OPTIMIZATION OF POTENTIALLY
RUNAWAY REACTIONS CARRIED OUT IN
PLUG FLOW REACTORS
Tesi di: Carlos Sierra (797299)
Scuola di Ingegneria Industriale e dell'Informazione
Corso di Laurea in Ingegneria della Prevenzione e della
Sicurezza nell’Industria di Processo
Anno Accademico 2013 – 2014
2
Tesi di Laurea Magistrale – Carlos Sierra
Introduction
• Plug flow reactors that are carrying out exothermic reactions
can generate, under certain operating conditions, an
uncontrolled temperature increase (thermal runaway) that could
lead to the triggering of other more exothermic reactions, such
as decomposition reactions.
• Thermal runaway studies on continuous reactors are normally
held for steady state operation, because all the dynamic
changes take place in a short period of time.
3
Tesi di Laurea Magistrale – Carlos Sierra
Scope
• Identify possible thermal runaway conditions for steady and
unsteady state operations in a plug flow reactor.
• Set a safety operation range for a case study of a high
exothermic reaction.
• Perform a criterion for thermal runaway conditions during
unsteady state operations.
4
Tesi di Laurea Magistrale – Carlos Sierra
Structure of the work
Theory model
Steady state simulations
Unsteady state simulations
Identify differences and establish a
criterion (if possible)
MATLAB
System solution
Equations for mass and energy balances
Solution method
Analytical
Numerical
Results analysis
Problem definition Runaway in PFR in unsteady state
ODE/PDE
NOT possible
5
Tesi di Laurea Magistrale – Carlos Sierra
Model assumptions:
• Constant diameter through all the reactor.
• No radial variations of velocity, concentrations, temperature or
reaction rate. Perfect mixing in the radial direction.
• Constant density along the reactor.
• Constant inlet velocity, which is equal to the axial velocity.
Model Plug Flow Reactor (PFR)
6
Tesi di Laurea Magistrale – Carlos Sierra
Unsteady State equations
Mass balance:
Dimensionless mass balance:
Energy balance:
Dimensionless energy balance:
Steady State equations: Neglected time dependent and diffusive terms
7
Tesi di Laurea Magistrale – Carlos Sierra
Thermal runaway
It is defined as a situation where a temperature rise changes the
conditions of a chemical reaction, in a way that causes a further
increase in temperature. This is a kind of uncontrolled positive
feedback.
8
Tesi di Laurea Magistrale – Carlos Sierra
Parametric sensitivity
• Parametric sensitivity
• Normalized sensitivity
It is defined as the system behavior with respect to changes in its
input parameters. By changing these values, the system
characteristics can achieve desired or undesired behaviors.
When a system operates in the parametrically sensitive region (the
region where small variations of a parameter make the system
becomes sensitive), its performance becomes unreliable and
changes with small variations in the parameters.
9
Tesi di Laurea Magistrale – Carlos Sierra
Numerical solution
MATLAB solver
Steady state simulations Unsteady state simulations
ODE solver
Ordinary Differential
Equation (ODE)
Partial Differential
Equation (PDE)
Function ODE15S
PDE solver
Function PDEPE
✔Useful in all cases ✘Not useful in all cases
Method of lines +
Finite difference
✔Useful in all cases
PDE Toolbox not considered
✘Useful for one dependent
variable system
10
Tesi di Laurea Magistrale – Carlos Sierra
Equation characteristics
• Initial conditions:
• Empty reactor.
• Reactor temperature equal to wall temperature
• Boundary conditions: Dankwerts type
11
Tesi di Laurea Magistrale – Carlos Sierra
Equation characteristics
• Boundary conditions: Dankwerts type
No change
Smooth step
change
12
Tesi di Laurea Magistrale – Carlos Sierra
Method of lines
• Approximation of the spatial derivatives using finite
differences.
Differentiation matrixes
Fornberg algorithm
Iteratively computing the weighting coefficients of finite difference
formulas of arbitrary order of accuracy on arbitrarily spaced
spatial grids.
• Time integration of the resulting semi-discrete (discrete in
space, but continuous in time) ODE, using ODE MATLAB
solvers.
13
Tesi di Laurea Magistrale – Carlos Sierra
Method of lines vs PDEPE
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
z[-]
X[-]
 =0[-]
 =0.15[-]
 =0.3[-]
 =0.45[-]
 =0.6[-]
 =0.75[-]
 =0.9[-]
 =1.05[-]
 =1.2[-]
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
z[-]
X[-]
 =0[-]
 =0.15[-]
 =0.3[-]
 =0.45[-]
 =0.6[-]
 =0.75[-]
 =0.9[-]
 =1.05[-]
 =1.2[-]
Solution for P_in = 1.5[kPa]
Solution for P_in = 2.0[kPa]
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
z[-]
X[-]
 =0[-]
 =0.15[-]
 =0.3[-]
 =0.45[-]
 =0.6[-]
 =0.75[-]
 =0.9[-]
 =1.05[-]
 =1.2[-]
Oscillation for P_in = 1.5[kPa]
Not possible solution for P_in = 2.0[kPa]
Method of lines PDEPE
14
Tesi di Laurea Magistrale – Carlos Sierra
Case study
Naphthalene oxidation to phthalic Anhydride
It is a chemical intermediate in the production of plasticizers for
polyvinyl chloride (PVC).
It was considered a gas phase reaction.
15
Tesi di Laurea Magistrale – Carlos Sierra
The gas-phase naphthalene oxidation to phthalic anhydride is a
highly exothermic reaction. This reaction is carried out in multi-
tubular reactors, cooled by molten salt that is passing around an
external jacket. As a catalyst, V2O5 is used.
Case study
16
Tesi di Laurea Magistrale – Carlos Sierra
Case study
Initial values for the studied input parameters:
• Inlet pressure
• Inlet temperature
• Wall temperature
• Inlet velocity
• Steady state analysis:
• Reactor temperature vs dimensionless reactor length.
• Normalized sensitivity for each input parameter.
• Unsteady state:
• Normalized sensitivity for each input parameter, for different
dimensionless times.
• Ratio between maximum reactor temperature and inlet temperature
vs conversion at that temperature (topological diagram).
17
Tesi di Laurea Magistrale – Carlos Sierra
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
625
725
825
925
1025
1125
1225
1325
1425
z[-]
T[K]
Pin
=1.5[kPa]
Pin
=1.6[kPa]
Pin
=1.7[kPa]
Pin
=1.8[kPa]
Pin
=1.9[kPa]
Pin
=2[kPa]
Pin
=2.1[kPa]
Pin
=2.2[kPa]
Results Steady State Inlet Pressure
Maximum rate of
change
High sensitivity
1.5 1.6 1.7 1.8 1.9 2 2.1 2.2
0
10
20
30
40
50
60
70
Pin
[kPa]
S(T,Pin
)
P
in
= 1.85, S[T,P
in
] = 58.6996
Increase inlet
pressure
Increase maximum
reactor temperature
Hot spot moving to
the left
18
Tesi di Laurea Magistrale – Carlos Sierra
Results Steady State Inlet Temperature
0 0.1 0.2 0.3 0.4 0.5
600
700
800
900
1000
1100
1200
1300
z[-]
T[K]
Tin
=620[K]
Tin
=630[K]
Tin
=640[K]
Tin
=650[K]
Tin
=660[K]
Tin
=670[K]
Tin
=680[K]
Tin
=690[K]
Tin
=700[K]
Tin
=710[K]
Tin
=720[K]
620 630 640 650 660 670 680 690 700 710 720
0
20
40
60
80
100
120
140
160
180
200
Tin
[K]
S(T,Tin
)
T
in
= 675, S[T,T
in
] = 161.1574
Maximum rate of
change
High sensitivity
Increase inlet
temperature
Increase maximum
reactor temperature
Hot spot moving to
the left
19
Tesi di Laurea Magistrale – Carlos Sierra
Results Steady State Wall Temperature
0 0.1 0.2 0.3 0.4 0.5
600
700
800
900
1000
1100
1200
1300
z[-]
T[K]
Tw
=580[K]
Tw
=590[K]
Tw
=600[K]
Tw
=610[K]
Tw
=620[K]
Tw
=630[K]
Tw
=640[K]
Tw
=650[K]
Tw
=660[K]
Tw
=670[K]
Tw
=680[K]
580 590 600 610 620 630 640 650 660 670 680
0
20
40
60
80
100
120
Tw
[K]
S(T,Tw
)
T
w
= 638, S[T,T
w
] = 104.3867
Maximum rate of
change
High sensitivity
Increase wall
temperature
Increase maximum
reactor temperature
Hot spot moving to
the left
20
Tesi di Laurea Magistrale – Carlos Sierra
Results Steady State Inlet Velocity
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
625
630
635
640
645
650
z[-]
T[K]
v0
=0.1[m/s]
v0
=0.3[m/s]
v0
=0.5[m/s]
v0
=0.7[m/s]
v0
=0.9[m/s]
v0
=1.1[m/s]
v0
=1.3[m/s]
v0
=1.5[m/s]
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5
-1.5
-1
-0.5
0
0.5
1
x 10
-4
v0
[m/s]
S(T,v0
)
v
0
[m/s] = 0.78, S[T,v
0
] = -9.1981e-05
Maximum rate of
change
Very low sensitivity
Increase inlet
velocity
Constant maximum
reactor temperature
Hot spot moving to
the right
Values close to zero
21
Tesi di Laurea Magistrale – Carlos Sierra
Results Unsteady State Inlet Pressure
1.5 1.6 1.7 1.8 1.9 2 2.1 2.2
0
40
80
120
160
200
Pin
[kPa]
S(T,Pin
)
 =0.1[-]
 =0.2[-]
 =0.3[-]
 =0.4[-]
 =0.5[-]
 =0.6[-]
 =0.7[-]
 =0.8[-]
 =0.9[-]
 =1[-]
Maximum rate of
change
No high sensitivity
before this point
Steady state
22
Tesi di Laurea Magistrale – Carlos Sierra
Results Unsteady State Inlet Pressure
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
1
1.2
1.4
1.6
1.8
2
2.2
2.4
Xmax
[-]
Tmax
/Tin
[-]
 =0.3[-]
 =0.4[-]
 =0.5[-]
Steady state
Safety region
Steady state limit
23
Tesi di Laurea Magistrale – Carlos Sierra
Results Unsteady State Inlet Temperature
620 630 640 650 660 670 680 690 700 710 720
0
50
100
150
200
250
Tin
[K]
S(T,Tin
)
 =0.1[-]
 =0.2[-]
 =0.3[-]
 =0.4[-]
 =0.5[-]
 =0.6[-]
 =0.7[-]
 =0.8[-]
 =0.9[-]
 =1[-]
Maximum rate of
change
No high sensitivity
before this point
Steady state
24
Tesi di Laurea Magistrale – Carlos Sierra
Results Unsteady State Inlet Temperature
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
1
1.2
1.4
1.6
1.8
Xmax
[-]
Tmax
/Tin
[-]
 =0.1[-]
 =0.2[-]
 =0.3[-]
Steady state
Safety region
Steady state limit
25
Tesi di Laurea Magistrale – Carlos Sierra
Results Unsteady State Wall Temperature
580 590 600 610 620 630 640 650 660 670 680
0
40
80
120
160
200
Tw
[K]
S(T,Tw
)
 =0.1[-]
 =0.2[-]
 =0.3[-]
 =0.4[-]
 =0.5[-]
 =0.6[-]
 =0.7[-]
 =0.8[-]
 =0.9[-]
 =1[-]
Maximum rate of
change
No high sensitivity
before this point Steady state
26
Tesi di Laurea Magistrale – Carlos Sierra
Results Unsteady State Wall Temperature
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
1
1.2
1.4
1.6
1.8
2
Xmax
[-]
Tmax
/Tin
[-]
 =0.2[-]
 =0.3[-]
 =0.4[-]
Steady state
Safety region
Steady state limit
27
Tesi di Laurea Magistrale – Carlos Sierra
Results Unsteady State Inlet Velocity
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5
-15
-10
-5
0
5
x 10
-3
vin
[m/s]
S(T,Tw
)
 =0.1[-]
 =0.2[-]
 =0.3[-]
 =0.4[-]
 =0.5[-]
 =0.6[-]
 =0.7[-]
 =0.8[-]
 =0.9[-]
 =1[-]
Values close to zero
No maximum rate of change
Steady state
Inlet velocity is not relevant under this analysis
28
Tesi di Laurea Magistrale – Carlos Sierra
Results Unsteady State Inlet Velocity
0 0.1 0.2 0.3
0.95
1
1.05
1.1
Xmax
[-]
Tmax
/Tin
[-]
 =0.2[-]
 =0.3[-]
 =0.4[-]
Steady state
No safety region
definition
Inlet velocity is not relevant under this analysis
Points concentration No change in reactor behavior
29
Tesi di Laurea Magistrale – Carlos Sierra
Conclusions
• The steady state case was the best approach from the point of view
of safety.
• The unsteady state maximum temperature values were close to the
steady state ones.
• In the unsteady state case the hot spots were moving along the
reactor, without describing a relevant higher temperature compared
to steady state case.
• A new concept is not provided for thermal runaway criterion in
unsteady state operation.
• The inlet pressure, the inlet temperature and the wall temperature
registered a visible peak sensitivity change over the selected
range; for the inlet velocity there was not reported this behavior.
30
Tesi di Laurea Magistrale – Carlos Sierra
Questions
Thanks for your attention!
31
Tesi di Laurea Magistrale – Carlos Sierra
Unsteady State equations
Mass balance:
Energy balance:
Steady State equations: Neglected time dependent (accumulation)
and diffusive terms
32
Tesi di Laurea Magistrale – Carlos Sierra
Dimensionless quantities
Conversion:
Dimensionless
temperature:
Dimensionless
axial coordinate:
Dimensionless
time:
Arrhenius
number:
Damkohler
number:
Stanton
number:
Mass Peclet
number:
Energy Peclet
number:
Dimensionless
adiabatic
temperature rise:
33
Tesi di Laurea Magistrale – Carlos Sierra
Steady State equations
Mass
balance:
Dimensionless
mass balance:
Energy
balance:
Dimensionless
energy balance:
Neglected time dependent and diffusive terms
34
Tesi di Laurea Magistrale – Carlos Sierra
Equation characteristics
• Stiff equation:
PDE for which certain numerical methods for solving the equation
are numerically unstable, unless the step size is taken to be
extremely small. This is evident when the equation includes some
terms that can lead to rapid variation in the solution.
T increases
exp() increases
35
Tesi di Laurea Magistrale – Carlos Sierra
PDEPE solver
• Gibbs phenomena:
Fourier series of a piecewise continuously differentiable behaves at
a jump discontinuity: the n-th partial sum of the Fourier series has
large oscillations near the jump, which might increase the maximum
of the partial sum above that of the function itself.
36
Tesi di Laurea Magistrale – Carlos Sierra
Slope limiters and numerical dissipation
Wouwer A, Saucez P, and Vilas C. Simulation of ODE/PDE models with MATLAB, OCTAVE and SCILAB:
scientific and engineering applications. Springer, 2014.
A slope limiter could be useful for oscillation problems in second
order approximations.
Algorithm more complex, and possible numerical dissipation.
First order approx
Second order approx
37
Tesi di Laurea Magistrale – Carlos Sierra
Further work
• Modelling of highly exothermic reactions with solvers as
COMSOL multi physics (CFD interface), for better
visualization of hot spot points and understanding of the
overall behavior along the reactor in the dynamical operation
condition.
• Calculations and model validation for reactions in solid and
liquid phases.

2014_12_Sierra

  • 1.
    OPTIMIZATION OF POTENTIALLY RUNAWAYREACTIONS CARRIED OUT IN PLUG FLOW REACTORS Tesi di: Carlos Sierra (797299) Scuola di Ingegneria Industriale e dell'Informazione Corso di Laurea in Ingegneria della Prevenzione e della Sicurezza nell’Industria di Processo Anno Accademico 2013 – 2014
  • 2.
    2 Tesi di LaureaMagistrale – Carlos Sierra Introduction • Plug flow reactors that are carrying out exothermic reactions can generate, under certain operating conditions, an uncontrolled temperature increase (thermal runaway) that could lead to the triggering of other more exothermic reactions, such as decomposition reactions. • Thermal runaway studies on continuous reactors are normally held for steady state operation, because all the dynamic changes take place in a short period of time.
  • 3.
    3 Tesi di LaureaMagistrale – Carlos Sierra Scope • Identify possible thermal runaway conditions for steady and unsteady state operations in a plug flow reactor. • Set a safety operation range for a case study of a high exothermic reaction. • Perform a criterion for thermal runaway conditions during unsteady state operations.
  • 4.
    4 Tesi di LaureaMagistrale – Carlos Sierra Structure of the work Theory model Steady state simulations Unsteady state simulations Identify differences and establish a criterion (if possible) MATLAB System solution Equations for mass and energy balances Solution method Analytical Numerical Results analysis Problem definition Runaway in PFR in unsteady state ODE/PDE NOT possible
  • 5.
    5 Tesi di LaureaMagistrale – Carlos Sierra Model assumptions: • Constant diameter through all the reactor. • No radial variations of velocity, concentrations, temperature or reaction rate. Perfect mixing in the radial direction. • Constant density along the reactor. • Constant inlet velocity, which is equal to the axial velocity. Model Plug Flow Reactor (PFR)
  • 6.
    6 Tesi di LaureaMagistrale – Carlos Sierra Unsteady State equations Mass balance: Dimensionless mass balance: Energy balance: Dimensionless energy balance: Steady State equations: Neglected time dependent and diffusive terms
  • 7.
    7 Tesi di LaureaMagistrale – Carlos Sierra Thermal runaway It is defined as a situation where a temperature rise changes the conditions of a chemical reaction, in a way that causes a further increase in temperature. This is a kind of uncontrolled positive feedback.
  • 8.
    8 Tesi di LaureaMagistrale – Carlos Sierra Parametric sensitivity • Parametric sensitivity • Normalized sensitivity It is defined as the system behavior with respect to changes in its input parameters. By changing these values, the system characteristics can achieve desired or undesired behaviors. When a system operates in the parametrically sensitive region (the region where small variations of a parameter make the system becomes sensitive), its performance becomes unreliable and changes with small variations in the parameters.
  • 9.
    9 Tesi di LaureaMagistrale – Carlos Sierra Numerical solution MATLAB solver Steady state simulations Unsteady state simulations ODE solver Ordinary Differential Equation (ODE) Partial Differential Equation (PDE) Function ODE15S PDE solver Function PDEPE ✔Useful in all cases ✘Not useful in all cases Method of lines + Finite difference ✔Useful in all cases PDE Toolbox not considered ✘Useful for one dependent variable system
  • 10.
    10 Tesi di LaureaMagistrale – Carlos Sierra Equation characteristics • Initial conditions: • Empty reactor. • Reactor temperature equal to wall temperature • Boundary conditions: Dankwerts type
  • 11.
    11 Tesi di LaureaMagistrale – Carlos Sierra Equation characteristics • Boundary conditions: Dankwerts type No change Smooth step change
  • 12.
    12 Tesi di LaureaMagistrale – Carlos Sierra Method of lines • Approximation of the spatial derivatives using finite differences. Differentiation matrixes Fornberg algorithm Iteratively computing the weighting coefficients of finite difference formulas of arbitrary order of accuracy on arbitrarily spaced spatial grids. • Time integration of the resulting semi-discrete (discrete in space, but continuous in time) ODE, using ODE MATLAB solvers.
  • 13.
    13 Tesi di LaureaMagistrale – Carlos Sierra Method of lines vs PDEPE 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 z[-] X[-]  =0[-]  =0.15[-]  =0.3[-]  =0.45[-]  =0.6[-]  =0.75[-]  =0.9[-]  =1.05[-]  =1.2[-] 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 z[-] X[-]  =0[-]  =0.15[-]  =0.3[-]  =0.45[-]  =0.6[-]  =0.75[-]  =0.9[-]  =1.05[-]  =1.2[-] Solution for P_in = 1.5[kPa] Solution for P_in = 2.0[kPa] 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 z[-] X[-]  =0[-]  =0.15[-]  =0.3[-]  =0.45[-]  =0.6[-]  =0.75[-]  =0.9[-]  =1.05[-]  =1.2[-] Oscillation for P_in = 1.5[kPa] Not possible solution for P_in = 2.0[kPa] Method of lines PDEPE
  • 14.
    14 Tesi di LaureaMagistrale – Carlos Sierra Case study Naphthalene oxidation to phthalic Anhydride It is a chemical intermediate in the production of plasticizers for polyvinyl chloride (PVC). It was considered a gas phase reaction.
  • 15.
    15 Tesi di LaureaMagistrale – Carlos Sierra The gas-phase naphthalene oxidation to phthalic anhydride is a highly exothermic reaction. This reaction is carried out in multi- tubular reactors, cooled by molten salt that is passing around an external jacket. As a catalyst, V2O5 is used. Case study
  • 16.
    16 Tesi di LaureaMagistrale – Carlos Sierra Case study Initial values for the studied input parameters: • Inlet pressure • Inlet temperature • Wall temperature • Inlet velocity • Steady state analysis: • Reactor temperature vs dimensionless reactor length. • Normalized sensitivity for each input parameter. • Unsteady state: • Normalized sensitivity for each input parameter, for different dimensionless times. • Ratio between maximum reactor temperature and inlet temperature vs conversion at that temperature (topological diagram).
  • 17.
    17 Tesi di LaureaMagistrale – Carlos Sierra 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 625 725 825 925 1025 1125 1225 1325 1425 z[-] T[K] Pin =1.5[kPa] Pin =1.6[kPa] Pin =1.7[kPa] Pin =1.8[kPa] Pin =1.9[kPa] Pin =2[kPa] Pin =2.1[kPa] Pin =2.2[kPa] Results Steady State Inlet Pressure Maximum rate of change High sensitivity 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 0 10 20 30 40 50 60 70 Pin [kPa] S(T,Pin ) P in = 1.85, S[T,P in ] = 58.6996 Increase inlet pressure Increase maximum reactor temperature Hot spot moving to the left
  • 18.
    18 Tesi di LaureaMagistrale – Carlos Sierra Results Steady State Inlet Temperature 0 0.1 0.2 0.3 0.4 0.5 600 700 800 900 1000 1100 1200 1300 z[-] T[K] Tin =620[K] Tin =630[K] Tin =640[K] Tin =650[K] Tin =660[K] Tin =670[K] Tin =680[K] Tin =690[K] Tin =700[K] Tin =710[K] Tin =720[K] 620 630 640 650 660 670 680 690 700 710 720 0 20 40 60 80 100 120 140 160 180 200 Tin [K] S(T,Tin ) T in = 675, S[T,T in ] = 161.1574 Maximum rate of change High sensitivity Increase inlet temperature Increase maximum reactor temperature Hot spot moving to the left
  • 19.
    19 Tesi di LaureaMagistrale – Carlos Sierra Results Steady State Wall Temperature 0 0.1 0.2 0.3 0.4 0.5 600 700 800 900 1000 1100 1200 1300 z[-] T[K] Tw =580[K] Tw =590[K] Tw =600[K] Tw =610[K] Tw =620[K] Tw =630[K] Tw =640[K] Tw =650[K] Tw =660[K] Tw =670[K] Tw =680[K] 580 590 600 610 620 630 640 650 660 670 680 0 20 40 60 80 100 120 Tw [K] S(T,Tw ) T w = 638, S[T,T w ] = 104.3867 Maximum rate of change High sensitivity Increase wall temperature Increase maximum reactor temperature Hot spot moving to the left
  • 20.
    20 Tesi di LaureaMagistrale – Carlos Sierra Results Steady State Inlet Velocity 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 625 630 635 640 645 650 z[-] T[K] v0 =0.1[m/s] v0 =0.3[m/s] v0 =0.5[m/s] v0 =0.7[m/s] v0 =0.9[m/s] v0 =1.1[m/s] v0 =1.3[m/s] v0 =1.5[m/s] 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 -1.5 -1 -0.5 0 0.5 1 x 10 -4 v0 [m/s] S(T,v0 ) v 0 [m/s] = 0.78, S[T,v 0 ] = -9.1981e-05 Maximum rate of change Very low sensitivity Increase inlet velocity Constant maximum reactor temperature Hot spot moving to the right Values close to zero
  • 21.
    21 Tesi di LaureaMagistrale – Carlos Sierra Results Unsteady State Inlet Pressure 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 0 40 80 120 160 200 Pin [kPa] S(T,Pin )  =0.1[-]  =0.2[-]  =0.3[-]  =0.4[-]  =0.5[-]  =0.6[-]  =0.7[-]  =0.8[-]  =0.9[-]  =1[-] Maximum rate of change No high sensitivity before this point Steady state
  • 22.
    22 Tesi di LaureaMagistrale – Carlos Sierra Results Unsteady State Inlet Pressure 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 1.2 1.4 1.6 1.8 2 2.2 2.4 Xmax [-] Tmax /Tin [-]  =0.3[-]  =0.4[-]  =0.5[-] Steady state Safety region Steady state limit
  • 23.
    23 Tesi di LaureaMagistrale – Carlos Sierra Results Unsteady State Inlet Temperature 620 630 640 650 660 670 680 690 700 710 720 0 50 100 150 200 250 Tin [K] S(T,Tin )  =0.1[-]  =0.2[-]  =0.3[-]  =0.4[-]  =0.5[-]  =0.6[-]  =0.7[-]  =0.8[-]  =0.9[-]  =1[-] Maximum rate of change No high sensitivity before this point Steady state
  • 24.
    24 Tesi di LaureaMagistrale – Carlos Sierra Results Unsteady State Inlet Temperature 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 1.2 1.4 1.6 1.8 Xmax [-] Tmax /Tin [-]  =0.1[-]  =0.2[-]  =0.3[-] Steady state Safety region Steady state limit
  • 25.
    25 Tesi di LaureaMagistrale – Carlos Sierra Results Unsteady State Wall Temperature 580 590 600 610 620 630 640 650 660 670 680 0 40 80 120 160 200 Tw [K] S(T,Tw )  =0.1[-]  =0.2[-]  =0.3[-]  =0.4[-]  =0.5[-]  =0.6[-]  =0.7[-]  =0.8[-]  =0.9[-]  =1[-] Maximum rate of change No high sensitivity before this point Steady state
  • 26.
    26 Tesi di LaureaMagistrale – Carlos Sierra Results Unsteady State Wall Temperature 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 1.2 1.4 1.6 1.8 2 Xmax [-] Tmax /Tin [-]  =0.2[-]  =0.3[-]  =0.4[-] Steady state Safety region Steady state limit
  • 27.
    27 Tesi di LaureaMagistrale – Carlos Sierra Results Unsteady State Inlet Velocity 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 -15 -10 -5 0 5 x 10 -3 vin [m/s] S(T,Tw )  =0.1[-]  =0.2[-]  =0.3[-]  =0.4[-]  =0.5[-]  =0.6[-]  =0.7[-]  =0.8[-]  =0.9[-]  =1[-] Values close to zero No maximum rate of change Steady state Inlet velocity is not relevant under this analysis
  • 28.
    28 Tesi di LaureaMagistrale – Carlos Sierra Results Unsteady State Inlet Velocity 0 0.1 0.2 0.3 0.95 1 1.05 1.1 Xmax [-] Tmax /Tin [-]  =0.2[-]  =0.3[-]  =0.4[-] Steady state No safety region definition Inlet velocity is not relevant under this analysis Points concentration No change in reactor behavior
  • 29.
    29 Tesi di LaureaMagistrale – Carlos Sierra Conclusions • The steady state case was the best approach from the point of view of safety. • The unsteady state maximum temperature values were close to the steady state ones. • In the unsteady state case the hot spots were moving along the reactor, without describing a relevant higher temperature compared to steady state case. • A new concept is not provided for thermal runaway criterion in unsteady state operation. • The inlet pressure, the inlet temperature and the wall temperature registered a visible peak sensitivity change over the selected range; for the inlet velocity there was not reported this behavior.
  • 30.
    30 Tesi di LaureaMagistrale – Carlos Sierra Questions Thanks for your attention!
  • 31.
    31 Tesi di LaureaMagistrale – Carlos Sierra Unsteady State equations Mass balance: Energy balance: Steady State equations: Neglected time dependent (accumulation) and diffusive terms
  • 32.
    32 Tesi di LaureaMagistrale – Carlos Sierra Dimensionless quantities Conversion: Dimensionless temperature: Dimensionless axial coordinate: Dimensionless time: Arrhenius number: Damkohler number: Stanton number: Mass Peclet number: Energy Peclet number: Dimensionless adiabatic temperature rise:
  • 33.
    33 Tesi di LaureaMagistrale – Carlos Sierra Steady State equations Mass balance: Dimensionless mass balance: Energy balance: Dimensionless energy balance: Neglected time dependent and diffusive terms
  • 34.
    34 Tesi di LaureaMagistrale – Carlos Sierra Equation characteristics • Stiff equation: PDE for which certain numerical methods for solving the equation are numerically unstable, unless the step size is taken to be extremely small. This is evident when the equation includes some terms that can lead to rapid variation in the solution. T increases exp() increases
  • 35.
    35 Tesi di LaureaMagistrale – Carlos Sierra PDEPE solver • Gibbs phenomena: Fourier series of a piecewise continuously differentiable behaves at a jump discontinuity: the n-th partial sum of the Fourier series has large oscillations near the jump, which might increase the maximum of the partial sum above that of the function itself.
  • 36.
    36 Tesi di LaureaMagistrale – Carlos Sierra Slope limiters and numerical dissipation Wouwer A, Saucez P, and Vilas C. Simulation of ODE/PDE models with MATLAB, OCTAVE and SCILAB: scientific and engineering applications. Springer, 2014. A slope limiter could be useful for oscillation problems in second order approximations. Algorithm more complex, and possible numerical dissipation. First order approx Second order approx
  • 37.
    37 Tesi di LaureaMagistrale – Carlos Sierra Further work • Modelling of highly exothermic reactions with solvers as COMSOL multi physics (CFD interface), for better visualization of hot spot points and understanding of the overall behavior along the reactor in the dynamical operation condition. • Calculations and model validation for reactions in solid and liquid phases.