7 14, 89,5 516, 44, 364,6 83,7 7 64, 9,6 153, 94,6 551,7 468,1 381, 91,5 8 84,9 4,7 16,6 99,1 583,8 491,3 395,9 98,4 9 34, 38,4 171, 13,1 613,1 51,4 49, 34,6 1 31,7 51, 179,1 16,7 64, 531,7 41,3 31, ,5 73,5 193, 113,1 689,1 566,3 44,7 319, ,7 93,3 5, 118,5 73,4 596,7 461,3 38, 16 47,1 311, 15,9 13,3 771,5 63,9 477,8 335, ,4 37,1 5,6 17,6 87, 648,6 49,7 34,1 451,9 341,9 34,4 131,4 84, 671,3 56,3 348,1 Temperatures may well lie above the ignition temperatures the mixtures presumably have in the precompressed state! Formula for temperature increase due to adiabatic compression: ( γ 1) γ H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 47 T final = T initial p final p initial

48 Autoignition temperatures of Propene/O /N Propene [Vol.-%] O [Vol.- %] N [Vol.- %] Autoignition Temperature (AIT) [ C] AIT [ C] at P initial = 5 bar abs for Propene/O mixtures AIT [ C] at P initial = 5 bar abs for stoichiometric (with respect to CO and H O- formation) Propene/O mixtures, i.e. Propene:O = 1 : AIT [ C] at P initial = 1 bar abs for stoichiometric (with respect to CO and H O- formation) Propene/O mixtures, i.e. Propene:O = 1 : (CHEMSAFE, DIN-method in open Erlenmeyer flask) N [Vol.-%] stoichiometric C 3H O -> 3 CO + 3 H O Propene C 3 H 6 [Vol.-%] 9 8 range of detonative explosion stoichiometric C 3H O -> 3 CO + 3 H O 7 6 range of deflagrative explosion, stoichiometric C 3 H O -> 3 CO + 3 H propene/air-mixtures O [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 48

49 Induction times of two mixture compositions lying in the yellow range of the explosion diagram at 5 bar abs Induction time, ms Temperature, K Calculation by A.A. Konnov, Department of Mechanical Engineering Vrije Universiteit Brussel (VUB), based on the rate coefficients for all elementary reactions C3H6-O-N=-4-4 C3H6-O-N= Propene C 3 H 6 [Vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton N [Vol.-%] stoichiometric C 3 H O -> 3 CO + 3 H O range of detonative explosion stoichiometric C 3 H O -> 3 CO + 3 H O Propene= 5 vol.-%, O = 95 Vol.-% 7 6 range of deflagrative explosion, stoichiometric C 3 H O -> 3 CO + 3 H propene/air-mixtures Propene= vol.-%, O = 4 vol.-%, N = 4 vol.-% O [Vol.-%]

50 Is heat explosion a realistic assumption for what happens in yellow range of the explosion triangle? - temperatures achieved by adiabatic compression lie well above the autoignition temperature, i. e. heat explosion is possible in principle - but: induction times calculated on the basis of the rate coefficients of the elementary reactions are still far too long (The induction time would have to be of the order of.1 ms or even less, otherwise the flame front of the initial deflagration would have run already through the thin shell of unreacted mixture before the heat explosion can develop). - including the influence of thermal radiation from the cental part of the sphere does not give a substancial reduction in induction times either (calculation of Hans Pasman) Not really clear what happens in the yellow range of the explosion triangle H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 5

51 Summary on explosion regimes of combustible/air/o mixtures ignited with a thermal ignition source in vessel-like geometry There are three explosion regimes: deflagration, heat explosion, detonation The detonative range in the l sphere is discernibly smaller than the potentially detonative range What really happens in the heat explosion regime is not yet really understood, maybe the pressure time curves are just due to failed DDT s The heat explosion regime becomes much smaller with increasing initial pressure Mixtures of type combustible/air seem to remain deflagrative even at elevated pressures and temperatures. However, the data collected so far do not allow to exclude that the detonative regime will expand down to the air line when going to extremely high initial pressures and large vessel volumes. H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 51

52 Introduction Explosion characteristics Course of explosion in l sphere Pressure load P det on the wall of the vessel - quantification of P det for four scenarios occuring in a pipe (no precompression) - quantification of P det for four scenarios occuring in a pipe (with precompression) - quantification of P det for Propene/O /N in vessel at P initial = 5 bar abs, T initial = C Comparison: detonative regimes in l vessels 1, 5, 5 l vessels Predetonation distances of Propene/O in tubes Did ignition source directly trigger a spherical detonation? H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 5

53 Schematic pressure-time trace of a detonative pressure pulse hot reaction products flamefront coupled with shockfront, v >> v soundl pressure sensor, flush with inner surface of wall unburned gas mixture Pressure/P initial Leading edge of detonation peak has almost infinite slope (rise time < 1µs) P initial von Neumann spike, about twice as high as P CJ width < 1 µs => much less than cycle time of any eigenfrequency of the containment => not noticed by the containment and not registered by the fastest available pressure sensors P CJ, about twice the deflagration pressure ratio of the gas mixture time width of Taylor expansion fan : ca. µs 5 µs in laboratory scale containments, longer in structures of larger dimensions (less than one quarter of the time interval over which the peak has already propagated) deflagration pressure ratio of gas mixture H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 53

54 Pressure-time diagrams recorded during a detonation in a long pipe at different locations (1/) Example: Propane/O, stoichiometric, P initial = 6 bar abs, T initial = C, pipe 5x, L = 5 m, side-on pressure P det 35*P initial (same value at all sensors) v = Δs/Δt =.945m/.375ms = 519 m/s location of ignition pipe: 5 x 15 mm P1 P P3 P4 197 mm 915 mm 3855 mm 5 mm H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 54

55 Pressure-time diagrams recorded during a detonation in a long pipe at different locations (/) Experiment: Ethylene/air stoichiometric,14 bar abs, C, side-on pressure, pipe: φ i = 44.3 mm, φ o = 48.3 mm, L = 8 m, turbulence enhancer in front of ignition source to reduce run-up distance P 3 1 Δs = m Δt = 1.33 ms v det =Δs/Δt=1936 m/s P Δs = 1.5 m Δt =.796 ms v det =Δs/Δt=1883 m/s P location of ignition P 3 P P 1 m 1.5 m.5 m H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 55

56 Qualitative consideration: differences between the pressures produced in a long pipe by a deflagration and by a detonation (/) Example: Propane/air, stoichiometric, T initial = C, explosion pressure ratio: 9.5 (simplification: no heat loss to wall) Detonative Explosion (predetonation stage necglected) Location of ignition hot reaction products flamefront coupled with shockfront, v >> v soundl unburnt gas mixture P ex /P initial P = 9.5*P initial P det = *9.5*P initial extension: in space ca. 1 cm In time ca. 5 µs P = P initial distance from left pipe end P is different at different locations in the pipe: - behind shockfront: P = 9.5*P initial (9.5 is explosion pressure ratio of mixture) - at shockfront: P = *9.5*P initial (i.e. about twice as high as behind shockfront) - in front of shockfront: P = P initial (reason: reaction gases would like to expand into region ahead of shockfront, but speed of sound in hot reaction gases only about 1 m/s whereas shockfront propagates with about 17 m/s) H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 56

57 Detonation pressures P det in long pipes (L >> L predet, no precompression) Stable Detonation, side-on pressure (i. e. in plane perpendicular to shockfront) detonation front pressure sensor, flushwithinner surface of wall P det = P initial * P CJ P CJ is the so-called Chapman-Jouguet-pressure ratio of the explosive gas mixture Stable Detonation, reflected pressure (i. e. in plane parallel to shockfront, e. g. blind flange) detonation front pressure sensor, flushwithinner surface of flangel P det = P initial * P CJ * F reflec Important: The pipe directly ahead of the blind flange experiences the reflected pressure as well. The length of this section corresponds to the axial extension of the taylor expansion fan. (In lab-scale equipment of up to 5 m length a section of about 3 pipe diameters from the blind flange back into the pipe is affected, in longer pipes this section increases). F reflec.5 H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 57

58 Detonation pressures P det in long pipes (L >> L predet, no precompression) (continued) Side-on pressure at location where Deflagration-to-Detonation transition (DDT) occurs P det = P initial * P CJ * F DDT 1.5 F DDT location of ignition flame front of initial first occurrence of ignition deflagration has accelerated by adiabatic compression to a very high speed in precompressed mixture; generation of shock front coupled with flame front, propagating leftwards Reflected pressure, if DDT occurs directly in front of blind flange (a very rare case!) flame front of initial deflagration has accelerated to a very high speed first occurrence of ignition by adiabatic compression in precompressed mixture; generation of shock front coupled with flame front, propagating leftwards P det = P initial * P CJ * F DDT * F reflec F reflec.5 Important: The pipe directly ahead of the blind flange experiences the reflected pressure as well. The length of this section corresponds to the axial extension of the taylor expansion fan. (In lab-scale equipment of up to 5 m length a section of about 3 pipe diameters from the blind flange back into the pipe is affected, in longer pipes this section increases). H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton F DDT

59 Chapman-Jouguet pressure ratios P CJ calculated by STANJAN Example: Propene/O /N (Chapman-Jouguet pressure ratio) pcj/pinitial P (Chapman-Jouguet CJ pressure divided by pinitial) Pcj/Pinitial at Pcj/Pinitial at 5 bar abs, C measured defl. pr.-ratios Pex/Pinitial at 5 bar abs, 5 C measured defl. pr.-ratios Pex/Pinitial at 5 bar abs, C Propene concentration in Propene/O -mixture [vol.-%] pex/pinitial (deflagration pressure ratio) Solid lines: Calculated Chapman-Jouguet pressure ratios for p initial = 5 bar abs (by Prof. M. Braithwaite with software package STANJAN) Symbols: Measured deflagration pressure ratios (scale on right side of plots) The values are just about half of the Chapman-Jouguet pressure ratios!! (Chapman-Jouguet pressure ratio) pcj/pinitial P (Chapman-Jouguet CJ pressure divided by pinitial) calculated Pcj/Pinitial at 5 bar abs, 5 C measured defl. pr.-ratios Pex/Pinitial at Propene conc. in stoich. Propene/O /N -mixture [vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton pex/pinitial (deflagration pressure ratio)

60 Chapman-Jouguet pressure ratios P CJ calculated by STANJAN Example: Methane/O /N (Chapman-Jouguet pressure ratio) pcj/pinitial P (Chapman-Jouguet CJ pressure divided by Pinitial ) calculated Pcj/Pinitial at calculated Pcj/Pinitial at 5 bar abs, C measured defl. Pr.-ratios Pex/Pinitial at 5 bar abs, 5 C measured defl. Pr.-ratios Pex/Pinitial at 5 bar abs, C Methane concentration in Methane/O -mixture [vol.-%] 17,5 15 1,5 1 7,5 5,5 pex/pinitial (deflagration pressure ratio) (Chapman-Jouguet pressure ratio) P PCJ Chapman-Jouguet CJ detonation pressure ratio bar abs, C Methane concentration in Methane/air-mixture [vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 6

61 Chapman-Jouguet pressure ratios P CJ calculated by STANJAN Example: H /O /N 1 pcj/pinitial (Chapman-Jouguet pressure divided by pinitial) calculated detonation pr.-ratios PCJ/Pinitial at 1 bar abs, C calculated detonation pr.-ratios PCJ/Pinitial at 5 bar abs, C calculated detonation pr.-ratios PCJ/Pinitial at 5 bar abs, C measured pex/pinitial at 1 bara bs, C H concentration in H /O -mixture [vol.-%] 7,5 5,5 pex/pinitial deflagration pressure ratio) 1 pcj/pinitial (Chapman-Jouguet pressure divided by pinitial) calculated detonation pr.-ratios PCJ/Pinitial at 1 bar abs, C calculated detonation pr.-ratios PCJ/Pinitial at 5 bar abs, C calculated detonation pr.-ratios PCJ/Pinitial at 5 bar abs, C measured pex/pinitial at 1 bar abs, C measured pex/pinitial at 5 bar abs, C measured pex/pinitial at 5 bar abs, C H concentration in H /Air-mixture [vol.-%] H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 61 7,5 5,5 pex/pinitial deflagration pressure ratio)

62 Chapman-Jouguet pressure ratios P CJ calculated on basis of detonation velocity meaning of symbols: P CJ = p det /p initial = (γ *M +1)/(γ+1) (γ * M )/(γ+1) = (D * MW)/(R*T initial *(γ+1)) = ρ/p initial * D /(γ+1) M = D/c Mach number D propagation speed of detonation [m/s] c = (γ * p initial /ρ).5 speed of sound in the unreacted gas mixture [m/s] ρ density of unreacted gas mixture [kg/m 3 ] MW mean molar mass of unreacted gas mixture [kg/mol] R=8.314 universal gas constant [J/(mol*K)] γ c p /c v, for diatomic gases γ = 1.4 T initial, p initial initial temperature [K], initial pressure [Pa] of gas mixture side on pressure of stable detonation p det Remarks: For using STANJAN various thermodynamic parameters of the reaction components must be known. If these are not available, the detonation velocity is experimentally easily aminable with high precision by recording the side-on pressure of the propagating detonation at different axial locations in the pipe with piezoelectric pressure sensors. Since the absolute signal value, which is actually the quantity one is looking for, is often prone to errors, one has to use the bypass over the speed. H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 6

63 General remarks on Chapman-Jouguet pressure ratios P CJ Chapman-Jouguet pressure ratios can be calculated by e.g. STANJAN (freeware, STANJAN Chemical Equilibrium Solver v4.1, Stanford University 3) If thermodynamic quantities required by STANJAN are not available for all components of the considered explosive gas mixture, P CJ can be calculated on the basis of the experimentally determined detonation speed. Chapman-Jouguet pressure ratios are with good precision twice as high as the explosion pressure ratio found for deflagrative explosion: - stoichiometric combustible/air- mixtures at C: 16 P CJ - stoichiometric combustible/o -mixtures at C: 3 P CJ 5 Chapman-Jouguet pressure ratio is inversely proportional to the absolute initial temperature. (Propagation speed of detonation is almost independent of T initial!) H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 63

64 Example: bulging out of a pipe end due to reflected pressure Experiment: Acetylene, 8 bar abs, C, pipe.9 x 8., i.e. φ i = 4.5 mm, material: St35.8, R p. = 3 N/mm, P Rp. =184 bar, L = 8 m, detonative ignition source to reduce run-up distance; side-on pressure ca. 4 bar, reflected pressure ca. 8 bar Increase of tube diameter directly in front of blind flange to 117 % of original value no increase in diameter in the section that has only been exposed to the sideon pressure H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 64

65 Example: Rupture of pipe end due to reflected pressure Experiment: Ethylene/air stoichiometric, 17 bar abs, C, pipe 76.1 x.6, i.e. φ i = 7.9, material: , R p. = 1 N/mm, P Rp. =154 bar, L = 8 m; side-on pressure ca. 34 bar, reflected pressure ca. 765 bar increase in diameter to 15 % of original value in the section that has only been exposed to the sideon pressure end of tube, which was exposed to the reflected pressure, is fragmented H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 65

66 Example: bulging out of a pipe wall due to the extra pressure at the point where the DDT occurs Experiment: Acetylene, 4 bar abs, C, pipe 3.17 x.5, φ i =.17 mm, material: , R p. 5 N/mm, P Rp. 115 bar, kaltgezogen, L = 1 m Localized bulging of tube is attributed to the location of the deflagration to detonation transition H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 66

67 Introduction Explosion characteristics Course of explosion in l sphere Pressure load P det on the wall of the vessel - quantification of P det for four scenarios occuring in a pipe (no precompression) - quantification of P det for four scenarios occuring in a pipe (with precompression) - quantification of P det for Propene/O /N in vessel at P initial = 5 bar abs, T initial = C Comparison: detonative regimes in l vessels 1, 5, 5 l vessels Predetonation distances of Propene/O in tubes H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 67

68 How does precompression ( pressure piling ) come into existence? During the initial deflagrative stage of the explosion each volume element that reacts through expands by a factor equal to the explosion pressure ratio at the expense of the volume available for the yet unreacted mixture. If this volume is large compared to the volume occupied by the reaction gases formed during initial deflagrative stage, the pressure in the yet unreacted mixture will almost stay the same as at the moment of ignition (P = P initial ). However, if this volume is small, the yet unreacted mixture will be precompressed. When the deflagration to detonation transition then occurs, the detonation will not propagate in a mixture at pressure P = P initial, but at P = F precomp * P initial F precomp may attain any value between 1 and the explosion pressure ratio of the gas mixture (up to 1 for combustible/air-mixtures, up to 4 for combustible/o - mixtures). The temperature rise that comes along with the fast and hence adiabatic compression will slightly alleviate the detonative pressure P det. This is accounted for by the Factor F temp (details see next slide) In the formulae given beforehand for P det, F precomp and F temp will appear as additional factors H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 68

69 Temperature factor F temp The temperature factor F temp accounts for the reduction of the Chapman-Jouguet pressure ratio due to the rise in temperature of the unburned mixture brought about by the precompression during the initial deflagrative stage of the explosion. I: Temperature T final of unburned mixture attained by adiabatic compression: T final ( γ 1) γ ( γ 1) final = T = γ initial Tinitial F p precomp initial p with: γ = c p /c v (γ = 1.4 for ideal gases) T initial = temperature at moment of ignition II: Chapman-Jouguet pressure ratio is inversely proportional to the temperature the gas mixture exhibits at the moment of detonation (here denoted by T final ). F temp results from the combined effect of I: and II: F temp = T T initial final = F 1 precomp ( γ 1) γ H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 69

70 Temperatures obtainable by adiabatic compression T initial = 5 C = 98 K T initial = C = 473 K P final /P initial T final [ C] for different values of γ = c p /c v γ = 1,4 γ = 1,3 γ = 1, γ = 1,1 P final /P initial T final [ C] for different values of γ = c p /c v γ = 1,4 γ = 1,3 γ = 1, γ = 1,1 1, 4,9 37,8 34, 3, 1, 5,3,3 14,6 7,9 1,5 61,6 54, 45,8 36, 1,5 58,1 46,4 33,1 17,8 9,3 76,7 61,5 44,4 33,6 8, 57,9 3, ,8 137,3 1,5 65, 4 49,9 378,3 3,9 63,5 6 4, 177,6 18,7 77, , 44, 364,6 83,7 8 66,8 8,5 148,4 87, 8 583,8 491,3 395,9 98,4 1 3,3 34, 164,4 94,4 1 64, 531,7 41,3 31, ,1 55,8 177,9 1, ,1 566,3 44,7 319, ,4 74,9 189,6 15, ,4 596,7 461,3 38, , 9,1, 11, ,5 63,9 477,8 335, ,6 37,6 9,4 114, , 648,6 49,7 34,1 48,4 31,9 18, 118,3 84, 671,3 56,3 348, ,5 353,4 36,6 16, ,5 71, 535,8 36, ,5 38,3 5,3 133, 3 977, 763,9 56,8 371, , 43,9 66, 138, , 81,4 58,5 38,5 4 58, 45,1 78,1 143, , 835,1 61,7 388, , 444,3 89, 148, ,5 865,6 619,1 395, , 46, 99, 15, ,4 893,6 634,9 4, Formula for temperature increase due to adiabatic compression: T final = T initial p final p initial ( γ 1) γ H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 7

71 Example 1: precompression in a pipe P 3 P 1 hot reaction products P flame front of initial deflagration first occurrence detonation front precompressed unburned gas mixture If the pipe is only slightly longer than the predetonation distance, the unburned gas which is remaining at the moment when the DDT occurs may have been precompressed by a factor F precomp by the initial deflagrative stage of the combustion. F precomp may attain any value between 1 and the explosion pressure ratio of the gas mixture (up to 1 for combustible/air-mixtures, up to 4 for combustible/o - mixtures). The temperature rise that comes along with the fast and hence adiabatic compression will slightly alleviate the detonative pressure P det. This is accounted for by the Factor F temp H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 71

72 Example : precompression in the vessel-pipe configuration vessel pipe, slightly longer than L predet Location of ignition hot, gaseous reaction products Precompressed, unburned gas mixture deflagrative flamefront If in the vessel the DDT does not happen because of the unfavourable L/D-ratio and if the volume of the pipe is small compared to the volume of the vessel, the precompression factor may come close to the highest possible value, i. e. the explosion pressure ratio. The highest possible value is usually not attained, because the available time is too short to pressurize the pipe up to the same pressure as in the vessel. This configuration occurs frequently in practical applications, because vessels often have nozzles onto which short pipe segments with a block valve are welded. H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 7

73 Example 3: precompression in empty vessels (i.e. without packing inside) vessel (here spherical, L/D = 1) Location of ignition hot reaction products Precompressed, unburned mixture First occurrence of detonation front Flamefront of initial deflagration If there is a transition from deflagration to detonation in the vessel, precompression will almost always occur, because the diameter is usually not much larger than the predetonation distance. The precompression factor may attain the highest possible value, i. e. the explosion pressure ratio. Reference: GCT report N, I H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 73

74 Detonation pressures P det in pipes with precompression (1/) Side-on pressure (i. e. in plane perpendicular to shockfront) detonation front pressure sensor, flushwithinner surface of wall P det = P initial * P CJ * F precomp * F temp Reflected pressure (i. e. in plane parallel to shockfront, e. g. blind flange) detonation front pressure sensor, flushwithinner surface of flangel P det = P initial * P CJ * F reflec * F precomp * F temp Important: about 3 pipe diameters from the blind flange back into the pipe this higher pressure also acts on the wall of the pipe! H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 74

75 Detonation pressures P det in pipes with precompression (/) Side-on pressure at location where Deflagration-to-Detonation transition (DDT) occurs P det = P initial * P CJ * F DDT * F precomp * F temp location of ignition flame front of initial first occurrence of ignition deflagration has accelerated by adiabatic compression to a very high speed in precompressed mixture; generation of shock front coupled with flame front, propagating leftwards Reflected pressure, if DDT occurs directly in front of blind flange (a very rare case!) P det = P initial * P CJ * F DDT * F reflec * F precomp * F temp flame front of initial deflagration has accelerated to a very high speed first occurrence of ignition by adiabatic compression in precompressed mixture; generation of shock front coupled with flame front, propagating leftwards Important: about 3 pipe diameters from the blind flange back into the pipe this higher pressure also acts on the wall of the pipe! H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 75

76 Detonation pressure P det acting on the wall of a vessel (almost always with precompression) P det = Initial pressure in vessel at moment of ignition (P initial ) x Chapman-Jouguet pressure ratio of the mixture at the temperature the mixture exhibits at the moment of ignition (P CJ ) x Precompression factor (F Precomp ) x Temperature factor (F Temp ) x Factor accounting for reflection of stable detonation at wall (F reflec ) x Factor accounting for extra pressure if DDT happens directly before wall (F DDT ), otherwise factor is 1 P det = P initial * P CJ * F precomp * F temp * F reflec * (F DDT or 1, depending on where DDT happened) H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 76

77 Example: Rupture of the wall of a pipe installed in the vessel-pipe configuration due to precompression detonation in this section, outer diameter is 1 % to 11 % of original value closed Valve (to vacuum pump) 1x pipe (1.4541) rupture pressure in hydraulic test: 348 bar rupture at location of DDT in precompressed mixture l sphere only deflagration in this section, outer diameter is not increased Rest of the bend has been burned off by outflowing hot reaction gases P initial = 5 bar abs, T initial = 5 C 99.5 vol.-% N O.5 vol.-% Tetradecane Ignition in center of sphere H.-P. Schildberg, 41st UKELG meeting, 13 May 8, Shell Technology Centre Thornton 77