modeling of electrical interactions with sofcs library/research/coal...cfc/bt/gtri perc/uic emi/vt...

PERC/UICCFC/BT/GTRI EMI/VT

Modeling of Electrical Interactions with SOFCs

Project Investigators• S.K. Mazumder

– University of Illinois (PERC/UIC)

• M. von Spakovsky & D. Nelson– Virginia Tech (EMI/VT)

• C. Haynes – Georgia Tech (CFC/BT/GTRI)

Industrial Collaborators• D. Herbison & V. Tchkalov

– Synopsys

• Joseph Hartvigsen – Ceramatech

SECA Modeling & Simulation Team – Integration Meeting

October 15, 2002

NETL, Pittsburgh


APU Overview

• Meet customer requirements, but also add new features– Pre-heat and Pre-cool– Engine-off electrical accessories– Remote and emergency power

• Peak power of 4-5 kWe for LDV• Could reduce peripherals fuel consumption by

50 %• Largest benefit for applications with high idle

times


Key APU RequirementsKey APU Requirements

• Common fuel, small, lightweight

• High system efficiency, low emissions

• Start-up in less than 3 minutes

• Maintenance, durability, cost consistent with application (commercial, RV, luxury, first applications?)

• Noise


SOFC APU Sizing

Typical vehicle APU application:

Maximum power = 5000 We (100 %)Average power = 1500 We (30 %)Minimum power = 500 We (10 %)


Transient ResponseDifferent time constants - fast to slow:

• Electrochemistry – instantaneous (sufficiently fast to assume quasi-stationary behavior)

• Stack electrical response – fraction of a second

• Thermal-hydraulic response – seconds (e.g., POX) to minutes (e.g., SMR)

• Load response with some energy storage - 1 sec

• Start up time - 10 sec to 3 min

• PES electrical response - msec


OBJECTIVES (Phases I and II)• Develop fully transient nonlinear, unified models for SOFC planar configurations, different PESs, and a variety of BOPS components

•Develop a prototypical software package (Phase II) for industry to understand the dynamics of SOFC stack, power electronics, and system interactions

• Implement models in the SaberDesigner and gProms dynamic simulation and optimization environments

• Demonstrate the feasibility of integrating these models into an overall systems-analysis and optimization tool (Phase I)

•Conduct parametric studies (Phase I) and optimizations (Phase II) to determine control strategies and their effects on cell reliability, efficiency, and power density; as well as system response and configuration, and component designs.

PERC/UICCFC/BT/GTRI EMI/VTPERC/UIC

SOFC Power-Electronics Subsystem (PES):

Fuel Processor

Fuel-Cell Stack

Water Management

Air System

Power-Electronics

System

Fuel In

Wat

er O

ut

Air In Heat Out E

xhau

st

Fuel

Application Load

PESSOFCSS

Thermal Management

Heat Out

Exhaust Out

BOPS


Technical Issues: PES

• To investigate the impact of critical parameters of a closed-loop PES, which can negatively affect the performance and integrity of a SOFC stack for a given application load {Phase I (set cases); Phase II (experimental verifications of Phase-I results and extension to generalized analysis)}

– Circuit and control parameters of the power converters– Topological architectures (standalone, cascaded, or distributed).– Switching schemes (that is, whether the converters are operated with

PWM, soft switching, interleaving etc.)– Application load (converter and stationary/non-stationary loads)

PERC/UIC


Impact of Parametric Variations on Ripple: An “Illustration” for a DC-DC Converter

By simply varying “only one” parameter (load in this case), the voltage and current ripples of the converter change drastically. In reality, more than one parameter can vary simultaneously.

VR

CR

VR

CR

VR

CR

VR

VR: Voltage rippleCR: Current ripple



• To investigate the steady-state and dynamic interaction problems due to the integration of the SOFCSS and PES for a given application load {Phase I (set cases); Phase II (experimental verifications of Phase-I results and extension to generalized analysis)}

– Fast- and slow-scale scale instabilities in ripple dynamics– Impact of variations in SOFC output voltage on PES– Effect of time-varying perturbations of the stationary/non-stationary

loads and PES on the SOFC



• To determine the criteria for the synthesis of an optimal power-electronic converter (for a given SOFC), which can increase the lifetime and efficiency of the fuel cell {Phase II (robust)}

– What type of converter topology should one choose, for a given application load, so that the performance and efficiency of the SOFC can be maximized? (A SOFC can be interfaced to an application load using multiple converter topologies. The severity of the interactions among the subsystems is not the same for all the topologies.)

– How should one optimize an existing power-electronics system so that it has minimal negative effect on the SOFC?


R&D Approach: PES

• Unified Model of PES• Power-electronic systems are “hybrid systems”; they comprise discontinuous

differential equations, discrete differential equations, functional differentialinclusions, digital automata, impulsive differential equations, non-smooth differential equations, ordinary and even partial differential equations

• Unified framework is an indexed collection of dynamical systems along with a map for transitions among them that can account for “any” dynamical model of a PES and predict fast- and slow-scale ripple dynamics

Conventional Averaged Models of PES cannot account for all ripple dynamics and in many cases predict incorrect stability results.


R&D Approach: PES

System Interaction using “Bifurcation” Analyses

When does a nominal system loose stability?

What is the mechanism of the instability?

What happens after the

instability (post-instability

dynamics)?For variations in one or more parameters, when

can one expect the ripple magnitude and frequency

of the PES to change?

How conservative should the PES design be?

• Are the “new” voltage and current ripples

dangerously high for the SOFC?


Previous Successes: SOFCSS - Modeling TSOFC Transients with Lagrangian Approach

• At “t*=0+” the load increases and current initially “spikes” up

• The reactants supply, however, does not change

• A new steady state is reached wherein the hydrogen profile is decreased along the anode due to reactant depletion

0 0.25 0.5

0.75 1

1.25 1.5

1.75 2

Par

tial P

ress

ure

(atm

)

0 0.2 0.4 0.6 0.8 1 Dimensionless Length Along Cell

t/T=0

t/T=0.333

t/T=0.667

t/T=1

Axial H2 Partial PressuresAs a Function of Time

t*


• Reactants’ inlet flow properties are the same

• The fuel elements’ exit properties depend upon their locations at t*=0+

• Steady state is regained when element 3 exits (t*=1), because every successive element will then pass along the cell “seeing” only the new operating potential

Fuel Cell

1234

Fuel Stream

Oxidant Stream

t = 0+




� ηelement(t+∆t) =ηfield(x+∆x, t +∆t)

Element properties Element properties were calculated by were calculated by using the proven steady using the proven steady state model on an state model on an instantinstant--byby--instant basis instant basis until a new electrical until a new electrical steady state was steady state was reachedreached

Fuel Cell

11

Fuel Stream

Oxidant Stream

x

t + ∆t t

x+ ∆x

t t + ∆t v1

v2


360

400

440

480

520

84%

87%

90%

93%

96%

0 0.2 0.4 0.6 0.8 1 t*=t/T

Current, Fuel Utilization TransientsDependence on Dimensionless Time

Fuel utilization

Current (Amps)

• Current spikes up, yet the fuel supply remains invariant due to the decoupling of the cell

• Fuel utilization thus increases; this causes current (and power) to decrease from t*=0+

values, until a new steady state “match” occurs at the new voltage (t*=1)



R&D Approach: SOFCSS- Need Enhancements to Transient Model

Resolution of operating environment via model(s)Extensive communication with SECA teammates

regarding cell sensitivity issues (e.g., “electrochemical fatigue” parameters??)

• Dynamic response to current rippleSuperposition applied to original Lagrangian

approach wherein multiple, periodic stimuli (as opposed to initial stimulus) serve as the “forcing function”

• Cell reliability under load variation


Previous Work Modeling the SOFCSS/BOPSPrevious Work Modeling the SOFCSS/BOPS

PEMFC based PEMFC based TES modelTES model

SOFC Stack SOFC Stack modelmodel

SOFC Stack Subsystem SOFC Stack Subsystem (SOFCSS) transient & (SOFCSS) transient &

steady state modelssteady state models

SOFC System SOFC System (SOFCSS + BOPS) (SOFCSS + BOPS) steady state modelsteady state model


Air

Combustion Gases

Water

Heat Recovery

Anode

CathodeB

E

AH

C

G

D

I

F

HX III

HX V

HX IV

HX I

HX II

HX VI

Methane

Previous Work: System (SOFCSS/BOPS) ModeledPrevious Work: System (SOFCSS/BOPS) Modeled

Assumptions:Assumptions:

• Pure methane

•Air humidity fixed at 65%and 298 °K, 1 atm

• Cell temperature 1273°K


Previous Work: BOPSPrevious Work: BOPS

PrePre--reforming modelreforming model: SMR and shift reactions – kinetic / equilibrium / geometry based

∫=

= −=

4

4

4

0

CHi,

XX

X CHBcrtubes

reactorCH

reactor rdX

)An(

nL

ρ

Gas turbine, combustor, heat exchangers Gas turbine, combustor, heat exchangers : thermodynamic / geo-metry based models

Fuel Processing andenergy recovery model

Pre-reforming CombustorGas TurbineHeatExchangers


Previous Work: System (SOFCSS/BOPS) ResultsPrevious Work: System (SOFCSS/BOPS) ResultsSimulation in the Simulation in the

Synthesis/Design modeSynthesis/Design modeFixed cell power at different

cell voltages

• The model works and can be used for trade-off analysis or large-scale optimization

• The model gives coherent and interesting results

The parameters studied:

•Methane conversion during pre-reforming

• Steam to carbon ratio

• Cell pressure

• Fuel utilization

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85

C e l l v o l t a g e [ V ]

SMR , 60%

SMR , 30%

Cell , 30%,60%

Influence of methane conversion

•High methane conversion during pre-reforming is not recommended

• There is a cell voltage where the total cost of the cell and the SMR is a minimum

Cell voltage [V]Le

ngth

[m

m]


400

900

1400

1900

2400

2900

3400

0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85

C e l l v o l t a g e [ V ]

SMR , 2.8

Cell , 3.5

SMR , 3.5

Cell , 2.8

Influence of the steam to carbon ratio

• Better to work at a low steam to carbon ratio

• However, lower limit of the steam to carbon ratio exists because of carbon deposition

Previous Work: System (SOFCSS/BOPS) ResultsPrevious Work: System (SOFCSS/BOPS) Results

• At higher pressure, equipment cost decreases

• At higher pressure, system efficiency decreases

Influence of the cell pressure

300

800

1300

1800

2300

2800

3300

3800

4300

0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85Cell voltage

Cel

l le

ngth

[m

m]

0.4

0.45

0.5

0.55

0.6

0.65

0.7

0.75

0.8

0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85

Cell voltage [V]

Eff

icie

ncy

Cell , 15 bar

SMR , 25 bar

SMR, 15 bar

Cell , 25 bar

15 bar

25 bar

44 CHCHmethane

cellGTsystemelec nMLHV

PP +=η

Cell voltage [V] Cell voltage [V]

Cell voltage [V]

Leng

th [

mm

]

Leng

th [

mm

]E

ffic

ienc

y


Previous Work: System (SOFCSS/BOPS) ResultsPrevious Work: System (SOFCSS/BOPS) ResultsInfluence of fuel utilization

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85

Cell voltage [V]

Eff

icie

ncy

Cell, 75%

Cell, 85%

System, 75%System, 85%

44 CHCHmethane

cellcellelec nMLHV

P=η

1200

1300

1400

1500

1600

1700

1800

1900

2000

0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85

Cell voltage [V]

Inle

t exp

and

er te

mp

erat

ure

[K]

Expander inlet temperature

75%

85%

0

2000

4000

6000

8000

10000

12000

0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85

Cell voltage [V]

SM

R le

ng

th [m

m]

75%

85%

200

400

600

800

1000

1200

1400

1600

0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85

Cell voltage [V]

Len

gth

[mm

]

85%

75%

Cell voltage [V]Cell voltage [V]

Cell voltage [V]

Cell voltage [V]

Eff

icie

ncy

SMR

leng

th (m

m)

Cel

l len

gth

(mm

)

Tem

pera

ture

(o K)


SupersonicPenetration

Combat Air Patrol

Loiter

Descend and Land

Warm-up and Takeoff

Accelerate and Climb

Subsonic Cruise Climb

Descend

CombatEscapeDashClimb

Subsonic Cruise Climb

Descend

Deliver expendables

Source: Mattingly et al, 1987

n System: Advanced Air-to-Air Fighter (PS-ECS-FLS-VC/PAOS-AFS) –Synthesis/design optimized with 553 degrees of freedom

Previous Work: Integrated System-level Synthesis / DesignOptimization - ILGO

n System: Fuel Cell Based Total Energy System (SS-FPS-EHPS)– Synthesis/design optimized with 39 degrees of freedom

Fuel ProcessingSub-system

(FPS)

ElectricHeat PumpSub-system

(EHPS)

Stack Sub-system(SS)

TESFPSE&

airQ&

EHPSE&

netE&

2Hn&

fueln&

EHPSQ& SSQ&

FPSQ&


R&D Approach: Phase I TasksR&D Approach: Phase I Tasks

• Planar SOFCSS model development, implementation and validation

• Parametric studies of best-practice control strategies

• Analysis of system stability and dynamics

• Refine models and couple with ILGO; determine optimal control strategies and analyze load profile variations on reliability, performance, and response

• Characterization of the PES interface with the SOFCSS

• Load profile development

• BOPS model development, implementation and validation

• Integration of the PES, SOFCSS, and BOPS models

Phase I:

Phase II:


R&D Approach

Collaboration with DOE Labs and SECA partners is critical to an effective Phase I !!!

• Integration within (i.e. UIC, Va. Tech and Ga. Tech) • Integration without (i.e. SECA teammates)

modeling of electrical interactions with sofcs library/research/coal...cfc/bt/gtri perc/uic emi/vt...

Documents