A Stability/Bifurcation Framework For Process Design
A Stability/Bifurcation Framework For Process Design
A Stability/Bifurcation Framework For Process Design
2
Centre for Process System Engineering,
Imperial College, London, SW7 2BY, UK
Princeton University
Motivation
• A large number of existing scientific, large-scale legacy codes
–Based on transient (timestepping) schemes.
• Enable legacy codes perform tasks such as bifurcation/stability analysis
–Efficiently locate multiple steady states and assess the stability of solution branches.
–Identify the parametric window of operating conditions
for optimal performance
bif. quantity
–Locate periodic solutions
•Autonomous, forced (PSA,RFR)
–Appropriate controller design.
• RPM: method of choice to build around existing time-stepping codes. parameter
Princeton University
Recursive Projection Method (RPM)
• Treats timstepping routine, as a
Reconstruct solution: “black-box”
Initial state un
u = p+q = PN(p,q)+QF – Timestepper evaluates un+1= F(un)
F.P.I. • Recursively identifies subspace of
Newton slow eigenmodes, P
ons
Timestepping
iter ard
iterations
ati
Princeton University
gPROMS:A General Purpose Package
Nonlinear Nonlinear
algebraic programming
equation solvers gPROMS solvers
Steady-state Steady-state &
& Dynamic Dynamic
Differential
Simulation Optimisation
algebraic
equation solvers Dynamic
gPROMS optimisation
Parameter Model solvers
Estimation
Data
Reconciliation
Maximum likelihood
estimation solvers
Princeton University
Mathematical solution methods in gPROMS
• Combined symbolic, structural & numerical techniques
symbolic differentiation for partial derivatives
automatic identification of problem sparsity
structural analysis algorithms •well-posedness
•DAE index analysis
• •consistency of DAE IC’s
Advanced features:
•automatic block triangularisation
exploitation of sparsity at all levels
support for mixed analytical/numerical partial derivatives
handling of symmetric/asymmetric discontinuities at all levels
• Component-based architecture for numerical solvers
open interface for external solver components
hierarchical solver architectures
• mix-and-match
• external solvers can be introduced at any level of the hierarchy
Princeton University
FitzHugh-Nagumo: An PDE-based Model
• Reaction-diffusion model in one dimension
• Employed to study issues of pattern formation
in reacting systems
– e.g. Beloushov-Zhabotinski ut u u u v
2 3
reaction
– u “activator”, v “inhibitor” v t δ 2
v ε (u a1v a0 )
– Parameters:
δ 4.0, a0 0.03, a1 2.0
– no-flux boundary conditions
– , time-scale ratio, continuation parameter
• Variation of produces turning points
and Hopf bifurcations
Princeton University
Bifurcation Diagrams
Around Hopf Around Turning Point
<u>
Princeton University
Eigenspectrum Around Hopf
Princeton University
Eigenvectors
= 0.02
Princeton University
Arc-length continuation with gPROMS
dy
System: f ( y, p)
dt y
f ( y; p*)
Det[ ]0
y p
Pseudo – arc length condition
0 f ( y, p) (1) ( y1 y0 )T ( p p0 )
( y y1 ) 1 ( p p1 ) S 0 (2)
S S
Princeton University
System Jacobian
dx
dx f ( x, y , p )
ODEs : f ( x, p ) DAEs : dt y y * ( x)
dt 0 g ( x, y , p )
1
f ( x, p ) f f g g
x x y y x
Jacobian of the ODE Stability matrix
Princeton University
Tubular Reactor: A DAE system
Dimensionless equations:
x1 1 x1
2
x1 x2
Pe1 Da (1 x1 ) exp[ ] (1)
t z 2
z 1 x2 /
x2 1 x2
2
x2 x2
Pe2 x 2 BDa (1 x1 ) exp[ ] x2 w (2)
t z 2
z 1 x2 /
Boundary Conditions:
x1 ( z 0, t ) x2 ( z 0, t ) (3)
Pe1 x1 0 Pe2 x2 0
z z
x1 ( z 1, t ) x2 ( z 1, t ) (4)
0 0
z z
Eqns (1)-(4): system of DAEs. Can also substitute to obtain system of ODEs.
Princeton University
Bifurcation/Stability with RPM-gPROMS
Princeton University
Eigenspectrum
1 6.00E-02
5.00E-02
0.5
4.00E-02
0
3.00E-02
-1 -0.5 0 0.5 1 1.5
-0.5 2.00E-02
1.00E-02
-1
0.00E+00
1 3.50E-02
3.00E-02
0.5 2.50E-02
2.00E-02 Im
0
1.50E-02
-1 -0.5 0 0.5 1 1.5
1.00E-02 Re
-0.5
5.00E-03
-1 0.00E+00
0 20 40 60 80 100 120
Da=0.110021
Princeton University
Stability Analysis without the Equations
1
-0.5
y(k) -1
Matrix-free ARNOLDI
C h o o s e q 1 w ith q 1 1
+
F o r j = 1 U n til C o n v e r g e n c e D O
( 1 ) C o m p u t e a n d s t o r e Aq
εq j
( 3 ) r j Aq j
j
h t, j q t
t1
Princeton University
Rapid Pressure Swing Adsorption
1-Bed 2-Step Periodic Adsorption Process
t=0 to T/2 •Isothermal operation t= T/2 to T
•Modeling Equations (Nilchan & Pantelides)
Ci(z=0)=PfYf/(RTf) Ci
Mass balance in ads. bed ( z 0) 0
z=L z
P(z=0)=Pf Ci qi (vCi ) 2 Ci
t b Di P(z=0)=Pw
t t z z 2
n
P
Ci
RT i 1
P 180v (1 b ) 2
Darcy’s law
z=0 z d p2 b3
qi
t
ki (mi pi qi ) Rate of ads .
Step 2: Depressurisation
Step 1 :
Pressurisation
Princeton University
Rapid Pressure Swing Adsorption
1-Bed 2-Step Periodic Adsorption Process
q , c (t=T)
Production of oxygen enriched air
Zeolite 5A adsorbent (300m) q ,c (t=0) q , c (t=T/2)
– T= 3s
Princeton University
Typical RPSA simulation results
(Nilchan and Pantelides, Adsorption, 4, 113-147, 1998)
50 0.5
45
c1(z=0.5) (mol/m )
3
40
0
-1 -0.5 0 0.5 1
35
30
-0.5
25
Time (s)
20
0 50 100 150 200 -1
Princeton University
PRM-gPROMS Spatial Profiles (t=T)
30
0.3
c1 mol/m3 q1 mol/kg
20 0.2
10 0.1
0
0
0 0.2 0.4 0.6 0.8 x 1
0 0.2 0.4 0.6 0.8 x 1
z z
90 0.3
c2 mol/m3 q2 mol/kg
60 0.2
30 0.1
0 0
0 0.2 0.4 0.6 0.8 x 1 0 0.2 0.4 0.6 0.8 x 1
z z
Princeton University
Leading Eigenvectors, =0.99484
0.16 0.0012
c1 q1
0.12 c1 q1
0.0008
0.08
0.0004
0.04
0 0
0 9 0 1 0 0 1 1 0 1 2 0 1 3 0 1 4 0 1 5 0 1 6 0 1 7 0
0.00E+00
-0.05 -2.00E-04
-0.1 -4.00E-04
c2 q2
c2 q2
-6.00E-04
-0.15
Princeton University
Conclusions
• Can construct a RPM-based computational framework around large-scale
timestepping legacy codes to enable them converge to unstable steady states and
efficiently perform bifurcation/stability analysis tasks.
– gPROMS was employed as a really good simulation tool
– communication with wrapper routines through F.P.I.
• Both for PDE and DAE-based systems.
• Have “brought to light” features of gPROMS for continuation around turning
points and information on the Jacobian and/or stability matrix at steady states of
systems.
• Employed matrix-free Arnoldi algorithms to perform stability analysis of steady
state solutions without having either the Jacobian or even the equations!
• Used the RPM-based superstructure to speed-up convergence and perform
stability analysis of an almost singular periodically-forced system
• Have enabled gPROMS to trace autonomous limit cycles
• Newton-Picard computational superstructure for autonomous limit cycles.
Princeton University
gPROMS
• General purpose commercial package for modelling, optimization and control of process
systems.
• Allows the direct mathematical description of distributed unit operations
• Operating procedures can be modelled
– Each comprising of a number of steps
• In sequence, in parallel, iteratively or conditionally .
• Complex processes: combination of distributed and lumped unit operations
– Systems of integral, partial differential, ordinary differential and algebraic equations
(IPDAEs).
– gPROMS solves using method of lines family of numerical methods.
• Reduces IPDAES to systems of DAEs.
– Time-stepping or pseudo-timestepping.
• Jacobians NOT explicitly available.
– Cannot perform systematic bifurcation/stability analysis studies.
Princeton University
Tracing limit cycles
Tracing Limit Cycles
continuation
(I) F.P.I R.P.M
within through
gPROMS FORTRAN
continuation F.P.I
(II)
through
FORTRAN F.P.I Getting system
tracing Jacobian
limit cycles through an FPI
within gPROMS
Princeton University
Tracing limit cycles
Tracing Limit Cycles
dy
SYSTEM: f ( y, p)
dt
d y
dt f ( y, p ) y(0) y (T ) 0
dT
0 G ( y (0), p) 0
dt
G ( y (0), p) yi (0) a 0
dy (0)
G ( y (0), p ) i 0
dt
Princeton University