Rovibrational internal energy excitation and ... - (cnr) - (bari)
Rovibrational internal energy excitation and ... - (cnr) - (bari) Rovibrational internal energy excitation and ... - (cnr) - (bari)
Rovibrational internal energy excitation and dissociation of molecular nitrogen in hypersonic flows Thierry MAGIN, 1 Marco PANESI, 2 Anne BOURDON, 3 Richard JAFFE, 4 and David SCHWENKE 4 Thanks to: Winifred Huo, 4 Galina Chaban, 4 Yen Liu, 4 and Christophe Laux 3 1 Aeronautics and Aerospace Department, von Karman Institute for Fluid Dynamics, Belgium 2 Institute for Computational Engineering and Sciences, The University of Texas at Austin 3 EM2C Laboratory, CNRS UPR 288 – Ecole Centrale Paris, France 4 NASA Ames Research Center Symposium in honour of Prof. Mario Capitelli on the occasion of his 70 th birthday Chemical Physics of Low Temperature Plasmas 31 Jan – 1 Feb 2011, University of Bari, Italy Prof. Capitelli’s symposium (U Bari) Rovibrational collisional model 31 Jan – 1 Feb 2011 1 / 18
- Page 2 and 3: Prof. Capitelli’s symposium (U Ba
- Page 4 and 5: Introduction Motivation Motivation:
- Page 6 and 7: NASA ARC database for N 2 + N syste
- Page 8 and 9: Rovibrational collisional model Mas
- Page 10 and 11: Rovibrational collisional model Bin
- Page 12 and 13: Results Simulation of internal ener
- Page 14 and 15: Results Rovibrational energy popula
- Page 16 and 17: Conclusion and further development
- Page 18: Prof. Capitelli’s symposium (U Ba
<strong>Rovibrational</strong> <strong>internal</strong> <strong>energy</strong> <strong>excitation</strong> <strong>and</strong> dissociation<br />
of molecular nitrogen in hypersonic flows<br />
Thierry MAGIN, 1 Marco PANESI, 2 Anne BOURDON, 3<br />
Richard JAFFE, 4 <strong>and</strong> David SCHWENKE 4<br />
Thanks to: Winifred Huo, 4 Galina Chaban, 4 Yen Liu, 4 <strong>and</strong> Christophe Laux 3<br />
1 Aeronautics <strong>and</strong> Aerospace Department, von Karman Institute for Fluid Dynamics, Belgium<br />
2 Institute for Computational Engineering <strong>and</strong> Sciences, The University of Texas at Austin<br />
3 EM2C Laboratory, CNRS UPR 288 – Ecole Centrale Paris, France<br />
4 NASA Ames Research Center<br />
Symposium in honour of Prof. Mario Capitelli on the occasion of his 70 th birthday<br />
Chemical Physics of Low Temperature Plasmas<br />
31 Jan – 1 Feb 2011, University of Bari, Italy<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 1 / 18
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 2 / 18
Outline<br />
Outline<br />
Introduction<br />
NASA ARC database for N 2 + N system<br />
<strong>Rovibrational</strong> collisional model<br />
1D shock-tube simulations<br />
Full Master equation<br />
Coarse graining model<br />
Conclusion<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 3 / 18
Introduction<br />
Motivation<br />
Motivation: developing high-fidelity nonequilibrium models<br />
⇒ Underst<strong>and</strong>ing thermo-chemical nonequilibrium effects is important<br />
For an accurate prediction of the radiative heat flux for reentries at<br />
v>10km/s (Moon <strong>and</strong> Mars returns)<br />
For a correct interpretation of experimental measurements<br />
In flight<br />
In ground wind-tunnels<br />
Calculated <strong>and</strong> measured intensity N 2 (1+) system<br />
⇒ St<strong>and</strong>ard nonequilibrium models for<br />
hypersonic flows were mainly<br />
developed in the 1980’s (correlation<br />
based)<br />
e.g. dissociation model of Park<br />
Multitemperature model:<br />
T = T r , T v = T e = T ele<br />
Average temperature √ T T v for<br />
fictitious Arrhenius rate coefficient<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 4 / 18
Introduction<br />
Objective<br />
Objective: developing reduced nonequilibrium models for<br />
reentry flows based on microscopic theory <strong>and</strong> applying<br />
them to macroscopic scale<br />
⇒ work at the interface between computational chemistry, experimental<br />
measurements, <strong>and</strong> CFD<br />
Computational methods<br />
Physico−chemical models<br />
Experimental data<br />
N 3 Potential Energy Surface<br />
NASA Ames Research Center<br />
Blast capsule simulation<br />
VKI COOLFluiD / Mutation<br />
Diagnostics in plasma jet<br />
VKI Plasmatron<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 5 / 18
NASA ARC database for N 2 + N system<br />
Internal <strong>energy</strong> <strong>excitation</strong> <strong>and</strong> dissociation derived from<br />
ab initio calculations<br />
Characterization of nonequilibrium air chemistry from first principles<br />
by the chemistry group of NASA Ames Research Center<br />
⇒ Papers AIAA 2008-1208, 2008-1209, 2009-1569, 2010-4517, RTO-VKI LS 2008<br />
The 9390 (v,J) rovibrational <strong>energy</strong> levels for N 2 are split into<br />
Bound levels below the dissociation <strong>energy</strong><br />
Quasi-bound or predissociated levels above the dissociation <strong>energy</strong><br />
but below the centrifugal barrier of the potential<br />
Potential curve of N 2 for J=0,20,40,. . .<br />
⇒ Complementary work at Bari (PHYS4ENTRY) <strong>and</strong> U Minnesota (MURI)<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 6 / 18
NASA ARC database for N 2 + N system<br />
A novel approach for nonequilibrium models...<br />
Computation from first principles at NASA ARC in 2 steps:<br />
1 Quantum chemistry calculations to generate realistic nuclear<br />
interaction potentials<br />
2 Classical trajectory method for the reaction cross-sections<br />
e.g. N 2 + N system:<br />
9390 (v,J) rovibrational <strong>energy</strong><br />
levels for N 2<br />
23 × 10 6 reaction mechanism<br />
N 2(v, J) + N ↔ N + N + N<br />
N 2(v, J) ↔ N + N<br />
N 2(v, J) + N ↔ N 2(v ′ , J ′ ) + N<br />
Experimental data are used only for<br />
validation<br />
⇒ This model has a wide application<br />
range (physics based)<br />
N 3 Potential Energy Surface<br />
NASA Ames Research Center<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 7 / 18
<strong>Rovibrational</strong> collisional model<br />
Master equation<br />
Detailed chemical mechanism coupled with a flow solver<br />
Full master eq. of conservation of mass for the 9390 rovibrational<br />
<strong>energy</strong> levels i = (v, J) for N 2 , <strong>and</strong> for N atoms coupled with eqs. of<br />
conservation of momentum <strong>and</strong> total <strong>energy</strong><br />
⎛ ⎞ ⎛ ⎞ ⎛ ⎞<br />
ρ i<br />
ρ i u M N2 ω i<br />
d<br />
⎜ρ N<br />
⎟<br />
dt ⎝ρu⎠ + d<br />
⎜ ρ N u<br />
⎟<br />
dx ⎝ρu 2 + p⎠ = ⎜M N ω N<br />
⎟<br />
⎝ 0 ⎠<br />
ρE<br />
ρuH<br />
0<br />
... but computationally too expensive for 3D CFD applications<br />
⇒ reduction of the chemical mechanism by lumping the <strong>energy</strong> levels i:<br />
e.g. vibrational state-to-state models (AIAA 2009-3837, 2010-4335)<br />
d<br />
dt ρ v + d<br />
dx (ρ v u) = M N2 ω v<br />
The <strong>energy</strong> levels are lumped for each v assuming a rotational <strong>energy</strong><br />
population following a Maxwell-Boltzmann distribution at T<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 8 / 18
<strong>Rovibrational</strong> collisional model<br />
Bin model<br />
Coarse graining model<br />
Another lumping scheme is to sort the (v,J) levels by <strong>energy</strong> <strong>and</strong><br />
include in a bin all levels with similar energies<br />
25x10 -19<br />
20x10 -19<br />
Energy [ J ]<br />
15x10 -19<br />
10x10 -19<br />
5x10 -19<br />
0.0<br />
0 2000 4000 6000 8000 10000<br />
Index [ - ]<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 9 / 18
<strong>Rovibrational</strong> collisional model<br />
Bin model<br />
Coarse graining model<br />
The <strong>energy</strong> for level i = (v, J) is constant in bin k <strong>and</strong> equal to the average<br />
<strong>energy</strong><br />
∑<br />
i∈I<br />
Ē k = k<br />
g i E i<br />
,<br />
ḡ k<br />
with the bin degeneracy ḡ k = ∑ i∈I k<br />
g i<br />
The <strong>energy</strong> level populations are assumed to be uniform within a bin<br />
n i<br />
¯n k<br />
= 1 ḡ k<br />
g i ,<br />
i ∈ I k<br />
The initial population of the bins is assumed to follow a Maxwell-Boltzmann<br />
distribution<br />
( )<br />
¯n k<br />
= ḡk −<br />
n N2<br />
¯Q(T int ) exp Ē k<br />
k B T int<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 10 / 18
<strong>Rovibrational</strong> collisional model<br />
Numerical method<br />
Numerical method<br />
The post-shock conditions are obtained from the Rankine-Hugoniot<br />
jump relations<br />
The 1D Euler eqs. for collisional model comprises<br />
Mass conservation eqs. for N<br />
Mass conservation eqs. for the 9390 rovibrational levels of N 2<br />
Momentum conservation eq.<br />
Total <strong>energy</strong> conservation eq.<br />
The backward reaction rate coefficients are based on microreversibility<br />
The conservative form is transformed into an ODE system solved by<br />
LSODE<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 11 / 18
Results<br />
Simulation of <strong>internal</strong> <strong>energy</strong> <strong>excitation</strong> <strong>and</strong> dissociation<br />
processes behind a strong shockwave in N 2 flow<br />
Simulations based on different models:<br />
Full master eq. model<br />
Coarse graining model<br />
Vibrational collisional model<br />
Multitemperature model for the 2-species mixture with simplified<br />
mechanism<br />
Free stream (1), post-shock (2), <strong>and</strong> LTE (3) conditions<br />
1 2 3<br />
T [K] 300 62,546 11,351<br />
p [Pa] 13 10,792 13,363<br />
u [km/s] 10 2.51 0.72<br />
x N 0.028 0.028 1<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 12 / 18
Results<br />
Temperature <strong>and</strong> composition profiles<br />
Temperatures T , T v (v = 1), T int (v = 0, J = 10)<br />
Temperature [K]<br />
70000<br />
60000<br />
50000<br />
40000<br />
30000<br />
20000<br />
T full CR: no pred.<br />
T int<br />
T full CR<br />
T int<br />
T vibrational CR<br />
T v<br />
T 2-species Park<br />
Mole fractions [-]<br />
1<br />
0.8<br />
0.6<br />
0.4<br />
N 2 full CR : no predis.<br />
N<br />
N 2 full CR<br />
N<br />
N 2 vibrational CR<br />
N<br />
T v<br />
2.0×10 -6 4.0×10 -6 6.0×10 -6 8.0×10 -6 10×10 -6<br />
10000<br />
0.2<br />
0<br />
2.0×10 -6 4.0×10 -6 6.0×10 -6 8.0×10 -6 10×10 -6<br />
Time [s]<br />
0<br />
Time [s]<br />
Free stream: T 1 = 300 K, p 1 = 13 Pa, u 1 = 10 km/s, x N1 ∼ 2.8%, 10 −5 s ↔ 2.5 cm<br />
⇒ Thermalization <strong>and</strong> dissociation occur after a larger distance for the<br />
full collisional model<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 13 / 18
Results<br />
<strong>Rovibrational</strong> <strong>energy</strong> population of N 2<br />
n(v, J) in function of E(v, J) at t = 2.6 × 10 −6 s (7mm)<br />
A rotational temperature T r (v) is introduced for each vibrational <strong>energy</strong> level v:<br />
PJ max(v)<br />
n(v, J)∆E(v, J)<br />
J=0<br />
=<br />
PJ max(v)<br />
n(v, J)<br />
J=0<br />
PJ max(v)<br />
J=0<br />
“ ”<br />
−∆E(v,J)<br />
g J ∆E(v, J) exp<br />
kTr (v)<br />
“ ”<br />
−∆E(v,J)<br />
kTr (v)<br />
PJ max(v)<br />
g<br />
J=0 J exp<br />
⇒ The assumption of equilibrium between the rotational <strong>and</strong><br />
translational modes is questionable...<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 14 / 18
Results<br />
Coarse graining model for 3D CFD applications<br />
Coarse-graining model: lumping the <strong>energy</strong> levels into bins as a<br />
function of their global <strong>energy</strong><br />
70000<br />
Temperature [K]<br />
60000<br />
50000<br />
40000<br />
30000<br />
20000<br />
2<br />
5<br />
10<br />
20<br />
40<br />
100<br />
150<br />
Full CR<br />
10000<br />
0<br />
0<br />
2×10 -6 4×10 -6 6×10 -6 8×10 -6 10×10 -6<br />
Time [s]<br />
Without predissociation reactions<br />
Free stream: T 1 = 300 K, p 1 = 13 Pa, u 1 = 10 km/s, x N1 ∼ 2.8%, 10 −5 s ↔ 2.5 cm<br />
⇒ The uniform distribution allows to describe accurately the <strong>internal</strong><br />
<strong>energy</strong> relaxation <strong>and</strong> dissociation processes for ∼20 bins<br />
A M-B distribution is being investigated for CFD applications<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 15 / 18
Conclusion <strong>and</strong> further development<br />
Conclusion<br />
A 1D rovibrational collisional model was developed to describe the<br />
<strong>internal</strong> <strong>energy</strong> relaxation <strong>and</strong> dissociation processes behind a strong<br />
shockwave in a nitrogen flow<br />
The 9390 rovibrational <strong>energy</strong> levels of the nitrogen molecule of the ab<br />
initio NASA Ames database are taken into account in the master eq.<br />
Thermalization <strong>and</strong> dissociation occur after a larger distance for the full<br />
collisional model than for the multitemperature model <strong>and</strong> vibrational<br />
collisional model (distinct rotational temperatures per vibrational level)<br />
The uniform distribution bin model allows to describe accurately the<br />
<strong>internal</strong> <strong>energy</strong> relaxation <strong>and</strong> dissociation processes based on a<br />
reduced number of eqs.<br />
Further development<br />
Lessons learned will allow to investigate the N 2 (v, J) + N 2 (v ′ , J ′ )<br />
system (AIAA 2010-4517)<br />
A coarse graining model based on a M-B distribution of the levels<br />
within a bin is being developed for CFD applications (AIAA 2010-4332)<br />
A sound Chapman-Enskog method is being derived for<br />
multitemperature models [ESA WG, D. Giordano]<br />
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 16 / 18
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 17 / 18
Prof. Capitelli’s symposium (U Bari) <strong>Rovibrational</strong> collisional model 31 Jan – 1 Feb 2011 18 / 18